Communication system

Abstract

According to the present invention, a communication terminal device uses artificial intelligence (AI) and / or machine learning (ML) to perform prediction processing regarding at least one among radio link failure (RLF), measurement processing, flight of the communication terminal device, overheating in the communication terminal device, in-device coexistence interference in the communication terminal device, and delay budget.

Description

Communication system

[0001] This disclosure relates to wireless communication technology.

[0002] In the 3GPP (3rd Generation Partnership Project), a standardization organization for mobile communication systems, the fifth-generation (hereinafter sometimes referred to as "5G") wireless access system is being studied as a successor to Long Term Evolution (LTE), one of the fourth-generation wireless access systems, and Long Term Evolution Advanced (LTE-A) (see Non-Patent Document 1) (for example, Non-Patent Document 2). The technology for the 5G radio section is called "New Radio Access Technology" ("New Radio" is abbreviated as "NR"). The NR system is being studied based on the LTE system and the LTE-A system.

[0003] For example, in Europe, the requirements for 5G have been summarized by a group called METIS (see Non-Patent Document 3). In the 5G wireless access system, compared with the LTE system, the system capacity is required to be 1000 times, the data transmission speed is 100 times, the data processing delay is 1 / 5, and the number of simultaneously connected communication terminals is 100 times, and further power consumption reduction and device cost reduction are required (see Non-Patent Document 3).

[0004] In order to meet such requirements, 3GPP is conducting a standardization study on 5G (see Non-Patent Documents 4 to 23).

[0005] As the access method of NR, OFDM (Orthogonal Frequency Division Multiplexing) is used in the downlink direction, and OFDM and DFT-s-OFDM (Discrete Fourier Transform-spread-OFDM) are used in the uplink direction. Also, like the LTE and LTE-A systems, the 5G system does not include circuit switching and is only a packet communication system.

[0006] NR allows for the use of higher frequencies compared to LTE, in order to improve transmission speed and reduce processing delay.

[0007] In noise reduction (NR), which sometimes uses higher frequencies than LTE, cell coverage is ensured by forming a narrow beam-shaped transmission and reception range (beamforming) and changing the direction of the beam (beam sweeping).

[0008] The decisions regarding the frame configuration in the NR system under 3GPP, as described in Non-Patent Document 1 (Chapter 5), will be explained using Figure 1. Figure 1 is an explanatory diagram showing the configuration of a radio frame used in an NR communication system. In Figure 1, one radio frame is 10 ms. The radio frame is divided into 10 subframes of equal size. In the NR frame configuration, one or more numerologies, i.e., one or more subcarrier spacings (SCS), are supported. In NR, regardless of the subcarrier spacing, one subframe is 1 ms, and one slot consists of 14 symbols. Furthermore, the number of slots contained in one subframe is one for a subcarrier spacing of 15 kHz, and the number of slots for other subcarrier spacings increases proportionally to the subcarrier spacing (see Non-Patent Document 11 (3GPP TS38.211)).

[0009] The decisions regarding the channel configuration in the NR system at 3GPP are described in Non-Patent Document 2 (Chapter 5) and Non-Patent Document 11.

[0010] A Physical Broadcast Channel (PBCH) is a channel used for downlink transmission from a base station (hereinafter sometimes simply referred to as "base station") to a communication terminal (hereinafter sometimes simply referred to as "communication terminal" or "terminal") such as a mobile terminal (hereinafter sometimes simply referred to as "mobile terminal"). The PBCH is transmitted together with a Downlink Synchronization Signal.

[0011] In NR, the downlink synchronization signal consists of a primary synchronization signal (P-SS) and a secondary synchronization signal (S-SS). The synchronization signal is transmitted from the base station as a synchronization signal burst (Synchronization Signal Burst; hereinafter sometimes referred to as SS burst) at a predetermined period and for a predetermined duration. The SS burst is composed of a synchronization signal block (Synchronization Signal Block; hereinafter sometimes referred to as SS block) for each beam of the base station.

[0012] The base station transmits SS blocks of each beam, switching beams, during the duration of the SS burst. The SS blocks consist of P-SS, S-SS, and PBCH.

[0013] The Physical Downlink Control Channel (PDCCH) is the channel used for downlink transmission from the base station to the communication terminal. The PDCCH carries Downlink Control Information (DCI). DCI includes resource allocation information for the Downlink Shared Channel (DL-SCH), one of the transport channels described later, resource allocation information for the Paging Channel (PCH), another transport channel described later, and HARQ (Hybrid Automatic Repeat reQuest) information related to the DL-SCH. In addition, DCI may include Uplink Scheduling Grant. DCI may also include Ack (Acknowledgement) / Nack (Negative Acknowledgement), which are response signals to uplink transmissions. Furthermore, to allow for flexible switching between DL / UL within a slot, DCI may include a Slot Format Indication (SFI). PDCCH, or DCI, is also called an L1 / L2 control signal.

