Mobile devices, network nodes, and methods

Abstract

This disclosure relates to the collection of data for the management and monitoring of AI / ML models. In one described embodiment, a method performed by user equipment (UE) includes receiving from a network node a first instruction for model management decisions and a second instruction for model performance reporting, and, if the first instruction does not indicate model activation and the second instruction enables model performance reporting, running the model and reporting model performance to the network node without using the output from the model to control communication with the network node. Various other methods, including corresponding network node methods, are disclosed.

Description

Technical Field

[0001] The present disclosure relates to a communication system.

Background Art

[0002] The present disclosure is particularly, but not limited to, related to wireless communication systems and devices operating according to 3rd Generation Partnership Project (3GPP (registered trademark)) standards or their equivalents or derivatives (including LTE-Advanced, next-generation or 5G networks, future generations, and the like). The present disclosure is particularly, but not necessarily exclusively, related to artificial intelligence and machine learning (AI / ML) models used in "New Radio" systems (also referred to as "next-generation" systems) and similar systems.

[0003] Recent developments in 3GPP standards are referred to as Evolved Packet Core (EPC) Long Term Evolution (LTE) and Evolved UMTS Terrestrial Radio Access Network (E-UTRAN), commonly known as "4G." In addition, the terms "5G" and "new radio" (NR) refer to evolving communication technologies expected to support a variety of applications and services. Various details of 5G networks are described in the "NGMN 5G White Paper" V1.0 by the Next Generation Mobile Network (NGMN) Alliance, which can be obtained, for example, from https: / / www.ngmn.org / 5g-white-paper.html. 3GPP intends to support 5G through the so-called 3GPP Next Generation (NextGen) Radio Access Network (RAN) and 3GPP NextGen Core Network.

[0004] Under the 3GPP standard, a NodeB (or eNB in ​​LTE, gNB in ​​5G) is a Radio Access Network (RAN) node (or simply an "access node," "access network node," or "base station") through which communication devices (user equipment, or "UE") connect to the core network and communicate with other communication devices or remote servers. For simplicity, this application uses the terms RAN node, base station, or access network node to refer to any such access node. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] The "NGMN 5G White Paper" V1.0 by the Next Generation Mobile Networks (NGMN) Alliance is available from https: / / www.ngmn.org / 5g-white-paper.html. [Overview of the project] [Problems that the invention aims to solve]

[0006] Some of the additional developments in 3GPP concern the use of artificial intelligence (AI) and machine learning (ML), often abbreviated as AI / ML. Predictions or inferences generated using AI / ML models can be used as part of various methods to improve the reliability or efficiency of communications in a network. For example, AI / ML models can be used to predict the path of a UE based on its previous mobility, for beam management, or in methods of encoding and transmitting information. Supported use cases may include, for example, energy saving, traffic steering, anomaly detection, quality of experience (QoE) optimization, mobility robustness optimization (MRO), RAN slice service level agreement (SLA) assurance, multiple-input multiple-output (MIMO) beamforming optimization, network slice subnet instance (NSSI) resource allocation, coverage and capacity optimization (CCO), mobility load balancing (MLB), RACH optimization, or UE transmit power optimization.

[0007] AI / ML models can be hosted at a base station, which can then use the AI / ML model to perform controls related to communication resources or the status of a UE (e.g., UE mobility control or UE Radio Resource Control (RRC) state control) based on inferences (e.g., decisions or predictions) generated. The base station can also transmit the inferences generated using the model to another node in the network for use by other nodes. Alternatively, an AI / ML model can be hosted at two nodes in the network, for example, a base station and a UE. In this case, both the base station and the UE can use the model to make decisions or predictions. For example, a UE can use the model as part of an encoding process to encode (and / or compress) channel state information (CSI) for transmission to the base station, and the base station can use the same model as part of a corresponding decoding (and / or restoration) process to decode the CSI received from the UE.

[0008] The inventors have explored ways to extend air interfaces with features that enable support for AI / ML-based algorithms for improved performance (such as improved throughput, robustness, accuracy, and reliability) and / or reduced complexity / overhead. The inventors have explored using AI / ML-based algorithms for CSI feedback enhancement (e.g., overhead reduction, improved accuracy, and prediction), beam management (e.g., beam prediction in the time domain and / or spatial domain for overhead and latency reduction, improved beam selection accuracy), and positioning accuracy enhancement for different scenarios (e.g., those with severe non-line-of-sight (NLOS) conditions).

[0009] More generally, there is a need for improved methods and devices for collecting data for AI-ML model management and monitoring in communication networks. [Means for solving the problem]

[0010] This disclosure aims to provide apparatus and methods that address at least one of the above needs and / or problems in part.

[0011] According to one embodiment, a method is provided which is performed by user equipment (UE), the method comprising transmitting UE capability information to a network node, the UE capability information indicating which reporting quantities or metrics the UE can support for model performance reporting.

[0012] In another embodiment, a method is provided which is performed by user equipment (UE), the UE receiving from a network node a first instruction for a model management decision and a second instruction for a model performance report, and, if the first instruction does not indicate model activation and the second instruction enables model performance reporting, running the model and reporting the model performance to the network node without using the output from the model to control communication with the network node.

[0013] In another embodiment, a method is provided which is performed by user equipment (UE), and includes receiving configuration data from a network node for model performance reporting of a plurality of model functions, executing a model of each function configured by the configuration data, and reporting the model performance of each model function according to the configuration data.

[0014] In another embodiment, a method is provided which is performed by user equipment (UE), the method being provided which includes: running a model to predict parameters relating to communication with a network node in a first time; obtaining at least one measurement of at least one signal relating to the parameters predicted by the model and received by the user equipment before and after the first time; and reporting the parameters predicted by the model and at least one measurement to the network node.

[0015] In another embodiment, a method is provided which is performed by user equipment (UE) and includes: running a model that predicts parameters related to communication with a network node; monitoring the model or at least one metric related to communication with the network node; and reporting to the network node if at least one metric satisfies at least one criterion.

[0016] In another embodiment, a method is provided which is performed by a network node and includes receiving UE capability information from user equipment (UE), the UE capability information indicating what reporting quantities or metrics the UE can support for model performance reporting.

[0017] In another embodiment, a method is provided which is performed by a network node, the method being provided which includes sending a first instruction for a model management decision and a second instruction for a model performance report to user equipment (UE), and, if the first instruction does not indicate model activation and the second instruction enables model performance reporting, receiving a model performance report from the UE regarding a model running on the UE without using the output from the model to control communication with the network node.

[0018] In another embodiment, a method is provided which is performed by a network node and includes transmitting configuration data for model performance reporting relating to a plurality of model functions to user equipment (UE), and receiving model performance data from the UE relating to the execution of a model for each function configured by the configuration data for each model function according to the configuration data.

[0019] In another embodiment, a method is provided which is performed by a network node, comprising configuring the UE to run a model in a first time that predicts parameters relating to communication between user equipment (UE) and the network node, the UE taking at least one measurement of at least one signal received by the user equipment before and after the first time, relating to the parameters predicted by the model, and receiving a report from the UE containing the parameters predicted by the model and at least one measurement.

[0020] In another embodiment, a method is provided which is performed by a network node, the method comprising configuring user equipment (UE) to run a model that predicts parameters relating to communication with the network node, the UE monitoring the model or at least one metric relating to communication with the network node, and receiving a report from the UE if at least one metric satisfies at least one criterion.

[0021] In another embodiment, user equipment (UE) is provided, which includes means for transmitting UE capability information to network nodes, and the UE capability information indicates which reporting quantities or metrics the UE can support for model performance reporting.

[0022] According to another aspect, there is provided a user equipment (UE) comprising means for receiving, from a network node, a first instruction for model management decision and a second instruction for model performance report, and means for executing the model without using the output from the model to control communication with the network node and reporting the model performance to the network node when the first instruction does not indicate activation of the model and the second instruction enables the model performance report.

[0023] According to another aspect, there is provided a user equipment (UE) comprising means for receiving, from a network node, configuration data for model performance report regarding a plurality of model functions, means for executing a model for each function configured by the configuration data, and means for reporting model performance for each model function according to the configuration data.

[0024] According to another aspect, there is provided a user equipment (UE) comprising means for executing, at a first time, a model for predicting parameters related to communication with a network node, means for obtaining at least one measurement of at least one signal related to parameters predicted by the model and received by the user equipment before and after the first time, and means for reporting the parameters predicted by the model and the at least one measurement to the network node.

[0025] According to another aspect, there is provided a user equipment (UE) comprising means for executing a model for predicting parameters related to communication with a network node, means for monitoring at least one metric related to communication with the model or the network node, and means for reporting to the network node when the at least one metric meets at least one criterion.

[0026] According to another aspect, a network node is provided that comprises means for receiving UE capability information from a user equipment (UE), the UE capability information indicating which reporting amount or metric the UE can support for model performance reporting.

[0027] According to another aspect, a network node is provided that comprises means for transmitting a first instruction for model management decision and a second instruction for model performance reporting to a user equipment (UE), where the first instruction does not indicate activation of the model, and means for receiving, from the UE, a report on model performance regarding a model executed on the UE without using the output from the model to control communication with the network node when the second instruction enables model performance reporting.

[0028] According to another aspect, a network node is provided that comprises means for transmitting configuration data for model performance reporting regarding a plurality of model functions to a user equipment (UE), and means for receiving, from the UE, model performance data regarding execution of a model for each function configured by the configuration data for each model function according to the configuration data.

[0029] According to another aspect, a network node is provided that comprises means for configuring the UE to execute, at a first time, a model for predicting parameters related to communication between the user equipment (UE) and the network node, the UE acquiring at least one measurement of at least one signal related to parameters predicted by the model and received by the user equipment around the first time, and means for receiving, from the UE, a report including the parameters predicted by the model and the at least one measurement.

[0030] In another embodiment, a means for configuring user equipment (UE) to run a model that predicts parameters related to communication with a network node is provided, wherein the network node includes means for monitoring the model or at least one metric related to communication with a network node, and means for receiving a report from the UE if at least one metric satisfies at least one criterion.

[0031] The various functional means defined above, which are part of the UE, may be provided by memory and one or more processors that execute instructions stored in memory. Similarly, the various functional means defined above, which are part of the network node, may be provided by memory and one or more processors that execute instructions stored in memory.

[0032] The disclosure may also provide a computer program product that includes computer-implementable instructions for causing a programmable computer to perform any of the methods described above. Computer-implementable instructions may be provided as signals or on a tangible computer-readable medium. [Brief explanation of the drawing]

[0033] Exemplary embodiments of this disclosure will be described, by reference to the accompanying drawings, as an example.

[0034] [Figure 1] This is a schematic diagram illustrating a mobile ("cellular" or "wireless") telecommunications system. [Figure 2] Figure 1 shows a typical frame structure that may be used in the communication system. [Figure 3] Figure 2 shows the resource grid of the subframe. [Figure 4] This describes the mobility procedure performed when a UE moves from a source base station (or cell) to a target base station (or cell). [Figure 5]This document presents a functional framework for AI / ML models. [Figure 6] This document demonstrates how to train AI / ML models and monitor their performance. [Figure 7] This demonstrates how base stations can configure UEs for AI / ML performance reporting and different reporting options. [Figure 8] This demonstrates how a base station can be configured to perform AI / ML management decisions and different reporting options. [Figure 9] Figure 1 is a schematic block diagram showing the main components of the UE for a telecommunications system. [Figure 10] Figure 1 is a schematic block diagram showing the main components of a base station for a telecommunications system. [Modes for carrying out the invention]

[0035] overview Next, with reference to Figures 1, 2, and 3, an exemplary telecommunications system will be described, for illustrative purposes only, using general terminology.

[0036] Figure 1 schematically shows a mobile ("cellular" or "wireless") communication system 1 to which exemplary embodiments of the present disclosure can be applied.

[0037] In communication system 1, user equipment (UE) 3-1, 3-2, 3-3 (such as mobile phones and / or other mobile devices) can communicate with each other via radio access network (RAN) nodes 5 operating according to one or more compatible radio access technologies (RATs). In the illustrated example, the (R)AN node 5 comprises a base station 5 or "gNB" 5 operating one or more associated cells 9. Communication via the base station 5 is typically routed via a core network 7 (which is, for example, a 5G core network or an evolved packet core network (EPC)).

[0038] As those skilled in the art will understand, three UE3 and one base station 5 are shown in Figure 1 for illustrative purposes, but the system, when implemented, typically includes other base stations 5 and UE3.

[0039] Each base station 5 controls one or more associated cells 9 directly or indirectly through one or more other nodes (such as home base stations, repeaters, remote radio heads, and distributed units). It will be understood that base stations 5 may be configured to support 4G, 5G, 6G, and / or any other 3GPP or non-3GPP communication protocols.

[0040] The UE3 and their serving base stations 5 are connected via appropriate air interfaces (such as the so-called "Uu" interface). Adjacent base stations 5 can be connected to each other via appropriate inter-base station interfaces (such as the so-called "X2" interface, "Xn" interface, etc.).

