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

Abstract

A method performed by a terminal according to at least one of embodiments disclosed in the present specification may comprise: acquiring information on predicted channel state information (CSI) and information on compressed CSI on the basis of an artificial intelligence / machine learning (AI / ML) model supporting both CSI compression and CSI prediction; and transmitting a CSI report on the basis of the information on the predicted CSI and the information on the compressed CSI, wherein on the basis that the performance of the AI / ML model is less than a threshold, the terminal adjusts past CSIs used as an input data set of the CSI compression before changing at least one of a configuration for the CSI compression and a configuration for the CSI prediction of the AI / ML model.

Description

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

[0001] The present disclosure relates to a wireless communication system, and more particularly, to a method and device for transmitting or receiving an uplink / downlink wireless signal by a terminal or a network in a wireless communication system.

[0002] Wireless communication systems are widely deployed to provide various types of communication services, such as voice and data. Typically, wireless communication systems are multiple access systems that support communication with multiple users by sharing available system resources (e.g., bandwidth, transmission power). Examples of multiple access systems include code division multiple access (CDMA), frequency division multiple access (FDMA), time division multiple access (TDMA), orthogonal frequency division multiple access (OFDMA), and single-carrier frequency division multiple access (SC-FDMA).

[0003] Recently, research is being conducted on CSI feedback based on AI / ML (artificial intelligence / machine learning) in NR standardization, and as one of the main research topics, CSI compression to reduce CSI overhead through AI / ML models and CSI prediction to predict CSI at a future point in time are being considered.

[0004] The technical task of the present disclosure is to provide a method and device for efficiently performing wireless signal transmission and reception processes. For example, a method and device for more accurate and efficient performance monitoring and AI / ML model management can be provided for an AI / ML model that simultaneously performs CSI compression and CSI prediction.

[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 includes obtaining information about predicted CSI and information about compressed CSI based on an AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and transmitting a CSI report based on the information about the predicted CSI and the information about the compressed CSI, wherein based on performance of the AI / ML model being below a threshold, the terminal can adjust past CSIs used as an input data set of the CSI compression prior to changing at least one of a setting of the AI / ML model for the CSI compression or a setting of the CSI prediction.

[0007] After adjusting the above past CSIs, based on the performance of the AI / ML model being above the threshold, the terminal can maintain the AI / ML model without changing the settings for the CSI compression or the settings for the CSI prediction.

[0008] Even after adjusting the past CSIs, based on the performance of the AI / ML model being below the threshold, the terminal may change at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0009] The terminal can determine whether the CSI compression causes a performance degradation of the AI / ML model or whether the CSI prediction causes a performance degradation of the AI / ML model by maintaining one of the settings for the CSI compression and the settings for the CSI prediction while changing the other.

[0010] The terminal may reconfigure the AI / ML model or switch to another AI / ML model based on the performance of the AI / ML model being below the threshold even after changing at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0011] The adjustment of the past CSIs may be performed based on at least one of a first mode for excluding all of the past CSIs from the input data set of the CSI compression; a second mode for replacing the past CSIs with CSIs accumulated from the present; and a third mode for selectively excluding some of the past CSIs from the input data set of the CSI compression.

[0012] The terminal may receive network signaling indicating that the performance of the AI / ML model is below the threshold. The network signaling may indicate at least one of the first mode, the second mode, and the third mode.

[0013] The terminal may determine whether the performance of the AI / ML model is below a threshold based on a second AI / ML model for CSI reconstruction or a proxy model for the second AI / ML model.

[0014] According to another aspect of the present disclosure, a non-transitory computer-readable recording medium having recorded thereon a program for performing the method described above may be provided.

[0015] According to another aspect of the present disclosure, a device comprises a memory configured to store instructions; and a processor configured to perform operations by executing the instructions, wherein the operations of the processor include obtaining information about predicted CSI and information about compressed CSI based on an AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and transmitting a CSI report based on the information about predicted CSI and the information about compressed CSI, wherein based on performance of the AI / ML model being below a threshold, the device can adjust past CSIs used as an input data set of the CSI compression prior to changing at least one of a setting of the AI / ML model for the CSI compression or a setting of the CSI prediction.

[0016] The above device may further include a transceiver.

[0017] The above device may be a terminal in a wireless communication system.

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

[0019] According to another aspect of the present disclosure, a method performed by a base station includes receiving a CSI report through an AI / ML (artificial intelligence / machine learning) model of a terminal that supports both CSI compression (channel state information) and CSI prediction; and obtaining information about predicted CSI and information about compressed CSI based on the CSI report, wherein based on performance of the AI / ML model being below a threshold, the base station can instruct the terminal to adjust past CSIs used as an input data set of the CSI compression before changing at least one of a setting of the AI / ML model for the CSI compression or a setting of the CSI prediction.

[0020] According to another aspect of the present disclosure, a non-transitory computer-readable recording medium having recorded thereon a program for performing the method described above may be provided.

[0021] According to another aspect of the present disclosure, a base station comprises a memory configured to store instructions; and a processor configured to perform operations by executing the instructions, wherein the operations of the processor include receiving a CSI report through an AI / ML (artificial intelligence / machine learning) model of a terminal that supports both CSI compression (channel state information) and CSI prediction; and obtaining information about predicted CSI and information about compressed CSI based on the CSI report, wherein based on a performance of the AI / ML model being below a threshold, the base station can instruct the terminal to adjust past CSIs used as an input data set of the CSI compression prior to changing at least one of a setting of the AI / ML model for the CSI compression or a setting of the CSI prediction.

[0022] According to the present disclosure, wireless signal transmission and reception can be efficiently performed in a wireless communication system. For example, when performance degradation occurs in an AI / ML model that simultaneously performs CSI compression and CSI prediction, the system distinguishes whether the performance degradation is due to the input data set or the CSI compression / CSI prediction function, enabling more accurate and efficient performance monitoring and AI / ML model management.

[0023] In addition to the technical effects described above, other technical effects can be inferred from the description below.

[0024] Figure 1 illustrates physical channels used in a 3GPP system, which is an example of a wireless communication system, and a general signal transmission method using the channels.

[0025] Figure 2 illustrates the structure of a radio frame.

[0026] Figure 3 illustrates a resource grid of slots.

[0027] Figure 4 illustrates an example of physical channels being mapped within a slot.

[0028] Figure 5 illustrates the PDSCH and ACK / NACK transmission process.

[0029] Figure 6 illustrates a PUSCH transmission process.

[0030] Figure 7 shows an example of a CSI-related procedure.

[0031] Figure 8 is a diagram to explain the concept of AI / ML / Deep learning.

[0032] Figures 9 to 12 illustrate various AI / ML models of deep learning.

[0033] Figure 13 is a diagram illustrating segmentation AI inference.

[0034] Figure 14 is a diagram illustrating a framework for 3GPP RAN Intelligence.

[0035] Figures 15 to 17 illustrate AI Model Training and Inference environments.

[0036] Figure 18 is a diagram to explain the concept of AI / ML-based CSI compression.

[0037] Figure 19 is a diagram to explain the concept of AI / ML-based CSI prediction.

[0038] Figure 20 is a diagram for explaining the operation of a terminal and a network according to one implementation.

[0039] Figure 21 illustrates a flow of a method performed by a terminal in a wireless communication system according to one embodiment.

[0040] Figure 22 illustrates a flowchart of a method performed by a base station in a wireless communication system according to one embodiment.

[0041] Figures 23 to 26 illustrate a communication system (1) and a wireless device applicable to the present disclosure.

[0042] The following technologies can be used in various wireless access 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 radio technologies such as UTRA (Universal Terrestrial Radio Access) or CDMA2000. TDMA can be implemented with radio 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 radio technologies such as IEEE 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802-20, and E-UTRA (Evolved UTRA). UTRA is a part of UMTS (Universal Mobile Telecommunications System). 3GPP (3rd Generation Partnership Project) LTE (long term evolution) is part of E-UMTS (Evolved UMTS) that uses E-UTRA, and LTE-A (Advanced) is an evolved version of 3GPP LTE. 3GPP NR (New Radio or New Radio Access Technology) is an evolved version of 3GPP LTE / LTE-A.

[0043] As more and more communication devices demand greater communication capacity, the need for improved mobile broadband communications compared to existing Radio Access Technology (RAT) is emerging. Furthermore, massive Machine Type Communications (MTC), which connects multiple devices and objects to provide diverse services anytime, anywhere, is also a key issue to be considered in next-generation communications. Furthermore, communication system design that considers reliability and latency-sensitive services / terminals is being discussed. Accordingly, the introduction of next-generation RATs, such as enhanced Mobile BroadBand Communication (eMBB), massive MTC, and Ultra-Reliable and Low Latency Communication (URLLC), is being discussed. For convenience, these technologies are referred to as NR (New Radio or New RAT) in this specification.

[0044] For clarity of explanation, the description will focus on 3GPP NR, but the technical idea of ​​the present disclosure is not limited thereto.

[0045] In this specification, the expression "setting" can be replaced with the expression "configure / configuration", and the two can be used interchangeably. In addition, conditional expressions (e.g., "if", "in a case", or "when", etc.) can be replaced with the expression "based on that ~~" or "in a state / status". In addition, the operation of the terminal / base station or the SW / HW configuration according to the satisfaction of the condition can be inferred / understood. In addition, if the process of the receiving (or transmitting) side can be inferred / understood from the process of the transmitting (or receiving) side in signal transmission / reception between wireless communication devices (e.g., base stations, terminals), the description thereof can be omitted. For example, signal determination / generation / encoding / transmission, etc. of the transmitting side can be understood as signal monitoring reception / decoding / determination, etc. of the receiving side. In addition, the expression that the terminal performs (or does not perform) a specific operation can also be interpreted as meaning that the base station operates while expecting / assuming (or expecting / assuming that the terminal does not perform) the specific operation. In addition, the expression that the base station performs (or does not perform) a specific operation can also be interpreted as meaning that the terminal operates while expecting / assuming (or expecting / assuming that the base station does not perform) the specific operation. In addition, the division and index of each section, embodiment, example, option, method, plan, etc. in the following description are for the convenience of explanation and should not be interpreted as meaning that each constitutes an independent invention or that each must be implemented only individually. In addition, in describing each section, embodiment, example, option, method, plan, etc., if there is no explicitly conflicting / opposing description, it can be inferred / interpreted that at least some of them can be combined and implemented together, or at least some can be implemented with the omission of each.

[0046] In a wireless communication system, a terminal receives information from a base station via the downlink (DL) and transmits it to the base station via the uplink (UL). The information transmitted and received between the base station and the terminal includes data and various control information, and various physical channels exist depending on the type and purpose of the information being transmitted and received.

[0047] Figure 1 is a drawing for explaining physical channels used in a 3GPP NR system and a general signal transmission method using them.

[0048] When a terminal is powered on again from a powered-off state or enters a new cell, it performs an initial cell search operation, such as synchronizing with the base station, in step S101. To this end, the terminal receives a Synchronization Signal Block (SSB) from the base station. The SSB includes a Primary Synchronization Signal (PSS), a Secondary Synchronization Signal (SSS), and a Physical Broadcast Channel (PBCH). Based on the PSS / SSS, the terminal synchronizes with the base station and obtains information such as a cell ID (cell identity). In addition, the terminal can obtain broadcast information within the cell based on the PBCH. Meanwhile, the terminal can check the downlink channel status by receiving a Downlink Reference Signal (DL RS) during the initial cell search phase.

[0049] After completing the initial cell search, the terminal can obtain more specific system information by receiving a physical downlink control channel (PDCCH) and a physical downlink shared channel (PDSCH) based on the physical downlink control channel information in step S102.

[0050] Thereafter, the terminal may perform a random access procedure such as steps S103 to S106 to complete connection to the base station. To this end, the terminal may transmit a preamble through a physical random access channel (PRACH) (S103) and receive a response message to the preamble through a physical downlink control channel and a corresponding physical downlink shared channel (S104). In the case of contention-based random access, a contention resolution procedure such as transmission of an additional physical random access channel (S105) and reception of a physical downlink control channel and a corresponding physical downlink shared channel (S106) may be performed.

[0051] The terminal that has performed the procedure as described above can then perform the general uplink / downlink signal transmission procedure, such as receiving a physical downlink control channel / physical downlink shared channel (S107) and transmitting a physical uplink shared channel (PUSCH) / physical uplink control channel (PUCCH) (S108). The control information that the terminal transmits to the base station is collectively referred to as uplink control information (UCI). UCI includes HARQ ACK / NACK (Hybrid Automatic Repeat and request Acknowledgement / Negative-ACK), SR (Scheduling Request), CSI (Channel State Information), etc. CSI includes CQI (Channel Quality Indicator), PMI (Precoding Matrix Indicator), RI (Rank Indication), etc. UCI is generally transmitted through PUCCH, but can be transmitted through PUSCH when control information and traffic data must be transmitted simultaneously. Additionally, UCI can be transmitted aperiodically via PUSCH upon request / instruction from the network.

[0052] Figure 2 illustrates the structure of a radio frame. In NR, uplink and downlink transmissions are organized into frames. Each radio frame is 10 ms long and is divided into two 5 ms half-frames (HF). Each half-frame is divided into five 1 ms sub-frames (SF). A sub-frame is divided into one or more slots, and the number of slots within a sub-frame depends on the subcarrier spacing (SCS). Each slot contains 12 or 14 OFDM (Orthogonal Frequency Division Multiplexing) symbols, depending on the cyclic prefix (CP). When a normal CP is used, each slot contains 14 OFDM symbols. When an extended CP is used, each slot contains 12 OFDM symbols.

[0053] Table 1 illustrates that when CP is normally used, the number of symbols per slot, the number of slots per frame, and the number of slots per subframe vary depending on the SCS.

[0054] SCS (15*2^u)N slot symb N frame,u slot N subframe,u slot 15KHz (u=0)1410130KHz (u=1)1420260KHz (u=2)14404120KHz (u=3)14808240KHz (u=4)1416016

[0055] * N slot symb : Number of symbols in the slot

[0056] * N frame,u slot : Number of slots in the frame

[0057] * N subframe,u slot : Number of slots in a subframe

[0058] Table 2 illustrates that when extended CP is used, the number of symbols per slot, the number of slots per frame, and the number of slots per subframe vary depending on the SCS.

[0059] SCS (15*2^u)N slot symb N frame,u slot N subframe,u slot 60KHz (u=2)12404

[0060] The structure of the frame is only an example, and the number of subframes, number of slots, and number of symbols in the frame can be varied.

[0061] In an NR system, OFDM numerology (e.g., SCS) may be set differently between multiple cells that are merged into a single terminal. Accordingly, the (absolute time) interval of a time resource (e.g., SF, slot, or TTI) (conveniently referred to as TU (Time Unit)) consisting of the same number of symbols may be set differently between the merged cells. Here, the symbol may include an OFDM symbol (or CP-OFDM symbol), an SC-FDMA symbol (or Discrete Fourier Transform-spread-OFDM, DFT-s-OFDM symbol).

[0062] Figure 3 illustrates a resource grid of a slot. A slot contains multiple symbols in the time domain. For example, in the case of a regular CP, one slot contains 14 symbols, but in the case of an extended CP, one slot contains 12 symbols. A carrier contains multiple subcarriers in the frequency domain. A Resource Block (RB) is defined as multiple (e.g., 12) consecutive subcarriers in the frequency domain. A Bandwidth Part (BWP) is defined as multiple consecutive Physical RBs (PRBs) in the frequency domain and can correspond to a single numerology (e.g., SCS, CP length, etc.). A carrier can contain up to N (e.g., 5) BWPs. Data communication is performed through activated BWPs, and only one BWP can be activated for a single terminal. Each element in the resource grid is referred to as a Resource Element (RE), to which one complex symbol can be mapped.

[0063] Figure 4 illustrates an example of how physical channels are mapped within a slot. A PDCCH can be transmitted in the DL control region, and a PDSCH can be transmitted in the DL data region. A PUCCH can be transmitted in the UL control region, and a PUSCH can be transmitted in the UL data region. GP provides a time gap between the base station and the terminal when switching from transmission mode to reception mode or from reception mode to transmission mode. Some symbols within a subframe at the time of transition from DL to UL can be set as GP.

[0064] Below, each physical channel is described in more detail.

[0065] The PDCCH carries Downlink Control Information (DCI). For example, the PCCCH (i.e., DCI) carries the transmission format and resource allocation of the downlink shared channel (DL-SCH), resource allocation information for the uplink shared channel (UL-SCH), paging information for the paging channel (PCH), system information on the DL-SCH, resource allocation information for upper layer control messages such as random access responses transmitted on the PDSCH, transmission power control commands, activation / deactivation of Configured Scheduling (CS), etc. The DCI includes a cyclic redundancy check (CRC), which is masked / scrambled with various identifiers (e.g., Radio Network Temporary Identifier, RNTI) depending on the owner or usage of the PDCCH. For example, if the PDCCH is for a specific terminal, the CRC is masked with a terminal identifier (e.g., Cell-RNTI, C-RNTI). If the PDCCH is for paging, the CRC is masked with the Paging-RNTI (P-RNTI). If the PDCCH is for system information (e.g., a System Information Block, SIB), the CRC is masked with the System Information RNTI (SI-RNTI). If the PDCCH is for a random access response, the CRC is masked with the Random Access-RNTI (RA-RNTI).