[0014] In NR, a time and frequency range is defined as a candidate range for PDCCH. This range is called the Control Resource Set (CORESET). The communication terminal monitors the CORESET and acquires PDCCH.

[0015] The Physical Downlink Shared Channel (PDSCH) is a channel used for downlink transmission from a base station to a communication terminal. The PDSCH is mapped to the Downlink Shared Channel (DL-SCH), which is a transport channel, and the PCH, which is also a transport channel.

[0016] The Physical Uplink Control Channel (PUCCH) is the channel used for uplink transmission from the communication terminal to the base station. The PUCCH carries Uplink Control Information (UCI). UCI includes Ack / Nack, which are response signals to downlink transmissions, Channel State Information (CSI), and Scheduling Request (SR). CSI consists of the Rank Indicator (RI), Precoding Matrix Indicator (PMI), and Channel Quality Indicator (CQI) reports. RI is the rank information of the channel matrix in MIMO (Multiple Input Multiple Output). PMI is the information of the precoding weight matrix used in MIMO. CQI is quality information indicating the quality of the received data or the quality of the communication channel. UCI may be carried by the PUCCH, which will be described later. PUCCH, or UCI, is also called the L1 / L2 control signal.

[0017] The Physical Uplink Shared Channel (PUSCH) is a channel used for uplink transmission from a communication terminal to a base station. The Uplink Shared Channel (UL-SCH), which is one of the transport channels, is mapped to the PUSCH.

[0018] A Physical Random Access Channel (PRACH) is a channel used for uplink transmission from a communication terminal to a base station. The PRACH carries a random access preamble.

[0019] The downlink reference signal (RS) is a well-known symbol in NR (Noise Reduction) communication systems. The following four types of downlink reference signals are defined: UE-specific reference signals: Demodulation Reference Signal (DM-RS), Phase Tracking Reference Signal (PT-RS), Positioning Reference Signal (PRS), and Channel State Information Reference Signal (CSI-RS). Measurements at the physical layer of the communication terminal include Reference Signal Received Power (RSRP) measurement and Reference Signal Received Quality (RSRQ) measurement.

[0020] Similarly, the uplink reference signals are also known symbols in NR (Noise Reduction) communication systems. The following three types of uplink reference signals are defined: Data Demodulation Reference Signal (DM-RS), Phase Tracking Reference Signal (PT-RS), and Sounding Reference Signal (SRS).

[0021] This section explains the transport channel described in Non-Patent Document 2 (Chapter 5). Of the downlink transport channels, the broadcast channel (BCH) broadcasts to the entire coverage of the base station (cell). The BCH is mapped to the physical broadcast channel (PBCH).

[0022] Downlink Shared Channels (DL-SCH) are subject to HARQ-based retransmission control. DL-SCH can broadcast to the entire coverage of a base station (cell). DL-SCH supports dynamic or semi-static resource allocation. Semi-static resource allocation is also called semi-persistent scheduling. DL-SCH supports discontinuous reception (DRX) for communication terminals to reduce power consumption. DL-SCH is mapped to the physical downlink shared channel (PDSCH).

[0023] Paging Channels (PCHs) support DRX for communication terminals to enable low power consumption for those terminals. PCHs are required to broadcast across the entire coverage of a base station (cell). PCHs are mapped to physical resources, such as Physical Downlink Shared Channels (PDSCHs), which are dynamically available for traffic.

[0024] Among the uplink transport channels, the Uplink Shared Channel (UL-SCH) is subject to retransmission control by HARQ. UL-SCH supports dynamic or quasi-static resource allocation. Quasi-static resource allocation is also called configured grant. UL-SCH is mapped to the physical uplink shared channel (PUSCH).

[0025] Random Access Channels (RACHs) are limited to control information. RACHs carry a risk of collisions. RACHs are mapped to Physical Random Access Channels (PRACHs).

[0026] This section explains HARQ. HARQ is a technology that improves the communication quality of a transmission line by combining Automatic Repeat Request (ARQ) and Forward Error Correction. HARQ has the advantage that error correction through retransmission works effectively even for transmission lines where the communication quality changes. In particular, it is possible to further improve quality by combining the reception result of the initial transmission and the reception result of the retransmission during retransmission.

[0027] Here is an example of how to retransmit data. If the receiving side is unable to correctly decode the received data, in other words, if a CRC (Cyclic Redundancy Check) error occurs (CRC = NG), the receiving side sends "Nack" to the sending side. Upon receiving "Nack," the sending side retransmits the data. If the receiving side is able to correctly decode the received data, in other words, if no CRC error occurs (CRC = OK), the receiving side sends "Ack" to the sending side. Upon receiving "Ack," the sending side sends the next data.

