Method performed by terminal or network and device therefor in wireless communication system

Abstract

A terminal according to an embodiment of the present disclosure may: receive configuration information for association between a first model for channel state information (CSI) compression and a second model for CSI reconstruction; construct a third model on the basis of the configuration information; and perform CSI reporting on the basis of an inference result acquired from a combination of the first model and the third model, wherein the third model may be configured in a format for inputting an output of the first model to the second model.

Description

A method performed by a terminal or network in a wireless communication system and an apparatus for the same

[0001] The present disclosure relates to a wireless communication system, and more specifically, to a method and apparatus for performing wireless communication between terminals or networks in a wireless communication system.

[0002] The 5G mobile communication system is a successor technology to LTE (Long Term Evolution) and is a new clean-slate type of mobile communication system characterized by high performance, low latency, and high availability. In the case of 5G NR, all available spectrum resources can be utilized, ranging from low-frequency bands below 1 GHz to intermediate frequency bands between 1 GHz and 10 GHz, and high-frequency (millimeter wave) bands above 24 GHz. Based on the underlying technology of 5G mobile communication, 6G mobile communication systems are being developed.

[0003] Advancements in computational processing technology and AI (artificial intelligence) / ML (machine learning) are leading to the intelligentization and sophistication of nodes and terminals constituting wireless communication networks. In next-generation wireless communication systems utilizing AI / ML, it is expected that various network decision parameter values ​​(e.g., transmit / receive power of each base station, transmit power of each terminal, precoder / beam of base stations and terminals, time / frequency resource allocation for each terminal, duplex method of each base station, etc.) can be rapidly optimized, derived, and applied based on diverse network environment parameters (e.g., distribution / location of base stations, distribution / location / material of buildings / furniture, location / movement direction / speed of terminals, weather information, etc.).

[0004] The technical problem to be solved by the present disclosure is to provide a method for efficiently performing a wireless signal transmission and reception process and an apparatus for such purpose. According to one embodiment, a method for compressing CSI feedback for downlink transmission and reporting it to a network is provided. For example, in an environment where a CSI feedback compression model (e.g., CSI encoder) on the terminal side and a CSI feedback restoration model (e.g., CSI decoder) on the base station side are learned / configured independently of each other, a method for collaboration between the terminal and the base station may be provided. For example, a Bridge AI / ML model, which is an additional AI / ML model located between the CSI feedback compression model (e.g., CSI encoder) on the terminal side and the CSI feedback restoration model (e.g., CSI decoder) on the base station side, may be provided.

[0005] In addition to the technical challenges described above, other technical challenges can be inferred from the description below.

[0006] According to one aspect of the present disclosure, a method performed by a terminal comprises receiving configuration information for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; configuring a third model based on said configuration information; and performing a CSI report based on an inference result obtained from the combination of said first model and said third model, wherein the third model may be configured in a format for inputting the output of said first model to said second model.

[0007] The above setting information may include at least one of identification information of the third model or information about the first data set for learning the third model.

[0008] The information regarding the first data set may include identification information of the first data set associated with the third model.

[0009] The above terminal configuring the above third model may include learning the above third model based on the above first data set.

[0010] The above-mentioned first model and the above-mentioned third model can be trained based on different data sets.

[0011] The above third model may be attached to the end of the above first model or combined as part of the above first model.

[0012] The above terminal can transmit a terminal capability report containing information about candidate third models supported by the terminal.

[0013] The above third model may be a reference model pre-processed based on an ensemble of candidate first models.

[0014] The first model above is a CSI encoder model, the second model above is a CSI decoder model, and the third model above may be a bridge model located between the CSI encoder model and the CSI decoder.

[0015] According to another aspect of the present disclosure, a computer-readable non-transient recording medium may be provided that records a program for performing the method described above.

[0016] An apparatus according to another aspect of the present disclosure comprises: a memory configured to store instructions; and a processor configured to perform operations by executing said instructions, wherein the operations of the processor include receiving configuration information for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; constructing a third model based on said configuration information; and performing a CSI report based on an inference result obtained from the combination of the first model and the third model, and the third model may be configured in a format for inputting the output of the first model to the second model.

[0017] The above device may further include a transceiver. The above device may be a terminal operating in a wireless communication system.

[0018] The above device may be a processing device configured to control a terminal operating in a wireless communication system.

[0019] According to another aspect of the present disclosure, a method performed by a base station comprises determining a third model for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; transmitting configuration information for the third model to a terminal; receiving a compressed CSI report from the terminal; and performing CSI reconstruction based on the compressed CSI report and the second model, wherein the compressed CSI report includes an inference result of the terminal obtained from the combination of the first model and the third model, and the third model may be a model configured in a format for inputting the output of the first model to the second model.

[0020] A base station according to another aspect of the present disclosure comprises: a memory configured to store instructions; and a processor configured to perform operations by executing said instructions, wherein the operations of the processor include determining a third model for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; transmitting configuration information for said third model to a terminal; receiving a compressed CSI report from said terminal; and performing CSI reconstruction based on said compressed CSI report and said second model, wherein the compressed CSI report includes an inference result of said terminal obtained from the combination of said first model and said third model, and said third model may be a model configured in a format for inputting the output of said first model to said second model.

[0021] According to the present disclosure, signal transmission and reception can be performed efficiently in a wireless communication system. For example, signaling overhead associated with CSI reporting can be reduced by compressing CSI feedback for downlink transmission and reporting it to the network. A method for collaboration between a terminal and a base station is provided even when the terminal-side CSI feedback compression model (e.g., CSI encoder) and the base station-side CSI feedback restoration model (e.g., CSI decoder) are independently trained or configured. This ensures the independence and autonomy of the implementation of the CSI feedback compression model (e.g., CSI encoder) and the base station-side CSI feedback restoration model (e.g., CSI decoder), and resolves proprietary issues regarding the disclosure of each model.

[0022] In addition to the technical effects described above, other technical effects can be inferred from the description below.

[0023] FIG. 1 illustrates an exemplary flexible network topology to which some of the examples of the present specification may be applied.

[0024] FIG. 2 illustrates an example of a communication system applicable to the present disclosure.

[0025] FIG. 3 illustrates an example of a wireless device that can be applied to the present disclosure.

[0026] FIG. 4 illustrates a communication procedure between a first node (e.g., a terminal) and a second node (e.g., a base station) applicable to the present disclosure.

[0027] Figure 5 illustrates a general functional architecture for an AI / ML model.

[0028] FIG. 6 illustrates a communication procedure between a first node (e.g., terminal) and a second node (e.g., base station) to which an AI / ML model is applied.

[0029] FIG. 7 shows an electromagnetic spectrum according to one embodiment of the present disclosure.

[0030] FIG. 8 illustrates an example of a procedure for transmitting system information for THz communication to which the present disclosure applies.

[0031] FIG. 9 illustrates a beam management procedure applicable to the present disclosure.

[0032] FIG. 10 shows an example of a sensing operation according to one embodiment of the present disclosure.

[0033] FIG. 11 illustrates a time / frequency resource for a sensing operation according to one embodiment of the present specification.

[0034] FIG. 12 illustrates a procedure related to a sensing operation according to one embodiment of the present specification.

[0035] Figure 13 illustrates a CSI Encoder for CSI compression and a CSI Decoder for CSI restoration.

[0036] FIG. 14 is a diagram illustrating a pre-learning step according to one embodiment.

[0037] FIG. 15 is a diagram illustrating a post-learning step according to one embodiment.

[0038] FIG. 16 is a diagram illustrating an inference step according to one embodiment.

[0039] FIG. 17 is a diagram illustrating the operation of a terminal and a base station according to one embodiment.

[0040] FIG. 18 is a diagram illustrating a compressed CSI report according to one embodiment.

[0041] FIG. 19 is a diagram illustrating the operation of a terminal and a base station according to one embodiment.

[0042] FIG. 20 illustrates the flow of a method performed at a terminal according to one embodiment.

[0043] FIG. 21 illustrates the flow of a method performed at a base station according to one embodiment.

[0044] In this specification, "A or B" may mean "only A," "only B," or "both A and B." Alternatively, in this specification, "A or B" may be interpreted as "A and / or B." For example, in this specification, "A, B or C" may mean "only A," "only B," "only C," or "any combination of A, B and C."

[0045] A slash ( / ) or a comma used in this specification may mean "and / or." For example, "A / B" may mean "A and / or B." Accordingly, "A / B" may mean "only A," "only B," or "both A and B." For example, "A, B, C" may mean "A, B or C."

[0046] In this specification, "at least one of A and B" may mean "only A," "only B," or "both A and B." Additionally, in this specification, the expressions "at least one of A or B" or "at least one of A and / or B" may be interpreted as synonymous with "at least one of A and B."

[0047] Additionally, in this specification, "at least one of A, B and C" may mean "only A," "only B," "only C," or "any combination of A, B and C." Also, "at least one of A, B or C" or "at least one of A, B and / or C" may mean "at least one of A, B and C."

[0048] Additionally, parentheses used in this specification may mean "for example." Specifically, when indicated as "control information (ABC)," "ABC" may be described as an example of "control information." For example, "control information" may include DEF as another example. In other words, "control information" in this specification is not limited to "ABC," and "ABC" may be described as an example of "control information." Also, when indicated as "control information (i.e., ABC)," "ABC" may be described as an example of "control information."

[0049] In addition, terms such as "first," "second," etc. in this specification are used solely for the purpose of distinguishing one component from another and are not used to limit the components, nor are they used to limit the order or importance of the components unless specifically limited. Accordingly, a first component in one embodiment of this specification may be referred to as a second component in another embodiment, and likewise, a second component in one embodiment may be referred to as a first component in another embodiment.

[0050] In the following explanation, 'when, if, in case of' can be replaced with 'based on'.

[0051] Technical features described individually within a single drawing in this specification may be implemented individually or simultaneously.

[0052] In this specification, a terminal is a user-side device (user equipment, UE) or a consumer-side device, and may also be referred to as a first node that receives / transmits signals from / to a base station / second node / IAB node / Transmission-Reception Point (TRP). A terminal may correspond to a physical node or a logical node. A terminal may correspond to a user-side endpoint or an intermediate point between other endpoints. In communication between two points not limited to endpoints (including one-to-one / many-to-one / one-to-many / many-to-many communication), a terminal may correspond to a served node. A terminal may be a fixed-location node or a non-fixed-location (or mobile) node.

[0053] In this specification, a Base Station (BS) is a device on the network side and may also be referred to as a second node / IAB node / x-NodeB (x-NodeB, where x may be an abbreviation related to Radio Access Technology (RAT)) / Transmission-Reception Point (TRP). A Base Station may correspond to a physical node or a logical node. A Base Station may correspond to an endpoint on the network side or an intermediate point between other endpoints. In communication between two points not limited to endpoints (including one-to-one / many-to-one / one-to-many / many-to-many communication), a Base Station may correspond to a serving node. A Base Station may be a node with a fixed location or a node with an indefinite location.

[0054] In this specification, higher layer parameters may be set for the terminal, pre-set, or pre-defined. For example, a base station may transmit higher layer parameters to the terminal. For example, the terminal may transmit parameters such as capability to the base station as higher layer parameters. For example, higher layer parameters may be transmitted via RRC (radio resource control) signaling or MAC (medium access control) signaling.

[0055] In this specification, information / state / parameters being "configured" or "pre-configured" may be interpreted as the information / state / parameters being provided / pre-provided to the terminal through pre-defined signaling (e.g., SIB, MAC, RRC) from the base station. In this specification, information / state / parameters being "defined" or "pre-defined" may be interpreted as being known or stored in advance by the base station and the terminal without signaling between the base station and the terminal.