[0066] The PDCCH consists of 1, 2, 4, 8, or 16 Control Channel Elements (CCEs) depending on the Aggregation Level (AL). A CCE is a logical allocation unit used to provide a PDCCH with a predetermined code rate depending on the radio channel status. A CCE consists of six Resource Element Groups (REGs). A REG is defined as one OFDM symbol and one (P)RB. The PDCCH is transmitted through a Control Resource Set (CORESET). A CORESET is defined as a set of REGs with a given numerology (e.g., SCS, CP length, etc.). Multiple CORESETs for a single UE can overlap in the time / frequency domain. A CORESET can be configured through system information (e.g., Master Information Block, MIB) or UE-specific upper layer (e.g., Radio Resource Control, RRC, layer) signaling. Specifically, the number of RBs and the number of OFDM symbols (up to 3) that constitute the CORESET can be set by upper layer signaling.

[0067] To receive / detect PDCCH, the UE monitors PDCCH candidates. PDCCH candidates represent the CCE(s) that the UE should monitor for PDCCH detection. Each PDCCH candidate is defined as 1, 2, 4, 8, or 16 CCEs depending on the AL. Monitoring involves (blind) decoding the PDCCH candidates. The set of PDCCH candidates that the UE monitors is defined as a PDCCH Search Space (SS). The search space includes a Common Search Space (CSS) or a UE-specific search space (USS). The UE can acquire DCI by monitoring PDCCH candidates in one or more search spaces configured by the MIB or higher-layer signaling. Each CORESET is associated with one or more search spaces, and each search space is associated with one COREST. The search space can be defined based on the following parameters.

[0068] - controlResourceSetId: Indicates the CORESET associated with the search space.

[0069] - monitoringSlotPeriodicityAndOffset: Indicates the PDCCH monitoring period (in slots) and the PDCCH monitoring interval offset (in slots).

[0070] - monitoringSymbolsWithinSlot: Indicates the PDCCH monitoring symbols within the slot (e.g., the first symbol(s) of the CORESET).

[0071] - nrofCandidates: AL={1, 2, 4, 8, 16} indicates the number of PDCCH candidates (one of 0, 1, 2, 3, 4, 5, 6, 8)

[0072] * An opportunity (e.g., time / frequency resource) for monitoring PDCCH candidates is defined as a PDCCH (monitoring) opportunity. One or more PDCCH (monitoring) opportunities can be configured within a slot.

[0073] Table 3 illustrates the characteristics of each search space type.

[0074] TypeSearch SpaceRNTIUse CaseType0-PDCCHCommonSI-RNTI on a primary cellSIB DecodingType0A-PDCCHCommonSI-RNTI on a primary cellSIB DecodingType1-PDCCHCommonRA-RNTI or TC-RNTI on a primary cellMsg2, Msg4 decoding in RACHType2-PDCCHCommonP-RNTI on a primary cellPaging DecodingType3-PDCCHCommonINT-RNTI, SFI-RNTI, TPC-PUSCH-RNTI, TPC-PUCCH-RNTI, TPC-SRS-RNTI, C-RNTI, MCS-C-RNTI, or CS-RNTI(s)UE SpecificUE SpecificC-RNTI, or MCS-C-RNTI, or CS-RNTI(s)User specific PDSCH decoding

[0075] Table 4 illustrates DCI formats transmitted via PDCCH.

[0076] DCI formatUsage0_0Scheduling of PUSCH in one cell0_1Scheduling of PUSCH in one cell1_0Scheduling of PDSCH in one cell1_1Scheduling of PDSCH in one cell2_0Notifying a group of UEs of the slot format2_1Notifying a group of UEs of the PRB(s) and OFDM symbol(s) where UE may assume no transmission is intended for the UE2_2Transmission of TPC commands for PUCCH and PUSCH2_3Transmission of a group of TPC commands for SRS transmissions by one or more UEs

[0077] DCI format 0_0 is used to schedule a TB-based (or TB-level) PUSCH, and DCI format 0_1 ​​can be used to schedule a TB-based (or TB-level) PUSCH or a CBG (Code Block Group)-based (or CBG-level) PUSCH. DCI format 1_0 is used to schedule a TB-based (or TB-level) PDSCH, and DCI format 1_1 can be used to schedule a TB-based (or TB-level) PDSCH or a CBG-based (or CBG-level) PDSCH (DL grant DCI). DCI format 0_0 / 0_1 may be referred to as UL grant DCI or UL scheduling information, and DCI format 1_0 / 1_1 may be referred to as DL grant DCI or DL ​​scheduling information. DCI format 2_0 is used to convey dynamic slot format information (e.g., dynamic SFI) to the terminal, and DCI format 2_1 is used to convey downlink pre-emption information to the terminal. DCI format 2_0 and / or DCI format 2_1 can be conveyed to the terminals within a group through the group common PDCCH, which is a PDCCH conveyed to the terminals defined as a group.

[0078] DCI format 0_0 and DCI format 1_0 may be referred to as fallback DCI formats, while DCI format 0_1 ​​and DCI format 1_1 may be referred to as non-fallback DCI formats. In the fallback DCI format, the DCI size / field configuration remains the same regardless of the terminal configuration. On the other hand, in the non-fallback DCI format, the DCI size / field configuration varies depending on the terminal configuration.

[0079] PDSCH carries downlink data (e.g., DL-SCH transport block, DL-SCH TB) and applies modulation methods such as Quadrature Phase Shift Keying (QPSK), 16 Quadrature Amplitude Modulation (QAM), 64 QAM, and 256 QAM. TB is encoded to generate a codeword. PDSCH can carry up to two codewords. Scrambling and modulation mapping are performed for each codeword, and modulation symbols generated from each codeword can be mapped to one or more layers. Each layer is mapped to resources along with a Demodulation Reference Signal (DMRS), generated as an OFDM symbol signal, and transmitted through the corresponding antenna port.

[0080] PUCCH carries Uplink Control Information (UCI). UCI includes:

[0081] - SR (Scheduling Request): Information used to request UL-SCH resources.

[0082] - HARQ(Hybrid Automatic Repeat reQuest)-ACK(Acknowledgement): This is a response to a downlink data packet (e.g., codeword) on the PDSCH. It indicates whether the downlink data packet was successfully received. One HARQ-ACK bit can be transmitted in response to a single codeword, and two HARQ-ACK bits can be transmitted in response to two codewords. The HARQ-ACK response includes a positive ACK (simply, ACK), a negative ACK (NACK), a DTX, or a NACK / DTX. Here, HARQ-ACK is used interchangeably with HARQ ACK / NACK and ACK / NACK.

[0083] - CSI (Channel State Information): Feedback information for the downlink channel. MIMO (Multiple Input Multiple Output)-related feedback information includes the Rank Indicator (RI) and Precoding Matrix Indicator (PMI).

[0084] Table 5 illustrates PUCCH formats. Depending on the PUCCH transmission length, they can be classified into Short PUCCH (formats 0 and 2) and Long PUCCH (formats 1, 3, and 4).

[0085] PUCCH formatLength in OFDM symbols N PUCCH symb Number of bitsUsageEtc01 - 2≤2HARQ, SRSequence selection14 - 14≤2HARQ, [SR]Sequence modulation21 - 2>2HARQ, CSI, [SR]CP-OFDM34 - 14>2HARQ, CSI, [SR]DFT-s-OFDM(no UE multiplexing)44 - 14>2HARQ, CSI, [SR]DFT-s-OFDM(Pre DFT OCC)

[0086] PUCCH format 0 carries UCI of up to 2 bits in size and is mapped and transmitted based on sequence. Specifically, the terminal transmits a specific UCI to the base station by transmitting one of multiple sequences through the PUCCH of PUCCH format 0. The terminal transmits the PUCCH of PUCCH format 0 within the PUCCH resources for the corresponding SR configuration only when transmitting a positive SR.

[0087] PUCCH format 1 carries UCI of up to 2 bits in size, and modulation symbols are spread in the time domain using an orthogonal cover code (OCC) (which is set differently depending on whether frequency hopping is used). DMRS are transmitted in symbols where modulation symbols are not transmitted (i.e., transmitted using Time Division Multiplexing (TDM).

[0088] PUCCH format 2 carries UCI with a bit size greater than 2 bits, and modulation symbols are transmitted by frequency division multiplexing (FDM) with DMRS. DM-RSs are located at symbol indices #1, #4, #7, and #10 within a given resource block with a density of 1 / 3. Pseudo Noise (PN) sequences are used for DM_RS sequences. Frequency hopping can be enabled for 2-symbol PUCCH format 2.

[0089] PUCCH format 3 does not multiplex terminals within the same physical resource blocks and carries UCI with a bit size greater than 2 bits. In other words, PUCCH resources in PUCCH format 3 do not include orthogonal cover codes. Modulation symbols are transmitted through time division multiplexing (TDM) with DMRS.

[0090] PUCCH format 4 supports multiplexing of up to four terminals within the same physical resource blocks and carries UCI with a bit size greater than 2 bits. In other words, PUCCH resources in PUCCH format 3 include orthogonal cover codes. Modulation symbols are transmitted through time division multiplexing (TDM) with DMRS.

[0091] At least one of one or more configured cells in a terminal may be configured for PUCCH transmission. At least the primary cell may be configured as a cell for PUCCH transmission. At least one PUCCH cell group may be configured in the terminal based on at least one cell configured for PUCCH transmission, and each PUCCH cell group includes one or more cells. The PUCCH cell group may be simply referred to as a PUCCH group. PUCCH transmission may be configured not only for the primary cell but also for the SCell, and the primary cell belongs to the primary PUCCH group, and the PUCCH-SCell configured for PUCCH transmission belongs to the secondary PUCCH group. For cells belonging to the primary PUCCH group, the PUCCH on the primary cell may be used, and for cells belonging to the secondary PUCCH group, the PUCCH on the PUCCH-SCell may be used.

[0092] PUSCH carries uplink data (e.g., UL-SCH transport block, UL-SCH TB) and / or uplink control information (UCI), and is transmitted based on a CP-OFDM (Cyclic Prefix - Orthogonal Frequency Division Multiplexing) waveform or a DFT-s-OFDM (Discrete Fourier Transform - spread - Orthogonal Frequency Division Multiplexing) waveform. When the PUSCH is transmitted based on a DFT-s-OFDM waveform, the UE transmits the PUSCH by applying transform precoding. For example, when transform precoding is disabled (e.g., transform precoding is disabled), the UE transmits the PUSCH based on the CP-OFDM waveform, and when transform precoding is enabled (e.g., transform precoding is enabled), the UE can transmit the PUSCH based on the CP-OFDM waveform or the DFT-s-OFDM waveform. PUSCH transmissions can be dynamically scheduled by UL grants in DCI, or semi-statically scheduled (configured grant) based on higher layer (e.g., RRC) signaling (and / or Layer 1 (L1) signaling (e.g., PDCCH)). PUSCH transmissions can be performed in a codebook-based or non-codebook-based manner.

[0093] Figure 5 illustrates an ACK / NACK transmission process. Referring to Figure 5, a terminal can detect a PDCCH in slot #n. Here, the PDCCH includes downlink scheduling information (e.g., DCI formats 1_0, 1_1), and the PDCCH indicates a DL assignment-to-PDSCH offset (K0) and a PDSCH-HARQ-ACK reporting offset (K1). For example, DCI formats 1_0, 1_1 can include the following information:

[0094] - Frequency domain resource assignment: Indicates the set of RBs allocated to the PDSCH.

[0095] - Time domain resource assignment: K0 (e.g., slot offset), indicates the starting position of the PDSCH within slot #n+K0 (e.g., OFDM symbol index), and the length of the PDSCH (e.g., number of OFDM symbols).

[0096] - PDSCH-to-HARQ_feedback timing indicator: Indicates K1

[0097] - HARQ process number (4 bits): Indicates the HARQ process ID (Identity) for data (e.g., PDSCH, TB)

[0098] - PUCCH resource indicator (PRI): Indicates the PUCCH resource to be used for UCI transmission among multiple PUCCH resources within the PUCCH resource set.

[0099] Afterwards, the terminal receives PDSCH from slot #(n+K0) according to the scheduling information of slot #n, and when reception of PDSCH is finished in slot #n1 (where, n+K0≤n1), UCI can be transmitted through PUCCH in slot #(n1+K1). Here, UCI may include HARQ-ACK response for PDSCH. In Fig. 5, for convenience, it is assumed that SCS for PDSCH and SCS for PUCCH are the same and slot # n1 = slot #n+K0, but the present invention is not limited thereto. If the SCSs are different, K1 can be indicated / interpreted based on the SCS of PUCCH.

[0100] When the PDSCH is configured to transmit at most 1 TB, the HARQ-ACK response may consist of 1 bit. When the PDSCH is configured to transmit at most 2 TB, the HARQ-ACK response may consist of 2 bits if spatial bundling is not configured, and may consist of 1 bit if spatial bundling is configured. When the HARQ-ACK transmission timing for multiple PDSCHs is designated as slot #(n+K1), the UCI transmitted in slot #(n+K1) includes HARQ-ACK responses for multiple PDSCHs.

[0101] Whether a UE should perform spatial bundling for a HARQ-ACK response can be configured (e.g., via RRC / higher layer signaling) for each cell group. For example, spatial bundling can be individually configured for each HARQ-ACK response transmitted over the PUCCH and / or each HARQ-ACK response transmitted over the PUSCH.

[0102] Spatial bundling can be supported when the maximum number of TBs (or codewords) that can be received at a time (or scheduled via 1 DCI) in the serving cell is 2 (or more than 2) (e.g., when the upper layer parameter maxNrofCodeWordsScheduledByDCI corresponds to 2-TB). Meanwhile, more than 4 layers can be used for 2-TB transmission, and up to 4 layers can be used for 1-TB transmission. Consequently, when spatial bundling is configured for the cell group, spatial bundling can be performed for serving cells that can schedule more than 4 layers among the serving cells in the cell group. On the serving cell, a terminal that wishes to transmit a HARQ-ACK response via spatial bundling can generate the HARQ-ACK response by performing a (bit-wise) logical AND operation on the A / N bits for multiple TBs.

[0103] For example, assuming that a terminal receives a DCI scheduling 2 TB and receives 2 TB via PDSCH based on the DCI, the terminal performing spatial bundling can generate a single A / N bit by logically ANDing the first A / N bit for the first TB and the second A / N bit for the second TB. Consequently, if both the first TB and the second TB are ACK, the terminal reports the ACK bit value to the base station, and if either TB is NACK, the terminal reports the NACK bit value to the base station.

[0104] For example, if only 1-TB is actually scheduled on a serving cell configured to receive 2-TB, the terminal can generate a single A / N bit by logically ANDing the A / N bit for the 1-TB with bit value 1. Consequently, the terminal reports the A / N bit for the 1-TB to the base station as is.

[0105] A base station / terminal has multiple parallel DL HARQ processes for DL ​​transmission. These multiple parallel HARQ processes allow DL transmissions to be performed continuously while waiting for HARQ feedback regarding the successful or unsuccessful reception of a previous DL transmission. Each HARQ process is associated with a HARQ buffer in the MAC (Medium Access Control) layer. Each DL HARQ process manages state variables such as the number of transmissions of MAC Physical Data Blocks (PDUs) in the buffer, HARQ feedback for MAC PDUs in the buffer, and the current redundancy version. Each HARQ process is identified by a HARQ process ID.

[0106] Figure 6 illustrates a PUSCH transmission process. Referring to Figure 6, a terminal can detect a PDCCH in slot #n. Here, the PDCCH includes uplink scheduling information (e.g., DCI formats 0_0 and 0_1). DCI formats 0_0 and 0_1 can include the following information.

[0107] - Frequency domain resource assignment: Indicates the set of RBs allocated to PUSCH.

[0108] - Time domain resource assignment: Slot offset K2 indicates the starting position (e.g., symbol index) and length (e.g., number of OFDM symbols) of the PUSCH within the slot. The starting symbol and length can be indicated through SLIV (Start and Length Indicator Value) or can be indicated separately.

[0109] Thereafter, the terminal can transmit a PUSCH in slot #(n+K2) according to the scheduling information of slot #n. Here, the PUSCH includes a UL-SCH TB.

[0110] CSI-related actions

[0111] Figure 7 shows an example of a CSI-related procedure.

[0112] The terminal receives configuration information related to CSI from the base station via RRC signaling (710). The configuration information related to CSI may include at least one of CSI-IM (interference management) resource-related information, CSI measurement configuration-related information, CSI resource configuration-related information, CSI-RS resource-related information, or CSI report configuration-related information.

[0113] - CSI-IM resources can be configured for interference measurement (IM) of the terminal. In the time domain, the CSI-IM resource set can be configured periodically, semi-persistently, or aperiodicly. The CSI-IM resources can be configured as Zero Power (ZP)-CSI-RS for the terminal. The ZP-CSI-RS can be configured separately from the Non-Zero Power (NZP)-CSI-RS.

[0114] - The UE may assume that the CSI-RS resource(s) for channel measurement configured for one CSI reporting and the CSI-IM / NZP CSI-RS resource(s) for interference measurement (when NZP CSI-RS resource(s) are used for interference measurement) are in a QCL relationship with respect to 'QCL-TypeD' per resource.