[0041] The core network 7 includes several logical nodes (or "functions") to support communication in the communication system 1. In this example, the core network 7 comprises a control plane function (CPF) 10 and one or more user plane functions (UPFs) 11. The CPF 10 comprises one or more access and mobility management functions (AMFs) 10-1, one or more session management functions (SMFs) 10-2, and a number of other functions 10-n.

[0042] Base station 5 is connected to the core network nodes via appropriate interfaces (or "reference points"), such as an N2 reference point between base station 5 and AMF10-1 for control signaling communications, and an N3 reference point between base station 5 and each UPF11 for user data communications. Each UE3 is connected to AMF10-1 via a logical non-access stratum (NAS) connection on an N1 reference point (similar to the S1 reference point in LTE). It will be understood that N1 communications are routed transparently through base station 5.

[0043] One or more UPF11s are connected to an external data network (e.g., an IP network such as the Internet) via reference point N6 for the communication of user data.

[0044] The AMF10-1 performs mobility management-related functions, maintains NAS signaling connections with each UE3, and manages UE registration. The AMF10-1 also manages paging. The SMF10-2 provides session management functions (which form part of the MME function in LTE) and combines several control plane functions (provided by the serving gateway and packet data network gateway in LTE). The SMF10-2 also assigns IP addresses to each UE3.

[0045] The base station 5 of communication system 1 is configured to operate at least one cell 9 on an associated TDD carrier operating in an unpaired spectrum. It will be understood that base station 5 may also operate at least one cell 9 on an associated FDD carrier operating in a paired spectrum.

[0046] Base station 5 is also configured for transmitting control information and user data via several downlink (DL) physical channels, and UE3 is configured to receive them and to transmit several physical signals. DL physical channels correspond to resource elements (REs) that carry information transmitted from higher layers, and DL physical signals correspond to REs used in the physical layer that do not carry information transmitted from higher layers.

[0047] Physical channels may include, for example, a physical downlink shared channel (PDSCH), a physical broadcast channel (PBCH), and a physical downlink control channel (PDCCH). The PDSCH carries data that shares its capacity on a time and frequency basis. The PDSCH can carry various data items, including, for example, user data, UE-specific upper-layer control messages mapped from higher channels, system information blocks (SIBs), and paging. The PDCCH carries downlink control information (DCI) to support several functions, including, for example, scheduling downlink transmissions on the PDSCH and uplink data transmissions on the physical uplink shared channel (PUSCH). The PBCH provides the Master Information Block (MIB) to the UE3. It also supports time and frequency synchronization, along with the PDCCH, which assists in cell acquisition, selection, and re-selection. UE3 may receive a Synchronization Signal Block (SSB), and UE3 may assume that the reception timings of the PBCH, Primary Synchronization Signal (PSS), and Secondary Synchronization Signal (SSS) are consecutive symbols, forming an SS / PBCH block. Base station 5 may transmit multiple synchronization signal (SS) blocks corresponding to different DL beams. The total number of SS blocks may be limited, for example, to a 5ms duration as an SS burst.The periodicity of SSB transmissions can be indicated to the UE using any appropriate signaling (e.g., using ssb-periodicityServingCell for each serving cell). The periodicity value of an SSB can be, for example, 20ms or more. For initial cell selection, UE3 may be configured to assume that SS bursts occur with a 2-frame periodicity. UE3 may also provide instructions (e.g., using ssb-PositionsInBurst) on which SSBs to transmit within a 5ms duration.

[0048] DL physical signals may include, for example, a reference signal (RS) and a synchronization signal (SS). The reference signal (sometimes known as a pilot signal) is a signal with a predetermined special waveform known to both the UE3 and the base station 5. The reference signal may include, for example, a cell-specific reference signal, a UE-specific reference signal (UE-RS), a downlink demodulation signal (DMRS), and a channel state information reference signal (CSI-RS).

[0049] Similarly, UE3 is configured to transmit control information and user data via several uplink (UL) physical channels corresponding to REs that carry information emitted from higher layers, and UL physical signals used in the physical layer that do not carry information emitted from higher layers, and base station 5 is configured to receive control information and user data via several UL physical channels corresponding to REs that carry information emitted from higher layers, and UL physical signals used in the physical layer that do not carry information emitted from higher layers. The physical channels may include, for example, PUSCH, physical uplink control channel (PUCCH), and / or physical random-access channel (PRACH). The UL physical signals may include, for example, demodulation reference signals (DMRS) for UL control / data signals, and / or sounding reference signals (SRS) used for UL channel measurement.

[0050] When UE3 first establishes a Radio Resource Control (RRC) connection with base station 5 via cell 9, it registers with the appropriate core network node (e.g., AMF, MME). UE3 is in a so-called RRC connected state, and the associated UE context is maintained by the network. When UE3 is in a so-called RRC idle state, or RRC inactive state, UE3 selects an appropriate cell to camp on so that the network (though not necessarily at the cell level) can recognize UE3's approximate location.

[0051] The base station 5 may also be a base station 5 divided between one or more distributed units (DUs) 50 and a central unit (CU) 60, where the CU 60 typically performs higher-level functions and communication with the next-generation core, and the DU 50 performs lower-level functions and communication with neighboring UEs 3 (i.e., within the cell operated by the base station 5) via an air interface. This type of base station 5 may be referred to as a “distributed” base station 5 or gNB 5. A distributed gNB 5 includes the following functional units:

[0052] gNB Central Unit (gNB-CU): A logical node that controls the operation of one or more gNB-DUs and hosts the Radio Resource Control (RRC) layer, Service Data Adaptation Protocol (SDAP) layer, and Packet Data Convergence Protocol (PDCP) layer of the gNB (or the RRC and PDCP layers of the en-gNB). The gNB-CU terminates the so-called F1 interface that connects to the gNB-DU.

[0053] A gNB Distributed Unit (gNB-DU) is a logical node that hosts the Radio Link Control (RLC), Medium Access Control (MAC), and Physical (PHY) layers of a gNB or en-gNB, and its operation is partially controlled by a gNB-CU. A single gNB-DU supports one or more cells. A single cell is supported by only one gNB-DU. A gNB-DU terminates the F1 interface connected to the gNB-CU.

[0054] gNB-CU-Control Plane (gNB-CU-CP): A logical node that hosts the control plane portion of the RRC and PDCP protocols for the gNB-CU for en-gNB or gNB. The gNB-CU-CP terminates the so-called E1 interface connected to the gNB-CU-UP and the F1-C (F1 control plane) interface connected to the gNB-DU.

[0055] gNB-CU-User Plane (gNB-CU-UP): A logical node that hosts the user plane portion of the PDCP protocol for gNB-CU for en-gNB, and the user plane portions of the PDCP protocol and SDAP protocol for gNB-CU for gNB. gNB-CU-UP terminates the E1 interface connected to gNB-CU-CP and the F1-U (F1 user plane) interface connected to gNB-DU.

[0056] When a distributed base station or a similar control plane-user plane (CP-UP) partition is employed, it will be understood that the control plane entity and the user plane entity may each include associated transceiver circuits, antennas, network interfaces, controllers, memory, operating systems, and communication control modules. If base station 5 comprises a distributed base station, the network interfaces also include E1 interfaces and F1 interfaces (F1-C for the control plane and F1-U for the user plane) for communicating signals between the respective functions of the distributed base stations.

[0057] Frame structure Referring to Figure 2, which shows a typical frame structure that may be used in communication system 1, the base station 5 and UE3 of communication system 1 communicate with each other using resources organized into frames with a length of 10 ms in the time domain. Each frame consists of 10 subframes of equal size, each 1 ms long. Each subframe is divided into one or more slots, each containing 14 orthogonal frequency-division multiplexing (OFDM) symbols of equal length.

[0058] As shown in Figure 2, communication system 1 supports multiple different numerologies (subcarrier spacing (SCS), slot length, and thus OFDM symbol length). Specifically, each numerology is identified by a parameter μ, where μ=0 represents 15kHz (corresponding to LTE SCS). Currently, the SCS for other values ​​of μ can actually be derived from μ=0 by scaling up by a power of 2 (i.e., SCS = 15 × 2). μ (kHz). The relationship between the parameter μ and SCS(Δf) is shown in Table 1. [Table 1]

[0059] Figure 3 shows the resource grid of the subframe shown in Figure 2. As illustrated, the subcarrier spacing and the number of OFDM symbols within the subframe vary depending on the numerology. One block shown in Figure 3 corresponds to one resource element and is the smallest unit of the resource grid, consisting of one subcarrier in the frequency domain and one OFDM symbol in the time domain. Resource block 25 is defined only for the frequency domain and is defined as 12 consecutive subcarriers in the frequency domain within one OFDM symbol.

[0060] System information and SIB It will be understood that transmissions in cell 9 of base station 5 may include one or more broadcast transmissions, one or more unicast transmissions for reception by UE3, and / or one or more multicast transmissions for reception by a group of UE3. System information (SI) transmitted within the cell may include "minimum SI" (MSI) and "other SI" (OSI). OSI may be broadcast on demand, for example, using a downlink shared channel (DL-SCH). OSI may also be broadcast in response to requests from UE3 in a Radio Resource Control (RRC) idle or RRC inactive state. OSI may also be requested by UE3 in an RRC connected state, for example, via one or more dedicated RRC transmissions.

[0061] The SI may include information to enable UE3 to complete cell selection (for example, setting UE3 to complete cell selection), information to enable UE3 to complete cell re-selection procedures, and information to enable UE3 to receive one or more paging messages transmitted in a cell. The SI may be broadcast using a Master Information Block (MIB) and one or more System Information Blocks (SIBs).

[0062] The MSI comprises an MIB and a system information block 1 (SIB1). The MIB includes information used by the UE3 to receive SIB1, such as the subcarrier interval for SIB1. The MIB provides information corresponding to the Control Resource Set (CORESET) and the search space. SIB1 is sometimes referred to as the “remaining MSI” (RMSI). SIB1 may be transmitted in a dedicated RRC message, and other SIBs (e.g., SIB2 to SIB9) may be transmitted using one or more other suitable RRC transmissions (e.g., another dedicated RRC message). The MIB and SIB1 can provide the UE3 with scheduling information instructions for receiving and decoding other SIBs such as SIB2 to SIB9, and can provide information used by the UE3 to receive one or more paging messages. The OSI may comprise, for example, SIB2 to SIB9 transmitted using DL-SCH in an SI message. The mapping of SIB2 to SIB9 to the corresponding SI messages may be provided to the UE3 by the base station 5. MIB and SIB1 through SIB9 are described in more detail, for example, in 3GPP TS 38.331. SIB2 provides information for intra-frequency, inter-frequency, and inter-system cell reselection. SIB3 provides cell-specific information for intra-frequency cell reselection. SIB4 provides information for inter-frequency cell reselection. SIB5 provides information for inter-system cell reselection for 4G (LTE). SIB6 and SIB7 provide information for earthquake and tsunami warning systems (ETWS). SIB8 provides information for commercial mobile alert service (CMAS) notifications, for example, to provide warning messages to UE3.SIB9 includes information regarding Coordinated Universal Time (UTC), Global Positioning System (GPS) time (for example, for GPS initialization), and local time.

[0063] SIBs may be broadcast periodically (for example, according to a predetermined periodic pattern), or alternatively, they may be provided "on demand" in response to a request from UE3, for example. For example, MIBs may be transmitted with a periodicity of 80ms and repetitions occurring within 80ms, while SIB1 may be transmitted with a periodicity of 160ms and a variable transmission repetition periodicity within 160ms (for example, 20ms). SIB1 may be used to indicate to UE3 which SIBs are transmitted periodically and which SIBs are available on demand in response to requests from UE3. UE3 may be configured to request on-demand SIBs using message 1 (message 1:MSG1), which may be referred to as an MSG1-based on-demand SI request, or message 3 (message 3:MSG3), which may be referred to as an MSG3-based on-demand SI request.

[0064] A physical broadcast channel (PBCH) may be used to broadcast the MIB. Base station 5 may transmit the PBCH along with a synchronization signal (SS) (e.g., a primary synchronization signal (PSS) and a secondary synchronization signal (SSS)) in an SS / PBCH block. The SS / PBCH block comprises four orthogonal frequency-division multiplexed (OFDM) symbols mapped to the PSS, SSS, and PBCH associated with a demodulation reference signal (DM-RS). In the frequency domain, the SS / PBCH block contains 240 consecutive subcarriers. When UE3 is in an RRC connection state, base station 5 may provide UE3 with instructions for the resources to be used for the SS / PBCH, for example, using dedicated signaling. SIB1 may be transmitted using a physical downlink shared channel (PDSCH). OSI may similarly be transmitted using a PDSCH, for example. When one or more beamformed transmissions are transmitted within a cell provided by base station 5, a portion of the SI (e.g., a portion of the SIB) may be transmitted using only a specific beam or using only a specific transmission / reception point (TRP).

[0065] UE Mobility Figure 4 outlines a mobility procedure that can be performed in the type of communication system 1 shown in Figure 1. In this example, a handover of UE3 from source base station 5-1 to target base station 5-2 is performed.