[0028] Here are some other examples of retransmission methods. If a CRC error occurs at the receiving end, the receiving end requests retransmission from the sending end. This request is made by toggling the NDI (New Data Indicator). Upon receiving the retransmission request, the sending end retransmits the data. If no CRC error occurs at the receiving end, no retransmission request is made. If the sending end does not receive a retransmission request within a specified time, it is assumed that no CRC error occurred at the receiving end.

[0029] This section explains the logical channel described in Non-Patent Document 1 (Chapter 6). The Broadcast Control Channel (BCCH) is a downstream channel for broadcasting system control information. The BCCH, being a logical channel, is mapped to the broadcast channel (BCH), which is a transport channel, or to the downstream shared channel (DL-SCH).

[0030] The Paging Control Channel (PCCH) is a downstream channel for transmitting changes to paging information and system information. The PCCH, being a logical channel, is mapped to the Paging Channel (PCH), which is a transport channel.

[0031] The Common Control Channel (CCCH) is a channel used to transmit control information between a communication terminal and a base station. The CCCH is used when a communication terminal does not have an RRC connection with the network. In the downlink direction, the CCCH is mapped to the downlink common channel (DL-SCH), which is a transport channel. In the uplink direction, the CCCH is mapped to the uplink common channel (UL-SCH), which is a transport channel.

[0032] A Dedicated Control Channel (DCCH) is a channel that transmits individual control information between a communication terminal and the network on a one-to-one basis. DCCH is used when a communication terminal has an RRC connection with the network. DCCH is mapped to the Uplink Shared Channel (UL-SCH) for uplink traffic and to the Downlink Shared Channel (DL-SCH) for downlink traffic.

[0033] A Dedicated Traffic Channel (DTCH) is a one-to-one communication channel to a communication terminal for transmitting user information. DTCHs exist for both uplink and downlink traffic. Uplink DTCHs are mapped to the Uplink Shared Channel (UL-SCH), and downlink DTCHs are mapped to the Downlink Shared Channel (DL-SCH).

[0034] Location tracking of communication terminals is performed in units of areas consisting of one or more cells. Location tracking is performed to track the location of communication terminals even when they are in standby mode, and to enable them to be called, in other words, to enable them to receive calls. This area used for location tracking of communication terminals is called a Tracking Area (TA).

[0035] In NR, calling communication terminals within a range smaller than the tracking area is supported. This range is called the RAN Notification Area (RNA). Paging of communication terminals in the RRC_INACTIVE state, as described later, is performed within this range.

[0036] In NR (Noise Reduction), carrier aggregation (CA), which involves aggregating two or more component carriers (CCs), is being considered to support wide frequency bandwidths. CA is described in Non-Patent Document 1.

[0037] When a CA (Carrier Acquisition) is configured, the UE (User Equipment) which is a communication terminal has a single RRC (Rapid Relay Control) connection to the network (NW). In the RRC connection, one serving cell provides NAS (Non-Access Stratum) mobility information and security input. This cell is called the primary cell (PCell). Depending on the capabilities of the UE, secondary cells (SCells) are configured together with the PCell to form a set of serving cells. A set of serving cells consisting of one PCell and one or more SCells is configured for one UE.

[0038] Furthermore, in 3GPP, in order to further increase communication capacity, there is dual connectivity (DC), in which a UE connects to two base stations for communication. DC is described in Non-Patent Documents 1 and 22.

[0039] In dual connectivity (DC) base stations, one is sometimes called the "Master Node (MN)" and the other the "Secondary Node (SN)". The serving cells composed of the master base station are sometimes collectively called the Master Cell Group (MCG), and the serving cells composed of the secondary base station are sometimes collectively called the Secondary Cell Group (SCG). In a DC, the primary cell within the MCG or SCG is called a Special Cell (SpCell or SPCell). A Special Cell in an MCG is called a PCell, and a Special Cell in an SCG is called a Primary SCG Cell (PSCell).

[0040] Furthermore, in NR, the base station pre-configures a portion of the carrier frequency band (hereinafter sometimes referred to as the Bandwidth Part (BWP)) for the UE, and the UE performs transmission and reception with the base station in the BWP, thereby reducing power consumption in the UE.

[0041] Furthermore, 3GPP is considering supporting services (or applications) using side-link (SL) communication (also called PC5 communication) in both EPS (Evolved Packet System) and 5G core systems (see Non-Patent Documents 1, 2, 26-28). SL communication takes place between terminals. Examples of services using SL communication include V2X (Vehicle-to-everything) services and proximity services. In SL communication, not only direct communication between terminals but also communication between UE and NW via relay has been proposed (see Non-Patent Documents 26, 28).

[0042] The physical channels used for SL (see Non-Patent Documents 2 and 11) will be described. The physical sidelink broadcast channel (PSBCH: Physical sidelink broadcast channel) carries information related to system synchronization and is transmitted from the UE.