[0115] - The CSI resource configuration may include at least one of a CSI-IM resource for interference measurement, an NZP CSI-RS resource for interference measurement, and an NZP CSI-RS resource for channel measurement. The CMR (channel measurement resource) may be an NZP CSI-RS for CSI acquisition, and the IMR (Interference measurement resource) may be an NZP CSI-RS for CSI-IM and IM.

[0116] - CSI-RS can be configured for one or more terminals. Different CSI-RS configurations may be provided for each terminal, or the same CSI-RS configuration may be provided to multiple terminals. CSI-RS can support up to 32 antenna ports. CSI-RS corresponding to N (N is 1 or greater) antenna ports can be mapped to N RE locations within a time-frequency unit corresponding to one slot and one RB. When N is 2 or greater, N-port CSI-RS can be multiplexed using CDM, FDM, and / or TDM schemes. CSI-RS can be mapped to REs other than REs to which CORESET, DMRS, and SSB are mapped. In the frequency domain, CSI-RS can be configured for the entire bandwidth, a portion of the bandwidth (BWP), or a portion of the bandwidth. CSI-RS may be transmitted in each RB within the bandwidth for which CSI-RS is configured (i.e., density = 1), or in every second RB (e.g., even or odd RB) (i.e., density = 1 / 2). When CSI-RS is used as a Tracking Reference Signal (TRS), a single-port CSI-RS may be mapped on three subcarriers in each resource block (i.e., density = 3). One or more CSI-RS resource sets may be configured for a UE in the time domain. Each CSI-RS resource set may include one or more CSI-RS configurations. Each CSI-RS resource set may be configured periodically, semi-persistently, or aperiodicly.

[0117] - The CSI report configuration may include configurations for feedback type, measurement resources, report type, etc. The NZP-CSI-RS resource set may be used for the CSI report configuration of the corresponding terminal. The NZP-CSI-RS resource set may be associated with CSI-RS or SSB. In addition, multiple periodic NZP-CSI-RS resource sets may be configured as TRS resource sets. (i) The feedback type may include a Channel Quality Indicator (CQI), a Precoding Matrix Indicator (PMI), a CSI-RS Resource Indicator (CRI), an SSB Resource block Indicator (SSBRI), a Layer Indicator (LI), a Rank Indicator (RI), a Layer 1-Reference Signal Received Strength (RSRP), etc. (ii) Measurement resources may include configurations for downlink signals and / or downlink resources on which the terminal performs measurements to determine feedback information. The measurement resources may be configured as ZP and / or NZP CSI-RS resource sets associated with CSI reporting configurations. The NZP CSI-RS resource set may include a CSI-RS set or an SSB set. For example, L1-RSRP may be measured for a CSI-RS set or an SSB set. (iii) Reporting types may include configurations for a time point at which the terminal performs reporting and an uplink channel, etc. The reporting time point may be configured as periodic, semi-persistent, or aperiodic. Periodic CSI reporting may be transmitted on PUCCH. Semi-persistent CSI reporting may be transmitted on PUCCH or PUSCH based on a MAC CE indicating activation / deactivation. Aperiodic CSI reporting may be indicated by DCI signaling.For example, the CSI request field of an uplink grant may indicate one of several report trigger sizes. Aperiodic CSI reports may be transmitted on the PUSCH.

[0118] The terminal measures CSI based on configuration information related to CSI. CSI measurement may include a procedure of receiving a CSI-RS (720) and computing the received CSI-RS to acquire CSI (730).

[0119] The UE can transmit a CSI report to the base station (740). For the CSI report, the time and frequency resources that the UE can use are controlled by the base station. The CSI (channel state information) can include at least one of a channel quality indicator (CQI), a precoding matrix indicator (PMI), a CSI-RS resource indicator (CRI), an SS / PBCH block resource indicator (SSBRI), a layer indicator (LI), a rank indicator (RI), L1-RSRP, and / or L-SINR.

[0120] The time domain behavior of CSI reporting supports periodic, semi-persistent, and aperiodic. i) Periodic CSI reporting is performed on short PUCCH and long PUCCH. The periodicity and slot offset of periodic CSI reporting can be configured by RRC, and refer to the CSI-ReportConfig IE. ii) SP (semi-periodic) CSI reporting is performed on short PUCCH, long PUCCH, or PUSCH. In case of SP CSI on short / long PUCCH, the periodicity and slot offset are configured by RRC, and CSI reporting is activated / deactivated by separate MAC CE / DCI. In case of SP CSI on PUSCH, the periodicity of SP CSI reporting is configured by RRC, but the slot offset is not configured by RRC, and SP CSI reporting is activated / deactivated by DCI (format 0_1). For SP CSI reporting on PUSCH, a separate RNTI (SP-CSI C-RNTI) is used. The initial CSI reporting timing follows the PUSCH time domain allocation value indicated in the DCI, and subsequent CSI reporting timings follow the cycle set by RRC. DCI format 0_1 ​​includes a CSI request field and can activate / deactivate a specific configured SP-CSI trigger state. SP CSI reporting has the same or similar activation / deactivation mechanism as the data transmission mechanism on the SPS PUSCH.iii) Aperiodic CSI reporting is performed on PUSCH and is triggered by DCI. In this case, information related to the triggering of aperiodic CSI reporting can be transmitted / indicated / configured via MAC-CE. For AP CSI with AP CSI-RS, the AP CSI-RS timing is configured by RRC, and the timing for AP CSI reporting is dynamically controlled by DCI.

[0121] CSI codebooks defined in the NR standard (e.g., PMI codebooks) can be broadly divided into Type I and Type II codebooks. Type I codebooks are primarily targeted at SU (Single User)-MIMO, which supports both high-order and low-order signals. Type II codebooks can primarily support MI-MIMO, which supports up to two layers. Compared to Type I, Type II codebooks can provide more accurate CSI, but may increase signaling overhead. Meanwhile, Enhanced Type II codebooks were introduced to address the CSI overhead shortcomings of existing Type II codebooks. Enhanced Type II codebooks were introduced by reducing the codebook payload by considering frequency-axis correlation.

[0122] CSI reporting via PUSCH can be configured as Part 1 and Part 2. Part 1 has a fixed payload size and is used to identify the number of information bits in Part 2. Part 1 is transmitted in its entirety before Part 2.

[0123] - For Type I CSI feedback, Part 1 contains the RI (if reported), the CRI (if reported), and the CQI of the first code word. Part 2 contains the PMI, and when RI > 4, Part 2 contains the CQI.

[0124] - For Type II CSI feedback, Part 1 contains the RI (if reported), CQI, and an indication of the number of non-zero WB amplitude coefficients per layer of Type II CSI. Part 2 contains the PMI of Type II CSI.

[0125] - For Enhanced Type II CSI feedback, Part 1 contains the RI (if reported), CQI, and the total number of non-zero WB amplitude coefficients for all layers of Enhanced Type II CSI. Part 2 contains the PMI of Enhanced Type II CSI.

[0126] If CSI reporting on PUSCH includes two parts and the CSI payload to be reported is less than the payload size provided by the PUSCH resources allocated for CSI reporting, the UE may omit part of Part 2 CSI.

[0127] Meanwhile, semi-persistent CSI reporting performed in PUCCH format 3 or 4 supports Type II CSI feedback, but only Part 1 of Type II CSI feedback.

[0128] QCL (quasi-co location)

[0129] Two antenna ports are quasi-co-located if the channel properties of one antenna port can be inferred from the channel properties of the other antenna port. The channel properties may include one or more of Delay spread, Doppler spread, Frequency / Doppler shift, Average received power, Received Timing / average delay, and Spatial RX parameters.

[0130] A terminal can configure a list of multiple TCI-State configurations via the upper layer parameter PDSCH-Config. Each TCI-State is associated with one or two DL reference signals and a QCL configuration parameter between the DM-RS port of the PDSCH. The QCL can include qcl-Type1 for the first DL RS and qcl-Type2 for the second DL RS. The QCL type can correspond to one of the following:

[0131] - 'QCL-TypeA': {Doppler shift, Doppler spread, average delay, delay spread}

[0132] - 'QCL-TypeB': {Doppler shift, Doppler spread}

[0133] - 'QCL-TypeC': {Doppler shift, average delay}

[0134] - 'QCL-TypeD': {Spatial Rx parameter}

[0135] Beam Management (BM)

[0136] The BM process is a process for acquiring and maintaining a set of BS (or transmission and reception point (TRP)) and / or UE beams that can be used for downlink (DL) and uplink (UL) transmission / reception, and may include the following processes and terms.

[0137] - Beam measurement: An operation in which a BS or UE measures the characteristics of a received beamforming signal.

[0138] - Beam determination: An operation in which a BS or UE selects its own transmit beam (Tx beam) / receive beam (Rx beam).

[0139] - Beam sweeping: An operation of covering a spatial domain using transmit and / or receive beams over a predetermined time interval in a predetermined manner.

[0140] - Beam report: An operation in which a UE reports information about a beamformed signal based on beam measurement.

[0141] The BM process can be divided into (1) a DL BM process using SSB or CSI-RS, and (2) a UL BM process using SRS (sounding reference signal). In addition, each BM process can include Tx beam sweeping to determine a Tx beam and Rx beam sweeping to determine an Rx beam.

[0142] At this time, the DL BM process may include (1) transmission of beamformed DL RSs (e.g., CSI-RS or SSB) by the BS and (2) beam reporting by the UE.

[0143] Here, the beam report may include preferred DL RS ID(s) and corresponding reference signal received power (RSRP). The DL RS ID may be an SSB Resource Indicator (SSBRI) or a CSI-RS Resource Indicator (CRI).

[0144] Positioning

[0145] Positioning may refer to determining the geographic location and / or velocity of a UE by measuring radio signals. Position information may be requested by a client (e.g., an application) associated with the UE and reported to the client. Furthermore, the location information may be contained within the core network or requested by a client connected to the core network. The location information may be reported in a standard format, such as cell-based or geographic coordinates, and may also include an estimated error value for the UE's position and velocity and / or the positioning method used for positioning.

[0146] LPP can be used as a point-to-point between a location server (E-SMLC and / or SLP and / or LMF) and a target device (UE and / or SET) to position the target device using position-related measurements obtained from one or more reference sources. Through LPP, the target device and the location server can exchange measurement and / or position information based on Signal A and / or Signal B.

[0147] NRPPa can be used to exchange information between a reference source (ACCESS NODE and / or BS and / or TP and / or NG-RAN node) and a location server.

[0148] The functions provided by the NRPPa protocol may include:

[0149] - E-CID Location Information Transfer. This function allows location information to be exchanged between the reference source and the LMF for E-CID positioning purposes.

[0150] - OTDOA Information Transfer. This function allows information to be exchanged between the reference source and the LMF for OTDOA positioning purposes.

[0151] - Reporting of General Error Situations. This feature allows reporting of general error situations for which no function-specific error message is defined.

[0152] The positioning methods supported by NG-RAN may include GNSS (Global Navigation Satellite System), OTDOA, E-CID (enhanced cell ID), barometric positioning, WLAN positioning, Bluetooth positioning, terrestrial beacon system (TBS), and UTDOA (Uplink Time Difference of Arrival). Among the above positioning methods, the position of the UE may be measured using any one of the positioning methods, but the position of the UE may also be measured using two or more positioning methods.

[0153] OTDOA (Observed Time Difference Of Arrival)

[0154] The OTDOA positioning method utilizes the timing measurements of downlink signals received by the UE from multiple TPs, including the eNB, ng-eNB, and PRS-dedicated TPs. The UE measures the timing of the received downlink signals using location assistance data received from a location server. Based on these measurement results and the geographic coordinates of neighboring TPs, the UE's location can be determined.

[0155] A UE connected to a gNB can request a measurement gap for OTDOA measurements from a TP. If the UE does not recognize the SFN for at least one TP in the OTDOA assistance data, the UE can use an autonomous gap to obtain the SFN of the OTDOA reference cell before requesting a measurement gap to perform Reference Signal Time Difference (RSTD) measurements.

[0156] Here, the RSTD can be defined based on the smallest relative time difference between the boundaries of two subframes received from the reference cell and the measurement cell, respectively. That is, it can be calculated based on the relative time difference between the start time of the subframe of the reference cell that is closest to the start time of the subframe received from the measurement cell. Meanwhile, the reference cell can be selected by the UE.

[0157] Accurate OTDOA measurement requires measuring the time of arrival (TOA) of signals received from three or more geographically dispersed TPs or base stations. For example, the TOA for TP 1, TP 2, and TP 3 can be measured, and based on the three TOAs, the RSTD for TP 1-TP 2, the RSTD for TP 2-TP 3, and the RSTD for TP 3-TP 1 can be calculated. Based on these TOAs, a geometric hyperbola can be determined, and the point where these hyperbolas intersect can be used to estimate the UE's location. Since each TOA measurement may have inaccuracies and / or uncertainties, the estimated UE's location can be known within a certain range depending on the measurement uncertainty.

[0158] E-CID (Enhanced Cell ID)

[0159] In the Cell ID (CID) positioning method, the location of the UE can be measured through geographic information of the UE's serving ng-eNB, serving gNB, and / or serving cell. For example, geographic information of the serving ng-eNB, serving gNB, and / or serving cell can be obtained through paging, registration, etc.

[0160] Meanwhile, the E-CID positioning method may utilize additional UE measurements and / or NG-RAN radio resources in addition to the CID positioning method to improve the UE position estimate. In the E-CID positioning method, some of the same measurement methods as the measurement control system of the RRC protocol may be used, but generally, additional measurements are not performed solely for UE position measurement. In other words, a separate measurement configuration or measurement control message may not be provided to measure the UE's position, and the UE may not expect to be requested to perform additional measurement operations solely for position measurement, and may report measurement values ​​obtained through measurement methods that the UE can generally measure.

[0161] For example, a serving gNB can implement an E-CID positioning method using E-UTRA measurements provided from the UE.

[0162] AI / ML (Artificial intelligence / machine learning)

[0163] Technological advancements in AI / ML are leading to the intelligence / advanced development of node(s) and terminal(s) that make up wireless communication networks. In particular, the intelligence of networks / base stations will enable the rapid optimization / derivation / application of various network / base station decision parameter values ​​(e.g., transmit / receive power of each base station, transmit power of each terminal, precoder / beam of base stations / terminals, time / frequency resource allocation for each terminal, duplexing method of each base station, etc.) based on various environmental parameters (e.g., distribution / location of base stations, distribution / location / material of buildings / furniture, etc., location / movement direction / speed of terminals, climate information, etc.). In line with this trend, many standardization organizations (e.g., 3GPP, O-RAN) are considering its introduction, and research on it is also actively underway.

[0164] AI / ML can be easily referred to as artificial intelligence based on deep learning in a narrow sense, but conceptually it is as shown in Figure 8.

[0165] - Artificial Intelligence: This can refer to all automation where machines can replace tasks that people would otherwise do.

[0166] - Machine Learning: Machines can learn patterns for decision-making from data without explicitly programming rules.

[0167] Deep Learning: An AI / ML model based on artificial neural networks. Machines simultaneously extract features from unstructured data and make judgments. The algorithms rely on multilayer networks of interconnected nodes for feature extraction and transformation, inspired by the biological nervous system, or neural networks. Common deep learning network architectures include deep neural networks (DNNs), recurrent neural networks (RNNs), and convolutional neural networks (CNNs).

[0168] Classification of AI / ML types based on various criteria

[0169] 1. Offline vs. Online

[0170] (1) Offline Learning: This follows a sequential process of database collection, learning, and prediction. In other words, collection and learning are performed offline, and the completed program can be installed on-site for use in prediction tasks. This offline learning method is used in most situations. In offline learning, the system does not learn incrementally; learning is performed using all available collected data and applied to the system without further training. If learning on new data is required, learning can be restarted using the entire new data set.

[0171] (2) Online Learning: Online learning leverages the continuous availability of data available for learning via the Internet. This method incrementally improves performance by learning from additional data. Learning is performed in real time on specific data (sets) collected online, enabling the system to quickly adapt to changing data.

[0172] To build an AI system, only online learning may be used, so that learning is performed using only real-time data, or offline learning may be performed using a predetermined data set, and then additional learning may be performed using additional real-time data (online + offline learning).

[0173] 2. Classification by AI / ML Framework Concept

[0174] (1) Centralized Learning: Training data collected from multiple different nodes are reported to a centralized node, and all data resources / storage / learning (e.g., supervised, unsupervised, reinforcement learning) are performed in one central node.

[0175] (2) Federated Learning: A collective AI / ML model is built based on data across distributed data owners. Instead of importing data into an AI / ML model, the AI / ML model is imported as a data source, allowing local nodes / individual devices to collect data and train their own copies of the AI / ML model, eliminating the need to report source data to a central node. In federated learning, the parameters / weights of the AI / ML model are simply sent back to the centralized node to support general AI / ML model training. The advantages of federated learning include increased computational speed and superior information security. This eliminates the need to upload personal data to a central server, preventing personal information leaks and misuse.

[0176] (3) Distributed Learning: This concept represents the concept of machine learning processes being scaled and distributed across a cluster of nodes. Training AI / ML models are split and shared across multiple nodes operating simultaneously to accelerate AI / ML model training.