[0066] In the optional step S401, UE3 performs a measurement. The measurement may be a measurement of the signal transmitted by source base station 5-1 or a measurement of the signal transmitted by target base station 5-2. The measurement may be a measurement of signal strength, which can be used as part of the determination that UE3 is being handed over from source base station 5-1 to target base station 5-2. In the optional step S402, UE3 sends a measurement report to source base station 5-1 providing instructions on the results of the measurement. The measurement report may be sent from UE3 to source base station 5-1 in an RRC message. In this example, source base station 5-1 uses the information provided in the measurement report to determine that UE3 is being handed over to target base station 5-2. However, it will be understood that the determination that a handover to target base station 5-2 is being performed may, alternatively (or additionally), be based on a measurement performed at source base station 5-1 or target base station 5-2. Alternatively, the determination of whether a UE3 handover is performed may be based on factors other than signal measurements, such as the level of congestion in the cell operated by source base station 5-1, or inferences (e.g., decisions or predictions) generated using an AI / ML model.

[0067] In step S403, source base station 5-1 sends a handover request to target base station 5-2 requesting a handover of UE3 from source base station 5-1 to target base station 5-2. The handover request may include, for example, the identification information of source base station 5-1, the cause value for the handover, the identification information of the target cell, UE3 context information (such as the maximum bitrate of UE3 or the security capabilities of UE3), and instructions for UE history information. If the handover was triggered by a measurement report received by source base station 5-1 in step S402, the cause value may indicate, for example, that the handover is desirable for radio reasons. Alternatively, if the handover was triggered at source base station 5-1 to reduce the load, the cause value may indicate that the handover is to reduce the load on the serving cell. The handover request message may also include instructions for the AMF 10-1 servicing UE3.

[0068] In step S404, the target base station 5-2 sends an acknowledgment of the handover request (which may be referred to as the “Handover Request Acknowledgment” message). The Handover Request Acknowledgment message includes instructions for handover configuration information for the handover to be forwarded to the UE3. The Handover Request Acknowledgment message may also include configuration information that enables the source base station 5-1 to begin forwarding the UE3's user plane data to the target base station 5-2.

[0069] The transmissions in steps S403 and S404 may be performed via the Xn interface between the source base station 5-1 and the target base station 5-2 (therefore, the handover procedure in this example may be referred to as the Xn-based handover procedure). Steps S401 to S404 may be referred to as the "handover preparation phase".

[0070] In step S405, the source (R)AN node transmits handover configuration information to the UE3. The handover configuration information may be, for example, an RRC configuration transmitted in an RRC configuration message or an RRC reconfiguration message. In step S406, the UE3 applies the received handover configuration and transmits an instruction to the target base station 5-2 that the handover configuration is complete. The message transmitted in step S405 may be, for example, an RRC reconfiguration complete message. Steps S405 and S406 may be referred to as the "handover execution phase".

[0071] Following the handover execution phase, UE3 can operate to send uplink transmissions (e.g., uplink data) to target base station 5-2 and receive downlink transmissions (e.g., downlink data) from target base station 5-2.

[0072] It will be understood that the mobility methods and handover procedures for UE3 are not limited to the example shown in Figure 4. For example, UE3 may be configured to perform a conditional handover (CHO) in which UE3 determines whether or not to hand over UE3 to a candidate cell based on one or more execution conditions.

[0073] Artificial intelligence (AI) / Machine learning (ML) Figure 5 illustrates the framework for AI / ML models and how the various entities within the framework can interact with each other.

[0074] The entity includes a data acquisition function 41, a model training function 43, a model inference function 45, a model management function 47, and a model storage function 49. The data acquisition function 41 provides input data (training data) to the model training function 43 and the model inference function 45, and also provides monitoring data to the model management function 47. The data to be collected may be, for example, mobility data (such as UE3 handover or UE3 location). For example, the data may be acquired by a base station 5 (for example, by receiving measurement reports from UE3, or by receiving data from another base station 5 or core network node / function) and transmitted to another base station 5 that generates the AI / ML model inference output (or, alternatively, the same base station 5 that acquires the data may generate the AI / ML model output).

[0075] The model training function 43 can perform ML model training, validation, and testing, and generate model performance metrics as part of the model testing procedure. The model inference function 45 provides AI / ML model inference output (e.g., prediction or decision). The output from the model inference function 45 triggers a node performing model inference (which may be, for example, base station 5 or UE3) to perform the corresponding action. The AI / ML model inference output may be, for example, a prediction of UE3 mobility (such as expected path, route or trajectory, inter-cell or inter-beam mobility, or expected handover), or one or more parameters to be used when encoding or decoding transmissions between base station 5 and UE3. The model management function 47 is a function or node that manages the AI / ML models used in the network. The management functions performed by the model management function 47 include monitoring the inference performance of the AI / ML model, enabling / disabling the AI / ML model for a particular function, selecting the AI / ML model for a particular function (if multiple models are available for that function), switching from the current AI / ML model to a different AI / ML model, and falling back to a previous AI / ML model for a particular function. The model memory function 49 maintains a record of available AI / ML models and a record of those deployed (in use) within the network. The functions shown in Figure 8 may be located together on a single node of the communication network (e.g., base station 5 or core network node / function), or they may be distributed among multiple network nodes (e.g., multiple base stations 5) or UE3.

[0076] In the context of this framework, the terms referenced by 3GPP include the following: AI / ML model training: The online or offline process for training AI / ML models. AI / ML Model Validation: A method for evaluating the quality (e.g., predictive accuracy) of an AI / ML model using a dataset that differs from the one used for training the model and can help in selecting model parameters that generalize beyond the training dataset. AI / ML model testing: A method for evaluating the performance of a final AI / ML model using a different dataset than the one used for model training and validation. Unlike model validation, model testing does not anticipate any subsequent adjustments to the model. AI / ML Model Inference: A method of generating a set of outputs based on a set of inputs using a trained AI / ML model. AI / ML Data Collection: Methods for collecting data by network nodes, management entities, and / or UE3 for training AI / ML models, for data analysis (e.g., model performance monitoring), and / or for generating inferences using trained AI / ML models. Model monitoring: A method for monitoring the inference performance (e.g., prediction accuracy) of AI / ML models. Model activation: Enabling an AI / ML model for a specific function. Model deactivation: Disabling an AI / ML model for specific functions. Model switching: Deactivating the currently active AI / ML model and activating a different AI / ML model for a specific function. Training data: Data used as input for the AI / ML model training function. Supervised learning: A method for training AI / ML models using labeled data. Unsupervised learning: A method for training AI / ML models using unlabeled data. Semi-supervised learning: A method for training AI / ML models using both labeled and unlabeled data. Reinforcement learning: A method of training AI / ML models using inputs and feedback resulting from the model's output in situations in which the model interacts. Inference data: Data used as input for the AI / ML model inference function to generate inferences. Model Deployment / Update: Methods for deploying AI / ML models to model inference functions (e.g., sending them to network nodes), or for delivering updated models to model inference functions.

[0077] Data collection 41 may be performed at various nodes of the communication network (for example, one or more base stations 5 or UE3).

[0078] Figure 6 illustrates how to train an AI / ML model and monitor its performance. As shown in Figure 6, the stored data / features 52 may first be extracted in the data extraction step S601. In the data validation step S602, a decision is made (for example, based on the extracted data) whether to proceed with training or retaining the AI / ML model. In the data preparation step S603, the data is prepared for use when training the AI / ML model. For example, the data may be cleaned (e.g., filtered), transformed, or modified in any other appropriate way. The data may also be split into training data, validation data, and test data sets in the data preparation step.

[0079] In the model training step S604, the AI / ML model is trained (or retrained) using the training data created in the data preparation step S603. It will be understood that any appropriate training method (e.g., supervised learning, unsupervised learning, or reinforcement learning) can be used to train the AI / ML model. In the model evaluation step S605, the AI / ML model is evaluated using a test dataset (which may be generated in the data preparation step S603) (e.g., the prediction accuracy of the AI / ML model is evaluated). In the model validation step S606, it is determined (e.g., based on the results of the model evaluation step S605) whether the AI / ML model is suitable for deployment in a communication network.

[0080] In the model provision step S607, the AI / ML model is deployed for use in the communication system 1. AI / ML model deployment may include compiling the trained AI / ML model, packaging the model into an executable format, and delivering the AI / ML model to the target device. For example, the AI / ML model may be sent to base station 5 and / or UE3 for use in base station 5 and / or UE3 to generate predictions or decisions using the AI / ML model as part of the prediction service step S608 shown in Figure 6. In the performance monitoring step S609, the performance of the deployed AI / ML model is monitored. The prediction performance of the AI / ML model may be monitored by comparing the predictions generated using the model with one or more measurements. For example, if the AI / ML model is used to predict the location of UE3, the prediction accuracy of the AI / ML model may be evaluated using measurements of the actual location of UE3. Alternatively, if the AI / ML model is used to determine the parameters used when encoding and decoding data transmitted between base station 5 and UE3, the model may be evaluated based on the performance of the encoding and / or decoding process. In the retraining trigger step S610, the AI / ML model is retrained (for example, because the prediction accuracy of the AI / ML model falls below an acceptable threshold accuracy, or because the performance of the method using inference from the AI / ML model falls below an acceptable threshold performance), and the method returns to the data extraction step S601.

[0081] Single-sided model and double-sided model The AI / ML model may be hosted at both base station 5 and UE3, at base station 5 only, or at UE3 only. When the AI / ML model is used only at UE3 or only at the base station, it may be referred to as a "one-sided" model. For example, UE3 may host an AI / ML model for generating time (e.g., time resources) for communication using a specific beam transmitted by base station 5. However, even if the model is a one-sided model, it will be understood that the model does not necessarily need to be trained at UE3. For example, the model may be trained at base station 5 or at another node in the network (e.g., a core network node / function) and then transmitted to UE3 for use at UE3. In other words, the AI / ML model may be trained at another network node and then transferred / deployed to UE3.

[0082] Alternatively, the AI / ML model may be a "two-sided" model in which one AI / ML model is hosted in UE3 and the corresponding AI / ML model is hosted in base station 5. The AI / ML model hosted in UE3 and the AI / ML model hosted in base station 5 may be the same AI / ML model. UE3 can use the AI / ML model to generate a first inference, and base station 5 can use the AI / ML model to generate a corresponding second inference. For example, the first inference may be an inference of parameters used to compress data (e.g., channel state information (CSI)) transmitted from UE3 to base station 5, and the second inference may be an inference of parameters used to decompress the data in base station 5. As with one-sided models, two-sided models (or multiple models) can be trained on any suitable network node and then transmitted to UE3 and base station 5.

[0083] Model Monitoring As described above, model monitoring can be performed by UE3 or by network nodes such as base station 5. In the case of UE monitoring, UE3 performs most of the model monitoring and provides the network (e.g., base station 5) with limited information about model monitoring performance so that the network (e.g., base station 5) can make model management decisions. It is also possible for UE3 to make autonomous model management decisions and notify the network (e.g., base station 5) of the decisions it has made so that the network (e.g., base station 5) can perform the necessary UE reconfiguration for radio operation. For example, if the UE decides to deactivate the AI / ML beam prediction model, the network (e.g., base station 5) may need to configure the transmission of an additional Reference Signal (RS) so that UE3 can still perform legacy beam measurements.

[0084] Network Model Monitoring In the case of network-based monitoring, UE3 reports detailed information (e.g., actual measurements, model predictions, etc.) to the network (e.g., base station 5), which then makes appropriate model management decisions based on the information provided. Model monitoring involves UE3 sending a set of information to the network (e.g., base station 5) that includes one or more of the following: 1) AI / ML prediction output, and 2) ground truth data, i.e., the results of measurements performed by UE3 that can be used by the network (e.g., base station 5) to determine whether the AI / ML prediction output is correct or incorrect. It is assumed that both of the above pieces of information are provided to the network (e.g., base station 5) by UE3 within a single message and call, such as a report "Performance Report".

[0085] The network model monitoring framework may involve the following stages: determining the UE capabilities for monitoring; determining the network (e.g., base station 5) configuration for performance reporting; determining the UE performance reporting procedure; and determining the content of the UE performance report. Each of these stages is described in more detail below.

[0086] UE capabilities for monitoring The inventors recognized that because the required UE reporting / measurements differ for different AI / ML functions / models, a UE capability report should be defined that indicates which reporting / measurement functions a UE supports. To address this issue, the inventors propose extending the current UE capability report to include information indicating which reporting quantities / metrics a UE3 can support for AI / ML performance reporting for each AI / ML function / model. Furthermore, because the measurements required for AI / ML performance reporting can consume significant resources, different UE3s may have different capabilities regarding how often a UE3 can perform measurements. Therefore, the inventors propose also indicating how often (time-periodically) a UE3 can perform one or more measurements for AI / ML performance reporting for each AI / ML function / model.

[0087] The inventors also propose defining dynamic UE capabilities. In this case, if a network (e.g., base station 5) configures monitoring / activation for UE3 for one or more AI / ML functions, UE3 can indicate whether or not it can support the monitoring procedure. This may depend on various internal UE factors such as power consumption, battery level, and internal capability for inference measurements, which can be determined based on the network-provided configuration that determines whether or not UE3 can support the monitoring procedure.