[0043] The physical sidelink control channel (PSCCH: Physical sidelink control channel) carries control information from the UE for sidelink communication and V2X sidelink communication.

[0044] The physical sidelink shared channel (PSSCH: Physical sidelink shared channel) carries data from the UE for sidelink communication and V2X sidelink communication.

[0045] The physical sidelink feedback channel (PSFCH: Physical sidelink feedback channel) carries HARQ feedback on the sidelink from the UE that received the PSSCH transmission to the UE that transmitted the PSSCH.

[0046] The transport channels used for SL (see Non-Patent Document 1) will be described. The sidelink broadcast channel (SL-BCH: Sidelink broadcast channel) has a predetermined transport format and is mapped to the PSBCH, which is a physical channel.

[0047] The sidelink shared channel (SL-SCH) supports notification transmission. The SL-SCH supports both UE autonomous resource selection and resource allocation scheduled by the base station. There is a risk of collision in UE autonomous resource selection, and there is no collision when the UE is allocated individual resources by the base station. Also, the SL-SCH supports dynamic link adaptation by changing the transmission power, modulation, and coding. The SL-SCH is mapped to the physical channel PSSCH.

[0048] The logical channels used for SL (see Non-Patent Document 2) will be described. The sidelink broadcast control channel (SBCCCH) is a sidelink channel for notifying sidelink system information from one UE to other UEs. The SBCCCH is mapped to the transport channel SL-BCH.

[0049] The sidelink traffic channel (STCH) is a one-to-many sidelink traffic channel for transmitting user information from one UE to other UEs. The STCH is used only by UEs with sidelink communication capabilities and UEs with V2X sidelink communication capabilities. One-to-one communication between two UEs with sidelink communication capabilities is also realized by the STCH. The STCH is mapped to the transport channel SL-SCH.

[0050] The sidelink control channel (SCCH) is a sidelink control channel for transmitting control information from one UE to other UEs. The SCCH is mapped to the transport channel SL-SCH.

[0051] In LTE, SL communication was limited to broadcast only. In NR, support for unicast and groupcast in addition to broadcast is being considered for SL communication (see Non-Patent Document 27 (3GPP TS23.287)).

[0052] In SL, unicast and groupcast communications support HARQ feedback (Ack / Nack), CSI reporting, and other features.

[0053] Furthermore, 3GPP is considering Integrated Access and Backhaul (IAB), which involves wirelessly conducting both the access link between the UE and the base station, and the backhaul link between base stations (see Non-Patent Documents 2, 20, and 29).

[0054] Several new technologies are required for mobile communication systems. For example, the introduction of AI (Artificial Intelligence) / ML (Machine Learning) into RANs is required to improve communication performance. Such new technologies are being discussed within 3GPP (see Non-Patent Documents 30, 31, 32, and 33).

[0055] 3GPP TS36.300 V18.0.03GPP TS38.300 V18.0.0“Scenarios, requirements and KPIs for 5G mobile and wireless system”、ICT-317669-METIS / D1.13GPP TR23.799 V14.0.03GPP TR38.801 V14.0.03GPP TR38.802 V14.2.03GPP TR38.804 V14.0.03GPP TR38.912 V16.0.03GPP RP-1721153GPP TS23.501 V18.4.03GPP TS38.211 V18.1.03GPP TS38.212 V18.1.03GPP TS38.213 V18.1.03GPP TS38.214 V18.1.03GPP TS38.321 V18.0.03GPP TS38.322 V18.0.03GPP TS38.323 V18.0.03GPP TS37.324 V17.0.03GPP TS38.331 V18.0.03GPP TS38.401 V18.0.03GPP TS38.413 V18.0.03GPP TS37.340 V18.0.03GPP TS38.423 V18.0.03GPP TS38.305 V18.0.03GPP TS23.273 V18.4.03GPP TR23.703 V12.0.03GPP TS23.287 V18.2.03GPP TS23.303 V17.1.03GPP TS38.340 V18.0.03GPP TR37.817 V17.0.03GPP RP-2136023GPP RP-2403233GPP RP-2400823GPP R2-24045973GPP R2-24050963GPP R2-24043723GPP R2-24045593GPP TR26.910V18.0.0“The E-model: a computational model for use in transmission planning”,ITU-T G.107,インターネット<URL: https: / / www.itu.int / rec / T-REC-G.107 / >

[0056] In mobile communication systems, the number of required KPIs (Key Performance Indicators), such as latency, reliability, connection density, user experience, and energy efficiency, is increasing, and consistent optimization of these is required. For this reason, technologies are being investigated to improve various KPIs by introducing AI / ML into mobile communication systems (see Non-Patent Documents 30, 31, 32, and 33). However, specific processing methods associated with the introduction of AI / ML have not yet been considered. As a result, a problem arises in that it is not possible to optimize various KPIs under various communication conditions.