[0056] The technology described in this specification can be used in various wireless communication systems such as CDMA (code division multiple access), FDMA (frequency division multiple access), TDMA (time division multiple access), OFDMA (orthogonal frequency division multiple access), and SC-FDMA (single carrier frequency division multiple access). CDMA can be implemented with wireless technologies such as UTRA (universal terrestrial radio access) or CDMA2000. TDMA can be implemented with wireless technologies such as GSM (global system for mobile communications) / GPRS (general packet radio service) / EDGE (enhanced data rates for GSM evolution). OFDMA can be implemented with wireless technologies such as IEEE (institute of electrical and electronics engineers) 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802-20, E-UTRA (evolved UTRA), LTE (long term evolution), and 5G NR.

[0057] The technology described in this specification can be implemented as 6G wireless technology and applied to various 6G systems. For example, 6G systems may have key factors such as eMBB (enhanced mobile broadband), URLLC (ultra-reliable low latency communications), mMTC (massive machine-type communication), AI (artificial intelligence) integrated communication, tactile internet, high throughput, high network capacity, high energy efficiency, low backhaul and access network congestion, and enhanced data security.

[0058] <Symbols, Abbreviations, Terms>

[0059] - BM: beam management

[0060] - CQI: channel quality indicator

[0061] - CRI: CSI-RS (channel state information - reference signal) resource indicator

[0062] - CSI: channel state information

[0063] - CSI-IM: channel state information - interference measurement

[0064] - CSI-RS: channel state information - reference signal

[0065] - DMRS: demodulation reference signal

[0066] - FDM: frequency division multiplexing

[0067] - FFT: fast Fourier transform

[0068] - IFDMA: interleaved frequency division multiple access

[0069] - IFFT: inverse fast Fourier transform

[0070] - L1-RSRP: Layer 1 reference signal received power

[0071] - L1-RSRQ: Layer 1 reference signal received quality

[0072] - MAC: medium access control

[0073] - NZP: non-zero power

[0074] - OFDM: orthogonal frequency division multiplexing

[0075] - PDCCH: physical downlink control channel

[0076] - PDSCH: physical downlink shared channel

[0077] - PMI: precoding matrix indicator

[0078] - RE: resource element

[0079] - RI: Rank indicator

[0080] - RRC: radio resource control

[0081] - RSSI: received signal strength indicator

[0082] - Rx: Reception

[0083] - QCL: quasi co-location

[0084] - SINR: signal to interference and noise ratio

[0085] - SSB (or SS / PBCH block): synchronization signal block (including primary synchronization signal, secondary synchronization signal and physical broadcast channel)

[0086] - TDM: time division multiplexing

[0087] - TRP: transmission and reception point

[0088] - TRS: tracking reference signal

[0089] - Tx: transmission

[0090] - UE: user equipment

[0091] - ZP: zero power

[0092] FIG. 1 illustrates an exemplary flexible network topology to which some of the examples of the present specification may be applied.

[0093] To compensate for incomplete areas of network coverage, a network topology in which the Split Radio Access Network (RAN) is configured more flexibly and resiliently may be considered. To this end, various nodes such as IAB nodes, relays, and RF repeaters, as exemplified in Fig. 1, may be applied, and NTN may be integrated. For example, an IAB node may correspond to a node that provides wireless backhaul. For example, a relay may refer to any intermediate point, and in the case of a sidelink relay where a terminal functions as a relay, it may collectively refer to a terminal-to-network (U2N) relay and a terminal-to-terminal (U2U) relay. For example, an RF repeater may correspond to a node that performs simple signal amplification and forwarding functions, and in the case of a network-controlled repeater, it may adjust transmit / receive settings based on information provided by the network as well as signal amplification and forwarding. For example, an NTN node may correspond to a satellite or aircraft that provides NTN coverage that is difficult for a terrestrial network to provide. In addition to these examples, various intermediate points can be introduced to improve network topology.

[0094] Referring to FIG. 1, a split RAN can support the division of a base station into one centralized unit (CU) and one or more distributed units (DU). The CU and DU may correspond to logical units. The CU may be further divided into a control plane (CP) portion and one or more user plane (UP) portions. Since a failure in the CU-CP affects not only the CU-UP but also the DU, various intermediate points may be introduced to compensate for this.

[0095] An intermediate point may correspond to a terminal or a base station depending on its relative relationship with other nodes. For example, an IAB node may include a mobile-termination (MT) portion and a DU. The MT can connect the IAB node to a donor node. The DU of the IAB node may serve other terminals or connect to other IAB nodes to provide multi-hop wireless backhaul to terminals. In other words, an IAB node may correspond to a base station in its relative relationship with user-side nodes and to a terminal in its relative relationship with network-side nodes.

[0096] In some examples of this specification, the description of a terminal may apply equally to an intermediate point corresponding to a terminal in relation to a network-side endpoint as well as to a user-side endpoint. Similarly, in some examples of this specification, the description of a base station may apply equally to an intermediate point corresponding to a base station in relation to a user-side endpoint as well as to a network-side endpoint. However, in most cases where there is no additional description of the operation of three or more entities, the communication entities in this specification are briefly described by the term terminal and / or base station (or first node and / or second node), wherein the term terminal and / or base station (or first node and / or second node) is interpreted to include or replace any endpoint or any intermediate point in relation to other nodes.

[0097] That is, for the sake of brevity of description in some examples of this specification, the subject of the operation may be referred to as a base station and / or terminal (or a first node and / or a second node). Additionally, the term base station and / or terminal (or a first node and / or a second node) may be interpreted or substituted as in the following examples: for example, the base station (or the first node) and the terminal (or the second node) may correspond to a first endpoint and a second endpoint, respectively; may correspond to an endpoint and an intermediate point, respectively; may correspond to an intermediate point and an endpoint, respectively; or may correspond to a first intermediate point and a second intermediate point, respectively.

[0098] In this specification, there may be no intermediate points between the base station and the terminal, or there may be one or more. If intermediate points exist, the intermediate points may correspond to IAB nodes, relays, RF repeaters, NTN (non-terrestrial network) nodes, or nodes supporting other functions. The intermediate points may be nodes with a fixed location or nodes with an indefinite location.

[0099] FIG. 2 illustrates a communication system applicable to the present disclosure.

[0100] The communication system (100) of FIG. 2 includes a wireless device (110), a network device (120), and a network (130). Here, the wireless device (110) refers to a device that performs communication using wireless access technology (e.g., LTE, LTE-A, LTE-A pro, NR, 5G, 5G-A, 6G) and may be referred to as a communication / wireless / 5G / 6G device. Although not limited thereto, the wireless device (110) may include a robot (110a), a vehicle (110b-1, 110b-2), an XR (extended reality) device (110c), a hand-held device (110d), a home appliance (110e), an IoT (Internet of Thing) device (110f), and an AI (artificial intelligence) device / server (110g). For example, the vehicle may include a vehicle equipped with wireless communication capabilities, an autonomous vehicle, a vehicle capable of performing inter-vehicle communication, etc. Here, the vehicle (110b-1, 110b-2) may include an unmanned aerial vehicle (UAV) (e.g., a drone). The XR device (110c) includes an augmented reality (AR) / virtual reality (VR) / mixed reality (MR) device and may be implemented in the form of a head-mounted device (HMD), a head-up display (HUD) equipped in a vehicle, a television, a smartphone, a computer, a wearable device, a home appliance, digital signage, a vehicle, a robot, etc. The portable device (110d) may include a smartphone, a smartpad, a wearable device (e.g., a smartwatch, smart glasses), a computer (e.g., a laptop, etc.). The home appliance (110e) may include a TV, a refrigerator, a washing machine, etc. The IoT device (110f) may include a sensor, a smart meter, etc. The wireless device (110) may correspond to a terminal (or first node) or an intermediate point.The network device (120) may correspond to a base station (or a second node) or another intermediate point. For example, the network device (120) may also be implemented as a wireless device (110), and a specific wireless device (120a) may operate as a network device (120) to another wireless device (110).

[0101] Wireless devices (110a to 110f) can be connected to a network (130) through a network device (120). AI technology may be applied to the wireless devices (110a to 110f), and the wireless devices (110a to 110f) can be connected to an AI server (110g) through the network (130). The network (130) can be configured using a 3G network, a 4G (e.g., LTE) network, a 5G (e.g., NR) network, or a 6G network. The wireless devices (110a to 110f) may communicate with each other through the network device (120) / network (130), but may also communicate directly (e.g., sidelink communication) without going through the network device (120) / network (130). For example, vehicles (110b-1, 110b-2) can communicate directly (e.g., V2V (vehicle to vehicle) / V2X (vehicle to everything) communication). Also, an IoT device (110f) (e.g., a sensor) can communicate directly with another IoT device (e.g., a sensor) or other wireless devices (110a to 110f).

[0102] Wireless communication / connection (150a, 150b, 150c) can be established between wireless devices (110a to 110f) / network devices (120) and between network devices (120). Here, wireless communication / connection can be established through various wireless access technologies such as uplink / downlink communication (150a), sidelink communication (150b) (or D2D communication), and communication between network devices (150c) (e.g., relay, IAB (integrated access backhaul)). Through wireless communication / connection (150a, 150b, 150c), wireless devices and network devices / wireless devices, and network devices and network devices can transmit / receive wireless signals to / from each other. For example, wireless communication / connection (150a, 150b, 150c) can transmit / receive signals through various physical channels. To this end, based on the various descriptions of the present disclosure, at least some of the following may be performed: a process for setting various configuration information for transmitting / receiving wireless signals, a process for various signal processing (e.g., channel encoding / decoding, modulation / demodulation, resource mapping / demapping, etc.), a resource allocation process, etc.

[0103] FIG. 3 illustrates an example of a wireless device that can be applied to the present disclosure.

[0104] Referring to FIG. 3, the wireless device (200) can transmit and receive wireless signals through various wireless access technologies (e.g., LTE, LTE-A, LTE-A pro, NR, 5G, 5G-A, 6G). The wireless device (200) includes at least one processor (202) and at least one memory (204), and may additionally include at least one transceiver (206) and / or at least one antenna (208).

[0105] The processor (202) controls the memory (204) and / or the transceiver (206) and may be configured to implement the descriptions, functions, procedures, proposals, methods, and / or sequences of operation disclosed in this document. For example, the processor (202) may process information within the memory (204) to generate a first information / signal and then transmit a wireless signal containing the first information / signal through the transceiver (206). Additionally, the processor (202) may receive a wireless signal containing a second information / signal through the transceiver (206) and then store information obtained from the signal processing of the second information / signal in the memory (204). The memory (204) may be connected to the processor (202) and may store various information related to the operation of the processor (202). For example, memory (204) may store software code containing instructions for performing some or all of the processes controlled by the processor (202) or for performing the descriptions, functions, procedures, proposals, methods, and / or sequences of operations disclosed in this document. Here, the processor (202) and memory (204) may be part of a communication modem / circuit / chip designed to implement wireless communication technology. A transceiver (206) may be connected to the processor (202) and may transmit and / or receive wireless signals through at least one antenna (208). The transceiver (206) may include a transmitter and / or receiver. The transceiver (206) may be interchangeable with a radio frequency (RF) unit. In this disclosure, a wireless device may mean a communication modem / circuit / chip.