[0088] The inventors have also recognized that there may be cases where the network (e.g., base station 5) would like the UE3 to support AI / ML performance reporting without the UE3 using the AI / ML model to make predictions used to control UE / network RAN ​​operation. This may occur if the network is testing or validating the AI / ML model before deployment. At this stage, the network (e.g., base station 5) may simply want to determine the predictive accuracy of the AI / ML model before deployment. To support this operation, the inventors propose that the network provide separate network instructions for 1) model management decisions and 2) AI / ML performance reporting. In this way, the network can use the first instruction to provide instructions for activating / deactivating / switching / falling back the AI / ML model and separately indicating whether the UE3 should perform AI / ML performance reporting. The instructions may be provided in different messages or as different fields within a single message.

[0089] For example, a network (e.g., base station 5) may request UE3 to run an AI / ML model for CSI-RS beam prediction and report the prediction results to the network. Simultaneously, the network may configure the UE to perform legacy CSI-RS measurements for normal radio operation (i.e., without any AI / ML model assistance). The network may send separate messages to UE3 for model management and performance reporting, one message for model activation / deactivation / switching / fallback and the other message for enabling / disabling AI / ML performance reporting. Alternatively, the network may send a single message for both model management and performance reporting, with one field in the message for model management decisions (activation / deactivation / switching / fallback) and the other field for enabling / disabling AI / ML performance reporting.

[0090] UE3 uses only the AI / ML model for RAN operation when the model is activated by the network (e.g., base station 5), and if it is not activated and AI / ML performance reporting is enabled, UE3 operates the model only for AI / ML performance reporting, and the AI / ML model is not used for RAN operation.

[0091] If multiple AI / ML models exist for the same AI / ML function, the network (e.g., base station 5) may want UE3 to activate one AI / ML model for RAN operation and run another AI / ML model for AI / ML performance reporting for test purposes for the same AI / ML function. We propose that UE3 may demonstrate the ability to run multiple AI / ML models simultaneously for the same AI / ML function. This ability may be AI / ML function / model specific. If UE3 can run multiple AI / ML models simultaneously, the network (e.g., base station 5) can configure different AI / ML models for activation (RAN operation) and AI / ML performance monitoring. UE3 then uses the AI / ML model activated for RAN operation and uses the AI / ML model only for performance reporting.

[0092] Network configuration for performance reporting The inventors propose that a network (e.g., base station 5) configures AI / ML performance reporting for each AI / ML function. To enable this, the network (e.g., base station 5) should include a function identifier (identifier:ID) in the configuration to indicate to UE3 which AI / ML functions require AI / ML performance reporting. Furthermore, since each AI / ML function may be associated with two or more AI / ML models or model configurations (e.g., input / output configurations), the network (e.g., base station 5) may also need to indicate the models / model configurations on which performance reporting should be performed. This can be achieved by including a model identifier (identifier:ID) in the configuration message sent by the network to UE3 that uniquely identifies the AI / ML model on which performance reporting should be initiated. Alternatively, the model configuration (e.g., input / output configuration) can be included in the configuration message to provide a complete description of the model that UE3 uses for performance monitoring.

[0093] The network (e.g., base station 5) can configure an independent radio bearer (RB) that UE3 should use for performance reporting, and the RB parameters and QoS parameters are specifically defined for data collection requirements for life cycle management (LCM). For example, a data radio bearer (DRB) may be defined so that the UE's performance reports are sent directly to an external server.

[0094] UE Measurement Procedure Performing measurements for AI / ML performance reporting may require the UE3 to run the AI / ML model and perform legacy measurements (to provide ground truth data for the AI / ML model), which can be significantly power-intensive for the UE3. Furthermore, the UE3 may take time to complete measurements for AI / ML inference and ground truth data measurement. Therefore, the UE3 must know how frequently it needs to perform the required measurements and when to perform them, so that the legacy measurements correspond to the AI / ML performance measurements. The inventors propose defining a “UE Measurement Opportunity Selection for AI / ML Performance Reporting” parameter that specifies the opportunities and periodicity of AI / ML model measurements. This information may be specified to the UE3 by the network (e.g., base station 5) when the network is configuring the UE for AI / ML performance reporting. Alternatively, UE3 may determine its own measurement opportunity periodicity based on its own capabilities or the periodicity of the reference signal / other signal it uses for measurement, or based on reported periodicity provided by the network (e.g., base station 5), or based on maximum / minimum periodicity values ​​configured by the network or pre-programmed in UE3.

[0095] Regarding the execution of ground truth data measurements, given that the ground truth data is provided to determine the accuracy of the AI / ML prediction output, UE3 must complete the ground truth data measurements within a defined period of time instance associated with the AI / ML prediction output. For example, if the AI / ML use case is CSI-RS beam prediction for time opportunity T, all CSI-RS measurements relevant to determining the performance of the AI / ML prediction output should be performed within the time period {T-Threshold, T+Threshold}. The time limit / threshold may be pre-programmed in UE3 or configured by the network (e.g., base station 5), and different values ​​may be defined for different AI / ML functions / models. If UE3 is unable to perform all measurements within the time limit, UE3 should indicate to the network in the performance report any valid measurements that could have been performed within the time limit.

[0096] UE Reporting Procedure Once UE3 has performed the necessary measurements, it needs to know when to report them. We propose that the network (e.g., base station 5) defines this using a “reporting opportunity configuration” that it transmits to UE3. One option (Alt-1) for this configuration is for the network (e.g., base station 5) to define the reporting periodicity to UE3. Since use case requirements for each AI / ML model may differ (e.g., model requirements for an AI / ML model for SSB-based beam prediction may have different required reporting periodicity compared to the reporting periodicity required by the AI / ML model used to determine CSI compression parameters), different periodicity values ​​are defined by the network (e.g., base station 5) for each AI / ML model.

[0097] Instead of the network (e.g., base station 5) defining the reporting periodicity to UE3, as an alternative (Alt-2), the network (e.g., base station 5) may define one or more triggers (for each AI / ML model) that define when UE3 should send AI / ML metric reports. The triggers can be based on one or more metrics and can cover either when the AI / ML model is not performing well or when it is performing well. For example, the network (e.g., base station 5) may set an AI / ML metric (e.g., AI / ML prediction accuracy) and set an event that UE3 reports to the network when the AI / ML performance metric is better than a first threshold, and / or an event that UE3 reports to the network when the AI / ML performance metric is worse than a second threshold. As a further alternative (Alt-3), the network (e.g., base station 5) may poll (instruct) UE3 to perform AI / ML performance reporting when it wants UE3 to do so. In response, UE3 compiles the reporting metrics and other relevant measurement results for reporting to the network (e.g., base station 5).

[0098] These different reporting options are shown in Figure 7. As shown, in step S701, base station 5 initiates the procedure by sending an AI / ML performance reporting configuration to UE3. UE3 then performs the configured measurements, some of which are labeled 71. The measurements are performed periodically at intervals between consecutive measurements. The upper dashed box shown in Figure 7 (labeled Alt-1) illustrates the first alternative described above, where UE3 performs measurements and then reports periodically (in steps S702-1 and S702-2) using the time between reports defined by the period labeled “Reporting Interval” in Figure 7. The middle dashed box shown in Figure 7 (labeled Alt-2) illustrates the second alternative described above, where in step S702-3, base station 5 configures UE3 to send a report in response to some event “Reporting Trigger” 73. As described above, this Reporting Trigger 73 may be that the performance quality of the AI / ML model is better than a threshold. The lower dashed box shown in Figure 6 (labeled Alt-3) illustrates the third alternative described above, in step S703, when base station 5 wants UE3 to report the AI / ML performance measurements that UE3 performs in steps S702-4, base station 5 sends a report polling message to UE3.

[0099] Reporting for UE Model Monitoring In the case of UE model monitoring, UE3 needs to determine whether the AI / ML model is functioning correctly, and this decision can be supported by information from the network (e.g., base station 5) or by predefined (pre-stored or pre-programmed) information or rules.

[0100] For example, UE3 may monitor one or more monitoring metrics related to an AI / ML model / function within a time window. The monitoring metrics are configured by the network or predefined for the AI / ML model / function. An exemplary monitoring metric could be the AI / ML model prediction error (other detailed metrics are described later). UE3 uses at least one criterion to determine what model management action (e.g., activation / deactivation) the UE should take. At least one criterion can be in the form of a threshold check of a monitoring metric, and at least one criterion and its associated parameters (e.g., threshold) can be preprogrammed in UE3 or configured for UE3 by the network (e.g., base station 5). Multiple criteria can be defined / configured, and each criterion can be associated with one type of model management decision (e.g., the network may provide one criterion for model activation, and different criteria for model deactivation). The time window in which the UE performs monitoring may be set by the network or selected by UE3.

[0101] If the criteria check is successful, UE3 will perform one of the following steps:

[0102] Option 1: UE3 makes model management decisions for the AI / ML model (e.g., activation / deactivation / switching) and communicates these decisions to the network (e.g., base station 5). The model management decisions made are those related to criteria considered successful. UE3 sends a report to base station 5 indicating that the criteria have been met and waits for a specific time delay before implementing the model management decision. This delay allows network time to perform any radio reconfiguration. The report may include an AI / ML model / function identifier and an identification of either the model management decision made by UE3 (e.g., activation / deactivation) or at least one criterion for success. Different delay values ​​can be specified / configured for different model management actions.

[0103] Option 2: UE3 sends a report to the network (e.g., base station 5), and UE3 awaits confirmation of a decision from the network. In this case, the report sent by UE3 to base station 5 may indicate the AI / ML model / function identifier and either a model management decision made by UE3 (e.g., activation / deactivation) or an identification of at least one successful criterion. The network response may be in the form of an acceptance or rejection message, or may include a model management decision to be made by UE3.

[0104] These different reporting options are shown in Figure 8. As illustrated, in step S801, base station 5 transmits a UE monitoring configuration to UE3 via RRC signaling, which is provided per AI / ML function / model. UE3 then performs the configured measurements, some of which are labeled 81. Measurements are performed periodically with measurement intervals between consecutive measurements. Each time one or more measurements are performed, UE3 checks whether one or more measurements meet at least one defined criterion. In the exemplary scenario shown in Figure 8, we see that the last measurement shown meets at least one criterion for reporting.

[0105] The upper dashed box shown in Figure 8 (labeled Option-1) represents the first option described above, and if the measurement indicates that the AI / ML model or function meets at least one criterion, UE3 reports to base station 5 (in step S802-1). As described above, the report may include the AI / ML model / function identifier and identification information of the model management decision made by UE3 or at least one successful criterion (e.g., activation / deactivation). UE3 then waits for a period of time (indicated as “Action Delay” in Figure 8) before implementing the determined management decision (activation / deactivation / switching / fallback) in step S803.

[0106] The lower dashed box in Figure 8 (labeled Option 2) illustrates the second option described above, in which, if the measurement indicates that the AI / ML model or function meets at least one criterion, UE3 reports to base station 5 in step S802-2 that it has the AI / ML model / function identifier and the identification of the model management decision made by UE3 (e.g., activation / deactivation) or the successful criterion of at least one criterion. UE3 then waits in step S804 to receive a network response message, either in the form of an acceptance or rejection message, or containing the model management decision made by UE3.

[0107] Lower-tier UE report Some AI / ML use cases may require low-latency performance reporting. For example, if UE-based performance reporting is used and UE3 determines, for example, that a beam prediction algorithm is not performing well, UE3 should immediately send the report to the network to ensure that the network can take action as quickly as possible. Therefore, lower-layer-based reporting is advantageous in such situations. To enable such reporting at the lower layers (e.g., MAC CE / UCI), a mechanism must be defined for how information is reported by UE3 to the network (e.g., base station 5). The inventors have proposed the following reporting mechanism.

[0108] Reporting via UCI If the number of bits for AI / ML performance reporting is greater than the threshold, new Uplink Control Information (UCI) may be defined (compared to the UCI used for ACK / NACK, CSI, and SR). If the number of bits is less than the threshold, the same UCI may be used for reporting, as the UCI is used for reporting ACK / NACK / CSI / SR reports. If two different UCIs are used by the UE, one for AI / ML performance reporting and the other for ACK / NACK / SR / CSI reporting, the network (e.g., base station 5) identifies the type of UCI based on a number of factors, including one or more of the following: - The UCI ID may be included within a UCI to indicate the type of UCI to the network. Each UCI type transmission is associated with a set of radio resources (e.g., time, frequency resources, and / or carrier). When a network receives a UCI, it can determine the type of UCI based on the radio resources from which the UCI was received. - Each UCI type transmission is associated with a set of physical channels. For example, a UCI for AI / ML performance reporting is transmitted within a configured "UL configuration grant" PUSCH resource, while a UCI for ACK / NACK / SR / CSI is transmitted over other PUSCH or PUSCH resources. Based on the physical channel on which the UCI is received, the network can determine the type of UCI.