[0057] In light of the above-mentioned issues, one of the objectives of this disclosure is to realize a communication system that can perform effective processing according to various communication conditions.

[0058] The communication system of this disclosure comprises a communication terminal device and a network node configured to communicate with the communication terminal device, wherein the communication terminal device is configured to perform predictive processing related to at least one of the following using AI and ML: radio link failure (RLF), measurement processing, flight of the communication terminal device, overheating in the communication terminal device, in-device coexistence (IDC) interference in the communication terminal device, and delay budget.

[0059] According to the above configuration, effective processing can be performed according to various communication conditions using at least one of AI and ML.

[0060] The purpose, features, aspects, and advantages of this disclosure will become clearer from the following detailed description and accompanying drawings.

[0061] This is an explanatory diagram showing the configuration of a wireless frame used in an NR communication system. This is a block diagram showing the overall configuration of an NR communication system 210 as discussed in 3GPP. This is a configuration diagram of a DC by a base station connected to an NG core. This is a block diagram showing the configuration of a mobile terminal 202 shown in Figure 2. This is a block diagram showing the configuration of a base station 213 shown in Figure 2. This is a block diagram showing the configuration of the 5GC section. This is a flowchart outlining the process from cell search to standby operation performed by a communication terminal (UE) in an NR communication system. This is a diagram showing an example of a cell configuration in an NR system. This is a connection configuration diagram showing an example of a terminal connection configuration in SL communication. This is a connection configuration diagram showing an example of a base station connection configuration that supports access and backhaul integration. This is a diagram showing RLF processing as defined in the 3GPP standard. This is a schematic diagram of the case where an AI / ML model is used for RLF processing in Embodiment 1. This is a diagram showing an example of RLF detection and prediction processing in Embodiment 1. This is a diagram showing another example of RLF detection and prediction processing in Embodiment 1. This is a schematic diagram of Embodiment 1 in which an AI / ML model is used to predict CQI < Qout detection and CQI > Qin detection. This is a diagram showing an example of communication quality detection prediction processing in Embodiment 1. This is a diagram showing another example of communication quality detection prediction processing in Embodiment 1. This is a schematic diagram of Embodiment 1 in which an AI / ML model is used to predict PLP detection. This is a diagram showing an example of PLP detection prediction processing in Embodiment 1. This is a schematic diagram of Embodiment 1 in which an AI / ML model is used to directly predict PLP detection. This is a diagram showing another example of PLP detection prediction processing in Embodiment 1. This is a schematic diagram of Embodiment 1 in which an AI / ML model is used to predict RLF detection prediction processing. This is a diagram showing another example of PLP detection prediction processing in Embodiment 1. This is a schematic diagram of Embodiment 1 in which an AI / ML model is used to predict T310 expiration. This is a diagram showing an example of T310 expiration prediction processing in Embodiment 1. This is a diagram showing another example of T310 expiration prediction processing in Embodiment 1. This is another schematic diagram of Embodiment 1, showing how to predict T310 expiration using an AI / ML model.This is another schematic diagram of Embodiment 1, showing how to predict T310 expiration using an AI / ML model. This is a diagram of Embodiment 1, showing an example of T310 expiration prediction processing. This is a diagram of Embodiment 1, showing another example of T310 expiration prediction processing. This is a diagram of Embodiment 1, showing an example of TX4 expiration prediction processing when a new timer (TX4) is provided instead of the timer (T310). This is a diagram of Embodiment 1, showing another example of TX4 expiration prediction processing when a new timer (TX4) is provided instead of the timer (T310). This is a diagram of Embodiment 2, showing the measurement processing defined in the 3GPP standard. This is a schematic diagram of Embodiment 2, showing how to use an AI / ML model for measurement processing. This is a diagram of Embodiment 2, showing an example of event trigger detection prediction processing. This is a diagram of Embodiment 2, showing another example of event trigger detection prediction processing. This is a schematic diagram of Embodiment 2, showing how to predict TTT continuation using an AI / ML model for measurement processing. This figure shows an example of TTT continuation prediction processing for Embodiment 2. This is a schematic diagram of the case where event entry prediction processing and TTT continuation prediction processing are performed for Embodiment 2. This figure shows another example of the case where event entry prediction processing and TTT continuation prediction processing are performed for Embodiment 2. This is a schematic diagram of event trigger detection prediction processing using an AI / ML model for Embodiment 2. This figure shows another example of event trigger detection prediction processing for Embodiment 2. This is a schematic diagram of flight prediction processing using an AI / ML model for Embodiment 3. This is a schematic diagram of flight prediction processing that predicts deviation using an AI / ML model for Embodiment 3. This is a schematic diagram of a method for determining deviation after flight prediction processing using an AI / ML model for Embodiment 3. This is a schematic diagram of overheat prediction processing using an AI / ML model for Embodiment 4. This is a schematic diagram of overheat prediction processing that predicts overheat detection using an AI / ML model for Embodiment 4. This is a schematic diagram of a method for determining overheat detection after UE temperature prediction processing using an AI / ML model for Embodiment 4. Embodiment 5 is a diagram representing the IDC processing defined in the 3GPP standard. Embodiment 5 is a schematic diagram of the IDC prediction processing using an AI / ML model.This is a schematic diagram of a method for analyzing interference factors after IDC prediction processing using an AI / ML model in Embodiment 5. This is a diagram showing an example of IDC prediction processing in Embodiment 5. This is a diagram showing an example of IDC prediction processing when a process is performed to wait for notification of interference factor prediction results in Embodiment 5. This is a schematic diagram showing another example of IDC prediction processing using an AI / ML model in Embodiment 5. This is a diagram showing another example of IDC prediction processing in Embodiment 5. This is a diagram showing an example of IDC prediction processing when a process is performed to wait for notification of IDC interference countermeasures in Embodiment 5. This is a diagram showing an example of IDC prediction processing when Phase2x processing is introduced in Embodiment 5. This is a schematic diagram of delay budget prediction processing in Embodiment 6. This is a schematic diagram of a method for determining poor wireless conditions and poor voice communication quality after prediction processing of wireless conditions and voice communication quality using an AI / ML model in Embodiment 6. This is a schematic diagram showing another example of delay budget prediction processing in Embodiment 6. This is a schematic diagram showing another example of delay budget prediction processing in Embodiment 6. This is a schematic diagram showing an example of delay budget prediction processing for multiple UEs performing voice communication in Embodiment 6. This is a diagram showing an example of delay budget prediction processing in Embodiment 6. This is a diagram showing an example of delay budget prediction processing when a waiting process is performed in Phase Y in Embodiment 6.