[0106] Hereinafter, hardware elements of the wireless device (200) will be described in more detail. Although not limited thereto, at least one protocol layer may be implemented by at least one processor (202). For example, at least one processor (202) may implement at least one layer (e.g., functional layers such as PHY (physical), MAC (media access control), RLC (radio link control), PDCP (packet data convergence protocol), RRC (radio resource control), and SDAP (service data adaptation protocol). At least one processor (202) may generate at least one PDU (Protocol Data Unit) and / or at least one SDU (service data unit) according to the descriptions, functions, procedures, proposals, methods and / or operation sequences disclosed in this document. At least one processor (202) may generate messages, control information, data, or information according to the descriptions, functions, procedures, proposals, methods and / or operation sequences disclosed in this document. At least one processor (202) may generate a signal (e.g., baseband signal) including a PDU, SDU, message, control information, data, or information according to the functions, procedures, proposals, and / or methods disclosed in this document and provide it to at least one transceiver (206). At least one processor (202) may receive a signal (e.g., baseband signal) from at least one transceiver (206) and may obtain a PDU, SDU, message, control information, data, or information according to the descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this document.

[0107] At least one processor (202) may be referred to as a controller, microcontroller, microprocessor, or microcomputer. At least one processor (202) may be implemented by hardware, firmware, software, or a combination thereof. For example, at least one application-specific integrated circuit (ASIC), at least one digital signal processor (DSP), at least one digital signal processing device (DSPD), at least one programmable logic device (PLD), or at least one field programmable gate array (FPGA) may be included in at least one processor (202). The descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this document may be implemented using firmware or software, and the firmware or software may be implemented to include modules, procedures, functions, etc. Firmware or software configured to perform the descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this document may be included in at least one processor (202) or stored in at least one memory (204) and driven by at least one processor (202). The descriptions, functions, procedures, proposals, methods, and / or flowcharts disclosed in this document may be implemented using firmware or software in the form of code, instructions, and / or sets of instructions.

[0108] At least one memory (204) may be connected to at least one processor (202) and may store various forms of data, signals, messages, information, programs, codes, instructions, and / or commands. At least one memory (204) may be composed of ROM (read-only memory), RAM (random access memory), EPROM (erasable programmable read-only memory), flash memory, hard drive, registers, cache memory, computer read storage media, and / or combinations thereof. At least one memory (204) may be located inside and / or outside of at least one processor (202). Additionally, at least one memory (204) may be connected to at least one processor (202) via various technologies, such as wired or wireless connections.

[0109] At least one transceiver (206) may transmit user data, control information, wireless signals / channels, etc., as mentioned in the methods and / or operation flowcharts, etc. of this document to at least one other device. At least one transceiver (206) may receive user data, control information, wireless signals / channels, etc., as mentioned in the descriptions, functions, procedures, proposals, methods and / or operation flowcharts, etc. disclosed in this document from at least one other device. For example, at least one transceiver (206) may be connected to at least one processor (202) and may transmit and receive wireless signals. For example, at least one processor (202) may control at least one transceiver (206) to transmit user data, control information, or wireless signals to at least one other device. Additionally, at least one processor (202) may control at least one transceiver (206) to receive user data, control information, or wireless signals from at least one other device. Additionally, at least one transceiver (206) may be connected to at least one antenna (208), and at least one transceiver (206) may be configured to transmit and receive user data, control information, wireless signals / channels, etc., as described in the descriptions, functions, procedures, proposals, methods, and / or operation sequence diagrams disclosed in this document through at least one antenna (208). In this document, at least one antenna may be a plurality of physical antennas or a plurality of logical antennas (e.g., antenna ports). At least one transceiver (206) may convert the received wireless signals / channels, etc., from RF band signals to baseband signals in order to process the received user data, control information, wireless signals / channels, etc., using at least one processor (202). At least one transceiver (206) may convert the processed user data, control information, wireless signals / channels, etc., from baseband signals to RF band signals using at least one processor (202).To this end, at least one transceiver (206) may include an (analog) oscillator and / or filter.

[0110] The components of the wireless device described with reference to FIG. 3 may be referred to by other terms in terms of their function. For example, the processor (202) may be referred to as the control unit, the transceiver (206) as the communication unit, and the memory (204) as the storage unit. In some cases, the communication unit may be used to mean at least a part of the processor (202) and the transceiver (206).

[0111] The structure of the wireless device described with reference to FIG. 3 can be understood as the structure of at least part of various devices. For example, the structure of the wireless device illustrated in FIG. 3 may be at least part of the various devices described with reference to FIG. 2 (e.g., robot (110a), vehicle (110b-1, 110b-2), XR device (110c), portable device (110d), home appliance (110e), IoT device (110f), AI device / server (110g)). Furthermore, according to various embodiments, the device may include other components in addition to the components illustrated in FIG. 3.

[0112] For example, the device may be a portable device such as a smartphone, smartpad, wearable device (e.g., smart watch, smart glasses), or portable computer (e.g., laptop, etc.). In this case, the device may further include at least one of a power supply unit that supplies power and includes a wired / wireless charging circuit, a battery, etc., an interface unit that includes at least one port for connection with another device (e.g., audio input / output port, video input / output port), and an input / output unit for inputting and outputting video information / signals, audio information / signals, data, and / or information input by a user.

[0113] For example, the device may be a mobile device such as a mobile robot, vehicle, train, manned / unmanned aerial vehicle (AV), or ship. In this case, the device may further include at least one of a drive unit comprising at least one of an engine, motor, power train, wheel, brake, and steering device of the device; a power supply unit that supplies power and includes a wired / wireless charging circuit, battery, etc.; a sensor unit that senses state information, environmental information, and user information of the device or its surroundings; an autonomous driving unit that performs functions such as path maintenance, speed control, and destination setting; and a position measurement unit that acquires position information of the moving body through a GPS (global positioning system) and various sensors.

[0114] For example, the device may be an XR device such as an HMD, a HUD (head-up display) equipped in a vehicle, a television, a smartphone, a computer, a wearable device, a home appliance, digital signage, a vehicle, a robot, etc. In this case, the device may further include at least one of a power supply unit that supplies power and includes a wired / wireless charging circuit, a battery, etc., an input / output unit that acquires control information, data, etc. from the outside and outputs a generated XR object, and a sensor unit that senses state information, environment information, and user information of the device or the surroundings of the device.

[0115] For example, the device may be a robot that can be classified into industrial, medical, household, military, etc., depending on the purpose or field of use. In this case, the device may further include at least one of a sensor unit that senses state information, environmental information, and user information of the device or its surroundings, and a drive unit that performs various physical actions, such as moving robot joints.

[0116] For example, the device may be an AI device such as a TV, projector, smartphone, PC, laptop, digital broadcasting terminal, tablet PC, wearable device, set-top box (STB), radio, washing machine, refrigerator, digital signage, robot, vehicle, etc. In this case, the device may further include at least one of an input unit that acquires various types of data from the outside, an output unit that generates output related to sight, hearing, or touch, a sensor unit that senses state information, environmental information, and user information of the device or its surroundings, and a training unit that learns a model composed of an artificial neural network using training data.

[0117] The structure of the wireless device exemplified in FIG. 3 may be understood as part of a terminal (or first node), or part of an intermediate point, or part of a base station (or second node). If the device exemplified in FIG. 3 is a base station (or second node), the device may further include a wired transceiver for front haul and / or back haul communication. However, if the front haul and / or back haul communication is based on wireless communication, at least one transceiver (206) exemplified in FIG. 3 is used for front haul and / or back haul communication, and the wired transceiver may not be included.

[0118] FIG. 4 illustrates a communication procedure between a first node (e.g., a terminal) and a second node (e.g., a base station) applicable to the present disclosure.

[0119] The second node of FIG. 4 supports dynamic spectrum sharing (DSS) and can provide connectivity to both nodes where 6G technology is implemented and nodes where pre-6G wireless communication technology (e.g., 5G, 4G) is implemented. That is, the first node of FIG. 4 may have 6G technology implemented or pre-6G wireless communication technology (e.g., 5G, 4G) implemented. Additionally, the first node and / or the second node may support full duplex mode as well as non-overlapping full duplex mode.

[0120] In FIG. 4, for the sake of simplicity of explanation, the first node and the second node are assumed to be a terminal and a base station, respectively, and the operation of the terminal (110) and the base station (120) transmitting and / or receiving data, and the operation performed prior to this, are illustrated. However, the operation of FIG. 4 is not limited to the operation between the terminal and the base station, but can be interpreted as the operation between the first node and the second node. Additionally, FIG. 4 illustrates the operation of direct transmission and reception of wireless signals between the terminal (110) and the base station (120), but there may be one or more intermediate points between the terminal (110) and the base station (120), and wireless signals may be transmitted and received via one or more intermediate points.

[0121] Referring to FIG. 4, the terminal (110) and the base station (120) can perform synchronization (401). For example, the terminal (110) performs an initial cell search operation. Specifically, the terminal (110) can detect a synchronization signal for at least one base station connection transmitted from the base station (120) according to a predefined rule. Here, the synchronization signal may include a plurality of synchronization signals classified according to structure or use (e.g., a first synchronization signal (e.g., a primary synchronization signal), a second synchronization signal (e.g., a secondary synchronization signal), etc.). Through this, the terminal (110) can identify the boundary of the unit (e.g., frame, subframe, slot and / or symbol) constituting the wireless signal transmission of the base station (120) and obtain information about the base station (120) (e.g., cell identifier).

[0122] The terminal (110) can obtain system information transmitted from the base station (120) (403). The system information is information related to the attributes, characteristics, and / or capabilities of the base station (120) required to connect to the base station (120) and use the service, and can be classified according to content (e.g., whether it is essential for connection), transmission structure (e.g., channel used, whether it is provided on-demand), etc., and can be classified, for example, into first system information (e.g., MIB (master information block), primary system information), second system information (e.g., SIB (system information block), secondary system information), etc. If necessary, the terminal (110) may transmit a signal requesting system information prior to receiving the system information. However, the request and provision of system information may be performed after the random access procedure described later.

[0123] A terminal (110) and a base station (120) can perform a random access procedure (405). The terminal (110) can transmit and / or receive at least one message for a random access procedure (e.g., random access preamble, RAR (random access response) message, etc.) based on information related to the channel for the random access procedure of the base station (120) obtained through system information (e.g., channel location, channel structure, structure of supported preamble, etc.). For example, the terminal (110) can transmit a first message (e.g., preamble, MSG1) through the channel for the random access procedure, receive a second message (e.g., RAR message, MSG2), transmit a third message (e.g., MSG3) containing information related to the terminal (110) (e.g., identification information) to the base station (120) using scheduling information included in the second message, and receive a fourth message (e.g., MSG4) for contention resolution and / or connection establishment. As another example, the first message and the third message can be transmitted and received as a single message, or the second message and the fourth message can be transmitted and received as a single message.

[0124] The terminal (110) and the base station (120) can perform signaling of control information (407). Here, the control information can be defined in various layers, such as a layer that controls the connection (e.g., a radio resource control (RRC) layer), a layer that handles mapping between logical channels and transmission channels (e.g., a media access control (MAC) layer), and a layer that handles physical channels (e.g., a physical (PHY) layer). For example, the terminal (110) and the base station (120) can perform at least one of signaling to establish a connection, signaling to determine settings related to communication, and signaling to indicate allocated resources.

[0125] The terminal (110) and the base station (120) can transmit and / or receive data (409). In other words, the terminal (110) and the base station (120) can process data based on the signaling of control information and transmit and / or receive data. For example, when transmitting data, the terminal (110) or the base station (120) can perform at least one of channel encoding, rate matching, scrambling, constellation mapping, layer mapping, waveform modulation, antenna mapping, and resource mapping on the information bits. Conversely, when receiving data, the terminal (110) or the base station (120) can perform at least one of signal extraction from resources, antenna-specific waveform demodulation, signal placement considering layer mapping, constellation demapping, descrambling, and channel decoding.