[0109] Reporting via MAC CE A new Logical Channel Identifier (LCID) value may be defined to indicate UE inference reporting for AI / ML. One option is to use a single LCID value for multiple AI / ML functions / models. Alternatively, different LCID values ​​may be defined for different AI / ML functions / models. A variable-size Medium Access Control (MAC) Control Element (CE) may be defined for reporting, and the size of the MAC CE may be reported by UE3 within the MAC CE or configured by the network (e.g., based on the reporting configuration).

[0110] The above options may only be possible when the reported content is below a certain threshold; therefore, if UE3 needs to send a larger payload, we propose one of the following options.

[0111] Option 1: UE3 transmits a portion of the payload to base station 5 via lower layers (e.g., only indicating model management decisions or model failures), and the remaining payload is transmitted to the base station via RRC signaling.

[0112] Option 2: If the payload size is below the threshold, UE3 sends the payload to base station 5 via lower layer reporting; otherwise, UE3 sends the report to base station 5 via RRC signaling.

[0113] Option 3: Different channel types (e.g., RRC, MAC CE, or UCI) are used by UE3 for different types of reports transmitted to base station 5. For example, UE3 may use RRC signaling to send periodic performance monitoring reports when the AI / ML algorithm is working properly, or UE3 may use UCI / MAC CE for reporting when it needs to report a failure in AI / ML operation or when the AI / ML function is not working properly.

[0114] The selection of the type of information transmitted over which channel can be pre-programmed into UE3 or configured by the network (e.g., base station 5).

[0115] Report content For UE reporting or benchmark checks, multiple different metrics specific to each AI / ML use case can be defined. The network (e.g., base station 5) may configure the metrics to be used by UE3, or they may be pre-programmed in UE3. Metrics can be of two types. Type 1: Performance metrics, e.g., user throughput, handover failure rate / success rate, beam failure rate, BLER, etc. Type 2: AI / ML Prediction Metric: How well is the AI / ML operation performing?

[0116] The network (e.g., base station 5) can be configured in UE3 to consider both types of metrics or only one type of metric for performance reporting. Some examples of AI / ML predictive metrics that may be supported for the functionality include: 1) The number / percentage of instances in which the AI / ML model was used to perform the required RAN procedure (compared to the legacy procedure). This information may be useful if UE3 determines not to use the AI / ML model for a particular function (e.g., beam prediction) when it observes that the accuracy or confidence level of the AI / ML output does not exceed a certain threshold. This may also be useful if the AI / ML model is not run by UE3 due to problems in the input feature data (e.g., missing or inaccurate input features). Additionally or alternatively, UE3 may be configured to provide the total number of instances in which the AI / ML model was not run. This information may be provided along with one or more reasons for not running the AI / ML model (e.g., low confidence or missing data in AI / ML prediction). UE3 may also be configured to provide the number of instances in which the AI / ML model was not run for each reason (e.g., low confidence or inaccurate / missing data). 2) In the case of a classification problem, a typical metric to be reported can be a confusion matrix, which is a table (shown below) where the rows of the matrix represent instances of the actual class and each column represents an instance of the predicted class, or vice versa. In the example below, the actual class can be positive or negative, and the prediction can be either positive or negative. Thus, if the prediction is positive and the actual class is positive, this is called a "true positive"; if the prediction is positive and the actual class is negative, this is called a "false positive"; if the prediction is negative and the actual class is positive, this is called a "false negative"; and if the prediction is negative and the actual class is negative, this is called a "true negative". The metrics that can be reported by UE3 can be either the confusion matrix or a set of parameters derived from the confusion matrix (e.g., true positive, false positive, etc.). [Table 2] 3) If the AI / ML model predicts N best objects (for example, N best beams, or N best cells, or N best SSB / CSI resources), possible metrics that UE3 can report include the number or percentage of times the best object predicted by the AI / ML model is among (or not among) the M best objects actually observed by UE, and / or the number or percentage of times the best object actually measured by UE3 is among the N best objects predicted by the AI / ML model. 4) When an AI / ML model predicts the value of a parameter (e.g., location or CSI intensity), possible metrics may include the number or percentage of times the predicted value of the parameter falls within (or falls outside) the threshold range of actual observations of the same parameter, and / or the number or percentage of times the observed value of the parameter falls within (or falls outside) the threshold range of the predicted value of the same parameter.

[0117] Handling mobility events As described above, UE3 may move from source base station 5-1 to target base station 5-2 (or, in practice, a change of cell operated by the same base station), and this affects the AI / ML models / functions and measurements that UE3 performs. When UE3 is handed over to a new cell or a new base station, the inventors propose that UE3 either discard all stored measurements for AI / ML performance reports related to the source cell, or, if the AI / ML model has not changed after the cell change / handover, UE3 may not discard stored measurements for AI / ML performance reports related to the source cell. If UE3 does not discard measurements from the source cell, UE3 may include reports for both the source and target cells in the next reporting opportunity. UE3 may also include a cell / network node identifier in each such report. For example, a report may include the following information: AI / ML Performance Report - 1: Cell / Network Node Identifier - X: Measurement - 1, Measurement - 2, ... AI / ML Performance Report - 2: Cell / Network Node Identifier - Y: Measurement - 1, Measurement - 2, ...

[0118] The new base station (or cell) can then transfer the relevant inference measurements to the source base station (or cell).

[0119] In the case of centralized data collection, a new base station (cell) can also forward data to a specific address (e.g., an AI / ML server), as long as this information is shared by a handover request message from the serving base station (cell) to the target base station (cell) before the handover. However, if the target base station (cell) can benefit from data measurements from the UE's monitoring history (e.g., its AI / ML monitoring results), these data / measurements can also be used by the target base station (cell) to assist in its AI / ML management decisions.

[0120] Actions for model switching As described above, if multiple AI / ML models exist for the same function, base station 5 may decide to switch the AI / ML model being used by UE3. When this happens, UE3 must do something with the measurements it has collected and the monitoring procedures it uses before the switch. One option is for UE3 to be configured to discard any stored measurements taken before the model switch and to restart the monitoring procedures for the new model. Alternatively, UE3 may be configured to retain any stored measurements before the model switch and to continue the monitoring procedures for the new AI / ML model, then, when it is time to report the AI / ML results, UE3 reports the results for the old AI / ML model along with any results from the new AI / ML model. UE3 may also include a model identifier with each such report, and therefore a report may include: AI / ML Performance Report - 1: Model Identifier - X: Measurement - 1, Measurement - 2, ... AI / ML Performance Report - 2: Model Identifier - Y: Measurement - 1, Measurement - 2, ...

[0121] User equipment Figure 9 is a schematic block diagram showing the main components of the UE3 shown in Figure 1. As shown, the UE3 has a transceiver circuit 310 capable of transmitting and receiving signals to and from a base station 5 via one or more antennas 330 (e.g., including one or more antenna elements). The UE3 has a controller 370 (which may be a microprocessor) that controls the operation of the UE3. The controller 370 is associated with memory 390 and coupled to the transceiver circuit 310. Although not necessarily required for its operation, the UE3 can, of course, have all the usual functions of a conventional UE3 (e.g., a user interface 350 such as a touchscreen / keypad / microphone / speaker to enable direct control and interaction with the user), which can be provided, as appropriate, by one or any combination of hardware, software, and firmware. The software may be pre-installed in memory 390 and / or downloaded, for example, via a telecommunications network or from a removable data storage device (RMD).

[0122] In this example, the controller 370 is configured to control the overall operation of the UE3 by program instructions or software instructions stored in memory 390. As shown in the figure, these software instructions include, among other things, the operating system 410, the communication control module 430, and the AI / ML module 450. The communication control module 430 is operable to control communication between the UE3 and one or more serving base stations 5 (and further UE3s and / or other communication devices connected to the base stations 5, such as the core network node 10).

[0123] The communication control module 430 is configured for the overall processing of uplink communications via the relevant uplink channels (e.g., via the physical uplink control channel (PUCCH), random access channel (RACH), and / or physical uplink shared channel (PUSCH)), including both dynamic and semi-static signaling (e.g., SRS). The communication control module 430 is also configured for the overall processing of downlink communications received via the relevant downlink channels (e.g., via the physical downlink control channel (PDCCH) and / or physical downlink shared channel (PDSCH)), including both dynamic and semi-static signaling (e.g., CSI-RS). The communication control module 430 is also responsible for, for example, determining where to monitor downlink control information (e.g., the locations of CSS / USS, CORESET, and related PDCCH candidates to monitor), determining resources used by the UE3 for transmitting / receiving UL / DL communications (including interleaved resources and resources subject to frequency hopping), managing frequency hopping on the UE side, determining how slots / symbols are configured (e.g., for UL, DL, or SBFD communications), determining which one or more bandwidth portions are configured for the UE3, determining how uplink transmissions should be encoded, and appropriately applying any SBFD-specific communication configurations. The communication control module 430 may be configured to control communications in any of the ways described above (e.g., to transmit measurement reports in any of the ways described above).

[0124] The AI / ML module 450 is operable to control the use of one or more AI / ML models in UE3 (for example, to generate one or more inferences using one or more models). The AI / ML module 450 is also configured to perform any of the AI / ML-related functions of UE3 in any of the manner described above, including making AI / ML decisions, performing AI / ML measurements, and reporting the measurement results to a network (e.g., base station 5).

[0125] base station Figure 10 is a schematic block diagram showing the main components of base station 5 of the communication system 1 shown in Figure 1. As shown, base station 5 has a transceiver circuit 510 for sending and receiving signals to and from communication devices (such as UE3) via one or more antennas 530 (such as single or multi-panel antenna arrays / massive antennas), and a core network interface 550 (including, for example, N2, N3, and other reference points / interfaces) for sending and receiving signals to and from network nodes in the core network 7. Although not shown, base station 5 may be coupled to other base stations via appropriate interfaces (such as the so-called "Xn" interface in NR). Base station 5 has a controller 570 (which may be a microprocessor) that controls the operation of base station 5. Controller 570 is associated with memory 590. Software may be pre-installed in memory 590 and / or downloaded, for example, via the communication system 1 or from a removable data storage device (RMD). In this example, the controller 570 is configured to control the overall operation of the base station 5 by program instructions or software instructions stored in memory 590. As shown in the figure, these software instructions include, among other things, the operating system 610, the communication control module 630, and the AI / ML module 650.

[0126] The communication control module 630 is operable to control communication between the base station 5, the UE3, and other network entities connected to the base station 5. The communication control module 630 is configured for overall control of receiving and decoding uplink communications over relevant uplink channels (e.g., via the physical uplink control channel (PUCCH), random access channel (RACH), and / or physical uplink shared channel (PUSCH)), including both dynamic and semi-static signaling (e.g., SRS). The communication control module 630 is also configured to handle overall transmission of downlink communications over relevant downlink channels (e.g., via the physical downlink control channel (PDCCH) and / or physical downlink shared channel (PDSCH)), including both dynamic and semi-static signaling (e.g., CSI-RS). The communication control module 630 is responsible for managing full-duplex (e.g., SBFD) communications, including, where appropriate, the separation of UL and DL communications through different physical antenna elements. The communication control module 630 is responsible for, for example, determining where to configure UE3 to monitor downlink control information (e.g., the locations of CSS / USS, CORESET, and associated PDCCH candidates to monitor), determining resources to be scheduled for UE transmission / reception of UL / DL communications (including interleaved resources and resources subject to frequency hopping), managing frequency hopping on the base station side, appropriately configuring slots / symbols (e.g., for UL, DL, or SBFD communications), configuring one or more bandwidth portions for UE3, and providing configuration signaling related to UE3.

[0127] The AI / ML module 650 may be configured to perform any of the AI / ML-related functions of UE3 in any of the ways described above, including running AI / ML models, deploying AI / ML models to UE3, making management decisions related to deployed AI / ML models, and configuring UE to perform AI / ML performance measurement and AI / ML performance reporting.

[0128] Examples of modifications and alternatives As those skilled in the art will understand, several modifications and substitutions can be made to the exemplary embodiments described above, while still benefiting from the disclosures embodied therein. While the above example illustrates the concept using an AI / ML model, it should be understood that the above method is also advantageous when the model is not an AI / ML model. Any other suitable type of model or function can be used to generate inference (e.g., decision or prediction).

[0129] For example, while specific terms for cellular communication generations (2G, 3G, 4G, 5G, 6G, etc.) may be used to refer to specific communication entities for clarity, it will be understood that the technical features described for a given entity are not limited to devices of that particular communication generation. Technical features can be implemented in any functionally equivalent communication entity, regardless of any differences in the terminology used to refer to them.

[0130] In the above description, the UE and base station are described as having several separate functional components or modules for ease of understanding. These modules may thus be provided for a specific application, for example, in which an existing system is modified to implement the present disclosure, but in other applications, for example, in a system designed from the outset with the features of the present invention in mind, these modules may be incorporated into the operating system or the entire code, and therefore these modules may not be identifiable as separate entities.

[0131] In the exemplary embodiments described above, several software modules have been described. As those skilled in the art will understand, software modules may be provided in compiled or uncompiled form and may be supplied as signals via a computer network or on a recording medium. Furthermore, some or all of the functions performed by this software may be performed using one or more dedicated hardware circuits. However, the use of software modules is preferred because it facilitates the updating of base stations or UEs to update their functions.