[0177] 3. Classification by learning method

[0178] (1) Supervised Learning: Supervised learning is a machine learning task that aims to learn a mapping function from input to output given a labeled data set. The input data is called training data and has known labels or outcomes. Examples of supervised learning include (i) Regression: Linear Regression, Logistic Regression, (ii) Instance-based Algorithms: k-Nearest Neighbor (KNN), (iii) Decision Tree Algorithms: CART, (iv) Support Vector Machines: SVM, (v) Bayesian Algorithms: Naive Bayes, and (vi) Ensemble Algorithms: Extreme Gradient Boosting, Bagging: Random Forest. Supervised learning can be further grouped into regression and classification problems, where classification predicts labels and regression predicts quantities.

[0179] (2) Unsupervised Learning: A machine learning task that aims to learn features that explain hidden structures in unlabeled data. The input data is unlabeled and has no known outcome. Some examples of unsupervised learning include K-means clustering, principal component analysis (PCA), nonlinear independent component analysis (ICA), and long-term memory (LSTM).

[0180] (3) Reinforcement Learning: In reinforcement learning (RL), an agent interacts with the environment through a trial-and-error process, aiming to optimize a long-term goal. It is a goal-oriented learning method based on interaction with the environment. Examples of RL algorithms include (i) Q-learning, (ii) multi-armed bandit learning, (iii) deep Q network, state-action-reward-state-action (SARSA), (iv) temporal difference learning, (v) actor-critic reinforcement learning, (vi) deep deterministic policy gradient, and (vii) Monte-Carlo tree search. Reinforcement learning can be further grouped into AI / ML model-based reinforcement learning and AI / ML model-free reinforcement learning. Model-based reinforcement learning is an RL algorithm that uses a predictive AI / ML model to obtain transition probabilities between states by using various dynamic states of the environment and the AI / ML model that leads to these states as rewards. Model-free reinforcement learning is a value- or policy-based RL algorithm that maximizes future rewards. It is computationally less complex in multi-agent environments / states and does not require an accurate representation of the environment. RL algorithms can also be categorized into value-based RL versus policy-based RL, and policy-based RL versus non-policy RL.

[0181] AI / ML models

[0182] Figure 9 illustrates an FFNN (Feed-Forward Neural Network) AI / ML model. Referring to Figure 9, the FFNN AI / ML model includes an input layer, a hidden layer, and an output layer.

[0183] Figure 10 illustrates an RNN (Recurrent Neural Network) AI / ML model. Referring to Figure 10, the RNN AI / ML model is a type of artificial neural network in which hidden nodes are connected by directed edges to form a cyclic structure (directed cycle), and is an AI / ML model suitable for processing data that appears sequentially, such as voice and text. One type of RNN is LSTM (Long Short-Term Memory), and LSTM is a structure that adds a cell state to the hidden state of the RNN. Specifically, in LSTM, an input gate, a forget gate, and an output gate are added to the RNN cell, and a cell state is added. In Figure 10, A represents a neural network, x t is the input value, h t represents the output value. Here, h t can mean a status value that represents the present based on time, and h t-1 can represent the previous state value.

[0184] Figure 11 illustrates a CNN (Convolution Neural Network) AI / ML model. CNN uses convolution operations commonly used in image processing and video processing to achieve two goals: reducing AI / ML model complexity and extracting good features. Referring to Figure 11, a kernel or filter refers to a unit / structure that applies weights to inputs within a specific range / unit. The kernel (or filter) can be modified through learning. The stride refers to the range of movement of the kernel within the input. The feature map refers to the result of applying the kernel to the input. Padding refers to a value added to adjust the size of the feature map. Multiple feature maps can be extracted to induce robustness to distortion and changes. Pooling refers to an operation (e.g., max pooling, average pooling) to reduce the size of the feature map by downsampling it.

[0185] Figure 12 illustrates an auto-encoder AI / ML model. Referring to Figure 12, an auto-encoder is a neural network that receives a feature vector x as input and outputs the same or similar vector x'. The input and output nodes have the same features, and it is a type of unsupervised learning. Since an auto-encoder reconstructs the input, the output can be referred to as a reconstruction. The loss function can be expressed as in Mathematical Formula 1.

[0186]

[0187] The loss function of the auto encoder exemplified in Figure 12 is calculated based on the difference between the input and the output, and based on this, the degree of loss of the input is identified, and the auto encoder performs an optimization process to minimize the loss.

[0188] Figure 13 is a diagram illustrating segmentation AI inference.

[0189] Figure 13 illustrates a case where, among split AI operations, the Model Inference function is performed collaboratively by an end device such as a UE and a network AI / ML endpoint.

[0190] In addition to the Model Inference function, the Model Training function, Actor, and Data Collection functions can each be split into multiple parts depending on the current task and environment, and performed by multiple entities collaborating.

[0191] For example, computationally intensive and energy-intensive parts may be performed at the network endpoint, while privacy-sensitive and latency-sensitive parts may be performed at the end device. In this case, the end device may execute the task / model from input data up to a specific part / layer, and then transmit the intermediated data to the network endpoint. The network endpoint then executes the remaining parts / layers and provides the inference outputs to one or more devices that perform the actions / tasks.

[0192] The following describes a functional framework for AI operations.

[0193] Below, to explain AI (or AI / ML) more specifically, the terms can be defined as follows.

[0194] - Data collection: Data collected from network nodes, management entities, or UEs as a basis for AI model training, data analysis, and inference.

[0195] - 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.

[0196] - AI / ML Training: An online or offline process of training an AI model by learning features and patterns that best represent the data and obtain a trained AI / ML model for inference.

[0197] - AI / ML Inference: The process of making predictions or inducing decisions based on collected data and the AI ​​model using a trained AI model.

[0198] Referring to FIG. 14, the Data Collection function (10) is a function that collects input data and provides processed input data to the Model Training function (20) and the Model Inference function (30).

[0199] Examples of input data may include measurements from UEs or other network entities, feedback from actors, and output from AI models.

[0200] The Data Collection function (10) performs data preparation based on input data and provides input data processed through data preparation. Here, the Data Collection function (10) does not perform data preparation specific to each AI algorithm (e.g., data pre-processing and cleaning, formatting, and transformation), but can perform data preparation common to AI algorithms.

[0201] After the data preparation process is performed, the Model Training function (10) provides training data (11) to the Model Training function (20) and provides inference data (Inference Data) (12) to the Model Inference function (30). Here, the Training Data (11) is data required as input for the AI ​​Model Training function (20). The Inference Data (12) is data required as input for the AI ​​Model Inference function (30).

[0202] The Data Collection function (10) may be performed by a single entity (e.g., UE, RAN node, network node, etc.) or may be performed by multiple entities. In this case, Training Data (11) and Inference Data (12) may be provided to the Model Training function (20) and Model Inference function (30), respectively, from multiple entities.

[0203] The Model Training function (20) is a function that performs AI model training, validation, and testing, which can generate model performance metrics as part of the AI ​​model testing process. If necessary, the Model Training function (20) also handles data preparation (e.g., data pre-processing and cleaning, forming, and transformation) based on the Training Data (11) provided by the Data Collection function (10).

[0204] Here, Model Deployment / Update (13) is used to initially deploy the trained, verified, and tested AI model to the Model Inference function (30) or to provide the updated model to the Model Inference function (30).

[0205] The Model Inference function (30) is a function that provides AI model inference output (16) (e.g., prediction or decision). If applicable, the Model Inference function (30) may provide model performance feedback (14) to the Model Training function (20). In addition, the Model Inference function (30) is also responsible for data preparation (e.g., data pre-processing and cleaning, forming, and transformation) based on the Inference Data (12) provided by the Data Collection function (10), if necessary.

[0206] Here, Output (16) refers to the inference output of the AI ​​model generated by the Model Inference function (30), and the details of the inference output may vary depending on the use case.

[0207] Model Performance Feedback (14) can be used to monitor the performance of the AI ​​model if available, and this feedback may be omitted.

[0208] The actor function (40) is a function that receives the output (16) from the model inference function (30) and triggers or performs a corresponding task / action. The actor function (40) can trigger tasks / actions for other entities (e.g., one or more UEs, one or more RAN nodes, one or more network nodes, etc.) or for itself.

[0209] Feedback (15) can be used to derive training data (11), inference data (12), or to monitor the performance of the AI ​​model, its impact on the network, etc.

[0210] Meanwhile, the definitions of training / validation / test in the data set used in AI / ML can be distinguished as follows.

[0211] - Training data: This refers to the data set for learning the model.

[0212] - Validation data: This refers to a data set used to validate a model that has already completed training. In other words, it refers to a data set typically used to prevent overfitting of the training data set.

[0213] It also refers to a data set for selecting the best model among the various models learned during the learning process. Therefore, it can be viewed as a type of learning.

[0214] - Test data: This refers to the data set for final evaluation. This data is unrelated to learning.

[0215] In the case of the above data set, if the training set is generally divided, the training data and validation data can be divided and used in a ratio of 8:2 or 7:3 within the entire training set, and if the test is included, it can be divided and used in a ratio of 6:2:2 (training: validation: test).

[0216] Depending on the capability of the AI / ML function between the base station and the terminal, the level of cooperation can be defined as follows, and variations due to combination of multiple levels or separation of any one level are also possible.

[0217] Cat 0a) No collaboration framework: AI / ML algorithms are purely implementation-based and do not require any changes to the wireless interface.

[0218] Cat 0b) This level corresponds to a framework with a modified wireless interface tailored to efficient implementation-based AI / ML algorithms, but without collaboration.

[0219] Category 1) involves inter-node support to improve the AI / ML algorithms of each node. This applies when the UE receives support from the gNB (for training, adaptation, etc.) and vice versa. At this level, model exchange between network nodes is not required.

[0220] Category 2) Joint ML tasks can be performed between the UE and gNB. This level requires the exchange of AI / ML model commands or network nodes.

[0221] The functions exemplified in FIG. 14 above may be implemented in a RAN node (e.g., a base station, a TRP, a central unit (CU) of a base station, etc.), a network node, an operation administration maintenance (OAM) of a network operator, or a UE.

[0222] Alternatively, two or more entities, such as a RAN, a network node, a network operator's OAM, or a UE, may cooperate to implement the functions exemplified in FIG. 14. For example, one entity may perform some of the functions of FIG. 14, and another entity may perform the remaining functions. In this way, since some of the functions exemplified in FIG. 14 are performed by a single entity (e.g., a UE, a RAN node, a network node, etc.), the transmission / provision of data / information between each function may be omitted. For example, if the Model Training function (20) and the Model Inference function (30) are performed by the same entity, the transmission / provision of Model Deployment / Update (13) and Model Performance Feedback (14) may be omitted.

[0223] Alternatively, any one of the functions illustrated in FIG. 14 may be performed through collaboration between two or more entities, including a RAN, a network node, a network operator's OAM, or a UE. This may be referred to as a split AI operation.

[0224] Figure 15 illustrates a case where the AI ​​Model Training function is performed by a network node (e.g., a core network node, an OAM of a network operator, etc.) and the AI ​​Model Inference function is performed by a RAN node (e.g., a base station, a TRP, a CU of a base station, etc.).

[0225] Step 1: RAN node 1 and RAN node 2 transmit input data (i.e., training data) for AI model training to the network node. Here, RAN node 1 and RAN node 2 can also transmit data collected from the UE (e.g., UE measurements related to RSRP, RSRQ, SINR of the serving cell and neighboring cells, UE location, speed, etc.) to the network node.

[0226] Step 2: Network nodes train the AI ​​model using the received training data.

[0227] Step 3: The network node distributes / updates the AI ​​Model to RAN Node 1 and / or RAN Node 2. RAN Node 1 (and / or RAN Node 2) may continue model training based on the received AI Model.

[0228] For convenience of explanation, we assume that the AI ​​Model is deployed / updated only to RAN node 1.

[0229] Step 4: RAN node 1 receives input data for AI Model Inference (i.e., Inference data) from UE and RAN node 2.

[0230] Step 5: RAN node 1 performs AI Model Inference using the received Inference data to generate output data (e.g., prediction or decision).

[0231] Step 6: If applicable, RAN node 1 may transmit model performance feedback to the network nodes.

[0232] Step 7: RAN node 1, RAN node 2, and the UE (or 'RAN node 1 and the UE', or 'RAN node 1 and the RAN node 2') perform actions based on the output data. For example, in the case of a load balancing operation, the UE may move from RAN node 1 to RAN node 2.

[0233] Step 8: RAN node 1 and RAN node 2 transmit feedback information to the network nodes.

[0234] Figure 16 illustrates a case where both the AI ​​Model Training function and the AI ​​Model Inference function are performed by RAN nodes (e.g., base stations, TRPs, CUs of base stations, etc.).

[0235] Step 1: UE and RAN node 2 transmit input data (i.e., training data) for AI model training to RAN node 1.

[0236] Step 2: RAN node 1 trains the AI ​​model using the received training data.

[0237] Step 3: RAN node 1 receives input data for AI Model Inference (i.e., Inference data) from the UE and RAN node 2.

[0238] Step 4: RAN node 1 performs AI Model Inference using the received Inference data to generate output data (e.g., prediction or decision).

[0239] Step 5: RAN node 1, RAN node 2, and the UE (or 'RAN node 1 and the UE', or 'RAN node 1 and the RAN node 2') perform actions based on the output data. For example, in the case of a load balancing operation, the UE may move from RAN node 1 to RAN node 2.

[0240] Step 6: RAN node 2 sends feedback information to RAN node 1.

[0241] Figure 17 illustrates a case where the AI ​​Model Training function is performed by a RAN node (e.g., a base station, a TRP, a CU of the base station, etc.) and the AI ​​Model Inference function is performed by a UE.

[0242] Step 1: The UE transmits input data (i.e., training data) for AI model training to the RAN node. Here, the RAN node can collect data (e.g., UE measurements related to RSRP, RSRQ, SINR of the serving cell and neighboring cells, UE location, speed, etc.) from various UEs and / or from other RAN nodes.

[0243] Step 2: The RAN node trains the AI ​​model using the received training data.

[0244] Step 3: The RAN node distributes / updates the AI ​​model to the UE. The UE may also continue model training based on the received AI model.

[0245] Step 4: Receive input data (i.e., Inference data) for AI Model Inference from the UE and RAN nodes (and / or from other UEs).

[0246] Step 5: The UE performs AI Model Inference using the received Inference data to generate output data (e.g., prediction or decision).

[0247] Step 6: If applicable, the UE may send model performance feedback to the RAN node.

[0248] Step 7: The UE and RAN nodes perform actions based on the output data.

[0249] Step 8: The UE transmits feedback information to the RAN node.

[0250] AI / ML model performance monitoring

[0251] Below, we describe a method for monitoring and managing the performance of AI / ML models that perform temporal / spatial / frequency-domain CSI compression.

[0252] In NR Rel-18, AI / ML-based NR air interfaces and their key use cases, including CSI feedback enhancement, beam management, and positioning accuracy enhancement, were studied. Specifically, for CSI feedback enhancement, AI / ML models were used to reduce overhead and improve accuracy in CSI reporting. Further research is planned for NR Rel-19, with the following considerations for improving CSI feedback:

[0253] For CSI compression (two-sided model), improvements have been made to the trade-off between performance and complexity / overhead. For example, expanding spatial / frequency compression to spatial / temporal / frequency compression, cell / site-specific models, and considering prediction in addition to CSI compression (compared to Rel-18 non-AI / ML-based methods).

[0254] For CSI prediction (one-side models), research on the complexity associated with performance gains compared to existing non-AI / ML-based methods. For example, cell- and site-specific models could be considered to improve performance.

[0255] Thus, CSI feedback improvement can be broadly classified into two categories: CSI compression and CSI prediction, and the proposals described below can be mainly related to CSI compression.

[0256] Figure 18 is a diagram to explain the concept of AI / ML-based CSI compression.

[0257] Referring to Figure 18, CSI compression can follow a two-sided model (from an inference perspective), where each terminal and base station have their own AI / ML models and operate as a pair. For convenience, the terminal model is referred to as the AI-encoder, and the base station model is referred to as the AI-decoder.

[0258] The terminal uses measured / estimated channel information (e.g., raw channel matrix or precoder type channel information vector (e.g., eigen vector)) as input to the AI-encoder, generates output, quantizes it, and feeds it back to the base station as AI / ML-based CSI.

[0259] The base station dequantizes AI / ML-based CSI fed back from the terminal and uses it as input to the AI ​​decoder to recover the output CSI. This CSI process compresses the channel information that the terminal must send through AI / ML, reducing feedback overhead. This process is called CSI compression.

[0260] Figure 19 is a diagram to explain the concept of AI / ML-based CSI prediction.

[0261] Figure 19 illustrates a UE-sided model that performs model inference by equipping only the terminal side with an AI / ML model. The terminal can apply multiple historical measurements as AI / ML inputs and estimate / predict one or more future CSIs as AI / ML model outputs.

[0262] As mentioned above, as a technique to improve the accuracy of CSI compression and effectively reduce the overhead associated with it, in addition to the spatial / frequency domain utilized in the existing eType-II CSI codebook, the temporal domain, i.e., the time domain, can be additionally considered to configure CSI. At this time, a technique that can improve the compression efficiency of the corresponding channel information while also obtaining benefits in terms of overhead is being considered (this is called TSF-domain CSI compression). For example, in addition to the previously considered current channel information, past channel information (e.g., historical CSI) can be utilized as input to the AI / ML encoder / decoder to more efficiently reflect the channel characteristics. From an overhead perspective, since the current channel information can be expressed by the degree of change based on the previous and / or historical CSI (e.g., delta CSI), there is the benefit of reducing the feedback overhead in a series of signaling for the corresponding CSI.