[0126] 6G System Core Technology

[0127] The 6G (wireless communication) system aims for (i) very high data rates per device, (ii) a very large number of connected devices, (iii) global connectivity, (iv) very low latency, (v) reduced energy consumption of battery-free IoT (internet of things) devices, (vi) ultra-reliable connectivity, and (vii) connected intelligence with machine learning capabilities. The vision of the 6G system can be four aspects: intelligent connectivity, deep connectivity, holographic connectivity, and ubiquitous connectivity.

[0128] As core implementation technologies for 6G systems, technologies such as artificial intelligence (AI), THz (Terahertz) communication, optical wireless technology, FSO backhaul network, massive MIMO technology, blockchain, 3D networking, quantum communication, unmanned aerial vehicles, cell-free communication, wireless information and energy transfer (WIET), integration of sensing and communication, integration of access backhaul networks, holographic beamforming, big data analysis, and large intelligent surface (LIS) can be adopted.

[0129] artificial intelligence

[0130] The introduction of AI into communications can streamline and enhance real-time data transmission. AI can determine how complex target tasks are performed using numerous analyses. In other words, AI can increase efficiency and reduce processing latency. Time-consuming tasks such as handover, network selection, and resource scheduling can be performed instantly using AI. AI can also play a significant role in M2M, machine-to-human, and human-to-machine communication. Furthermore, AI can enable rapid communication in Brain-Computer Interfaces (BCI). AI-based communication systems can be supported by metamaterials, intelligent structures, intelligent networks, intelligent devices, intelligent cognitive radios, self-sustaining wireless networks, and machine learning.

[0131] The following describes a functional framework for AI / ML operations.

[0132] Below, to provide a more specific explanation of AI (or AI / ML), terms may be defined as follows.

[0133] - Data collection: Data collected from network nodes, management entities, or terminals, serving as a basis for AI model training, data analysis, and inference.

[0134] - AI Model: A data-driven algorithm that applies AI technology to generate a set of outputs containing predictive information and / or decision parameters based on a set of inputs.

[0135] - AI / ML Training: An online or offline process of training an AI model by learning features and patterns that best represent data and acquire an AI / ML model trained for inference.

[0136] - AI / ML Inference: A process of making predictions or deriving decisions based on collected data and AI models using trained AI models.

[0137] Life Cycle Management (LCM) procedures for AI / ML models (i.e., model training, model deployment, model inference, model monitoring, model updating, etc.) can be classified into functionality-based LCM and model-based LCM. In functionality-based LCM, AI / ML models may not be identifiable within the network, and the network can direct the activation, deactivation, fallback, or switching of AI / ML functionality. In model-ID (identifier)-based LCM, AI / ML models can be identified within the network, and the network or terminal can activate, deactivate, select, or switch AI / ML models via the model ID.

[0138] Figure 5 illustrates a general functional architecture for an AI / ML model.

[0139] In particular, Figure 5 illustrates a general functional architecture related to both Functionality-based LCM and Model-based LCM. Some functions or some data / information / command flows (i.e., arrows) illustrated in Figure 5 may be omitted.

[0140] Referring to FIG. 5, a general functional framework may be configured to include a data collection function (10), a model training function (20), a management function (30), an inference function (40), and a model storage function (50).

[0141] The Data Collection function (10) is a function that provides input data to the Model Training function (20), Management function (30), and Inference function (40). The Data Collection function (10) performs data preparation based on raw data and can provide input data processed through data preparation. Examples of raw data may include received data / measurement data from terminals or other network entities, inference / output of AI / ML models, etc. The Data Collection function (10) may be performed by a single entity (e.g., terminal, network node, etc.) but may also be performed by multiple entities.

[0142] Here, training data (11) refers to data required as input for the AI / ML model training function (20). monitoring data (12) refers to data required as input for the management (30) of the AI / ML model or AI / ML function. inference data (13) refers to data required as input for the AI / ML inference function (30).

[0143] The Model Training function (20) is a function that performs AI / ML model training, validation, and testing, which can generate model performance metrics that can be used as part of the AI / ML model testing procedure. If necessary, the Model Training function (20) can perform data preparation (e.g., data pre-processing and cleaning, forming and transformation) based on the Training Data (11) delivered from the Data Collection function (10).

[0144] Trained / Updated Model (21): If there is a Model Storage function (50), it is used to transfer trained, validated, and tested AI / ML models to the Model Storage function (50) or to transfer updated versions of the models to the Model Storage function (50).

[0145] The Management function (30) is a function that supervises the operation of an AI / ML model or an AI / ML function. Additionally, the Management function (30) may make decisions to ensure appropriate inference operations based on data received from the Data Collection function (10) (i.e., Monitoring Data (12)) and / or data received from the Inference function (40) (i.e., Inference Output (41)).

[0146] Management Instruction (32) is information required as input to manage the Inference function (40). The relevant information may include the selection / (de)activation / switching of an AI / ML model or an AI / ML-based function, and may also include a fallback to a non-AI / ML operation (i.e., not relying on the inference process).

[0147] A Model Transfer / Delivery Request (33) can be used to request model(s) from Model Storage (50).

[0148] Performance Feedback / Retraining Request (31) refers to information required as input to Model Training function (20) (e.g., for the purpose of retraining or updating the model).

[0149] The inference function (40) is a function that provides output from the process of applying an AI / ML model or AI / ML function using data (i.e., inference data (13)) provided by the data collection (10) as input. Data preparation (e.g., data preprocessing and cleaning, formatting and transformation) may also be performed based on the inference data (13) delivered by the data collection (10). If necessary, the inference function (40) may also perform data preparation (e.g., data pre-processing and cleaning, forming and transformation) based on the inference data (13) provided by the data collection function (10).

[0150] Inference Output (41) is data used in the Management function (30) to monitor the performance of an AI / ML model or AI / ML function. Inference Output (41) may include the inference output of an AI / ML model generated by the Inference function (30), and the details of the inference output may vary depending on the use case.

[0151] The Model Storage function (50) is a function that stores a trained / updated model that can be used to perform the Inference function (40). The Model Storage function (50) exemplified in FIG. 5 can be used as a reference point (if any) applicable to protocol termination, model transmission / delivery, and related processes. Additionally, the Model Storage function (50) is an example and is not intended to restrict the storage location of the actual AI / ML model, and may be omitted.

[0152] Model Transfer / Delivery (51) is used to transfer an AI / ML model to an inference function.

[0153] Cooperation levels can be defined as follows based on the capability of AI / ML functions among multiple nodes, and variations resulting from the combination of multiple levels or the separation of any one level are also possible.

[0154] Cat 0a) No collaboration framework: AI / ML algorithms are based on pure implementation and do not require changes to the wireless interface.

[0155] Cat 0b) This level corresponds to a framework that involves a wireless interface modified to fit efficient implementation-based AI / ML algorithms but without cooperation.

[0156] Cat 1) Inter-node support is involved to improve the AI / ML algorithms of each node. For example, this applies when a specific node receives support from other nodes (for training, adaptation, etc.) and vice versa. At this level, model exchange between network nodes is not required.

[0157] Cat 2) Collaborative AI / ML tasks can be performed among multiple nodes. This level requires the exchange of AI / ML model commands or network nodes.

[0158] FIG. 5 is a diagram illustrating an overall functional framework for an AI / ML model, and all functions and / or all data / information / command signals illustrated in FIG. 5 may not be performed within a specific node, and only some may be performed.

[0159] AI / ML models can be classified into one-side models and two-side models depending on whether training and / or inference are performed on a single node or jointly / sequentially on multiple nodes.

[0160] A one-side model can refer to an AI / ML model where inference is performed entirely by a single node (e.g., a terminal or a network). Here, the training of the AI / ML model can also be performed entirely by a single node. The training and inference of the AI / ML model may be performed by the same node, or they may be performed by different nodes.

[0161] A two-side model can refer to an AI / ML model in which joint inference is performed across multiple nodes (e.g., terminals and networks). Joint inference means that inference is performed collaboratively across multiple nodes; for example, the first part of the inference may be performed by the first node, and the remainder by the second node. Two-side models can be classified into various types as follows, depending on the training method of the AI / ML model.

[0162] - First type: An AI / ML model can be trained on a single node. In this case, joint training can be performed. The trained model can then be distributed to other nodes / entities.

[0163] - Second type: Joint training of AI / ML models can be performed on multiple nodes / entities (e.g., networks and terminals). Joint training can mean that model generation (e.g., CSI generation) and model reconstruction (CSI compression by sub-use cases) are trained in the same loop for forward activation and backward gradient. In this type, joint training can include both simultaneous training (i.e., model generation training and model reconstruction training are performed simultaneously) and sequential training (i.e., model reconstruction training is performed after model generation training).

[0164] - Third Type: Separate training of AI / ML models can be performed at multiple nodes (e.g., networks and terminals). Separate training may mean that training starts sequentially at one node and continues at another node. In this case, if the first node performs the AI / ML model first and shares the training data with the second node, the second node can perform the AI / ML model using the shared training data. For example, training for the CSI generation part may be performed by the terminal, while CSI reconstruction may be performed by the network.

[0165] FIG. 6 illustrates a communication procedure between a first node (e.g., terminal) and a second node (e.g., base station) to which an AI / ML model is applied.

[0166] The operations described below may be explained / interpreted based on an AI / ML model as shown in FIG. 6 below, even without separate mention (i.e., without explicit mention of being by / based on / for an AI / ML model). Furthermore, unless specifically limited, the AI / ML model may correspond to a one-side model in which inference is performed entirely by a single node or a two-side model in which joint inference is performed by multiple nodes.

[0167] First signaling (601): In the following description, the signaling (e.g., information / data / channel / signal, etc.) or set of signaling between a specific node (e.g., terminal, network, etc.) and another node may be interpreted as the signaling or set of signaling of the first signaling (601) used to perform an operation based on an AI / ML model, even if not otherwise mentioned. For example, it may correspond to training data for training (i.e., creation and / or reconstruction) of the AI / ML model of FIG. 5, or to inference data used for inference of the AI / ML model, or to feedback for the AI / ML model. If, in this specification, signaling between nodes is not required prior to an operation based on an AI / ML model, the first signaling (601) may be omitted. In this specification, if a one-side model is used, the unidirectional / bidirectional signaling (set) in this specification may correspond to the signaling of the first signaling (601). Additionally, when a two-side model is used in the present specification, unidirectional / bidirectional signaling in the present specification may correspond to the first signaling (601), and repetitive signaling operation may also correspond to the first signaling (601).

[0168] For example, in AI / ML model-based beam management, when a base station predicts (i.e., infers) high-quality beam(s) based on an AI / ML model, the base station can receive quality / intensity information for multiple beams from the terminal. Additionally, when a terminal predicts (i.e., infers) high-quality beam(s) based on an AI / ML model, the terminal can receive multiple beams from the base station.

[0169] AI / ML model-based operation (602): In the following description, an operation (e.g., computation, selection, prediction, etc.) at a specific node (e.g., terminal, network, etc.) or a common operation (e.g., computation, selection, prediction, etc.) at multiple nodes (e.g., terminal, network, etc.) may correspond to an AI / ML model-based operation (602) based on one or more functions in the functional framework of the AI / ML model, even without separate mention. For example, it may correspond to the training (i.e., creation and / or reconstruction) of the AI / ML model of FIG. 5 or to the inference of the AI / ML model. When a one-side model is used, an operation performed by a single node in this specification may correspond to an AI / ML model-based operation (602), and when a two-side model is used, a common operation performed by multiple nodes in this specification may correspond to an AI / ML model-based operation (602).