[0132] Each of the controllers described above may include, but is not limited to, any suitable form of processing circuitry, including, for example, one or more hardware-implemented computer processors, microprocessors, central processing units (CPUs), arithmetic logic units (ALUs), input / output (IO) circuits, internal memory / cache (programs and / or data), processing registers, communication buses (such as control buses, data buses, and / or address buses), direct memory access (DMA) functions, hardware or software-implemented counters, pointers, and / or timers. Various other modifications will be obvious to those skilled in the art and will not be described in further detail here.

[0133] One or more of the base stations 5 may be a “distributed” base station having a central unit ("CU") and one or more individual distributed units ("DU").

[0134] In this disclosure, user equipment (or "UE," "mobile station," "mobile device," or "wireless device") is an entity connected to a network via a wireless interface.

[0135] Please note that this disclosure is not limited to dedicated communication devices, but may apply to any device having communication capabilities, as described in the following paragraphs.

[0136] The terms “User Equipment” or “UE,” “Mobile Station,” “Mobile Device,” and “Wireless Device” (terms used by 3GPP) are generally considered synonymous with each other and include standalone mobile stations such as terminals, cell phones, smartphones, tablets, cellular IoT devices, IoT devices, and machines. The terms “Mobile Station” and “Mobile Device” will also be understood to include devices that remain stationary for extended periods.

[0137] UE may be items of equipment for production or manufacture and / or items of energy-related machinery, such as equipment or machinery (for example, boilers, engines, turbines, solar panels, wind turbines, hydroelectric generators, thermal generators, nuclear generators, batteries, nuclear systems and / or related equipment, heavy electrical machinery, pumps including vacuum pumps, compressors, fans, blowers, hydraulic equipment, pneumatic equipment, metalworking machinery, manipulators, robots and / or their application systems, tools, molds or dies, rolls, conveying equipment, elevators, material handling equipment, textile machinery, sewing machinery, printing and / or related machinery, paper conversion machinery, chemical machinery, mining machinery and / or construction machinery and / or related equipment, machinery and / or equipment for agriculture, forestry and / or fisheries, safety and / or environmental protection equipment, tractors, precision bearings, chains, gears, power transmission equipment, lubrication equipment, valves, pipe fittings and / or application systems for any of the aforementioned equipment or machinery, etc.).

[0138] UE may be an item of transport equipment, such as (for example, transport equipment such as railway cars, automobiles, motorcycles, bicycles, trains, buses, carts, rickshaws, ships and other vessels, aircraft, rockets, satellites, drones, balloons, etc.). UE may also be an item of information and communication equipment, such as (for example, information and communication equipment such as electronic computers and related equipment, communication and related equipment, electronic components, etc.).

[0139] UE may include, for example, refrigerators, refrigerator applications, commercial and / or service industry equipment items, vending machines, automated service machines, office machines or equipment, and household appliances and electronic devices (such as audio equipment, video equipment, loudspeakers, radios, televisions, microwave ovens, rice cookers, coffee machines, dishwashers, washing machines, dryers, electronic fans or related equipment, vacuum cleaners, etc.).

[0140] The UE may be an electrical application system or device, for example, such as an X-ray system, particle accelerator, radioisotope equipment, sound wave equipment, electromagnetic application equipment, power application equipment, etc.

[0141] UE may include, for example, electronic lamps, lighting fixtures, measuring instruments, analyzers, testers, or measuring or detecting equipment (such as smoke detectors, human alarm sensors, motion sensors, wireless tags, etc.), watches or clocks, laboratory equipment, optical devices, medical equipment and / or systems, weapons, tableware, hand tools, etc.

[0142] The UE may be, for example, a wirelessly equipped personal digital assistant or related equipment (such as a wireless card or module designed to be attached to or inserted into another electronic device, such as a personal computer or electrical measuring instrument).

[0143] The UE may be part of a device or system that uses various wired and / or wireless communication technologies to provide the following uses, services, and solutions related to the Internet of Things (IoT).

[0144] Internet of Things (IoT) devices (or "Things") may comprise appropriate electronics, software, sensors, network connectivity, etc., that enable these devices to collect and exchange data with each other and with other communication devices. IoT devices may comprise automated equipment that follows software instructions stored in internal memory. IoT devices may operate without requiring human supervision or interaction. IoT devices may also remain stationary and / or inactive for extended periods. IoT devices may be implemented as part of (generally) stationary equipment. IoT devices may also be incorporated into non-stationary equipment (e.g., such as a vehicle) or attached to animals or people being monitored / tracked.

[0145] It will be understood that IoT technology can be implemented on any communication device that can connect to a communication network to send / receive data, regardless of whether such communication device is controlled by human input or by software instructions stored in memory.

[0146] It should be understood that IoT devices are sometimes referred to as Machine-Type Communication (MTC) devices or Machine-to-Machine (M2M) communication devices. It should be understood that a UE may support one or more IoT or MTC applications. Some examples of MTC applications are listed in the table below. This list is not exhaustive and is intended to illustrate some examples of machine-type communication applications. [Table 3]

[0147] Applications, services, and solutions may include Mobile Virtual Network Operator (MVNO) services, emergency radio communication systems, Private Branch eXchange (PBX) systems, PHS / digital cordless telecommunications systems, Point of Sale (POS) systems, incoming advertising systems, Multimedia Broadcast and Multicast Service (MBMS), Vehicle to Everything (V2X) systems, train radio systems, location-related services, disaster / emergency wireless communication services, community services, video streaming services, femtocell application services, Voice over LTE (VoLTE) services, billing services, wireless on-demand services, roaming services, activity monitoring services, telecommunications carrier / communication network selection services, function restriction services, Proof of Concept (PoC) services, personal information management services, and ad-hoc network / delay-tolerant networking (DTN) services.

[0148] Furthermore, the UE categories described above are merely examples of applications of the technical concepts and embodiments described herein. It goes without saying that these technical concepts and embodiments are not limited to the UEs described above and can be modified in various ways.

[0149] Various other modifications are obvious to those skilled in the art and will not be described in further detail here.