[0062] Embodiment 1. Figure 2 is a block diagram showing the overall configuration of the NR communication system 210 discussed in 3GPP. Figure 2 will be explained. The radio access network is called NG-RAN (Next Generation Radio Access Network) 211. A mobile terminal device (hereinafter referred to as "Mobile Terminal (User Equipment: UE)") 202, which is a communication terminal device, can communicate wirelessly with a base station device (hereinafter referred to as "NR base station (NG-RAN NodeB: gNB)") 213 and transmits and receives signals wirelessly. The NG-RAN 211 is composed of one or more NR base stations 213.

[0063] Here, "communication terminal equipment" includes not only mobile terminal equipment such as portable mobile phone terminals, but also stationary devices such as sensors. In the following explanation, "communication terminal equipment" may sometimes be simply referred to as "communication terminal."

[0064] The Access Stratum (AS) protocol is terminated between UE202 and NG-RAN211. Examples of AS protocols used include RRC (Radio Resource Control), SDAP (Service Data Adaptation Protocol), PDCP (Packet Data Convergence Protocol), RLC (Radio Link Control), MAC (Medium Access Control), and PHY (Physical layer). RRC is used in the control plane (hereinafter sometimes referred to as C-plane, C-Plane, or CP), SDAP is used in the user plane (hereinafter sometimes referred to as U-plane, U-Plane, or UP), and PDCP, MAC, RLC, and PHY are used in both the C-plane and U-plane.

[0065] The control protocol RRC (Radio Resource Control) between UE202 and NR base station 213 performs functions such as broadcasting, paging, and RRC connection management. The states of NR base station 213 and UE202 in RRC are RRC_IDLE, RRC_CONNECTED, and RRC_INACTIVE.

[0066] In RRC_IDLE mode, functions such as PLMN (Public Land Mobile Network) selection, System Information (SI) broadcasting, paging, cell re-selection, and mobility are performed. In RRC_CONNECTED mode, the mobile terminal has an RRC connection and can send and receive data with the network. In RRC_CONNECTED mode, functions such as handover (HO) and neighbor cell measurement are performed. In RRC_INACTIVE mode, while the connection between the 5G core unit 214 and the NR base station 213 is maintained, functions such as System Information (SI) broadcasting, paging, cell re-selection, and mobility are performed.

[0067] The gNB 213 is connected via an NG interface to a 5G core unit (hereinafter sometimes referred to as the "5GC unit") 214, which includes an Access and Mobility Management Function (AMF), a Session Management Function (SMF), or a User Plane Function (UPF). Control information and / or user data are communicated between the gNB 213 and the 5GC unit 214. The NG interface is a collective term for the N2 interface between the gNB 213 and the AMF 220, the N3 interface between the gNB 213 and the UPF 221, the N11 interface between the AMF 220 and the SMF 222, and the N4 interface between the UPF 221 and the SMF 222. Multiple 5GC units 214 may be connected to a single gNB 213. The gNB213 units are connected via an Xn interface, and control information and / or user data are communicated between them.