[0170] For example, in an AI / ML model-based BM, a base station can predict (i.e., infer) high-quality beam(s) based on an AI / ML model by using quality / intensity information for multiple beams received from a terminal as inference data. Additionally, a terminal can measure multiple beams received from a base station and predict (i.e., infer) high-quality beam(s) based on an AI / ML model by using the measurement results as inference data.

[0171] Second signaling (603): In the following description, signaling (e.g., information / data / channel / signal, etc.) or a set of signaling between a specific node (e.g., terminal, network, etc.) and another node may be interpreted as the second signaling (603) or a set of signaling generated as a result of an operation based on an AI / ML model, even without separate mention. For example, it may correspond to the output resulting from the inference of the AI / ML model of FIG. 5. If signaling between nodes is not required as a result of an operation based on an AI / ML model in this specification, the second signaling (603) may be omitted. If a one-side model is used in this specification, the unidirectional / bidirectional signaling (set) in this specification may correspond to the second signaling (603). Additionally, when a two-side model is used in this specification, unidirectional / bidirectional signaling in this specification may correspond to the second signaling (603), and repetitive signaling operation may also correspond to the second signaling (603).

[0172] For example, in an AI / ML model-based BM, the base station may transmit beam(s) predicted based on the AI / ML model as candidates to the terminal so that the terminal can determine the optimal beam. Additionally, the terminal may report the beam(s) predicted based on the AI / ML model to the base station to request the base station to transmit candidate beams as candidates for determining the optimal beam.

[0173] THz communication

[0174] Data transmission rates can be increased by expanding bandwidth. This can be achieved by using sub-THz communication with wide bandwidth and applying advanced large-scale MIMO technology. THz waves, also known as sub-millimeter radiation, generally refer to a frequency band between 0.1 THz and 10 THz with corresponding wavelengths ranging from 0.03 mm to 3 mm. The 100 GHz–300 GHz band range (Sub-THz band) is considered the primary portion of the THz band for cellular communication. Adding the Sub-THz band to the mmWave band increases 6G cellular communication capacity. Among the defined THz bands, the 300 GHz–3 THz band is located in the far-infrared (IR) frequency band. Although the 300 GHz–3 THz band is part of the broadband, it lies at the boundary of the broadband and immediately following the RF band. Therefore, this 300 GHz–3 THz band exhibits similarities to RF.

[0175] FIG. 7 illustrates an electromagnetic spectrum according to one embodiment of the present disclosure. The embodiment of FIG. 7 may be combined with various embodiments of the present disclosure. Key characteristics of THz communication include (i) a widely available bandwidth to support very high data transmission rates, and (ii) high path loss occurring at high frequencies (highly directional antennas are indispensable). The narrow beam width generated by highly directional antennas reduces interference. The small wavelength of THz signals allows a much larger number of antenna elements to be integrated into devices and BSs operating in this band. This enables the use of advanced adaptive array techniques that can overcome range limitations.

[0176] Transmitting system information (i.e., information related to the attributes, characteristics, and / or capabilities of the BS required to use the service, etc.) (e.g., MIB, SIB, etc.) in the THz frequency band can be inefficient because, in the case of high frequency bands, beam sweeping must be performed more frequently to cover the entire area of ​​the cell as the beam width becomes narrow. In particular, transmitting system information in this manner is even more inefficient when there are not many users in the cell. Accordingly, a system information transmission procedure as shown in FIG. 8 below may be used.

[0177] FIG. 8 illustrates an example of a procedure for transmitting system information for THz communication to which the present disclosure applies. Although this example is written with THz conditions in mind, it is also applicable to 6G communication environments where THz is not applied. Furthermore, the procedure exemplified in FIG. 8 can be combined with various embodiments of the present disclosure described below. For example, the embodiments described below may be performed based on the system information obtained by the procedure exemplified in FIG. 8.

[0178] Referring to FIG. 8, the base station can transmit system information of cell #1 through cell #2 (801). That is, the base station provides at least two cells, cell #1 uses the THz frequency band, and cell #2 uses a frequency band other than the THz frequency band. Here, the system information may include at least one information / state / parameter / setting generated at the higher layer and the physical layer, respectively. For example, at least one information / state / parameter / setting generated at the higher layer may include at least one of SFN, control information setting for SIB1 (e.g., PDCCH configuration for SIB1, etc.), information related to cell selection / entry (e.g., cell barring, cell re-selection, etc.), and subcarrier spacing, and at least one information / state / parameter / setting generated at the physical layer may include at least one of SFN, half frame indicator, and SSB index. However, this is merely an example, and system information may include information, status, parameters, and settings related to Cell #1 / Cell #2 generated at various types of physical layers / upper layers. To this end, as an example, Cell #1 and Cell #2 may have a secondary cell and primary cell relationship.

[0179] The UE can obtain synchronization for cell #1 (803). Synchronization can be obtained by detecting a synchronization signal. Generally, synchronization is obtained prior to receiving system information, but since the system information for cell #1 is received in cell #2, synchronization for cell #1 can be obtained after receiving system information. For example, the UE can obtain synchronization based on system information. However, unlike FIG. 8, synchronization may be obtained before step 801 according to other examples.

[0180] The UE can transmit a signal to connect to Cell #1 (805). For example, the signal may include information for connecting to Cell #1 (e.g., a random access preamble). The structure of the signal and the resources for transmitting the signal (e.g., a channel) can be identified through system information. Subsequently, the UE and the base station can perform a connection procedure to Cell #1 and perform communication (807). In this process, operations according to various embodiments described below may be performed.

[0181] The procedure described with reference to FIG. 8 may be performed when the UE (801) first connects to cell #1 of the base station. Alternatively, a similar procedure may be performed when the UE (801) handovers to cell #1 of the base station. However, in the case of a handover, the system information of cell #1 may be received from a cell of a different base station rather than cell #2 of the base station.

[0182] Communication in the THz band is expected to experience severe path loss, and to overcome this, terminals and base stations must use very sharp beams. The use of sharp beams means that terminals and base stations must perform beam control along with beamforming, and the number of beams used becomes very large. Therefore, it takes a very long time to align the transmit and receive beams between the base station and the terminal. In addition, if the beam alignment between the base station and the terminal is misaligned due to the movement of the terminal, time is frequently required to realign the beams, which may result in an unstable link. Accordingly, a beam management procedure as shown in Fig. 9 below may be used.

[0183] FIG. 9 illustrates a beam management procedure applicable to the present disclosure. FIG. 9 illustrates an example of a procedure for searching and / or selecting beams for THz communication, but is not limited to a THz environment and is applicable to a 6G communication environment. Additionally, the procedure exemplified in FIG. 9 may be combined with various embodiments of the present disclosure described below. Here, a beam may be interpreted as 'spatial (configuration) information', 'spatial domain filter', 'spatial domain transmit filter', 'spatial domain receive filter', or / and a term having an equivalent technical meaning capable of distinguishing a beam (e.g., Reference signal, SSB (Synchronization Signal Block) Index, TRP (transmission reception point), panel, cell, TP (transmission point), base station, control resource-related information (e.g., CORESET (control resource set)-related information, etc.).

[0184] Referring to FIG. 9, the base station can configure resources for beam management (901). Here, the resources may include at least one of time-frequency resources, channels, and spatial resources (e.g., antenna ports). For example, the base station may utilize a beam search signal (BSS) that is spatially separated from existing downlink signals / channels for beam search. Here, the BSS may be transmitted based on a dedicated port for beam search. The dedicated port may be a port different from the port for transmitting existing downlink signals / channels (e.g., synchronization signals (e.g., SSB, etc.), data channels (e.g., PDSCH, etc.)). BSS is a term defined for convenience of explanation, and the technical concept according to the present embodiment is not limited to the term BSS itself. That is, a signal transmitted based on a dedicated port defined / configured for beam search may be included in the technical concept according to the present embodiment.

[0185] The base station can transmit measurement signals using multiple transmission beams (903). For example, the measurement signals may include at least one of a reference signal and a synchronization signal. At this time, the measurement signals may be transmitted as many times as the number of beams required for measurement, and may be transmitted using a multi-beam transmission method that forms multiple beams simultaneously to reduce sweeping time. Here, multi-beam transmission may be performed based on at least one of a multi-panel, a sub-array, and a true time delay (TTD).

[0186] The UE can transmit a feedback signal to the base station (905). The feedback signal indicates at least one beam selected by the UE. The UE can select at least one preferred beam based on the received measurement signals. The UE and the base station can perform communication (907). At this time, the UE and the base station can perform communication using the previously selected beam. If channel reciprocity is established, the UE's transmission beam can also be determined through operations 903 and 905, so the UE's transmission can also be performed using the beam selected in operation 905. If channel reciprocity is not established, a procedure including the transmission of the UE's measurement signals and the transmission of the base station's feedback signal may be performed first to determine the UE's transmission beam. In operation 907, operations according to various embodiments described below may be performed.

[0187] Integrated Sensing and Communication (ISAC)

[0188] Wireless sensing is a technology that utilizes radio frequencies to determine the instantaneous linear velocity, angle, and distance (range) of an object, thereby obtaining information about the characteristics of the environment and / or objects within that environment. Since radio frequency sensing capabilities do not require connecting to objects via devices within a network, they can provide services for determining object locations without the need for devices. The ability to obtain range, velocity, and angle information from radio frequency signals can provide a wide range of new functions, such as various object detection, object recognition (e.g., vehicles, humans, animals, UAVs), and high-precision localization, tracking, and activity recognition. Wireless sensing services can provide information to various industries (e.g., unmanned aerial vehicles, smart homes, V2X, factories, railways, public safety, etc.) that enable applications such as intruder detection, assisted vehicle steering and navigation, trajectory tracking, collision avoidance, traffic management, and health and traffic management. In some cases, wireless sensing may utilize non-3GPP type sensors (e.g., radar, cameras) to further support 3GPP-based sensing. For example, the operation of a wireless sensing service, that is, the sensing operation, may depend on the transmission, reflection, and scattering processing of wireless sensing signals. Therefore, wireless sensing can provide an opportunity to enhance existing communication systems from communication networks into wireless communication and sensing networks.

[0189] FIG. 10 illustrates an example of a sensing operation according to an embodiment of the present disclosure. The embodiment of FIG. 10 may be combined with various embodiments of the present disclosure. Specifically, FIG. 10(a) illustrates an example of sensing using a sensing receiver and a sensing transmitter located at the same position (e.g., monostatic sensing), and FIG. 10(b) illustrates an example of sensing using a separated sensing receiver and a sensing transmitter (e.g., bistatic sensing).

[0190] For example, in a wireless communication system based on a 6G network of the present specification, referring to FIG. 10(a), the sensing transmitter and the sensing receiver may be configured to be included in a single base station (i.e., the same base station) or a single terminal (i.e., the same terminal). Alternatively, referring to FIG. 10(b), the sensing transmitter and the sensing receiver may be configured to be included in different base stations, in different terminals, or in a terminal and a base station, respectively.

[0191] In this regard, based on whether the sensing transmitter and the sensing receiver are each included in a base station or a terminal, the following six types of sensing modes can be defined.

[0192] - Mode 1: A mode in which the sensing transmitter and sensing receiver are included in a single base station (e.g., base station-based sensing mode in monostatic mode)

[0193] - Second mode: A mode in which the sensing transmitter is included in the first base station and the sensing receiver is included in a second base station different from the first base station (e.g., base station-based sensing mode in bistatic mode)

[0194] - 3rd Mode: A mode in which the sensing transmitter is included in the base station and the sensing receiver is included in the terminal (e.g., base station-terminal sensing mode)

[0195] - 4th Mode: A mode in which the sensing transmitter is included in the terminal and the sensing receiver is included in the base station (e.g., terminal-base station sensing mode)

[0196] - 5th Mode: A mode in which the sensing transmitter and the sensing receiver are contained in a single terminal (e.g., terminal-based sensing mode in monostatic mode)

[0197] - 6th mode: A mode in which the sensing transmitter is included in the first terminal and the sensing receiver is included in a second terminal different from the first terminal (e.g., terminal-based sensing mode in bistatic mode)

[0198] In a wireless communication system based on a 6G network of the present specification, one or more of the six types of sensing modes described above may be utilized independently or in combination.