[0150] For example, some or all of the above embodiments may also be described as follows, but are not limited to the following. (Note 1) A method performed by user equipment (UE), the method is: Sending UE capability information to network nodes Includes, UE capability information indicates which reporting quantities or metrics the UE can support for model performance reporting. method. (Note 2) UE capability information is the method described in Appendix 1, which indicates the UE's ability to perform measurements for model performance reporting for a model or each model. (Note 3) The method according to Appendix 1 or 2, further comprising receiving a configuration from a network node to perform a model performance report, and transmitting the configuration in response to receiving the configuration. (Note 4) UE capability information includes information indicating the UE's ability to run multiple models simultaneously for the same function, as described in any one of the appendices 1 to 3. (Note 5) A method performed by user equipment (UE), the method is: The network node receives a first instruction for model management decisions and a second instruction for model performance reporting. If the first instruction does not indicate model activation, and the second instruction enables model performance reporting, then the model will be run, and model performance will be reported to the network node without using the output from the model to control communication with the network node. Methods that include... (Note 6) If the first instruction indicates model activation and the second instruction enables model performance reporting, the method is the method described in Appendix 5, which includes running the model, using the output from the model to control communication with network nodes, and reporting model performance to the network nodes. (Note 7) The first and second instructions are received in the same message or in different messages, as described in Appendix 5 or 6. (Note 8) If a first instruction indicates a first model and a second instruction indicates enabling model performance reporting for a second model different from the first model, the method is the method described in any one of the appendices 5 to 7, comprising: running the first model and using the output from the first model to control communication with a network node; and running the second model and reporting the model performance of the second model to the network node without using the output from the second model to control communication with the network node. (Note 9) The first and second models are for the same function, as described in Appendix 8. (Note 10) A method performed by user equipment (UE), the method is: Receiving configuration data from network nodes for model performance reporting regarding multiple model functions, This involves running the model for each function configured by the configuration data, Report the model performance for each model function according to the configuration data, Methods that include... (Note 11) The configuration data is as described in Appendix 10, including a function identifier for each function for which a model performance report is required. (Note 12) The method according to Appendix 11, further comprising a model identifier for at least one model function that identifies which of several models or model configurations associated with at least one model function is being executed and which model performance report is being executed. (Note 13) The method according to Appendix 11 or 12, further comprising model configuration data for at least one model function that constitutes a model that is executed and from which model performance reporting is performed. (Note 14) Configuration data includes data for configuring a wireless bearer between the UE and network nodes, used for model performance reporting, as described in any one of the methods in Appendix 11 to 13. (Note 15) A method performed by user equipment (UE), the method is: In the first time, we will run a model to predict parameters related to communication with network nodes, Obtaining at least one measurement of at least one signal relating to a parameter predicted by the model and received by the user's equipment before and after a first time period, Report the parameters predicted by the model and at least one measurement to the network nodes, Methods that include... (Note 16) The method described in Appendix 15, which includes repeating the execution and acquisition at different times, and reporting, for each iteration of execution and acquisition, the parameters predicted by the model and at least one measurement. (Note 17) Executing and repeating is the method described in Appendix 16, which is performed periodically. (Note 18) The method described in Appendix 17, comprising receiving configuration data from a network node that defines the periodicity of the actions that are performed and retrieved. (Note 19) The method according to Appendix 17, wherein the UE determines, based on the capabilities of the UE or based on the periodicity of receiving at least one signal used for measurement, the opportunity to perform and acquire is repeated. (Note 20) If the UE is unable to obtain all measurements within the reporting interval, the report shall indicate the measurements that the UE was able to perform within the reporting interval, as described in any one of the items in Appendix 15 to 19. (Note 21) Obtaining means obtaining at least one measurement within a predetermined period of time, as described in any one of the appendices 15 to 20. (Note 22) The method described in Appendix 21, wherein the predetermined period is defined in configuration data pre-stored in the UE or received from the network node. (Note 23) The specified period depends on the functionality of the model, as described in Appendix 21 or 22. (Note 24) The method described in any one of the appendices 15 to 23, which includes receiving configuration data from a network node indicating one or more reporting opportunities for the UE to perform reporting. (Note 25) The configuration data defines the periodicity for reporting, and the periodicity depends on the model's capabilities, as described in Appendix 24. (Note 26) The method described in any one of the appendices 15 to 23, which includes receiving configuration data from a network node that indicates one or more triggers causing the UE to perform reporting when the conditions are met. (Note 27) The method described in Appendix 26, wherein the trigger defines a model prediction metric, and the reporting is performed depending on the model prediction metric. (Note 28) The configuration data is configured in the manner described in Appendix 26 to perform reporting when the model's predictive metrics are better or worse than a threshold. (Note 29) The UE performs reporting in response to receiving a request from a network node, as described in any one of Annexes 15 to 28. (Note 30) A method performed by user equipment (UE), the method is: This involves running a model that predicts parameters related to communication with network nodes, Monitoring at least one metric related to communication with the model or network node, Report to the network node if at least one metric meets at least one criterion, A method that includes this. (Note 31) The method according to Appendix 30, comprising receiving configuration data from a network node that defines at least one metric and / or at least one criterion. (Note 32) At least one metric and / or at least one criterion is pre-stored within the UE, as described in Appendix 30. (Note 33) The method described in any one of the appendices 30 to 32, wherein multiple possible model control decisions exist, and at least one criterion is provided for each possible model control decision. (Note 34) The method described in any one of the appendices 30 to 33, wherein at least one criterion is a threshold check of one or more metrics. (Note 35) Monitoring is performed over a defined time window, as described in any one of the appendices 30 to 34. (Note 36) The method described in any one of the appendices 30 to 35, further comprising making a model control decision based on at least one metric and at least one criterion. (Note 37) Model management decisions are one of the following: model activation, model deactivation, model switching, and model fallback, as described in Appendix 36. (Note 38) The report shall be in the manner described in Appendix 36 or 37, indicating the model control decisions made by the UE. (Note 39) The method described in Appendix 38, wherein the UE waits for a certain period after reporting a model management decision before implementing the model management decision. (Note 40) Reporting is the method described in Appendix 38 or 39, which indicates a model control decision by showing at least one criterion met by at least one metric being monitored. (Note 41) The method described in any one of the appendices 30 to 35, further comprising, after reporting, receiving instructions for a model management decision from a network node and implementing the model management decision. (Note 42) The report is to include a model identifier and is prepared according to the method described in any one of the appendices 30 to 41. (Note 43) The metrics relate to the performance of communication with network nodes and are selected from a group including user throughput, handover failure rate / success rate, beam failure rate, and block error rate (BLER), as described in any one of the items in Appendix 30 to 42. (Note 44) The metric is a method described in any one of the appendices 30 to 42, relating to the performance of the model, indicating how well the model is functioning. (Note 45) The metric is as described in Appendix 44, including an indication of the number of instances in which the model was used to control communication with network nodes. (Note 46) The metric is as described in Appendix 44 or 45, including an indication of the number of instances in which the model was not used to control communication with network nodes. (Note 47) The report is as described in Appendix 46, indicating the number of instances in which the model was not used to control communication with network nodes, along with one or more reasons why the model was not used. (Note 48) If there are multiple reasons why the model is not used, report the number of instances in which the model was not used to control communication with network nodes, as described in Appendix 47, for each reason. (Note 49) The model is a classification model, and the metrics include the model's confusion matrix, as described in any one of the appendices 30 to 44. (Note 50) The report is the method described in Appendix 49, including a confusion matrix or parameters derived from a confusion matrix. (Note 51) The model is configured to predict measurements of N best objects, and the metric includes an indication of the number of instances of the best objects predicted by the model that are among the M best objects observed by the UE, as described in any one of the appendices 30 to 44. (Note 52) The model is configured to predict measurements of N best objects, and the metric is the method described in any one of the appendices 30 to 44, which includes an indication of the number of instances of the best objects measured by the UE that are among the N best objects predicted by the model. (Note 53) The model is configured to predict the value of a parameter, and the metric includes an indication of the number of instances in which the predicted value of the parameter falls within a threshold range for measurement of that parameter, as described in any one of the appendices 30 to 44. (Note 54) The model is configured to predict parameter values, and the metric includes an indication of the number of instances in which the measurement of the parameter falls within a threshold range of the parameter's predicted value from the model, as described in any one of the items in Appendix 30 to 44. (Note 55) For low latency reporting, reporting is done via uplink control information UCI messages, as described in any one of the items in Appendix 5 to 54. (Note 56) For low latency reporting, the reporting is performed via a Medium Access Control (MAC) Control Element (CE) as described in any one of the items in Appendix 5 to 54. (Note 57) MAC CE is the method described in Appendix 56, which includes a Logical Channel Identifier (LCID) indicating that the message contains a model report. (Note 58) The report is for multiple models, and the LCID values ​​are provided for each model, or a single LCID value is used for multiple models, as described in Appendix 57. (Note 59) The method described in any one of the appendices 5 to 58, wherein, depending on the amount of data transmitted to the network node in the report, the UE transmits part of the data in an Uplink Control Information (UCI) message or a Medium Access Control (MAC) Control Element (CE) message, and transmits part of the data in a Radio Resource Control (RRC) message. (Note 60) The method according to any one of the items in Appendix 5 to 58, wherein, depending on the amount of data transmitted to the network node in the report, the UE transmits the data in an uplink control information UCI message or a medium access control (MAC) control element (CE) message if the amount of data is less than a threshold amount, and transmits the data in a radio resource control (RRC) message otherwise. (Note 61) Reporting is performed via an uplink control information UCI message, a medium access control (MAC) control element (CE) message, or a radio resource control (RRC) message, depending on the type of report being reported, as described in any one of the items in Appendix 5 to 60. (Note 62) The method described in any one of the appendices 5 to 61, wherein when the UE changes to a target cell of the network node with which the UE communicates, it discards stored measurements relating to the source cell of the network node. (Note 63) The method described in any one of Annexes 5 to 61, wherein when the UE changes to a target cell of a network node with which the UE communicates, the stored measurements relating to the source cell of the network node are not discarded, and the stored measurements relating to the source cell are included in the report. (Note 64) The report is the method described in Appendix 63, which includes measurements relating to the source cell in the report and measurements relating to the target cell in the report. (Note 65) If the model is changed, the UE discards any measurements stored before the model change and resumes performance monitoring after the model change, as described in any one of the appendices 5 to 64. (Note 66) If the model is changed, the UE retains any stored measurements prior to the model change, as described in any one of the items in Appendix 5 to 64. (Note 67) The report is the method described in Appendix 66, which includes measurements of the model before the switchover and measurements of the model after the switchover. (Note 68) A method performed by a network node, the method is Receiving UE capability information from user equipment (UE). Includes, UE capability information indicates which reporting quantities or metrics the UE can support for model performance reporting. method. (Note 69) UE capability information is the method described in Appendix 68, which indicates the UE's ability to perform measurements for model performance reporting for a model or for each model. (Note 70) The method further comprises sending a configuration for performing a model performance report to the UE, and receiving the configuration, which is performed in response to sending the configuration, as described in Appendix 68 or 69. (Note 71) UE capability information includes information indicating the UE's ability to run multiple models simultaneously for the same function, as described in any one of the appendices 68 to 70. (Note 72) A method performed by a network node, the method is Sending the first instruction for model management decisions and the second instruction for model performance reporting to user equipment (UE), If the first instruction does not indicate model activation, and the second instruction enables model performance reporting, then the UE will receive model performance reports regarding the model running on the UE without using the output from the model to control communication with the network node. Methods that include... (Note 73) If a first instruction indicates model activation and a second instruction enables model performance reporting, the method is the method as described in Appendix 72, which includes using the output from the model to control communication between the UE and the network node, and receiving a model performance report from the UE regarding the model running on the UE. (Note 74) The first and second instructions are transmitted in the same message or in different messages, as described in Appendix 72 or 73. (Note 75) If a first instruction indicates a first model and a second instruction indicates enabling model performance reporting for a second model different from the first model, the method is the method according to any one of the appendices 72 to 74, comprising using the output from the first model run by the UE to control communication between the UE and the network node, and receiving a model performance report from the UE regarding the second model run on the UE, without using the output from the second model to control communication between the UE and the network node. (Note 76) The first and second models are for the same function, as described in Appendix 75. (Note 77) A method performed by a network node, the method is Send configuration data for model performance reporting regarding multiple model functions to user equipment (UE), Regarding the execution of the model for each function configured by the configuration data for each model function according to the configuration data, receiving model performance data from the UE, A method that includes this. (Note 78) The configuration data is as described in Appendix 77, including a function identifier for each function for which a model performance report is required. (Note 79) The method according to Appendix 78, further comprising a model identifier for at least one model function that identifies which of several models or model configurations associated with at least one model function is being executed and which model performance report is being executed. (Note 80) The method according to any one of the appendices 77 to 79, further comprising model configuration data for at least one model function that constitutes a model that is executed and for which model performance reporting is performed. (Note 81) Configuration data includes data for configuring a wireless bearer between the UE and network nodes used for model performance reporting, as described in any one of the items in Appendix 77 to 80. (Note 82) A method performed by a network node, the method is The user equipment (UE) is configured to run a model that predicts parameters related to communication between the UE and network nodes in a first time period, The UE obtains at least one measurement of at least one signal relating to a parameter predicted by the model and received by the user's equipment before and after a first time period, Receiving a report from the UE that includes the parameters predicted by the model and at least one measurement, A method that includes this. (Note 83) The method described in Appendix 82, wherein the UE repeatedly runs the model, obtains measurements corresponding to different time periods, and reports include the parameters predicted by the model and at least one measurement for each iteration. (Note 84) The method described in Appendix 83, comprising sending configuration data to the UE that defines the periodicity from which the UE should run the model and from which at least one measurement should be taken. (Note 85) If the UE is unable to obtain all measurements within the reporting interval, the received report shall include instructions for the measurements that the UE could have performed within the reporting interval, as described in Appendix 83 or 84. (Note 86) The method according to any one of the appendices 82 to 85, wherein at least one measurement is taken within a predetermined period of time of the first time. (Note 87) The method described in Appendix 86, which includes sending configuration data to the UE in order to define a predetermined period. (Note 88) The specified period depends on the functionality of the model, as described in Appendix 86 or 87. (Note 89) The method described in any one of the appendices 82 to 88, which includes sending configuration data to the UE indicating one or more reporting opportunities for the UE to submit a report. (Note 90) The method described in Appendix 82 to 88, which, when satisfied, includes sending configuration data to the UE indicating one or more triggers that cause the UE to send a report. (Note 91) The trigger defines and reports the model prediction metric, which is performed depending on the model prediction metric, as described in Appendix 90. (Note 92) The configuration data is configured in the manner described in Appendix 90 to perform reporting when the model's predictive metrics are better or worse than a threshold. (Note 93) The method described in any one of the appendices 82 to 92, including sending a request to the UE to send a report to a network node. (Note 94) A method performed by a network node, the method is This involves configuring user equipment (UE) to run a model that predicts parameters related to communication with network nodes, The UE monitors at least one metric related to communication with the model or network node, If at least one metric meets at least one criterion, we will receive a report from the UE, Methods that include... (Note 95) The method described in Appendix 94, which includes sending configuration data to the UE that defines at least one metric and / or at least one criterion. (Note 96) The method described in Appendix 94 or 95, wherein multiple possible model control decisions exist, and at least one criterion is provided for each possible model control decision. (Note 97) At least one criterion defines a threshold check for one or more metrics, as described in any one of the provisions of Appendix 94 to 96. (Note 98) Monitoring is performed over a defined time window, as described in any one of the appendices 94 to 97. (Note 99) The method described in any one of Annexes 94 to 98, wherein the UE makes a model control decision based on at least one metric and at least one criterion, and the report indicates the model control decision made by the UE. (Note 100) The method described in Appendix 99, further comprising adjusting the radio parameters in accordance with model control decisions made by the UE. (Note 101) The report, as described in Appendix 99 or 100, demonstrates a model control decision by indicating at least one criterion met by at least one monitored metric. (Note 102) The method described in any one of the appendices 94 to 98, further comprising making a model management decision based on a report, sending instructions for the model management decision to the UE, and implementing the model management decision. (Note 103) The report shall include a model identifier and shall be as described in any one of the items in Appendix 94 to 102. (Note 104) The metrics relate to the performance of communication with network nodes and are selected from a group including user throughput, handover failure rate / success rate, beam failure rate, and block error rate (BLER), as described in any one of the items in Appendix 94 to 103. (Note 105) The metric is a method described in any one of the appendices 94 to 103, relating to the performance of the model, indicating how well the model is functioning. (Note 106) The metric is as described in Appendix 105, including an indication of the number of instances used by the model to control communication with network nodes. (Note 107) The metric is as described in Appendix 105 or 106, including an indication of the number of instances in which the model was not used to control communication with network nodes. (Note 108) The report is as described in Appendix 107, indicating the number of instances in which the model was not used to control communication with network nodes, along with one or more reasons why the model was not used. (Note 109) If there are multiple reasons why the model was not used, the report should indicate, for each reason, the number of instances in which the model was not used to control communication with network nodes, as described in Appendix 108. (Note 110) The model is a classification model, and the metrics are those described in any one of the appendices 94 to 105, including the model's confusion matrix. (Note 111) The report is the method described in Appendix 110, including a confusion matrix or parameters derived from a confusion matrix. (Note 112) The model is configured to predict measurements of N best objects, and the metric includes an indication of the number of instances of the best objects predicted by the model that are among the M best objects observed by the UE, as described in any one of Appendix 94 to 105. (Note 113) The model is configured to predict measurements of N best objects, and the metric is as described in any one of the appendices 94 to 105, including an indication of the number of instances of the best objects measured by the UE that are among the N best objects predicted by the model. (Note 114) The model is configured to predict the value of a parameter, and the metric includes an indication of the number of instances in which the predicted value of the parameter falls within a threshold range for measurement of that parameter, as described in any one of the appendices 94 to 105. (Note 115) The model is configured to predict parameter values, and the metric includes an indication of the number of instances in which the parameter measurement falls within a threshold range of the parameter's predicted value from the model, as described in any one of Appendix 94 to 105. (Note 116) For low latency reporting, the report is received via uplink control information UCI messages, as described in any one of the appendices 72 to 115. (Note 117) For low latency reporting, the reporting is performed via a Medium Access Control (MAC) Control Element (CE) as described in any one of the appendices 72 to 115. (Note 118) MAC CE is the method described in Appendix 117, which includes a Logical Channel Identifier (LCID) indicating that the message contains a model report. (Note 119) The report is for multiple models, and the LCID values ​​are provided for each model, or a single LCID value is used for multiple models, as described in Appendix 118. (Note 120) The method described in any one of the appendices 72 to 115, wherein, depending on the amount of data in the report, some of the data is received in uplink control information UCI messages or medium access control (MAC) control element (CE) messages, and some of the data is received in radio resource control (RRC) messages. (Note 121) The method described in any one of the appendices 72 to 115, wherein if the amount of data reported is below a threshold, the report is received in an Uplink Control Information (UCI) message or a Medium Access Control (MAC) Control Element (CE) message; otherwise, the report is received in a Radio Resource Control (RRC) message. (Note 122) The reporting is performed via an uplink control information UCI message, a medium access control (MAC) control element (CE) message, or a radio resource control (RRC) message, depending on the type of report being reported, as described in any one of the items in Appendix 72 to 121. (Note 123) The report is provided in any one of the following clauses, from 72 to 122, including measurements relating to the source cell and measurements relating to the target cell. (Note 124) If the UE switches models, the report shall include measurements for the model before the switch and measurements for the model after the switch, as described in any one of the items in Appendix 72 to 122. (Note 125) User equipment (UE), Means for transmitting UE capability information to network nodes Equipped with, UE capability information indicates which reporting quantities or metrics the UE can support for model performance reporting. UE. (Note 126) User equipment (UE), Means for receiving a first instruction for model management decisions and a second instruction for model performance reporting from a network node, If the first instruction does not indicate model activation, and the second instruction enables model performance reporting, means for running the model and reporting model performance to the network node without using the output from the model to control communication with the network node are provided. A UE equipped with (Note 127) User equipment (UE), A means for receiving configuration data from network nodes for model performance reporting regarding multiple model functions, A means for running the model for each function configured by the configuration data, A means for reporting model performance for each model function according to configuration data, A UE equipped with (Note 128) User equipment (UE), In the first time period, a means for running a model that predicts parameters related to communication with network nodes, Means for obtaining at least one measurement of at least one parameter-related signal, which is predicted by a model and received by a user device before or after a first time period, Means for reporting the parameters predicted by the model and at least one measurement to the network node, A UE equipped with (Note 129) User equipment (UE), A means for running a model that predicts parameters related to communication with network nodes, Means for monitoring at least one metric related to communication with a model or network node, A means for reporting to a network node if at least one metric meets at least one criterion, A UE equipped with (Note 130) Network node, Means for receiving UE capability information from user equipment (UE). Equipped with, UE capability information indicates which reporting quantities or metrics the UE can support for model performance reporting. Network node. (Note 131) Network node, Means for transmitting a first instruction for model management decisions and a second instruction for model performance reporting to user equipment (UE), If the first instruction does not indicate model activation and the second instruction enables model performance reporting, a means for receiving model performance reports from the UE regarding the model running on the UE, without using the output from the model to control communication with network nodes, A network node equipped with these features. (Note 132) Network node, A means for transmitting configuration data for model performance reporting regarding multiple model functions to user equipment (UE), A means for receiving model performance data from the UE regarding the execution of a model for each function configured by configuration data for each model function according to the configuration data, A network node equipped with these features. (Note 133) Network node, In the first time, means for configuring a UE to run a model that predicts parameters related to communication between user equipment (UE) and network nodes, The UE has means to obtain at least one measurement of at least one signal related to a parameter predicted by the model and received by the user's equipment before and after a first time period, A means for receiving a report from the UE that includes parameters predicted by the model and at least one measurement, A network node equipped with these features. (Note 134) Network node, A means for configuring user equipment (UE) to run a model that predicts parameters related to communication with network nodes, The UE comprises means for monitoring at least one metric related to communication with a model or a network node, means for receiving a report from the UE when the at least one metric meets at least one criterion, and a network node.