[0068] The 5GC unit 214 is a higher-level device, specifically a higher-level node, and controls the connection between the NR base station 213 and the mobile terminal (UE) 202, and distributes paging signals to one or more NR base stations (gNB) 213 and / or LTE base stations (E-UTRAN NodeB: eNB). The 5GC unit 214 also performs mobility control in the idle state. The 5GC unit 214 manages the tracking area list when the mobile terminal 202 is in the idle state, in the inactive state, and in the active state. The 5GC unit 214 initiates the paging protocol by sending a paging message to a cell belonging to the tracking area where the mobile terminal 202 is registered.

[0069] The gNB213 may constitute one or more cells. If one gNB213 constitutes multiple cells, each cell is configured to communicate with the UE202.

[0070] The gNB213 may be divided into a Central Unit (CU) 215 and a Distributed Unit (DU) 216. One CU 215 is configured within the gNB213. One or more DU 216 are configured within the gNB213. One DU 216 constitutes one or more cells. The CU 215 is connected to the DU 216 by an F1 interface, and control information and / or user data are communicated between the CU 215 and the DU 216. The F1 interface consists of an F1-C interface and an F1-U interface. The CU 215 is responsible for the functions of the RRC, SDAP, and PDCP protocols, and the DU 216 is responsible for the functions of the RLC, MAC, and PHY protocols. One or more TRPs (Transmission Reception Points) 219 may be connected to the DU 216. The TRP219 transmits and receives wireless signals to and from the UE.

[0071] CU215 may be divided into a C-plane CU (CU-C) 217 ​​and a U-plane CU (CU-U) 218. One CU-C 217 is configured within CU215. One or more CU-U 218s are configured within CU215. CU-C 217 is connected to CU-U 218 by an E1 interface, and control information is communicated between CU-C 217 and CU-U 218. CU-C 217 is connected to DU216 by an F1-C interface, and control information is communicated between CU-C 217 and DU216. CU-U 218 is connected to DU216 by an F1-U interface, and user data is communicated between CU-U 218 and DU216.

[0072] In a 5G communication system, the Unified Data Management (UDM) function and Policy Control Function (PCF) described in Non-Patent Document 10 (3GPP TS23.501) may be included. The UDM and / or PCF may be included in the 5GC unit 214 in Figure 2.

[0073] In a 5G communication system, a Location Management Function (LMF) as described in Non-Patent Document 24 (3GPP TS38.305) may be provided. The LMF may be connected to a base station via an AMF, as disclosed in Non-Patent Document 25 (3GPP TS23.273).

[0074] In a 5G communication system, the Non-3GPP Interworking Function (N3IWF) described in Non-Patent Document 10 (3GPP TS23.501) may be included. In non-3GPP access with a UE, the N3IWF may terminate the Access Network (AN) with the UE.

[0075] Figure 3 shows the configuration of a DC (Dual Connectivity) connected to the NG core. In Figure 3, solid lines indicate U-Plane connections, and dashed lines indicate C-Plane connections. In Figure 3, the master base station 240-1 may be a gNB or an eNB. Similarly, the secondary base station 240-2 may be a gNB or an eNB. For example, in Figure 3, a DC configuration in which the master base station 240-1 is a gNB and the secondary base station 240-2 is an eNB may be called NG-EN-DC. Figure 3 shows an example in which the U-Plane connection between the 5GC unit 214 and the secondary base station 240-2 is performed via the master base station 240-1, but it may also be performed directly between the 5GC unit 214 and the secondary base station 240-2. Furthermore, in Figure 3, instead of the 5GC unit 214, an EPC (Evolved Packet Core), which is a core network connected to the LTE system and the LTE-A system, may be connected to the master base station 240-1. A U-Plane connection may also be directly established between the EPC and the secondary base station 240-2.

[0076] Figure 4 is a block diagram showing the configuration of the mobile terminal 202 shown in Figure 2. The transmission process of the mobile terminal 202 shown in Figure 4 will now be explained. First, control data from the control unit 310 and user data from the application unit 302 are sent to the protocol processing unit 301. Buffering of the control data and user data may be performed. The buffers for the control data and user data may be provided in the control unit 310, the application unit 302, or the protocol processing unit 301. The protocol processing unit 301 performs protocol processing such as SDAP, PDCP, RLC, MAC, etc., for example, determining the destination base station in DC, etc., and adding headers in each protocol. The data that has undergone protocol processing is passed to the encoder unit 304, where encoding processing such as error correction is performed. There may be data that is output directly from the protocol processing unit 301 to the modulation unit 305 without undergoing encoding processing. The data encoded by the encoder unit 304 is then modulated by the modulation unit 305. Precoding in MIMO may be performed in the modulation unit 305. The modulated data is converted into a baseband signal, then output to the frequency conversion unit 306, where it is converted to a wireless transmission frequency. The transmission signal is then sent from antennas 307-1 to 307-4 to the base station 213. Figure 4 illustrates the case with four antennas, but the number of antennas is not limited to four.