[0199] In relation to the sensing operation in FIG. 10, the sensing transmitter may transmit a sensing signal for sensing one or more objects (and / or the environment surrounding the objects). For example, the sensing signal may correspond to a radio (frequency) signal defined to be transmittable by a base station / terminal in a wireless communication system based on a 6G network of the present specification. The sensing receiver may receive a signal that is scattered / reflected by one or more objects (and / or the environment surrounding the objects) from the sensing signal transmitted from the sensing transmitter. In the sensing receiver, sensing data may be derived from the scattered / reflected signal, and sensing results may be generated / obtained through processing of the sensing data. Here, the sensing result may include characteristic information (e.g., location, distance, speed, angle, etc.) about one or more objects (and / or the environment surrounding the objects). The sensing result thus generated / acquired may be utilized for wireless sensing services (e.g., detection, tracking, etc. of objects and / or environments) provided by a wireless communication system based on a 6G network of the present specification, or may be provided / disclosed to a trusted third party.

[0200] Additionally, the sensing operation in FIG. 10 is described using a representative example of operation in a wireless communication system based on a 6G network, but it can be extended and applied to cases where terminals / base stations / signals based on previous generations (e.g., 4G, 5G, etc.) networks are utilized.

[0201] Additionally, with respect to the wireless sensing described in this specification, in a wireless communication system based on a 6G network of this specification, time / frequency resources for sensing operations and time / frequency resources for general communication (e.g., UL / DL / sidelink-based communication, etc.) may be scheduled / configured separately.

[0202] FIG. 11 illustrates a time / frequency resource for a sensing operation according to one embodiment of the present specification. The embodiment of FIG. 11 may be combined with various embodiments of the present disclosure.

[0203] Referring to FIG. 11, the time / frequency resources (hereinafter, sensing resources) for the aforementioned sensing operation (e.g., sensing operation based on FIG. 10) can be set / assigned separately from the time / frequency resources (hereinafter, communication resources) for general communication.

[0204] For example, as illustrated in FIG. 11, sensing resources may be configured / assigned in units of symbols in the time domain and / or in units of resource blocks in the frequency domain. Resources other than those configured / assigned to the sensing resources may be utilized as resources for general communication. That is, sensing resources and communication resources may be configured / assigned based on time-division multiplexing (TDM) and / or frequency-division multiplexing (FDM) methods in terms of base station / terminal operation. Additionally or alternatively, unlike that illustrated in FIG. 10, sensing resources may be configured / assigned based on other units in the time domain (e.g., slot, frame, absolute time (ms, us), etc.) and / or other units in the frequency domain (e.g., subcarrier, carrier, absolute frequency (MHz, GHz), etc.).

[0205] Additionally or alternatively, in relation to the setup / allocation / scheduling of resources for general communication described herein, it may be necessary to consider the relationship between said resources and the aforementioned sensing resources. For example, when setting / allocating resources for general communication according to the embodiments of the present disclosure, said resources may be set / allocated to rate-match or puncturing resource areas corresponding to the sensing resources. For example, when scheduling resources for general communication according to the embodiments of the present disclosure, said resources may be scheduled so as not to overlap with resource areas corresponding to the sensing resources. If resources for general communication and resource areas corresponding to the sensing resources are set / allocated / scheduled to overlap according to the embodiments of the present disclosure, either one or both operations may be dropped, skipped, or postponed based on priority, predefined rules, etc. That is, in the embodiments of this specification, resources related to general communication (e.g., resources for signals / channels related to UL / DL / Sidelink-based data / control, etc.) may be configured / assigned / scheduled so as not to overlap with the aforementioned sensing resources.

[0206] Additionally, various channel modeling methods may be applied in relation to the wireless sensing described herein. Channel modeling related to sensing may mean constructing a path for transmitting and receiving sensing signals and / or scattered / reflected signals by considering the object to be sensed and / or the environment to which the object belongs. Since channel modeling may be related to the performance / requirements of sensing in a wireless communication system, it may be an important matter for verifying the validity of the sensing function.

[0207] Channels related to sensing can be classified into channels between an object (e.g., target of interest) and a sensing transmitter / receiver, and channels between the environment to which the object belongs and a sensing transmitter / receiver. In this regard, channel modeling related to sensing can be classified based on the sensing mode (e.g., the six types of modes mentioned above), whether it is an object or an environment, and / or sensing scenarios. For example, channel modeling for a target in a base station / terminal-based monostatic sensing mode, channel modeling for a target in a base station / terminal-based bistatic sensing mode, channel modeling for an environment in a base station / terminal-based monostatic sensing mode, and channel modeling for an environment in a base station / terminal-based bistatic sensing mode can be optimized and configured differently. For example, when various sensing scenarios are classified, they can be divided into channel modeling for detection, location, and tracking scenarios, channel modeling for motion recognition, and channel modeling for imaging / environment reconstruction scenarios. Additionally, channel modeling related to sensing may be based on statistical channel modeling techniques and / or deterministic channel modeling techniques. For example, modeling for sensing in a wireless communication system based on a 6G network of this specification may be based on stochastic geometry channel modeling techniques and / or hybrid with ray tracing channel modeling techniques. Here, the stochastic geometry channel model may be based on various statistical characteristics of the channel state. Furthermore, the hybrid channel model may be based on both ray tracing techniques and stochastic techniques.In the case of a hybrid approach, channels for objects requiring high accuracy and consistency (e.g., targets of interest) can be modeled using ray tracing techniques, while channels for the environment can be modeled using probabilistic techniques.

[0208] FIG. 12 illustrates a procedure related to a sensing operation according to one embodiment of the present specification. The embodiment of FIG. 12 may be combined with various embodiments of the present disclosure.

[0209] For example, in a wireless communication system based on a 6G network of the present specification, in the case of a sensing operation in which a terminal participates, the base station may need to verify (1205) the terminal's capability for the sensing operation. In this regard, the terminal may be configured to report capability information to the base station regarding whether it supports the sensing operation. Additionally, or alternatively, if the terminal is defined in advance in the specification as supporting the sensing operation, the procedure may be omitted. Furthermore, in the case of a sensing operation in which only the base station participates, the base station may be configured to report capability information regarding whether it supports the sensing operation to the entity setting / controlling its sensing operation (e.g., a network entity at the upper level / layer of the base station).

[0210] For example, a base station may perform signaling with a terminal to exchange configuration information related to a sensing operation. For example, the base station may set / instruct the terminal information regarding the mode of the sensing operation (e.g., based on the six types of modes mentioned above), the subject of the sensing operation (e.g., a sensing transmitter, a sensing receiver), the resource of the sensing operation (e.g., a sensing resource as shown in FIG. 11), the target of utilization of the sensing result (e.g., a type of wireless sensing service based on a 6G network, a trusted third party), and channel modeling for sensing (e.g., a channel between the base station / terminal and an object / environment) (1210). For example, the base station may receive such information from a network entity at the upper level / layer of the base station.

[0211] For example, a base station and / or terminal may perform a sensing operation on information set / instructed (1215). For example, the base station and / or terminal may perform procedures such as transmitting a sensing signal as in FIG. 9 described above, receiving scattered / reflected signals, deriving sensing data, obtaining a sensing result through processing the sensing data, and providing the sensing result, as a role of a sensing transmitter and / or sensing receiver. For example, in the operation of the base station / terminal described in this specification, the sensing result provided through the sensing operation may be utilized.

[0212] Inter-vendor operability enhancement for two sided AI / ML models

[0213] Recently, AI / ML-based CSI feedback has been studied in NR standardization, and detailed use cases for CSI feedback include CSI compression and CSI prediction to reduce CSI overhead.

[0214] CSI compression is based on a two-sided model structure consisting of a UE-side model acting as a CSI encoder and a NW-side model acting as a CSI decoder. Due to proprietary / privacy issues regarding AI / ML models, UE / NW vendors prefer not to expose their respective AI / ML model information to other vendors; consequently, it is not expected that the structure of the two-sided model or its configuration parameters will be standardized or mutually shared between UE / NW.

[0215] If AI / ML models are implemented / trained on a per-vendor basis, the inter-vendor 2-side AI / ML models may not function correctly or may exhibit performance below the desired level. For example, a problem may arise where the output of the first vendor's CSI encoder is not correctly reconstructed by the second vendor's CSI decoder.

[0216] In this regard, as a method for inter-vendor collaboration in CSI compression based on a two-sided AI / ML model, we propose introducing a Bridge AI / ML model, which is an additional AI / ML model located between the UE side model operating as a CSI encoder and the NW side model operating as a CSI decoder.

[0217] In this specification, a model may mean a set of one or more AI / ML layers having learnable parameters.

[0218] The following considers AI / ML-based CSI reporting.

[0219] Figure 13 illustrates a CSI Encoder for CSI compression and a CSI Decoder for CSI restoration.

[0220] Referring to FIG. 13, the terminal is equipped with an AI-based CSI encoder and the base station is equipped with an AI-based CSI decoder, and the terminal and the base station each perform AI / ML model inference. A model in which AI / ML model inference is performed at two nodes in this manner is called a two-sided model. A configuration like that shown in FIG. 13 can be intended to reduce overhead through CSI compression and / or improve system performance.

[0221] In other words, two-sided AI / ML refers to a system where AI / ML models are deployed or configured on both the terminal and the base station (or network), respectively, to perform inference. For example, an AI / ML model in the form of an auto-encoder can be configured. On the terminal side (CSI encoder), the AI / ML model inference output can be calculated using channel information (e.g., channel matrix, channel covariance matrix, channel eigenvector) as input, or using information that has undergone pre-processing. The terminal can feed this output information back to the base station, and depending on the implementation, post-processing may or may not be performed.

[0222] Additionally, depending on the implementation, the base station (CSI decoder) may or may not pre-process the received feedback information. The base station inputs this into an AI / ML model and calculates an inference output. Depending on the implementation, the base station may or may not post-process the inference output, and through this series of processes, the final CSI can be decoded.

[0223] Theoretically, it is assumed that the models of a two-sided AI / ML model are generally trained under the same conditions. These identical training conditions include the training process being conducted using the same training data, model parameters being updated based on the same gradient update policy, and the interface between the terminal-side and base station-side models (such as the size, dimension, and attribute of intermediate features or edges) being consistent. Taking the two-sided AI / ML model of Fig. 13 as an example, for the terminal-side CSI encoder model and the base station-side CSI decoder model to be properly trained for CSI feedback between the terminal and the base station, the output features of the forward CSI decoder model are obtained based on the output features from the CSI encoder model and the cost function of the final stage is calculated; conversely, the CSI encoder model parameters must be updated based on the derivative at the CSI decoder input calculated by the reverse chain rule.

[0224] However, these conditions may be difficult to guarantee due to privacy and proprietary issues of the models. For example, while the coupled training and updating of the CSI decoder and CSI encoder described above requires collaboration between the terminal vendor equipped with the CSI encoder and the base station vendor equipped with the CSI decoder, the AI / ML models for the CSI encoder and the CSI decoder may possess vendor-specific know-how and privacy / proprietary issues. Consequently, at least one of the CSI decoder or CSI encoder may not be disclosed.