[0263] Meanwhile, from the perspective of life-cycle management (LCM) and performance monitoring for AI / ML models, it is necessary to consider introducing a method that utilizes historical CSI.

[0264] For example, when an offline pre-trained AI / ML model is transmitted to a terminal and / or base station to perform CSI compression operations, a mismatch may occur between the training data and actual field data due to variables in various external situations after the AI / ML model is deployed in the actual operating environment. This may result in a deterioration in the final output performance when the AI / ML model is applied. To resolve this, a series of processes, such as updating the AI / ML model or changing to another (appropriate) AI / ML model, is life-cycle management, and performance monitoring of the AI / ML model can serve as a standard during this process.

[0265] However, if the performance of an AI / ML model for TSF (temporal, spatial, frequency)-domain CSI compression utilizing historical CSI deteriorates, it may be ambiguous whether (i) the performance deteriorates because the AI / ML model does not fit the currently deployed real-world environment, or (ii) the AI / ML model is suitable for the currently deployed real-world environment but the performance deteriorates due to a problem with the historical CSI. Problems with the historical CSI may include, but are not limited to, cases where the historical CSI does not properly reflect the current instantaneous channel characteristics, or cases where an error propagation problem occurs due to errors in some of the specific channel information in the past.

[0266] It is clear that the final output performance should be the standard for monitoring AI / ML models that perform TSF-domain CSI compression, but it is also necessary to design a performance monitoring method that simultaneously considers the intermediate performance and / or historical CSI of the model.

[0267] The standard for monitoring AI / ML model performance can be to determine whether the performance degradation of the AI / ML model is due to a degradation of intermediate performance, or whether the environment, scenario, and assumptions of the offline trained AI / ML model do not match the actual operating environment.

[0268] In addition, conditions for performing fine-tuning or AI / ML model update / transfer based on monitored performance are required. Here, fine-tuning refers to cases where the input data used for existing training / inference for the current AI / ML model is continuously used / utilized to update the model. Otherwise, the AI / ML model can be reset and updated. In this case, it is possible to update only the coefficients while keeping the model structure (e.g. number of layers / nodes) as is, or to update the model structure as well.

[0269] For example, if the final output performance is poor, immediately updating / transferring the AI / ML model can result in overhead / latency issues. If issues arise in the intermediate performance and / or historical CSI, fine-tuning the performance of the intermediate module through fine-tuning may potentially improve the final output performance. Therefore, fine-tuning may be more appropriate than updating / transferring the AI / ML model. If the intermediate performance is good, it may be because the currently applied AI / ML model itself is not suitable for the real environment, resulting in poor final output performance. Therefore, performing a model update / transfer may be a more appropriate action.

[0270] In this specification, the performance of historical CSI may be interpreted to mean at least some of the following:

[0271] - Channel estimation accuracy: Channel estimation accuracy can be expressed based on the degree to which information such as channel information, including CQI, PMI, and RI, as well as second order statistics for the channel, change above / below a certain level depending on the instance(s).

[0272] - Channel prediction accuracy: Even if channel information such as CQI, PMI, and RI change depending on the instance(s), prediction accuracy can be expressed based on the degree of similarity between the channel information predicted using historical CSI and the channel information through current measurement.

[0273] Below, we propose solutions to resolve the ambiguity surrounding the underlying cause of AI / ML model performance degradation for TSF-domain CSI compression. For example, monitoring final output performance alone cannot determine whether the AI / ML model is unsuitable for the currently deployed environment, or whether the AI / ML model is suitable for the currently deployed environment but exhibits historical CSI issues. Therefore, we propose solutions to address this ambiguity.

[0274] According to one embodiment of the present disclosure, a two-step or multi-metric-based performance monitoring method is proposed for performance monitoring of AI / ML TSF-domain CSI compression, which first checks the performance of historical CSI as intermediate performance. Using this method, the signaling / processing overhead required to update or transfer AI / ML models when performing LCM through performance monitoring can be reduced. Furthermore, the efficiency and robustness of TSF-domain CSI compression itself can be improved through management of historical CSI.

[0275] Proposal 1. AI / ML Performance Monitoring Considering TSF-domain CSI Compression

[0276] For example, a performance monitoring entity (at UE-side or NW-side) of a model can determine an X value related to the final output performance of the AI / ML model (wherein the X value can be obtained through comparison with a specific metric and / or ground truth CSI, or can be a value signaled to the terminal by the network).

[0277] (1) Case 1: If the X value is above a certain level (e.g. X=0.8 or 80%), the current AI / ML model is maintained.

[0278] (2) Case 2: If the X value is below a certain level, the performance related to historical CSI can be evaluated based on at least one of the following options. For example, the performance monitoring entity may be a NW-side monitoring entity.

[0279] (i) Option 1: If the X value is below a certain level and the network does not signal to the UE that the output performance has degraded.

[0280] Figure 20 is a diagram illustrating the operation of a terminal and a network according to Option 1. The contents described in Figure 20 are not limited to Option 1 and may also be applied to other options / cases / proposals depending on the embodiment.

[0281] Referring to FIG. 20, the network can signal information to the terminal that output performance has deteriorated (1905).

[0282] Meanwhile, the network may indicate mode information for historical CSI processing during the signaling and / or through separate signaling. The mode information may include, for example, at least one of the following:

[0283] Mode 1: This mode performs CSI compression using only instantaneous channel input, without reflecting the currently applied historical CSI. For example, the terminal can exclude data related to historical CSI from data input to an AI / ML model.

[0284] Mode 2: This mode performs CSI compression by accumulating historical CSI from the present or a specific point in time, without reflecting the currently applied historical CSI. For example, a terminal can stop using historical CSI and replace it with CSI accumulated from the present.

[0285] - Mode 3: CSI compression can be performed by excluding significantly outdated historical CSI from the currently applied historical CSI. For example, the terminal can reduce the time range associated with past CSI.

[0286] The terminal may transmit an ACK for the signaling to the network (1910).

[0287] After ACK, the terminal can report output information of the encoder (e.g., AI / ML model of the terminal) configured according to the above mode instruction as a CSI report (1920).

[0288] The network can reconstruct CSI and monitor AI / ML model performance (1925). For example, the network can decode the output of an encoder (e.g., an AI / ML model on the terminal) through its own AI / ML model and, based on this, determine whether performance degradation related to historical CSI is occurring.

[0289] (ii) Option 2: If the X value is below a certain level and the network signals to the UE that the output performance has degraded.

[0290] For example, the network / terminal can determine the value of X through the Option 1 operation described above and act as one of the following candidates.

[0291] - Candidate 1) If the X value is above a certain level, it operates as Case 1.

[0292] - Candidate 2) If the X value is (still) below a certain level, it is not a problem with the historical CSI, but rather the AI / ML model itself is judged to be unsuitable for the currently deployed environment, so follow-up actions related to LCM (e.g. model fine-tuning / update) are performed.

[0293] Meanwhile, the above-described X-value-based operation can be performed by applying one or more of the following criteria to Case 1 and / or Case 2.

[0294] - Performs an action whenever the corresponding X value is instantaneously above / below a certain level.

[0295] - Performs an action when the corresponding case occurs n times during a specific time duration / window.

[0296] - If the corresponding case occurs m times in a row, the action is performed.

[0297] Proposal 1 is a performance monitoring method for TSF-domain CSI compression using an AI / ML model. The monitoring entity performing performance monitoring can be either UE-side or NW-side. It uses specific metrics and / or ground truth CSI information from model output to determine whether the model is functioning properly.

[0298] If the performance monitoring entity is NW (network)-side, performance can be checked based on the model output without separate signaling, so the X value can be judged on the NW-side and then actions can be taken according to the case above.

[0299] When the performance monitoring entity is UE-side, the following methods can be used for the UE to know the model output performance in the network. (i) In the simplest example, the network can signal the output result to the UE. Or (ii) the reconstruction model of the NW-side can be transmitted to the UE, and the UE can perform performance monitoring by extracting the estimated model output based on it. Or (iii) the UE can perform performance monitoring by calculating the model output (without signaling from the network) through a proxy model for the NW-side model it has. If the similarity in terms of input distribution between the historical CSI and the current channel information is below a certain level or the difference in the X value through a certain metric is determined to be above a certain level, the UE can also operate by applying the proposed method above to the encoder-decoder configuration between the UE and the base station through signaling in an event-triggered manner.

[0300] In this way, when the monitoring entity obtains the X value for the model output, it can perform subsequent actions in monitoring according to the level of the value.

[0301] For example, if the current X value is below a certain level, as in Option 1 of Case 2, and the network has not signaled to the UE that the performance is poor, it is necessary to first determine whether there is a problem with the historical CSI currently applied to the AI / ML model. For example, the NW-side monitoring entity can signal to the UE that the output performance is below a certain level (1905). In addition, the impact of the historical CSI can be confirmed together with the signaling or through a separate signaling, and a mode for processing the historical CSI as input data can be indicated.

[0302] For example, in the case of Mode 3, channel information with a large degree of outdation is excluded from the accumulated historical CSI. For example, the oldest i historical CSIs are excluded, specific i historical CSIs are selectively excluded, or a new time window for extracting historical CSIs is set based on the present.

[0303] As another example of Mode 3, if the terminal performs monitoring for separate historical CSI (e.g., data distribution based monitoring), and some of the data distributions of the historical CSIs are different from the latest CSI measurement (e.g., channels outside the coherent time window), the historical CSIs with different distributions (e.g., channels outside the coherent time window) can be input to the AI / ML model with zero-padding or truncated.

[0304] For example, a terminal may report to a base station a value that can represent the correlation of historical CSIs, e.g., correlation or average value, and the base station may utilize the reported value for mode selection, etc.

[0305] For example, a signal from the NW-side that performance has degraded can be mapped (implicit / explicit) to a specific mode and in a preset manner, and the specific mode mapped to it can be indicated (implicit / explicit) through that signaling.

[0306] The UE sends an ACK (1910) for the corresponding signaling, and performs encoding of the UE's channel information based on the application of the same mode between the UE and the network. The encoder output thus generated is reported to the network (1920), and the base station can decode the encoder output to produce the CSI output and the corresponding X value.

[0307] Even after performing changes / updates from a historical CSI perspective, if it is still option 2 in Case 2 and the X value is still below a certain level (e.g. candidate 2), the network / terminal can determine that the deployed environment and the AI / ML model do not fit well and perform LCM operation. For example, the network / terminal can perform fine-tuning for the AI / ML model or switch / transfer to another model.

[0308] In the above Cases, when the X value is judged to be above or below a certain level, this can be performed based on the instantaneous X value, or in order to prevent excessive fluctuation in terms of robustness for the monitoring operation, the Case can be operated when a certain level of the X value appears m times in a row or n times during a certain time duration / window.

[0309] Meanwhile, considering the target CSI of CSI compression and whether to utilize past CSI information of the UE-side and / or NW-side, TSF-domain CSI compression can be divided into five detailed cases as shown in Table 6 below.

[0310] CaseTarget CSI slot(s) Whether the terminal uses past CSI information Whether the network uses past CSI information 0Present slotNoNo1Present slotYesNo2Present slotYesYes3Future slot(s)YesNo4Future slot(s)YesYes5Present slotNoYes

[0311] Past CSI information on the UE side may include past model inputs (e.g., past CSI measurements) and / or any information derived therefrom. Past CSI information on the NW side may include past CSI feedback instances and / or any information derived therefrom.

[0312] For Cases 3 and 4, the terminal can perform prediction as a separate step or with compression. Similarly, the network can perform prediction as a separate step or with reconstruction.

[0313] The Target CSI Slot may refer to the slot corresponding to the reported CSI feedback. The Present Slot may refer to the most recent CSI-RS measurement slot used to generate the CSI report. The Future Slot may include one or more slots after the current slot, and may also include the current slot.

[0314] In particular, in cases such as Case 3 / 4 where the target CSI slot(s) are future slot(s), the CSI information can be configured by considering both compression and prediction for the target future channel information and / or current channel information. In this case, the respective functions of compression and prediction can be encoded / decoded in a sequential manner by applying separate AI / ML models, or the CSI can be encoded / decoded in a manner in which compression and prediction are jointly performed through a single AI / ML model.

[0315] In this way, in Case 3 / 4, the impact of performing CSI compression and CSI prediction individually and performing CSI compression and CSI prediction in a joint manner needs to be considered from the life cycle management (LCM) perspective.

[0316] This is because data collection, training, monitoring, and / or model control aspects may differ depending on whether CSI compression and prediction are performed separately or jointly. Specifically, when considering model monitoring, if the final output performance corresponding to jointly performing CSI compression and CSI prediction degrades, it can be difficult to determine whether the problem occurred in CSI compression or CSI prediction.

[0317] For example, if compression and prediction for CSI are separate methods that consist of separate AI / ML models, you can identify and fine-tune and update the model after identifying any performance degradation for a specific model among compression and prediction based on the output results of each model.

[0318] On the other hand, in the case of joint CSI compression and prediction, since a single AI / ML model performs both functions simultaneously, it may be difficult to determine the performance degradation of compression and prediction based solely on the final output of a single AI / ML model. For example, monitoring the final output performance cannot determine whether the AI / ML model is unsuitable for the currently deployed environment, or whether the AI / ML model is suitable for the currently deployed environment but has problems in historical CSI (e.g., problems with past CSIs used as input data sets for CSI compression), or which function of the two functions caused the problem when performing compression and prediction.

[0319] Therefore, a model monitoring method is needed to resolve this ambiguity problem. In one embodiment of this specification, for performance monitoring of AI / ML TSF-domain CSI compression and prediction, 1) the performance of historical CSI (e.g., past CSIs used as input data sets for CSI compression) is first checked / addressed as intermediate performance, and then, if necessary, 2) a multi-metric-based performance monitoring method is proposed according to the functions of compression and prediction. Using this method, the signaling / processing overhead required to update or transfer AI / ML models when performing LCM through performance monitoring can be reduced. In addition, the efficiency and robustness of TSF-domain CSI compression and prediction itself can be improved by managing historical CSI and compression / prediction functions.

[0320] Proposal 1-1. Performance monitoring considering joint CSI compression and prediction in the AI / ML TSF domain.

[0321] A model monitoring entity (e.g., entity at UE-side or NW-side) can determine an X value related to the final output performance of the AI / ML model. Here, the X value may be obtained through comparison with a specific metric and / or ground truth CSI, or may be a value signaled to the terminal by the network. Alternatively, the X value may be obtained based on a proxy model at the UE-side, or may be obtained based on a reconstruction model if the reconstruction model is transferred to the UE-side via network signaling. Alternatively, the X value may be a value related to the (estimated) output performance in other ways.

[0322] For example, the X value in CSI compression and prediction may be expressed as a single CSI value for present and / or future slot(s), or as a CSI value for each slot.

[0323] For example, the estimated output performance of compression and prediction for one present slot (at t0) and two future slots (at t1 and t2) can be a single value computed based on a (weighted) average of the estimation accuracies for the three slots under consideration, or the estimation accuracies expressed for each slot (e.g., X = {X_t0, X_t1, X_t2}).

[0324] The X value may not necessarily be just one value, and for example, multiple X values ​​(e.g. X1, X2, etc.) may be used.

[0325] Alternatively, a pair of output performance P1 in the current slot related to compression performance and output performance P2 in the future slot(s) related to prediction performance (e.g. weighted average depending on prediction window size) may be used. For example, when a single X value is represented by 2 bits, the 2 bits may be mapped to codepoint '00' if both P1 performance and P2 performance are above a certain threshold, '01' if only P1 performance is satisfied, '10' if only P2 performance is satisfied, and '11' if all are below the threshold, and the operations described below may be performed. However, the X value being represented by 2 bits is just one example and may be composed of multiple 3-bits or more.

[0326] According to one embodiment, the UE / NW side Model monitoring entity may maintain the current AI / ML model if the performance of the AI / ML model is above a certain level (e.g., threshold), but may first check the performance related to historical CSI (e.g., historical CSIs as an input data set for CSI compression) if the performance is below the certain level. In other words, the Model monitoring entity may first check whether the performance degradation is caused by historical CSI data used in CSI compression before checking for a problem with the CSI compression / CSI prediction function. The certain level may be, but is not limited to, a value set in advance, based on network signaling, or a parameter determined based on the channel environment. Case 1a and Case 2a are described as examples of such operations.

[0327] (1) Case 1a: If the X value is above a certain level, or if there are n or more X values ​​above a certain level among the X values ​​for each of multiple slots, the current AI / ML model can be maintained. For example, assuming a certain level is 0.8, if X is above 0.8, the current AI / ML model can be maintained. Alternatively, if more than 80% of the X values ​​are above a certain level, the current AI / ML model can be maintained.

[0328] (2) Case 2a: If the X value is below a specific level or there are n or more X values ​​below a specific level among the X values ​​for each of multiple slots, the Model monitoring entity can check the performance related to historical CSI. The performance related to historical CSI can be checked based on at least one of the following options. The Model monitoring entity may be, but is not limited to, an NW-side monitoring entity.