[0151] This application claims priority based on UK Patent Application No. 2305560.1 filed on April 14, 2023, and incorporates the entire disclosure thereof herein.

Description of Signs

[0152] 1 Communication system 3 User equipment 5 Base station 7 Core network 9 Cell 10 Control plane function 11 User plane function 20 External data network 41 Data collection 43 Model training 47 Model management 45 Model inference 49 Model storage 310 Transceiver circuit 330 Antenna 350 User interface 370 Controller 390 Memory 410 Operating system 430 Communication control module 450 AI / ML module 510 Transceiver circuit 530 Antenna 55 Contact network interface 570 Controller 590 Memory 610 Operating system 630 Communication control module 650 AI / ML module It should be noted that there may be some inaccuracies in the original text, such as "Contact network interface" in which might be a misspelling. It is recommended to double-check the original text for more accurate translation.

Claims

1. A method performed by user equipment (UE), Each function must manage at least one model, Transmitting performance information for each function, or for each of the at least one of the aforementioned models, to the network node. The performance information includes, Each function or each model has at least one predictive output, Information for the network node to determine whether the at least one prediction output is correct, including, method.

2. Monitoring at least one metric related to the at least one model It further includes, Managing the at least one model is performed on the basis that the at least one metric satisfies at least one criterion. If the at least one metric satisfies the at least one criterion, the performance information is transmitted. The method according to claim 1.

3. The performance information includes information indicating an action to manage the at least one model based on the fact that the at least one metric satisfies at least one criterion. The method according to claim 2.

4. The at least one metric and the at least one criterion are defined for each action of managing the at least one model. The method according to claim 2 or 3.

5. The aforementioned monitoring is performed within a time window. The method according to any one of claims 2 to 4.

6. A specific delay, or Receiving a response from the aforementioned network node Performing the action that manages the at least one model after at least one of the above The method according to any one of claims 2 to 5, further comprising:

7. The aforementioned specific delay depends on the action that manages the at least one model based on the fact that the at least one metric satisfies at least one criterion. The method according to claim 6.

8. The aforementioned performance information is, Information indicating the at least one criterion that is satisfied by the at least one metric, or Information indicating at least one of the aforementioned models Including at least one of the following: The method according to any one of claims 2 to 7.

9. The at least one metric is, User throughput, handover failure rate / success rate, Beam failure rate, or Block Error Rate (BLER) Performance related to at least one of the following: The method according to any one of claims 2 to 8.

10. The at least one metric is related to the prediction, including how well the prediction for the at least one mode is being performed. The method according to any one of claims 2 to 9.

11. The at least one metric is, The number of at least one models used to determine the at least one predictive output, or The number of at least one model that was not used to determine the aforementioned at least one predictive output. Including at least one of the following: The method according to any one of claims 2 to 10.

12. The aforementioned performance information is, The number of the at least one model that was not used to determine the at least one predictive output, One or more reasons why at least one of the aforementioned models was not used Including at least one of the following: The method according to claim 11.

13. The performance information includes, for each of the one or more reasons, the number of the at least one model that was not used to determine the at least one predictive output. The method according to claim 12.

14. The aforementioned at least one model includes a classification model, The at least one metric includes a confusion matrix associated with the at least one model, The method according to any one of claims 2 to 13.

15. The performance information includes information corresponding to the confusion matrix, The method according to claim 14.

16. The aforementioned at least one prediction output includes the measurement of N best objects, The at least one metric is, The number or proportion of times the best object predicted by at least one of the models is included in the M best objects measured by the UE, or The number or proportion of times the best object measured by the UE is included in the N best objects predicted by the at least one model Including at least one of the following: The method according to any one of claims 2 to 15.

17. The aforementioned at least one prediction output includes the value of a parameter, The at least one metric is, The number or proportion of times the value of the parameter predicted by the at least one model falls within the measurement range of the parameter, or The number or percentage of times the value of the parameter measured by the UE falls within the range of the parameter value predicted by the at least one model. Including at least one of the following: The method according to any one of claims 2 to 15.

18. Transmitting the aforementioned performance information means Uplink Control Information (UCI), or Media Access Control (MAC) Control Element (CE) Executed via at least one of the following: The method according to any one of claims 1 to 17.

19. The MAC CE includes a Logical Channel Identifier (LCID) indicating that the performance information includes an inference report relating to at least one model. The method according to claim 18.

20. The MAC CE includes each Logical Channel Identifier (LCID) indicating that the performance information includes an inference report relating to the corresponding model among the at least one model. The method according to claim 18.

21. Depending on the type of performance information, the transmission of the performance information is performed via the UCI or MAC CE and / or Radio Resource Control (RRC) message. The method according to claim 18.

22. When the UE changes from a source cell to a target cell, the performance information associated with the source cell is discarded. The method according to any one of claims 1 to 21, further comprising:

23. If at least one model is not changed when the source cell is changed to the target cell, the performance information associated with the source cell is not discarded. The method according to claim 22.

24. The performance information includes both information about the source cell and information about the target cell. The method according to claim 23.

25. If at least one model used for management is changed, Before modifying at least one model, discard the performance information relating to at least one model, After modifying at least one of the aforementioned models, the at least one model is remanaged. The method according to any one of claims 1 to 24, further comprising:

26. If at least one model for management is changed, the performance information related to the at least one model before the change is retained. The method according to any one of claims 1 to 24, further comprising:

27. The performance information includes both information relating to the at least one model before the modification of the at least one model and information relating to the at least one model after the modification of the at least one model. The method according to claim 26.

28. Receiving from the network node a first instruction for managing the at least one model and a second instruction for transmitting the performance information, If the first instruction does not indicate the activation of at least one model, and the second instruction enables the transmission of the performance information, Executing at least one of the aforementioned models, The process involves transmitting the performance information without using at least one predictive output from the aforementioned at least one model, The method according to any one of claims 1 to 27, further comprising:

29. If the first instruction indicates activation of the at least one model and the second instruction enables the transmission of the performance information, the method, Executing at least one of the aforementioned models, The process involves transmitting the performance information using at least one predictive output from the at least one model, including, The method according to claim 28.

30. The first instruction and the second instruction are received within the message or within each of the messages. The method according to claim 28 or 29.

31. If the first instruction indicates activation of a first model and the second instruction indicates that it is possible to transmit the performance information for a second model different from the first model, the method, Executing the first model described above, The process involves transmitting the performance information for the first model using at least one predictive output from the first model, Executing the second model described above, The process involves transmitting the performance information for the second model without using at least one predictive output from the second model, including, The method according to any one of claims 28 to 30.

32. The first model and the second model are for the same function. The method according to claim 31.

33. For each of the aforementioned functions, or for each of the at least one of the aforementioned models, transmit capability information to the network node indicating what quantity or metric the UE can support in transmitting the performance information. The method according to any one of claims 1 to 32, further comprising:

34. The capability information indicates, for each of the functions or for each of the at least one model, how often the UE can perform measurements to transmit the performance information. The method according to claim 33.

35. Receiving configuration information from the network node for performing the transmission of the performance information. It further includes, The aforementioned transmission is performed in response to the receipt of the configuration information. The method according to claim 33 or 34.

36. The aforementioned capability information indicates the UE's ability to simultaneously execute multiple models for the same function. The method according to any one of claims 33 to 35.

37. Receiving configuration information from the network node for transmitting the performance information for each function or each model. It further includes, Managing at least one of the aforementioned models is performed by the configuration information, The transmission of the aforementioned performance information is performed by the aforementioned configuration information. The method according to any one of claims 1 to 36.

38. The configuration information includes the respective function identifiers for each function that is required to transmit the performance information. The method according to claim 37.

39. The configuration information indicates which of the configurations of the at least one model or the at least one model associated with the at least one model is performed, and whether the transmission of the performance information is performed. The method according to claim 38.

40. The aforementioned configuration information is, A model identifier that identifies which of the at least one model or the configuration of the at least one model associated with the at least one model is executed, and which of the functions of transmitting the performance information is executed. Model configuration information of at least one model that constitutes each model that is executed and which transmits the performance information, or Wireless bearer configuration information for configuring a wireless bearer between the UE and the network node, used to transmit the performance information. Including at least one of the following: The method according to any one of claims 37 to 39.

41. Managing the at least one model is performed by running the at least one model and predicting the at least one predictive output, and the method is To predict the aforementioned at least one predictive output, perform at least one measurement at at least one measurement opportunity. including, The method according to any one of claims 1 to 40.

42. The aforementioned at least one measurement opportunity is Periodicity and / or opportunity configured by the aforementioned network nodes, The periodicity and / or occasions stored in the UE, The capabilities of the aforementioned UE, The periodicity of the signal used in the at least one measurement, or Periodicity of transmitting the performance information Defined by at least one of the following: The method according to claim 41.

43. The at least one measurement within a specific duration is used solely to predict the at least one predictive output. The method according to claim 41 or 42.

44. The aforementioned specific duration is The threshold configured by the aforementioned network node, The threshold stored in the UE, or The at least one measurement opportunity Defined by at least one of the following: The method according to claim 43.

45. At least one of the specified duration or the threshold is defined for each function of the at least one model, or for each of the at least one models. The method according to claim 44.

46. Transmitting the aforementioned performance information means An opportunity for transmitting the aforementioned performance information, The periodicity inherent in transmitting the aforementioned performance information, A trigger that, when the condition is met, causes the UE to transmit the performance information, or Request from the aforementioned network node Executed based on at least one of the following: The method according to any one of claims 1 to 45.

47. The trigger defines a metric for predicting the execution of the at least one model, The transmission of the performance information is performed based on a comparison of the metric and the threshold. The method according to claim 46.

48. The opportunity, the periodicity, and the trigger are defined for each function of the at least one model, or for each of the at least one models. The method according to claim 46 or 47.

49. The aforementioned opportunity, the aforementioned periodicity, and the aforementioned trigger are configured by the network node. The method according to any one of claims 46 to 48.

50. A method performed by a network node, wherein the method is This includes receiving performance information from user equipment (UE) for each function, or for each of at least one model for each function, wherein the at least one model is managed by the UE. The aforementioned performance information is, Each function or each model has at least one predictive output, Information for the network node to determine whether the at least one prediction output is correct, including, method.

51. User equipment (UE), A means for managing at least one model for each function, Means for transmitting performance information for each function, or for each of the at least one of the models, to a network node, Equipped with, The aforementioned performance information is, Each function or each model has at least one predictive output, Information for the network node to determine whether the at least one prediction output is correct, including, UE.

52. Network node, The system includes means for receiving performance information from user equipment (UE) for each function, or for each of at least one model for each function, wherein the at least one model is managed by the UE. The aforementioned performance information is, Each function or each model has at least one predictive output, Information for the network node to determine whether the at least one prediction output is correct, including, Network node.

Images