[0225] As such, when one of the two models is not exposed to the other, it is difficult to perform proper model training using a pre-offline training method. Therefore, recursive online training at the point where the terminal and the base station connect (e.g., a full training method that shares intermediate training outputs or the opposite model in real time) has been proposed as an alternative. However, this online training method has significant issues regarding traffic overhead and network latency.

[0226] In this disclosure, the concept of a Bridge model is newly defined, and a method is proposed to resolve these issues by combining and applying the Bridge model to a terminal side model.

[0227] For convenience in describing operations, the term 'few shot learning' is used, but it can be understood as terms such as model fine-tuning or model re-training; similarly, the term 'bridge model' can be understood as a model configured in addition to the terminal part / side model, or a sub-model that constitutes the terminal part / side model.

[0228] Few-shot learning is an AI / ML training framework that updates a pre-trained model for a specific task by utilizing a small amount of additional training data to ensure the model performs well on a new task. To explain the operating principle more specifically through model-agnostic meta-learning, a type of few-shot learning, the training model is not trained as a task-specific model targeting a particular target task but can instead perform training on a diverse task ensemble. The initial parameters of the model obtained through this process allow the model to rapidly converge and adapt to a new task using only a small amount of training data for that new task. For understanding model-agnostic meta-learning, "Y. Yuan, G. Zheng, K.-K. Wong, B. Ottersten and Z.-Q. Luo, "Transfer Learning and Meta Learning-Based Fast Downlink Beamforming Adaptation," in IEEE Transactions on Wireless Communications, March 2021, can be referenced.

[0229] The key contents to be described below regarding the major functions and roles of the Bridge model newly defined / proposed in this disclosure may include the following.

[0230] In an AI / ML two-sided model composed of a UE-side encoder and an NW-side decoder, the UE-side encoder and NW-side decoder are characterized by being trained integrally using the same training data to generate the intended NW-side decoder output (final result) for a specific UE-side encoder input.

[0231] - The Bridge model may be a UE-side partial / extension model that is subsequently trained and applied at the time of deployment of the UE-side encoder and NW-side decoder to resolve parameter mismatch and interface mismatch that occur because the UE-side model is not trained integrally with the NW-side model. Regarding interface mismatch, cases may also be included where the quantization of the UE-side model output mismatches with the dequantization of the NW-side model input.

[0232] The Bridge model combines with the rear section (head) of the existing UE-side model possessed by the UE to form a new UE-side model and can play the role of generating a new UE-side model output.

[0233] The Bridge model can be initially trained in the Network beforehand and delivered to the UE; while final training is performed in the UE, it may not operate independently in place of the UE-side model. (e.g., in a two-sided model, the Bridge model cannot independently process the input to the UE-side model.)

[0234] - For the final training of the Bridge model in the UE, the NW can provide the UE with a small amount of training dataset.

[0235] For the sake of convenience of explanation, with reference to FIGS. 14 through 16, the learning and inference operations of two-sided AI / ML between a terminal and a base station are explained by dividing them into a pre-learning stage, a post-learning stage, and an inference stage. In FIGS. 14 through 16, blocks indicated by thick lines represent the learning model targeted by the terminal (or base station) in the corresponding stage, and hatched boxes represent direct or indirect models whose model parameters are updated during learning.

[0236] FIG. 14 is a diagram illustrating a pre-learning step according to one embodiment.

[0237] Referring to FIG. 14, in the first pre-training stage, the terminal and the base station each configure individual two-sided AI / ML models and perform training on their respective sides. The terminal can obtain a terminal-side encoder model (E1) by performing training from E1 to D1. The base station obtains a base station-side decoder (D2) model and a bridge (B) model by training from E2 to D2 and E2' to B to D2'. At this time, the intermediate connected quantization / dequantization blocks (Q1 / DQ1 and Q2 / DQ2) may be included in or omitted from the AI / ML training depending on the implementation, and are indicated by dotted lines. Here, the E1 and E2 models, and / or the D1 and D2 models, may be models that differ structurally or have different training parameters. Additionally, the base station's E2' and D2' may each be a batch combination of models of various structures supported by the base station according to the cell-specific configuration; thus, although they are depicted as multiple models in FIG. 14, E2' and D2' may each be models of a single structure. The E2 model is one of the E2' models, and the D2 model may refer to one of the selected models of the D2' models.

[0238] The Bridge model (B) is a model that operates on the terminal during model inference, but does not match the existing terminal-side model (E1) and may represent a partial model that is coupled to the terminal-side model. The Bridge model (B) may be a set of some layers at the back of the entire terminal-side model during model inference, or it may be an additional / another set of layers that can be coupled to the entire terminal-side model.

[0239] As illustrated in FIG. 14, the Bridge model (B) can be trained by combining it with various two-sided AI / ML ensembles at the base station. Such various two-sided AI / ML ensembles may mean that the aforementioned E2' model or D2' model is applied in multiple batches, or that training data ensembles obtained in various configurations are used for training. The Bridge model may have a predefined reference structure between the base station and the terminal. In this case, it may be a model composed of convolution layers and / or fully-connected layers (dense layers) to scalably handle the model input size.

[0240] FIG. 15 is a diagram illustrating a post-learning step according to one embodiment.

[0241] In the second stage, the post-learning stage, the terminal and the network share their capabilities and configurations, and the base station instructs / transmits an appropriate Model B to the terminal. At this time, the base station instructs / transmits cell-specific reference learning data (e.g., model input / output data for terminal-side few-shot learning) to the terminal to learn Model B. Upon receiving this instruction / transmission, the terminal may update the terminal-side encoder model structure by combining it into an E1→B structure.

[0242] If Model B is a model with a reference structure and the terminal knows the structure of Model B in advance without base station instruction / transmission and pre-applies part or all of the Model B structure into Model E1 (e.g., if Model B is structurally part of Model E1), the base station may instruct / transmit only the initial parameter information of Model B to the terminal.

[0243] In this stage of learning, the hatched B model parameters can be updated on the terminal side using the aforementioned reference training data.

[0244] FIG. 16 is a diagram illustrating an inference step according to one embodiment.

[0245] Referring to Fig. 16, the final step is an inference step in which two-sided AI / ML is actually applied to a task. In the second step, the terminal transmits data generated through terminal-side model (E1→B) inference and quantization (Q2) using an updated model including model B to the base station. The base station can complete the two-sided AI / ML operation by inferring the base station-side model (D2) and performing inverse quantization (DQ2) on the received data.

[0246] Through this method, the base station can match two-sided AI / ML models between the base station and the terminal without acquiring terminal side model information or corresponding intermediate outputs, and at the same time, the terminal can rapidly adapt to a cell-specific environment / model.

[0247] The above cell-specific data / environment / model is not limited to cell-specific (UE-common) signaling in terms of signaling, and may be provided through UE-specific signaling or through UE-group-specific signaling.

[0248] Proposals for the operation between a terminal and a base station for the learning and inference operations summarized above are described below.

[0249] Proposal 1

[0250] According to one embodiment, the terminal may report to the base station whether to apply / support a bridge to the terminal side model. And / or, the base station may instruct / request the terminal side model whether to apply a bridge.

[0251] (1) Proposal 1-1. A terminal with a Bridge applied or a terminal capable of applying it can report information about the bridge model (e.g., model ID, structure information, input / output information) to a base station.

[0252] (2) Proposal 1-2. For a terminal that has a bridge applied or can apply a bridge, the base station may set / instruct the terminal to configure information for the bridge model.

[0253] In cases where structural or parameter mismatches are expected between the terminal side model and the base station side model for reasons mentioned above, in order to combine a bridge compatible with the base station side model into the terminal side model, the terminal or the base station itself must provide bridge model handling information to the other party. This is because the bridge models supported by each terminal may differ, and the bridge models supported by each base station may differ.

[0254] To support efficient two-sided AI / ML configuration / scheduling of the base station, for terminals expected to apply a bridge, information regarding the bridge model may be reported to the base station first, regardless of a request from the base station. The information regarding the bridge model may include, for example, at least one of the bridge's model ID, information regarding the bridge model's reference structure, and / or information regarding the bridge model's input / output. Conversely, the base station may provide the terminal with configuration information regarding a bridge model compatible with the base station side model, which may include, for example, the bridge's model ID, information regarding the bridge model's reference structure, and / or information regarding the bridge model's input / output, as well as at least one of the bridge model's initial parameter information, training dataset (including ground-truth / annotation data), validation dataset information, and monitoring metric information.

[0255] The various examples listed in the information on the Bridge model or the configuration information on the bridge model above may be included in part or in whole in the information on the Bridge model or the configuration information on the bridge model.

[0256] The terminal may report the status of its side model to the base station, separate from reporting whether a bridge is applied. This status may include information regarding the input / output of the terminal side model, such as input / output size or dimensions, as well as at least one of the following: compression ratios for the input and output or additional information regarding the input (e.g., temporal / spatial / frequency-domain (TSF) properties for a CSI compression-related model). Additionally, it may include information indicating whether the terminal side model is a privacy or proprietary model. This information may be additionally required for purposes such as selecting a base station side model, generating a final output from the base station side model output, or determining a bridge model.

[0257] The terminal may report the activation or deactivation of bridge operations to the base station, separate from reporting whether a bridge is applied. This may be because the terminal supports one or more terminal side models, or because it may be functionally possible to apply one or more bridges connected to a single terminal side model, requiring the activation or deactivation of two-sided AI / ML operations for each specific bridge. Additionally, when evaluating the output of a two-sided AI / ML model (or terminal side model output) to update the bridge model or switch to fallback mode, the process of reporting such bridge operation activation or deactivation to the base station may be involved.

[0258] Proposal 2

[0259] According to one embodiment, when a bridge model is applied to a terminal side model, the terminal can distinguish between a terminal side model update and a bridge model update based on differences in the input / output size of the model and / or the model update method on the LCM.

[0260] (1) Proposal 2-1. The input size of the Bridge model is set independently of the input size of the terminal side model or the output size of the base station side model, but the output size of the Bridge model can be set to be the same as the output size of the terminal side model and / or the input size of the base station side model.

[0261] (2) Proposal 2-2. The terminal side model is performed using the model update method on the LCM, and the application of the bridge model can be performed using the partial model update method on the LCM.

[0262] (3) Proposal 2-3. The bridge model is a model with a reference structure that is defined in advance, and the bridge model ID may not match part or all of the ID with the terminal side model ID.

[0263] In Model ID-based LCM, the bridge model has a reference structure and can be distinguished by whether the bridge model ID and the terminal side model ID do not match in part or in whole.

[0264] However, in a Functionality-based LCM, when the bridge application process to the terminal is performed via model transfer / delivery, it is necessary to distinguish it from the terminal-side model update process. If the process proceeds through the same model transfer / delivery, the application of the bridge model can be distinguished from the terminal-side model update by the fact that the new model input information is configured independently of the existing model's input information. Furthermore, to prevent such confusion, a partial model update process can be defined that is distinct from the existing model update process when applying the bridge model to the terminal.

[0265] The Bridge model can be defined as one or more common reference types that are independent of a specific configuration among the input information of the terminal side model. For example, the reference types may include type-A, which is independent of # of TX antenna, type-B, which is independent of # of rank, etc.

[0266] Proposal 3

[0267] According to one embodiment, a model ID distinguishing one or more Bridge models and / or a dataset ID distinguishing one or more datasets may be defined, and linkage information between them may be set from a base station to a terminal.

[0268] The model ID of the Bridge model may be separate from the model ID of the existing terminal side model or base station side model, and the reference model type information may also be defined separately from the model ID.

[0269] For example, a model configuration set from a base station to a terminal is defined and may include a dataset ID.

[0270] For example, a dataset configuration set from the base station to the terminal is defined and may include a model ID.