[0329] Meanwhile, among the X values ​​for each of the plurality of slots, n(>1) X values ​​may include an X value indicating performance for the present slot, or may be composed only of X values ​​indicating performance for the future slot(s). For example, information about the composition of the n X values ​​may be predefined or set through network signaling.

[0330] Meanwhile, when an AI / ML model's performance is assessed as falling below a certain level, terminal / network behavior may vary depending on whether this performance degradation is the first to occur or whether there was a prior performance degradation signaling the degradation. For example, terminal / network behavior may be as in, but not limited to, options 1a / 2a below.

[0331] 1) Option 1a: If n or more of the X values ​​or X values ​​per slot are below a certain level and the network does not signal to the terminal that the output performance has deteriorated.

[0332] Referring again to Fig. 20, the operation of the network and terminal according to Option 1a of Proposal 1-1 is described. Fig. 20 is not limited to Option 1a of Proposal 1-1 and may be referenced for other embodiments.

[0333] Referring to FIG. 20, the network can signal information to the terminal that the performance of the AI / ML model has deteriorated (1905).

[0334] For example, information indicating a mode for the output of an AI / ML model may be provided through signaled information or separately from the signaling above. The indicative modes may include, but are not limited to, at least one of the following:

[0335] - Mode 1: This mode performs CSI compression using only instantaneous channel input, without reflecting the currently applied historical CSI. For example, when CSI compression is being performed by inputting both historical CSI and instantaneous (current) channel data into an AI / ML model, Mode 1 can be designated, which no longer applies historical CSI and only uses instantaneous (current) channel data.

[0336] - Mode 2: This is a mode for performing CSI compression and prediction by newly accumulating historical CSI from the present or a specific point in time without reflecting the currently applied historical CSI. For example, when CSI compression is being performed by inputting historical CSI and instantaneous (current) channel data together into an AI / ML model, Mode 2 can be designated to flush / discard data for accumulated historical CSI, newly accumulate CSI from the present or a specific point in time, and use the newly accumulated CSI as historical CSI. For example, when Mode 2 is designated, only the current instantaneous (current) channel data is temporarily used until new historical CSI data is accumulated, and thereafter, CSI compression can be performed by using both newly accumulated historical CSI data and current instantaneous (current) channel data from a specific point in time.

[0337] - Mode 3: This mode performs CSI compression and prediction by excluding significantly outdated historical CSI among the currently applied historical CSI. For example, rather than flushing / discarding all accumulated historical CSI, CSI compression can be performed by partially selecting historical CSI and flushing / discarding it, and using the remaining historical CSI together with the current instantaneous (present) channel data. For example, the historical CSI to be excluded can be determined by the terminal or indicated by the network. If the terminal determines the historical CSI to be excluded on its own, the terminal can report the historical CSI to be excluded to the network (e.g., when reporting CSI), and the network can exclude the historical CSI from the input of its AI / ML model when reconstructing (restoring) CSI (e.g., 1925).

[0338] The terminal can transmit an ACK for signaling from the network (1910).

[0339] The terminal can perform inference by applying the indicated mode to the AI / ML model (CSI encoder AI / ML model).

[0340] The terminal can transmit the output of the AI / ML model (CSI encoder AI / ML model) obtained based on the mode through a CSI report (1920).

[0341] The network decodes the corresponding encoder output and can monitor performance based on this (1925). The network can reconstruct (restore) the CSI reported by the terminal through the CSI decoder AI / ML model. When reconstructing (restores) the CSI, the network can utilize the mode information previously indicated to the terminal. For example, if the past CSI indicates unused Mode 1, the network can perform CSI restoration without inputting the past CSI into the CSI decoder AI / ML model for CSI restoration. Alternatively, if the past CSI information is updated or some of the past CSI is excluded, such as in Mode 2 / 3, the network can perform CSI restoration by updating or excluding the past CSI information.

[0342] Network monitoring entities can monitor performance related to historical CSI based on restored CSI and determine whether performance has deteriorated. Since the network has at least partially deprecated or updated historical CSI through previously indicated modes, monitoring entities can determine whether performance has deteriorated or improved based on performance monitored after the mode instruction.

[0343] For example, a network monitoring entity can check the X value through monitoring. If the X value or n or more of the X values ​​per slot are above a certain level, the current AI / ML model settings can be maintained according to Case 1a.

[0344] 2) Option 2a: If n or more of the X values ​​or X values ​​per slot are below a certain level, and the network signals to the terminal that the output performance has deteriorated.

[0345] For example, option 2a described below can be performed as a follow-up operation to option 1a described above.

[0346] Even after signaling to the terminal that output performance has degraded via Option 1a, monitored output performance may still fall below a certain level. For example, if the monitoring entity finds that n or more of the X values ​​or X values ​​per slot are (still) below a certain level, this may indicate a problem with the compression and / or prediction capabilities of the AI / ML model, rather than a problem with historical CSI.

[0347] Referring again to Fig. 20, the operation of the network and terminal according to Option 2a of Proposal 1-1 is described. Fig. 20 is not limited to Option 2a of Proposal 1-1 and may be referenced for other embodiments. For example, after the procedure illustrated in Fig. 20 is performed once for Option 1a, the procedure illustrated in Fig. 20 may be performed once more for Option 2a. In this case, the detailed operations performed at each step of Fig. 20 may be configured differently for each option.

[0348] Referring to FIG. 20, the network can signal information to the terminal informing it of a degradation in output performance (1905).

[0349] For example, according to Option 2a, at least one of the following states regarding factor(s) affecting CSI compression / prediction can be set / indicated via the signaling or via a separate signaling:

[0350] - State 1: Compression only (eg prediction window size=0)

[0351] For example, State 1 may be indicated to perform only CSI compression without performing CSI prediction. As an example of a method for disabling CSI prediction, the size of the CSI prediction window may be set to 0. The CSI prediction window may be a set of future time points from the current time point at which CSI prediction will be performed.

[0352] - State 2: Prediction only

[0353] For example, State 2 may be indicated to perform only CSI prediction without performing CSI compression, as opposed to State 1.

[0354] For example, in State 2, CSI compression can be performed by setting values ​​such as compression ratio to the highest / lowest to focus on prediction-related performance, or the accuracy of CSI estimation can be improved by increasing the reported feedback payload (e.g., reporting uncompressed, high-accuracy CSI).

[0355] - State 3: Compression and prediction

[0356] For example, in State 3, Compression and prediction are performed, but the settings for each can be changed.

[0357] For example, compression ratio and / or prediction window size for all slots or per slot can be set / indicated for State 3.

[0358] The terminal can send an ACK for signaling (1905) (1910).

[0359] Afterwards, the terminal can transmit encoder output information configured according to the indicated state instructions through a CSI report (1920).

[0360] The network can reconstruct / decode the CSI reported by the terminal using the network's CSI encoder AI / ML model (1925). The network's monitoring entity can then assess the performance of the joint CSI compression and prediction AI / ML model based on the reconstructed / decoded CSI. For example, the network's monitoring entity can monitor performance using existing monitoring metrics, or it can monitor performance based on a separate metric specific to a given state.

[0361] The terminal / network can then perform subsequent actions based on the monitored performance metrics.

[0362] For example, if the performance level exceeds a certain level, the terminal can continue operating in that state. Alternatively, the terminal can operate in that state, but perform a CSI report using the compression ratio and prediction window set before the state at specific intervals / instances. If the performance(s) exceed a certain level, the terminal can fall back to the previous state. Alternatively, performance metrics can be measured while changing states according to a predefined or directed method.

[0363] If the performance level is below a certain level, the AI / ML model is determined to be unsuitable for the currently deployed environment, and subsequent LCM-related actions (e.g. model transfer, etc.) may be performed.

[0364] Although the above examples have focused on monitoring performance in the network, the present disclosure is not limited thereto, and performance monitoring may be performed at the terminal. When the terminal performs monitoring, the terminal may operate by determining whether n or more of the corresponding X values ​​or X values ​​per slot are below a specific level. For example, based on signaling using an event-triggered method (e.g., MAC-CE, DCI, SR PUCCH, etc.), whether there is a problem with past CSI between the terminal and the base station and the state of the CSI configuration may be indicated, and subsequent operations may be performed based on this. In addition, signaling using an event-triggered method may replace the network signaling in the above Option 1a / 2a, in which case the UE / NW-side confirmation operation using ACK may be omitted.

[0365] Meanwhile, for Case 1a and / or Case 2a or for an operation performed based on n or more X values ​​among the X values ​​per slot, at least one of the following criteria may be applied.

[0366] - The corresponding action can be triggered instantaneously for each event in which n X values(es) occur above / below a certain level.

[0367] - An action can be triggered when events are detected m times during a specific time duration / window. For example, the value of m can be predefined or set via network signaling.

[0368] - An action can be triggered when y consecutive events are detected. For example, the y value can be predefined or set via network signaling.

[0369] Below, more specific embodiments of the above-described proposal 1-1 are described.

[0370] Proposal 1-1 above describes a performance monitoring method for TSF-domain CSI compression and prediction using an AI / ML model. The monitoring entity performing performance monitoring can be located on either the UE side or the NW side. It uses specific metrics and / or ground truth CSI information from model output to determine whether the model is operating properly.

[0371] For NW-side monitoring entities, performance information can be obtained based on the model output it decodes (without separate signaling from the terminal). Therefore, after determining monitoring performance information / values ​​on the NW-side, actions can be taken according to cases 1a / 2a described above.

[0372] In the case of a UE-side monitoring entity, the following methods can be used for the terminal to know the final model output performance in the network. (i) In the simplest way, the network can signal the corresponding output result to the terminal. (ii) Alternatively, the network can transmit information about the NW-side reconstruction (decoder) model to the terminal, and the terminal can perform performance monitoring by extracting an estimated model output based on this. (iii) Alternatively, the terminal can perform performance monitoring by calculating the model output through a proxy model it has (without signaling from the network). When the terminal performs monitoring in this way, if the difference in similarity in terms of input distribution between the historical CSI and the current channel information exceeds a certain level, or if the difference in the performance monitoring value through a predetermined metric exceeds a certain level, the proposed method can be applied to the encoder-decoder configuration between the terminal and the base station through signaling in an event-triggered manner.

[0373] Through the above method, the monitoring entity can obtain the X value for the model output and perform subsequent actions based on the level of the value. The X value can be a single value considering the present and / or future slot(s), or a value expressed for each slot.

[0374] As in Option 1a of Case 2a, if the current X value or n or more of the X values ​​per slot are below a certain level and the network has not signaled to the terminal that the performance monitoring performance is poor, it is possible to first determine whether there is a problem with the historical CSI currently applied to the AI / ML model. At this time, the signaling on the NW side that the performance has deteriorated can be mapped to a specific mode and a preset method (implicit / explicit), and in this case, performing the signaling can have the same meaning as indicating a specific mode. The terminal that receives such signaling can send an ACK for the signaling to notify that the same mode is applied between the terminal and the network, and the terminal can perform the terminal's channel information encoding based on the mode. The encoder output generated in this way is reported to the network, and the network can decode the encoder output to produce the CSI output and the corresponding X value.

[0375] Based on this, even if changes / updates are performed from a historical CSI perspective, if n or more of the X values ​​or X values ​​per slot are still below a certain level, as in Option 2a in Case 2a, the monitoring entity can determine that there is no problem from a historical CSI perspective, but the AI / ML model for the joint CSI compression and prediction is not performing the compression and / or prediction functions properly. In this case, fine-tuning the AI / ML model or operations such as model transfer may be performed, but in order to maximize the use of the currently deployed model, it may be desirable to identify whether the model has a problem with compression or prediction and then perform compression and prediction based on the model within the maximum available range. This has the advantage of reducing the signaling overhead due to LCM while maintaining the output performance above a certain level. In a state where it is not known which factor between compression and prediction is the main cause of the performance degradation, the monitoring entity can obtain subsequent CSI reports by setting / instructing operations for each state as in the above proposal. A monitoring entity that obtains a CSI report can apply monitoring metrics appropriate to the state to determine whether there is a performance issue when operating in that state.Meanwhile, state instructions are not limited to a single state instruction, and pre-defined methods or state orders can be specified for multiple states. Performance monitoring can be performed by variably applying variables related to the performance of each state and the compression ratio / prediction window to be applied to the state. If performance is determined to be at a certain level when a specific state and / or a specific compression ratio / prediction window is applied, subsequent CSI reports can be generated / reported according to the state. Alternatively, the previously set compression and prediction before applying the state can be transmitted through a CSI report at a specific cycle or instance, or the previously set compression and prediction can be configured to be included in the CSI report according to the indicated state. Through this, it is possible to determine whether the AI / ML model operates properly in the preset method at that point in time. Accordingly, if necessary, the CSI report can be performed by falling back to the operation set in the original AI / ML model.

[0376] In the above cases, when an event occurs in which n or more of the X values ​​or X values ​​per slot occur, the corresponding action may be performed instantaneously, or in order to prevent excessive fluctuation in terms of robustness of the monitoring action, the corresponding action may be performed when events appear m times consecutively or when events occur y times during a specific time duration / window.

[0377] Meanwhile, in the TSF-domain CSI compression described above, we are considering a use case where CSI compression as well as CSI prediction can be performed simultaneously using an AI / ML model, as in Case 3. In particular, when utilizing CSI from a temporal-domain perspective, not only can the current target CSI be efficiently expressed based on past CSI information in terms of CSI compression, but future CSI can also be inferred using the past CSI. Therefore, it can be effective to consider compression and prediction simultaneously when considering the temporal-domain.

[0378] In terms of the temporal domain, the following options can be considered for LCM and training data collection for Case 3.

[0379] 1) Option 1: Target CSI for learning is calculated based on the predicted CSI of the future slot.

[0380] 2) Option 2: Target CSI for learning is calculated based on the measured CSI of the future slot.

[0381] Note: In the inference phase, the input of the CSI generation part is calculated based on the predicted CSI.

[0382] The following options may be considered for setting monitoring labels:

[0383] 1) Option 1: Monitoring labels are derived based on the predicted CSI of future slots. The CSI prediction output is used as input to the CSI generation unit. This applies only to monitoring CSI compression. CSI prediction can be monitored separately.

[0384] 2) Option 2: The monitoring label is calculated based on the measured CSI of the future slot.

[0385] - Option 2a: The CSI prediction output is used as input to the CSI generation unit. This corresponds to end-to-end monitoring of CSI prediction and compression.

[0386] - Option 2b: The measured CSI of the future slot is used as input to the CSI generator for monitoring purposes. This applies only to monitoring CSI compression. CSI predictions can be monitored separately.

[0387] In order for the NW (Network) to collect data for learning, the following needs to be considered:

[0388] - Data format: Codebook-based Rel-16 eType2 or Rel-18 eType2 for PMI prediction, number of samples in the report, whether channel information or precoder information is required for Temporal Case 3.

[0389] - Rank / Layer configuration and subband number configuration

[0390] - Ground-truth reporting mechanism

[0391] - Additional information reports on the sample, e.g. data quality, data quality definitions and related parameters.

[0392] - Whether improvements are needed for CSI-RS and SRS configurations

[0393] - Related information report that captures additional conditions on the terminal (UE) side

[0394] - Time domain aspect configuration and reporting for Temporal Cases 2 and 3, e.g., correlation between input and output CSI.

[0395] - Details of CSI measurements

[0396] When a UE collects data for learning, the following needs to be considered:

[0397] - NW configuration or UE request, e.g. RS configuration / transmission for data collection

[0398] - Whether improvements are needed for CSI-RS configuration

[0399] - Time domain aspect configuration for Temporal Case 2 / 3, e.g., correlation between input and output CSI.

[0400] - The need for ID configuration and how to configure ID

[0401] In the above proposal, when considering the model monitoring aspect in an operation such as case 3, it is ambiguous to diagnose which function caused the problem when the final output performance of jointly performing compression and prediction deteriorates. Therefore, as a representative solution to this problem, a root cause identification method was described that changes the operation to compression only or prediction only and then monitors again to diagnose which aspect the problem occurs.

[0402] At this time, when the performance of either compression or prediction is degraded, before performing model update / switch / transfer, other measures to improve monitoring performance may be needed, such as data collection aspects (e.g., model training / fine-tuning based on the quality of sample data), or changes in the ground truth label settings corresponding to changes in the compression ratio and prediction window due to degraded monitoring performance.

[0403] Considering the TSF-domain CSI compression and prediction method through the AI / ML model as described above (e.g., Case 3 / 4), when the performance of the AI / ML model deteriorates from the perspective of performance monitoring for the model, it is possible to identify which aspect of CSI compression / prediction is causing the performance degradation and, based on this, perform a model update to improve the performance of the model. At this time, we propose an operation method that considers the quality of the (training) data and a method for changing the association between the input data and the output data (directly / indirectly related to the compression ratio and / or prediction window) and the related ground truth label settings.

[0404] This approach not only reduces the signaling / processing overhead required to update or switch / transfer AI / ML models, but also improves the efficiency and robustness of TSF-domain CSI compression and prediction itself through management of historical CSI and compression / prediction functions.

[0405] Proposal 1-2. A method for performing LCM operations of a model based on the quality of sample data when monitoring the performance of joint CSI compression and prediction operations in the AI / ML TSF domain.