[0271] Among the bridge-related information that a base station can provide to a terminal, at least one of the linkage and / or extensible configurations of various individual pieces of information, such as model structure / parameters, the corresponding training dataset, training metric, and validation dataset, can be configured from the base station to the terminal. Therefore, assuming a model ID-based LCM, bridge configurations of this form can be proposed.

[0272] The above proposals 1 / 2 / 3 may be implemented alone or based on at least a combination of some of them.

[0273] Next, the UE-triggered case and the NW-triggered case are described, respectively, as an explanation of exemplary terminal and base station operations.

[0274] FIG. 17 is a diagram illustrating terminal operation according to one embodiment.

[0275] Referring to FIG. 17, the terminal can request terminal side model information (e.g., AI / ML capability) from the base station (1705).

[0276] The terminal can report information about the terminal side model (e.g., at least one of input / output information, additional information related to input, privacy / proprietary status) to the base station (1710).

[0277] The terminal can request information on the Bridge model that the terminal has applied / can apply from the base station (1715).

[0278] The terminal can report information about the Bridge model that the terminal has applied / can apply (e.g., model ID, structure information, input / output information, etc.) to the base station (1720).

[0279] The terminal can receive configuration information (e.g., model ID, reference structure information, input / output information, initial parameters information, training dataset information, validation dataset information, monitoring metric information) for a Bridge model selected by the base station from the base station (1725).

[0280] The terminal can perform inference through a terminal side model (update model) combined with a Bridge model (1730).

[0281] The terminal can transmit a CSI report to the base station containing the output of the inferred update model and / or the result calculated based on the monitoring metric (1735).

[0282] The terminal operation and / or base station operation described in the example of FIG. 17 may be performed by the device (200) of FIG. 3. For example, one or more processors (202) of the device (200) of FIG. 3 may be configured to perform the operation according to FIG. 17. Furthermore, one or more memories (204) of the device (200) may store instructions for performing the method in the example of FIG. 17 or in various examples of the foregoing specification when executed by one or more processors (202).

[0283] In the above terminal / base station operation, (some) specific steps may be omitted.

[0284] FIG. 18 is a diagram illustrating a compressed CSI report according to one embodiment. FIG. 18 may correspond to an example of a terminal inference operation in FIG. 17 and a report based thereon.

[0285] Referring to FIG. 18, the base station can transmit CSI-RS to the terminal (1805).

[0286] The terminal can report compressed CSI to the base station through inference based on a UE side model combined with a Bridge model as input values ​​for CSI-RS measurements (1810).

[0287] After receiving the compressed CSI of the terminal, the base station can restore the CSI and perform PDSCH scheduling / transmission by referring to the restored CSI (1815).

[0288] According to the present disclosure, inter-vendor collaboration issues caused by privacy / proprietary constraints of the terminal-side model can be resolved through a bridge provided by a base station to the terminal. The proposed method can expect a reduction in network overhead and latency compared to general two-sided AI / ML offline training methods.

[0289] FIG. 19 is a diagram illustrating the operation of a terminal and a base station according to one embodiment. It may correspond to a generalized example of the procedure described in FIG. 17 / 18.

[0290] Referring to FIG. 19, the terminal and the base station can transmit and receive at least one of information related to the UE side model and / or information related to the NW side model (1905). For example, the NW can transmit information related to the NW-side model to the UE via RRC, or the UE can provide information related to the UE-side model to the NW via an AI / ML applicability report or a UE capability report.

[0291] The terminal and the base station can transmit and receive information related to the bridge model (1910). For example, the UE can request the bridge model from the NW, or conversely, the NW can instruct / set the bridge model for the UE.

[0292] - In Functionality-based Life Cycle Management (LCM), the bridge model application phase can be designated as a partial model update for the UE.

[0293] - In a Model-ID based LCM, the bridge model can be designated as a Model-ID, which may be a second Model-ID distinct from the encoder / decoder Model-ID (the first Model-ID). For example, the first Model-ID format and the second Model-ID format may differ in part or entirely.

[0294] - In a Model-ID based LCM, the bridge model can be designated as a model-ID, which can be distinguished from the model-pairing ID. (The model-pairing ID is information regarding the compatibility of the encoder and decoder and can be shared between the UE and the NW.) For example, after the model-pairing ID is transmitted from the terminal to the network, the bridge model-ID can be transmitted from the network to the terminal. For instance, after the model-pairing ID is transmitted from the network to the terminal, the terminal determines a model mismatch and requests a bridge model, and the network can transmit the model-ID to the terminal.

[0295] The network can transmit dataset information to a terminal for training a bridge model (1915). According to one embodiment, linkage information can be established between the dataset configuration and the model configuration on the RRC. For example, a model-ID may be included in the dataset configuration or a dataset-ID may be included in the model configuration. For example, when the Bridge model is designated as a (standardized) reference model, model type information may be defined separately from model-D.

[0296] The terminal can perform bridge model training based on the transmitted data set (1920).

[0297] The terminal can perform inference through a terminal side model (update model) combined with a Bridge model (1930).

[0298] The terminal can transmit a CSI report to the base station based on inference (1935).

[0299] FIG. 20 illustrates the flow of a method performed at a terminal according to one embodiment. FIG. 20 is an example of implementation for at least some of the embodiments described above, and the above description may be referenced to aid in understanding FIG. 20 unless otherwise noted.

[0300] Referring to FIG. 20, the terminal can receive configuration information for linking between a first model for CSI (channel state information) compression and a second model for CSI reconstruction (2005).

[0301] The terminal can configure a third model based on the above setting information (2010).

[0302] The terminal can perform CSI reporting based on the inference result obtained from the combination of the first model and the third model (2015).

[0303] The above third model can be configured in a format for inputting the output of the above first model into the above second model.

[0304] The above setting information may include at least one of identification information of the third model or information about the first data set for learning the third model.

[0305] The information regarding the first data set may include identification information of the first data set associated with the third model.

[0306] The above terminal configuring the above third model may include learning the above third model based on the above first data set.

[0307] The above-mentioned first model and the above-mentioned third model can be trained based on different data sets.

[0308] The above third model may be attached to the end of the above first model or combined as part of the above first model.

[0309] The above terminal can transmit a terminal capability report containing information about candidate third models supported by the terminal.

[0310] The above third model may be a reference model pre-processed based on an ensemble of candidate first models.

[0311] The first model above is a CSI encoder model, the second model above is a CSI decoder model, and the third model above may be a bridge model located between the CSI encoder model and the CSI decoder.

[0312] FIG. 21 illustrates the flow of a method performed at a base station according to one embodiment. FIG. 21 is an example of implementation for at least some of the embodiments described above, and the above description may be referenced to aid in understanding FIG. 21 unless otherwise noted.

[0313] Referring to FIG. 21, the base station can determine a third model for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction (2105).

[0314] The base station can transmit configuration information for the third model to the terminal (2110).

[0315] The base station can receive a compressed CSI report from the terminal (2115).

[0316] The base station can perform CSI reconstruction based on the compressed CSI report and the second model (2120).

[0317] The above compressed CSI report may include the inference result of the terminal obtained from the combination of the first model and the third model.

[0318] The above third model may be a model configured in a format for inputting the output of the above first model into the above second model.

[0319] The above setting information may include at least one of identification information of the third model or information about the first data set for learning the third model.

[0320] The information regarding the first data set may include identification information of the first data set associated with the third model.

[0321] The above-mentioned first model and the above-mentioned third model can be trained based on different data sets.

[0322] The above third model may be attached to the end of the above first model or combined as part of the above first model.

[0323] The base station can receive a terminal capability report containing information about candidate third models supported by the terminal.

[0324] The above third model may be a reference model pre-processed based on an ensemble of candidate first models.

[0325] The first model above is a CSI encoder model, the second model above is a CSI decoder model, and the third model above may be a bridge model located between the CSI encoder model and the CSI decoder.

[0326] The embodiments described above are combinations of the components and features of the present disclosure in a specific form. Each component or feature should be considered optional unless otherwise explicitly stated. Each component or feature may be implemented in a form not combined with other components or features. Additionally, it is possible to construct embodiments of the present disclosure by combining some components and / or features. The order of operations described in the embodiments of the present disclosure may be changed. Some components or features of one embodiment may be included in another embodiment, or may be replaced with corresponding components or features of another embodiment. It is obvious that embodiments may be constructed by combining claims that are not explicitly related in the claims, or that they may be included as new claims by amendment after filing.

[0327] It is obvious to those skilled in the art that the present disclosure may be embodied in other specific forms without departing from the features of the present disclosure. Accordingly, the foregoing detailed description should not be interpreted restrictively in all respects and should be considered illustrative. The scope of the present disclosure shall be determined by a reasonable interpretation of the appended claims, and all modifications within the equivalent scope of the present disclosure are included within the scope of the present disclosure.

[0328] The present disclosure may be used in a terminal, base station, or other equipment of a wireless mobile communication system.

Claims

In a method performed by a terminal, Receive configuration information for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; Configure a third model based on the above setting information; and It includes performing CSI reporting based on the inference results obtained from the combination of the first model and the third model, and A method in which the above-mentioned third model is configured in a format for inputting the output of the above-mentioned first model into the above-mentioned second model. In Article 1, A method in which the above setting information includes at least one of identification information of the third model or information about a first data set for learning the third model. In Article 2, A method in which information regarding the first data set includes identification information of the first data set associated with the third model. In Clause 2, constituting the third model is, A method comprising training the third model based on the first data set. In Article 1, A method in which the first model and the third model are learned based on different data sets. In Article 1, A method in which the third model is attached to the end of the first model or combined as part of the first model. In Article 1, A method further comprising transmitting a terminal capability report containing information on candidate third models supported by the terminal. In Article 1, The above third model is a reference model pre-processed based on an ensemble of candidate first models, method. In Article 1, The above first model is a CSI encoder model, and The above second model is a CSI decoder model, and The above third model is a bridge model located between the above CSI encoder model and the above CSI decoder, a method. A computer-readable non-transitory recording medium storing a program for performing the method described in claim 1. Regarding the device, Memory configured to store instructions; and A processor configured to perform operations by executing the above instructions, comprising: The operations of the above processor are, Receive configuration information for linking a first model for CSI (channel state information) compression and a second model for CSI reconstruction; Configure a third model based on the above setting information; and It includes performing CSI reporting based on the inference results obtained from the combination of the first model and the third model, and The above third model is a device configured in a format for inputting the output of the above first model to the above second model. In Article 11, It further includes a transmitter and receiver, The above device is a device that is a terminal operating in a wireless communication system. In Article 11, The above device is a processing device configured to control a terminal operating in a wireless communication system. In a method performed by a base station, Determine a third model for linking the first model for CSI (channel state information) compression and the second model for CSI reconstruction; Transmitting configuration information for the above-mentioned third model to the terminal; Receive a compressed CSI report from the above terminal; and It includes performing CSI reconstruction based on the compressed CSI report and the second model, and The above compressed CSI report includes the inference result of the terminal obtained from the combination of the first model and the third model, and The above third model is a method in which the output of the above first model is configured in a format for inputting the output of the above second model. In the case of a base station, Memory configured to store instructions; and A processor configured to perform operations by executing the above instructions, comprising: The operations of the above processor are, Determine a third model for linking the first model for CSI (channel state information) compression and the second model for CSI reconstruction; Transmitting configuration information for the above-mentioned third model to the terminal; Receive a compressed CSI report from the above terminal; and It includes performing CSI reconstruction based on the compressed CSI report and the second model, and The above compressed CSI report includes the inference result of the terminal obtained from the combination of the first model and the third model, and The above third model is a base station, which is a model configured in a format for inputting the output of the above first model to the above second model.

Images