[0406] (1) Quantification of sample data quality

[0407] For example, at least one of the following may be used to quantify the quality of sample data.

[0408] (i) (Prediction) confidence information: Confidence information can be calculated based on the difference between the predicted CSI and the actual CSI.

[0409] (ii) Compression reconstruction error: After restoring compressed CSI, compression reconstruction error information can be calculated based on the difference between the restored CSI and the actual CSI.

[0410] (iii) Temporal consistency information: Temporal consistency information can indicate the extent to which current and past CSI follow a continuous pattern. For example, information on the mutual continuity / correlation between CSI information at each time point in the time domain can be used to quantify the quality of sample data.

[0411] (iv) Feedback dependency information: The extent to which the model contributed to reducing CSI feedback overhead.

[0412] (v) combination of (i) to (iv) (e.g., combination of feedback overhead vs. prediction / compression accuracy)

[0413] (2) Apply sample data during subsequent model (re-)training based on at least one of the above quality information (i) to (iv).

[0414] Based on the results of quantifying the quality of sample data as above, samples with a quality above a certain threshold can be used.

[0415] For example, only sample(s) whose quality(s) exceed a certain threshold can be used to train an AI / ML model.

[0416] If the value of the quality(s) is below a certain threshold, action can be taken based on at least one of the following:

[0417] (i) If the (Prediction) confidence falls below a certain level, the terminal / network can adjust the prediction / measurement window based on the degree of the corresponding quality value. Alternatively, the network can request direct CSI feedback from the terminal for each instance.

[0418] (ii) If the compression reconstruction error exceeds a certain level, the terminal / network may adjust the compression ratio or apply additional compensation depending on the quality value. Alternatively, the network may request uncompressed CSI feedback from the terminal instead of compressed CSI.

[0419] (iii) If temporal consistency falls below a certain level, the reliability of CSI predictions decreases, so the terminal / network can adjust the measurement duration / window based on the level of the corresponding quality value. Alternatively, the network can directly request CSI feedback from the terminal.

[0420] (iv) If feedback dependency falls below a certain level, it may indicate that the AI / ML model's feedback overhead mitigation effect is reduced. Therefore, the terminal / network may fall back to the existing Rel-16 / 18 codebook operation, or apply the AI / ML model up to a certain number of future instances based on the current CSI, and then use the existing codebook for feedback only for subsequent instances.

[0421] Meanwhile, in cases where multiple or a combination of the above-described quality information (i) to (iv) is used, operation may be performed based on a final single quality value obtained by applying a weight to each quality and then adding them together.

[0422] Additionally, the quality can be calculated on a sample-by-sample basis or on an instance-by-instance basis.

[0423] (3) Method of transmitting quality information of sample data

[0424] The terminal can transmit quality(s) information of sample data to the network, and for this purpose, at least one of the following channels can be used.

[0425] 1) PUCCH-based: While it enables rapid exchange of quality information, it may have limitations in conveying detailed quality information. The payload capacity available via PUCCH may also limit the size of the quality information that can be transmitted. In this case, the UE can prioritize the transmission of prediction-related information, such as (prediction) confidence and temporal consistency. For example, prediction-related information, such as (prediction) confidence and temporal consistency, can be given higher priority and transmitted first compared to other data quality information and / or information that can be conveyed via PUSCH.

[0426] The terminal can request fallback operation via the PUCCH based on the corresponding quality value information. Alternatively, the terminal can request fallback operation via the PUCCH by including a separate indicator.

[0427] 2) PUSCH-based: Can transmit detailed quality(s) information, but UCI overhead may increase.

[0428] When transmitted via PUCCH, PUSCH may include information about quality(s) related to compression and its detailed adjustments (e.g. compression ratio, window adjustment related settings, etc.).

[0429] Additionally, the terminal may transmit sample data (e.g. CSI for ground truth label purposes) corresponding to the quality value through PUSCH.

[0430] 3) Adaptive CSI feedback-based

[0431] Based on CSI / sample quality information, the terminal can directly decide whether to perform CSI feedback or use / update / change the corresponding AI / ML model.

[0432] For example, if the predicted CSI has high reliability, the CSI feedback (for ground truth label purposes) for the corresponding future instance(s) may be omitted. In this case, the terminal may transmit an indicator indicating that the current CSI feedback for the future instance is omitted.

[0433] (4) Method for setting / adjusting ground truth labels (related to the above quality values)

[0434] The terminal / network can perform subsequent monitoring and model (re-)training by adjusting the compression / prediction level of the model based on the degree of performance degradation, and dynamically adjust the association between input and output CSI. Furthermore, the terminal / network can also adjust the ground truth label for the future instance.

[0435] The terminal / network may act according to at least one of the following, depending on the level of quality value and / or monitoring performance degradation:

[0436] 1) If it is determined that there is a problem from a compression perspective, the terminal / network can adjust the compression ratio and use it as model input, and the corresponding ground truth label can also be adjusted or added to match the compression ratio.

[0437] 2) If a problem is determined from a prediction perspective, the terminal / network can determine a correction value that ensures the prediction has a certain level of quality or higher, or a time-domain correction value that will remain valid for a certain number of future instances, based on prediction-related confidence and / or consistency information. Based on the correction value, the terminal / network can then modify the corresponding instance of the ground truth label and / or the measurement duration / window.

[0438] 3) If both perspectives are determined to be problematic, the terminal / network may perform at least one of model retraining using the original CSI, ground truth label correction corresponding to the corresponding quality(s) values, and / or model fallback / switch.

[0439] A two-bit signal can be used to indicate a problematic status. For example, '00' could mean a compression issue, '01' could mean a prediction issue, '10' could mean issues on both sides, and '11' could mean no issues on either side, but is not limited to these.

[0440] Below, we describe more specific examples of proposals 1-2 described above.

[0441] Proposal 1-2 refers to a method of performing performance monitoring and LCM operation based on the quality(s) value of data samples during joint CSI compression and prediction considering the TSF-domain.

[0442] For example, the quality(s) described above may be determined for each sample or for each corresponding time instance of a sample. Signaling and required information may vary depending on the quality value and / or the level of performance degradation when monitoring performance degrades.

[0443] Additionally, signaling and required information can be differentiated depending on whether the monitoring entity is a terminal or a base station.

[0444] When the terminal is the subject of performance monitoring, it can detect model performance degradation and report it to the base station, allowing the base station to apply an optimized CSI feedback method. For example, if the level of performance degradation is marginal, the current AI / ML model-based CSI compression and prediction can be maintained, while quality information for samples can be reported via PUCCH. On the other hand, if the degree of performance degradation exceeds a certain level, the compression ratio and / or the duration / window of the measurement / prediction for prediction can be changed, and subsequent monitoring can be performed. In this case, detailed CSI quality reporting based on PUSCH can be performed, including association information for input CSI / output CSI and configuration changes for the corresponding ground truth label.

[0445] When the base station is the subject of performance monitoring, the terminal can support the base station's performance monitoring by transmitting metrics related to CSI quality, etc., via PUCCH. Alternatively, the base station can analyze CSI compression / prediction performance and transmit association change information for input-output CSI to the terminal. The terminal can also determine whether to (re-)train the model and notify the base station. Alternatively, the base station can determine whether to change the AI / ML model and fallback and transmit this information to the terminal via RRC messages.

[0446] Proposal 2. Multi-metric performance monitoring considering AI / ML TSF-domain CSI compression.

[0447] In Proposal 2, we propose a performance monitoring method based on multiple performance metrics when performing performance monitoring for TSF-domain CSI compression (rather than a sequential method that first checks historical CSI and then monitors output performance).

[0448] A monitoring entity (at UE-side or NW-side) can determine the X value related to the final output performance of the AI / ML model and the Y value related to the historical CSI performance (where X and Y values ​​can be obtained through comparison with specific metrics and / or ground truth CSI).

[0449] (1) Case A: If the values ​​of X and Y are above a certain level (e.g., X=0.8 or 80%, Y=0.8 or 80%), the current AI / ML model can be maintained. The specific levels (thresholds) for the values ​​of X and Y can be the same or different.

[0450] (2) Case B

[0451] (i) Case B-1: If the X value is above a certain level or the Y value is below a certain level, the monitoring entity signals that the performance related to the historical CSI has deteriorated and can operate according to Option 1 of Case 2 of Proposal 1 described above. Through this, the level of the Y value can be improved, which is expected to improve the final output performance.

[0452] (ii) Case B-2: If the X value is below a certain level but the Y value is above a certain level, the historical CSI characteristics are calculated correctly, but there may be a problem from a decoding perspective or the model may not be suitable, so the terminal / network may perform LCM operation but may perform fine-tuning or update of the NW-side decoder as a priority.

[0453] Since this method infers the characteristics of the channel input from the UE well, it can be used as is in subsequent model fine-tuning / update to reduce signaling overhead, or it can be advantageous in improving the efficiency of LCM operation by performing fine-tuning / update for the decoder stage first rather than performing fine-tuning / update for both the encoder and decoder.

[0454] (3) Case C: If both X and Y values ​​are below a certain level, the network / terminal may determine that this is not only a problem with historical CSI, but also that the AI / ML model itself is not suitable for the currently deployed environment, and may perform follow-up actions related to LCM (e.g., model fine-tuning / update).

[0455] Meanwhile, in order to perform an operation based on the X and Y values ​​in Case B-1, Case B-2 and / or Case C, at least one of the following criteria may be applied.

[0456] - It operates whenever the corresponding X, Y values ​​are instantaneously above / below a certain level.

[0457] - Action is taken when the corresponding case occurs n1, n2 times during the same / different time duration / window for the corresponding X, Y values.

[0458] - Action is taken when the corresponding Case occurs consecutively the same / different number of times for the corresponding X and Y values.

[0459] Proposal 2 is a method that considers multiple metrics in the monitoring entity. For example, one of the two monitoring metrics may be for monitoring the performance of historical CSI based on the encoder output from the UE, and the other may be for monitoring the output performance by decoding the corresponding encoder output. Accordingly, multiple values ​​for each metric may exist, such as X and Y, corresponding to each metric and / or including multiple metrics. Similar to Proposal 1, cases may be distinguished and operated based on the level of the corresponding X and Y values. In this case, the X and Y values ​​may be set to be the same or different. Cases are distinguished based on the level of the X and Y values, and the operation according to the case can simultaneously check the impact on historical CSI and output performance, enabling more efficient performance monitoring and LCM operation design than a sequential performance verification method.

[0460] Figure 21 illustrates a flow of a method performed by a terminal in a wireless communication system according to one embodiment.

[0461] Referring to FIG. 21, the terminal can obtain information about predicted CSI and information about compressed CSI based on an AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction (2105).

[0462] The terminal can transmit a CSI report based on information about the predicted CSI and information about the compressed CSI (2110).

[0463] Based on the performance of the AI / ML model being below a threshold, the terminal may adjust past CSIs used as an input data set of the CSI compression before changing at least one of the settings for the CSI compression or the settings for the CSI prediction of the AI / ML model.

[0464] After adjusting the above past CSIs, based on the performance of the AI / ML model being above the threshold, the terminal can maintain the AI / ML model without changing the settings for the CSI compression or the settings for the CSI prediction.

[0465] Even after adjusting the past CSIs, based on the performance of the AI / ML model being below the threshold, the terminal may change at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0466] The terminal can determine whether the CSI compression causes a performance degradation of the AI / ML model or whether the CSI prediction causes a performance degradation of the AI / ML model by maintaining one of the settings for the CSI compression and the settings for the CSI prediction while changing the other.

[0467] The terminal may reconfigure the AI / ML model or switch to another AI / ML model based on the performance of the AI / ML model being below the threshold even after changing at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0468] The adjustment of the past CSIs may be performed based on at least one of a first mode for excluding all of the past CSIs from the input data set of the CSI compression; a second mode for replacing the past CSIs with CSIs accumulated from the present; and a third mode for selectively excluding some of the past CSIs from the input data set of the CSI compression.

[0469] The terminal may receive network signaling indicating that the performance of the AI / ML model is below the threshold. The network signaling may indicate at least one of the first mode, the second mode, and the third mode.

[0470] The terminal may determine whether the performance of the AI / ML model is below a threshold based on a second AI / ML model for CSI reconstruction or a proxy model for the second AI / ML model.

[0471] Figure 22 illustrates a flowchart of a method performed by a base station in a wireless communication system according to one embodiment.

[0472] The base station can receive CSI reports through the terminal's AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction (2205).

[0473] The base station can obtain information about predicted CSI and information about compressed CSI based on the above CSI report (2210).

[0474] Based on the performance of the AI / ML model being below a threshold, the base station may instruct the terminal to adjust past CSIs used as an input data set of the CSI compression before changing at least one of the settings for the CSI compression or the settings for the CSI prediction of the AI / ML model.

[0475] After adjusting the above past CSIs, the AI / ML model can be maintained without changing the settings for the CSI compression or the settings for the CSI prediction based on the performance of the AI / ML model being above the threshold.

[0476] Even after adjusting the past CSIs, if the performance of the AI / ML model is below the threshold, the base station may instruct the terminal to change at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0477] By maintaining one of the settings for the CSI compression and the settings for the CSI prediction while changing the other, it can be determined whether the CSI compression causes a performance degradation of the AI / ML model or whether the CSI prediction causes a performance degradation of the AI / ML model.

[0478] The base station may instruct the terminal to reconfigure the AI / ML model or switch to another AI / ML model based on the performance of the AI / ML model being below the threshold even after changing at least one of the settings for the CSI compression or the settings for the CSI prediction.

[0479] The adjustment of the past CSIs may be performed based on at least one of a first mode for excluding all of the past CSIs from the input data set of the CSI compression; a second mode for replacing the past CSIs with CSIs accumulated from the present; and a third mode for selectively excluding some of the past CSIs from the input data set of the CSI compression.

[0480] The base station may transmit network signaling to the terminal, indicating that the performance of the AI / ML model is below the threshold. The network signaling may indicate at least one of the first mode, the second mode, and the third mode.

[0481] Fig. 23 illustrates a communication system (1) applicable to the present disclosure.

[0482] Referring to FIG. 23, a communication system (1) includes a wireless device, a base station, and a network. Here, the wireless device refers to a device that performs communication using a wireless access technology (e.g., 5G NR (New RAT), LTE (Long Term Evolution)) and may be referred to as a communication / wireless / 5G device. Although not limited thereto, the wireless device may include a robot (100a), a vehicle (100b-1, 100b-2), an XR (eXtended Reality) device (100c), a hand-held device (100d), a home appliance (100e), an IoT (Internet of Things) device (100f), and an AI device / server (400). For example, the vehicle may include a vehicle equipped with a wireless communication function, an autonomous vehicle, a vehicle capable of performing vehicle-to-vehicle communication, etc. Here, the vehicle may include an Unmanned Aerial Vehicle (UAV) (e.g., a drone). XR devices include AR (Augmented Reality) / VR (Virtual Reality) / MR (Mixed Reality) devices, and can be implemented in the form of HMD (Head-Mounted Device), HUD (Head-Up Display) installed in a vehicle, television, smartphone, computer, wearable device, home appliance, digital signage, vehicle, robot, etc. Mobile devices can include smartphone, smart pad, wearable device (e.g., smart watch, smart glass), computer (e.g., laptop, etc.), etc. Home appliances can include TV, refrigerator, washing machine, etc. IoT devices can include sensors, smart meters, etc. For example, base stations and networks can also be implemented as wireless devices, and a specific wireless device (200a) can act as a base station / network node to other wireless devices.

[0483] Wireless devices (100a to 100f) can be connected to a network (300) via a base station (200). Artificial Intelligence (AI) technology can be applied to the wireless devices (100a to 100f), and the wireless devices (100a to 100f) can be connected to an AI server (400) via the network (300). The network (300) can be configured using a 3G network, a 4G (e.g., LTE) network, a 5G (e.g., NR) network, etc. The wireless devices (100a to 100f) can communicate with each other via the base station (200) / network (300), but can also communicate directly (e.g., sidelink communication) without going through the base station / network. For example, vehicles (100b-1, 100b-2) can communicate directly (e.g., V2V (Vehicle to Vehicle) / V2X (Vehicle to Everything) communication). In addition, IoT devices (e.g., sensors) can communicate directly with other IoT devices (e.g., sensors) or other wireless devices (100a to 100f).

[0484] Wireless communication / connection (150a, 150b, 150c) can be established between wireless devices (100a~100f) / base stations (200), and base stations (200) / base stations (200). Here, wireless communication / connection can be achieved through various wireless access technologies (e.g., 5G NR) such as uplink / downlink communication (150a), sidelink communication (150b) (or D2D communication), and communication between base stations (150c) (e.g., relay, IAB (Integrated Access Backhaul). Through wireless communication / connection (150a, 150b, 150c), wireless devices and base stations / wireless devices, and base stations and base stations can transmit / receive wireless signals to each other. For example, wireless communication / connection (150a, 150b, 150c) can transmit / receive signals through various physical channels. To this end, at least some of various configuration information setting processes for transmitting / receiving wireless signals, various signal processing processes (e.g., channel encoding / decoding, modulation / demodulation, resource mapping / demapping, etc.), and resource allocation processes can be performed based on various proposals of the present disclosure.

[0485] Figure 24 illustrates a wireless device applicable to the present disclosure.

[0486] Referring to FIG. 24, the first wireless device (100) and the second wireless device (200) can transmit and receive wireless signals through various wireless access technologies (e.g., LTE, NR). Here, {the first wireless device (100), the second wireless device (200)} can correspond to {the wireless device (100x), the base station (200)} and / or {the wireless device (100x), the wireless device (100x)} of FIG. 23.

[0487] A first wireless device (100) includes one or more processors (102) and one or more memories (104), and may further include one or more transceivers (106) and / or one or more antennas (108). The processor (102) controls the memories (104) and / or the transceivers (106), and may be configured to implement the descriptions, functions, procedures, proposals, methods, and / or operational flowcharts disclosed in this document. For example, the processor (102) may process information in the memory (104) to generate first information / signal, and then transmit a wireless signal including the first information / signal via the transceiver (106). In addition, the processor (102) may receive a wireless signal including second information / signal via the transceiver (106), and then store information obtained from signal processing of the second information / signal in the memory (104). The memory (104) may be connected to the processor (102) and may store various information related to the operation of the processor (102). For example, the memory (104) may perform some or all of the processes controlled by the processor (102), or may store software code including commands for performing the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. Here, the processor (102) and the memory (104) may be part of a communication modem / circuit / chip designed to implement a wireless communication technology (e.g., LTE, NR). The transceiver (106) may be connected to the processor (102) and may transmit and / or receive wireless signals via one or more antennas (108). The transceiver (106) may include a transmitter and / or a receiver. The transceiver (106) may be used interchangeably with an RF (Radio Frequency) unit. In the present disclosure, a wireless device may also mean a communication modem / circuit / chip.

[0488] The second wireless device (200) includes one or more processors (202), one or more memories (204), and may further include one or more transceivers (206) and / or one or more antennas (208). The processor (202) controls the memories (204) and / or the transceivers (206), and may be configured to implement the descriptions, functions, procedures, proposals, methods, and / or operational flowcharts disclosed in this document. For example, the processor (202) may process information in the memory (204) to generate third information / signals, and then transmit a wireless signal including the third information / signals via the transceivers (206). Furthermore, the processor (202) may receive a wireless signal including fourth information / signals via the transceivers (206), and then store information obtained from signal processing of the fourth information / signals 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, the memory (204) may perform some or all of the processes controlled by the processor (202), or may store software code including commands for performing the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. Here, the processor (202) and the memory (204) may be part of a communication modem / circuit / chip designed to implement wireless communication technology (e.g., LTE, NR). The transceiver (206) may be connected to the processor (202) and may transmit and / or receive wireless signals via one or more antennas (208). The transceiver (206) may include a transmitter and / or a receiver. The transceiver (206) may be used interchangeably with an RF unit. In the present disclosure, a wireless device may also mean a communication modem / circuit / chip.

[0489] Hereinafter, the hardware elements of the wireless device (100, 200) will be described in more detail. Although not limited thereto, one or more protocol layers may be implemented by one or more processors (102, 202). For example, one or more processors (102, 202) may implement one or more layers (e.g., functional layers such as PHY, MAC, RLC, PDCP, RRC, SDAP). One or more processors (102, 202) may generate one or more Protocol Data Units (PDUs) and / or one or more Service Data Units (SDUs) according to the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. One or more processors (102, 202) may generate messages, control information, data, or information according to the descriptions, functions, procedures, proposals, methods, and / or operation flowcharts disclosed in this document. One or more processors (102, 202) can generate signals (e.g., baseband signals) including PDUs, SDUs, messages, control information, data or information according to the functions, procedures, proposals and / or methods disclosed herein, and provide the signals to one or more transceivers (106, 206). One or more processors (102, 202) can receive signals (e.g., baseband signals) from one or more transceivers (106, 206) and obtain PDUs, SDUs, messages, control information, data or information according to the descriptions, functions, procedures, proposals, methods and / or operational flowcharts disclosed herein.

[0490] One or more processors (102, 202) may be referred to as a controller, a microcontroller, a microprocessor, or a microcomputer. One or more processors (102, 202) may be implemented by hardware, firmware, software, or a combination thereof. For example, one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), one or more Digital Signal Processing Devices (DSPDs), one or more Programmable Logic Devices (PLDs), or one or more Field Programmable Gate Arrays (FPGAs) may be included in one or more processors (102, 202). The descriptions, functions, procedures, proposals, methods, and / or operational flowcharts 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. The descriptions, functions, procedures, suggestions, methods and / or operation flowcharts disclosed in this document may be implemented using firmware or software configured to perform one or more processors (102, 202) or stored in one or more memories (104, 204) and executed by one or more processors (102, 202). The descriptions, functions, procedures, suggestions, methods and / or operation flowcharts disclosed in this document may be implemented using firmware or software in the form of codes, instructions and / or sets of instructions.

[0491] One or more memories (104, 204) may be coupled to one or more processors (102, 202) and may store various forms of data, signals, messages, information, programs, codes, instructions, and / or commands. The one or more memories (104, 204) may be configured as ROM, RAM, EPROM, flash memory, hard drives, registers, cache memory, computer-readable storage media, and / or combinations thereof. The one or more memories (104, 204) may be located internally and / or externally to the one or more processors (102, 202). Additionally, the one or more memories (104, 204) may be coupled to the one or more processors (102, 202) via various technologies, such as wired or wireless connections.

[0492] One or more transceivers (106, 206) can transmit user data, control information, wireless signals / channels, etc., as mentioned in the methods and / or flowcharts of this document, to one or more other devices. One or more transceivers (106, 206) can receive user data, control information, wireless signals / channels, etc., as mentioned in the descriptions, functions, procedures, proposals, methods and / or flowcharts of this document, from one or more other devices. For example, one or more transceivers (106, 206) can be connected to one or more processors (102, 202) and can transmit and receive wireless signals. For example, one or more processors (102, 202) can control one or more transceivers (106, 206) to transmit user data, control information, or wireless signals to one or more other devices. Additionally, one or more processors (102, 202) may control one or more transceivers (106, 206) to receive user data, control information, or wireless signals from one or more other devices. Additionally, one or more transceivers (106, 206) may be coupled to one or more antennas (108, 208), and one or more transceivers (106, 206) may be configured to transmit and receive user data, control information, wireless signals / channels, or the like, as referred to in the descriptions, functions, procedures, proposals, methods, and / or operational flowcharts disclosed herein, via one or more antennas (108, 208). In this document, one or more antennas may be multiple physical antennas or multiple logical antennas (e.g., antenna ports). One or more transceivers (106, 206) can convert received user data, control information, 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 one or more processors (102, 202).One or more transceivers (106, 206) may convert user data, control information, wireless signals / channels, etc. processed by one or more processors (102, 202) from baseband signals to RF band signals. For this purpose, one or more transceivers (106, 206) may include an (analog) oscillator and / or filter.

[0493] Figure 25 illustrates another example of a wireless device applicable to the present disclosure. The wireless device may be implemented in various forms depending on the use case / service (see Figure 23).

[0494] Referring to FIG. 25, the wireless device (100, 200) corresponds to the wireless device (100, 200) of FIG. 24 and may be composed of various elements, components, units / units, and / or modules. For example, the wireless device (100, 200) may include a communication unit (110), a control unit (120), a memory unit (130), and an additional element (140). The communication unit may include a communication circuit (112) and a transceiver(s) (114). For example, the communication circuit (112) may include one or more processors (102, 202) and / or one or more memories (104, 204) of FIG. 24. For example, the transceiver(s) (114) may include one or more transceivers (106, 206) and / or one or more antennas (108, 208) of FIG. 24. The control unit (120) is electrically connected to the communication unit (110), the memory unit (130), and the additional elements (140) and controls the overall operation of the wireless device. For example, the control unit (120) may control the electrical / mechanical operation of the wireless device based on the program / code / command / information stored in the memory unit (130). In addition, the control unit (120) may transmit information stored in the memory unit (130) to an external device (e.g., another communication device) via a wireless / wired interface through the communication unit (110), or store information received from an external device (e.g., another communication device) via a wireless / wired interface in the memory unit (130).

[0495] The additional element (140) may be configured in various ways depending on the type of the wireless device. For example, the additional element (140) may include at least one of a power unit / battery, an input / output unit (I / O unit), a driving unit, and a computing unit. Although not limited thereto, the wireless device may be implemented in the form of a robot (Fig. 23, 100a), a vehicle (Fig. 23, 100b-1, 100b-2), an XR device (Fig. 23, 100c), a portable device (Fig. 23, 100d), a home appliance (Fig. 23, 100e), an IoT device (Fig. 23, 100f), a digital broadcasting terminal, a hologram device, a public safety device, an MTC device, a medical device, a fintech device (or a financial device), a security device, a climate / environmental device, an AI server / device (Fig. 23, 400), a base station (Fig. 23, 200), a network node, etc. Wireless devices may be mobile or stationary depending on the use / service.

[0496] In FIG. 25, various elements, components, units / parts, and / or modules within the wireless device (100, 200) may be entirely interconnected via a wired interface, or at least some may be wirelessly connected via a communication unit (110). For example, within the wireless device (100, 200), the control unit (120) and the communication unit (110) may be wired, and the control unit (120) and a first unit (e.g., 130, 140) may be wirelessly connected via the communication unit (110). In addition, each element, component, unit / part, and / or module within the wireless device (100, 200) may further include one or more elements. For example, the control unit (120) may be composed of a set of one or more processors. For example, the control unit (120) may be composed of a set of a communication control processor, an application processor, an electronic control unit (ECU), a graphics processing processor, a memory control processor, etc. As another example, the memory unit (130) may be composed of RAM (Random Access Memory), DRAM (Dynamic RAM), ROM (Read Only Memory), flash memory, volatile memory, non-volatile memory, and / or a combination thereof.

[0497] Figure 26 illustrates a vehicle or autonomous vehicle applicable to the present disclosure. The vehicle or autonomous vehicle may be implemented as a mobile robot, a car, a train, a manned or unmanned aerial vehicle (AV), a ship, or the like.

[0498] Referring to FIG. 26, a vehicle or autonomous vehicle (100) may include an antenna unit (108), a communication unit (110), a control unit (120), a driving unit (140a), a power supply unit (140b), a sensor unit (140c), and an autonomous driving unit (140d). The antenna unit (108) may be configured as a part of the communication unit (110). Blocks 110 / 130 / 140a to 140d correspond to blocks 110 / 130 / 140 of FIG. 25, respectively.

[0499] The communication unit (110) can transmit and receive signals (e.g., data, control signals, etc.) with external devices such as other vehicles, base stations (e.g., base stations, road side units, etc.), and servers. The control unit (120) can control elements of the vehicle or autonomous vehicle (100) to perform various operations. The control unit (120) can include an ECU (Electronic Control Unit). The drive unit (140a) can drive the vehicle or autonomous vehicle (100) on the ground. The drive unit (140a) can include an engine, a motor, a power train, wheels, brakes, a steering device, etc. The power supply unit (140b) supplies power to the vehicle or autonomous vehicle (100) and can include a wired / wireless charging circuit, a battery, etc. The sensor unit (140c) can obtain vehicle status, surrounding environment information, user information, etc. The sensor unit (140c) may include an IMU (inertial measurement unit) sensor, a collision sensor, a wheel sensor, a speed sensor, an incline sensor, a weight detection sensor, a heading sensor, a position module, a vehicle forward / backward sensor, a battery sensor, a fuel sensor, a tire sensor, a steering sensor, a temperature sensor, a humidity sensor, an ultrasonic sensor, an illuminance sensor, a pedal position sensor, etc. The autonomous driving unit (140d) may implement a technology for maintaining a driving lane, a technology for automatically controlling speed such as adaptive cruise control, a technology for automatically driving along a set path, a technology for automatically setting a path and driving when a destination is set, etc.

[0500] For example, the communication unit (110) can receive map data, traffic information data, etc. from an external server. The autonomous driving unit (140d) can generate an autonomous driving route and driving plan based on the acquired data. The control unit (120) can control the drive unit (140a) so that the vehicle or autonomous vehicle (100) moves along the autonomous driving route according to the driving plan (e.g., speed / direction control). During autonomous driving, the communication unit (110) can irregularly / periodically acquire the latest traffic information data from an external server and can acquire surrounding traffic information data from surrounding vehicles. In addition, during autonomous driving, the sensor unit (140c) can acquire vehicle status and surrounding environment information. The autonomous driving unit (140d) can update the autonomous driving route and driving plan based on newly acquired data / information. The communication unit (110) can transmit information regarding the vehicle location, autonomous driving route, driving plan, etc. to the external server. External servers can predict traffic information data in advance using AI technology or other technologies based on information collected from vehicles or autonomous vehicles, and provide the predicted traffic information data to the vehicles or autonomous vehicles.

[0501] The embodiments described above are combinations of components and features of the present disclosure in a predetermined form. Each component or feature should be considered optional unless explicitly stated otherwise. Each component or feature may be implemented without being combined with other components or features. Furthermore, it is also possible to form 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 self-evident that claims that do not have an explicit citation relationship in the patent claims may be combined to form embodiments or incorporated as new claims through post-application amendments.

[0502] It will be apparent to those skilled in the art that the present invention can be embodied in other specific forms without departing from the spirit or scope of the invention. Therefore, the above detailed description should not be construed as limiting in any respect, but rather as illustrative. The scope of the present invention should be determined by a reasonable interpretation of the appended claims, and all modifications within the scope of equivalents are intended to be included within the scope of the present invention.

[0503] The present disclosure may be used in a terminal, base station, or other equipment of a wireless mobile communication system.

Claims

1. In a method performed by a terminal, Obtain information about predicted CSI and information about compressed CSI based on an AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and Including transmitting a CSI report based on information about the predicted CSI and information about the compressed CSI, A method wherein, based on the performance of the AI / ML model being below a threshold, the terminal adjusts past CSIs used as an input data set of the CSI compression before changing at least one of the settings for the CSI compression or the settings for the CSI prediction of the AI / ML model.

2. In paragraph 1, A method wherein the terminal maintains the AI / ML model without changing the settings for the CSI compression or the settings for the CSI prediction based on the performance of the AI / ML model being above the threshold after adjusting the past CSIs.

3. In paragraph 1, A method in which the terminal changes at least one of the settings for the CSI compression or the settings for the CSI prediction based on the performance of the AI / ML model being below the threshold even after adjusting the past CSIs.

4. In paragraph 3, A method wherein the terminal determines whether the CSI compression causes a performance degradation of the AI / ML model or whether the CSI prediction causes a performance degradation of the AI / ML model by changing one of the settings for the CSI compression and the settings for the CSI prediction while maintaining the other.

5. In paragraph 3, A method in which the terminal reconfigures the AI / ML model or switches to another AI / ML model based on the performance of the AI / ML model being below the threshold even after changing at least one of the settings for the CSI compression or the settings for the CSI prediction.

6. In paragraph 1, The above adjustments to past CSIs are: A first mode that excludes all past CSIs from the input data set of the CSI compression; A second mode that replaces the past CSIs with CSIs accumulated from the present; and A method performed based on at least one of the third modes for selectively excluding some of the past CSIs from the input data set of the CSI compression.

7. In paragraph 6, Further comprising receiving network signaling that the performance of the AI / ML model is below the threshold, A method wherein the network signaling indicates at least one of the first mode, the second mode and the third mode.

8. In paragraph 1, A method in which the terminal determines whether the performance of the AI / ML model is below a threshold based on a second AI / ML model for CSI reconstruction or a proxy model for the second AI / ML model.

9. A non-transitory computer-readable recording medium having recorded thereon a program for performing the method described in paragraph 1.

10. In the device, a memory configured to store instructions; and A processor configured to perform operations by executing the above instructions, The operations of the above processor are: Obtain information about predicted CSI and information about compressed CSI based on an AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and Including transmitting a CSI report based on information about the predicted CSI and information about the compressed CSI, A device that adjusts past CSIs used as an input data set of the CSI compression prior to changing at least one of the settings for the CSI compression of the AI / ML model or the settings for the CSI prediction based on the performance of the AI / ML model being below a threshold.

11. In paragraph 10, Including a transmitter and receiver, The above device is a terminal in a wireless communication system.

12. In paragraph 10, The above device is a processing device configured to control a terminal in a wireless communication system.

13. In a method performed by a base station, Receiving CSI reports through the terminal's AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and Including obtaining information about predicted CSI and information about compressed CSI based on the above CSI report, A method wherein, based on the performance of the AI / ML model being below a threshold, the base station instructs the terminal to adjust past CSIs used as an input data set of the CSI compression before changing at least one of the settings for the CSI compression or the settings for the CSI prediction of the AI / ML model.

14. A non-transitory computer-readable recording medium having recorded thereon a program for performing the method described in Article 13.

15. At the base station, a memory configured to store instructions; and A processor configured to perform operations by executing the above instructions, The operations of the above processor are: Receiving CSI reports through the terminal's AI / ML (artificial intelligence / machine learning) model that supports both CSI compression (channel state information) and CSI prediction; and Including obtaining information about predicted CSI and information about compressed CSI based on the above CSI report, A base station instructing the terminal to adjust past CSIs used as an input data set of the CSI compression before changing at least one of the settings for the CSI compression of the AI / ML model or the settings for the CSI prediction based on the performance of the AI / ML model being below a threshold.

Images