[0077] Furthermore, the reception processing of the mobile terminal 202 is performed as follows: A radio signal from the base station 213 is received by antennas 307-1 to 307-4. The received signal is converted from the radio reception frequency to a baseband signal by the frequency conversion unit 306, and demodulation processing is performed by the demodulation unit 308. Weight calculation and multiplication processing may also be performed in the demodulation unit 308. The demodulated data is passed to the decoder unit 309, where decoding processing such as error correction is performed. The decoded data is passed to the protocol processing unit 301, where protocol processing such as MAC, RLC, PDCP, and SDAP is performed, for example, operations such as header removal in each protocol. Of the data that has undergone protocol processing, control data is passed to the control unit 310, and user data is passed to the application unit 302.

[0078] The series of processes performed by the mobile terminal 202 are controlled by the control unit 310. Therefore, although the control unit 310 is omitted in Figure 4, it is connected to each of the units 302, 304 to 309.

[0079] Each part of the mobile terminal 202, for example, the control unit 310, the protocol processing unit 301, the encoder unit 304, and the decoder unit 309, are implemented by a processing circuit that includes, for example, a processor and memory. For example, the control unit 310 is implemented by the processor executing a program that describes a series of processes for the mobile terminal 202. The program that describes a series of processes for the mobile terminal 202 is stored in memory. Examples of memory include non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), and flash memory. Each part of the mobile terminal 202, for example, the control unit 310, the protocol processing unit 301, the encoder unit 304, and the decoder unit 309, may also be implemented by a dedicated processing circuit such as an FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit), or DSP (Digital Signal Processor). In Figure 4, the number of antennas used by the mobile terminal 202 for transmission and the number of antennas used for reception may be the same or different.

[0080] Figure 5 is a block diagram showing the configuration of the base station 213 shown in Figure 2. The transmission process of the base station 213 shown in Figure 5 will now be explained. The EPC communication unit 401 transmits and receives data between the base station 213 and the EPC. The 5GC communication unit 412 transmits and receives data between the base station 213 and the 5GC (5GC unit 214, etc.). The other base station communication unit 402 transmits and receives data with other base stations. The EPC communication unit 401, the 5GC communication unit 412, and the other base station communication unit 402 each exchange information with the protocol processing unit 403. Control data from the control unit 411, as well as user data and control data from the EPC communication unit 401, the 5GC communication unit 412, and the other base station communication unit 402, are sent to the protocol processing unit 403. Buffering of control data and user data may be performed. A buffer for control data and user data may be provided in the control unit 411, the EPC communication unit 401, the 5GC communication unit 412, or the other base station communication unit 402.

[0081] The protocol processing unit 403 performs protocol processing such as SDAP, PDCP, RLC, and MAC, for example, routing of transmission data in DCs, and adding headers for each protocol. The processed data is passed to the encoder unit 405, where it undergoes encoding processing such as error correction. There may be data that is output directly from the protocol processing unit 403 to the modulation unit 406 without undergoing encoding processing. Data may also be sent from the protocol processing unit 403 to the other base station communication unit 402. For example, in a DC, data sent from the 5GC communication unit 412 or the EPC communication unit 401 may be sent to another base station, such as a secondary base station, via the other base station communication unit 402. The encoded data is then modulated in the modulation unit 406. Precoding in MIMO may be performed in the modulation unit 406. The modulated data is converted into a baseband signal and then output to the frequency conversion unit 407, where it is converted to a wireless transmission frequency. Subsequently, a transmission signal is sent from antennas 408-1 to 408-4 to one or more mobile terminals 202. Figure 5 illustrates the case where there are four antennas, but the number of antennas is not limited to four.

[0082] Furthermore, the reception processing at base station 213 is performed as follows: A radio signal from one or more mobile terminals 202 is received by antennas 408-1 to 408-4. The received signal is converted from the radio reception frequency to a baseband signal by the frequency conversion unit 407, and demodulation processing is performed by the demodulation unit 409. The demodulated data is passed to the decoder unit 410, where decoding processing such as error correction is performed. The decoded data is passed to the protocol processing unit 403, where protocol processing such as MAC, RLC, PDCP, and SDAP is performed, for example, operations such as header removal in each protocol. Of the data that has undergone protocol processing, control data is passed to the control unit 411, the 5GC communication unit 412, the EPC communication unit 401, or the other base station communication unit 402, and user data is passed to the 5GC communication unit 412, the EPC communication unit 401, or the other base station communication unit 402. Data sent from another base station communication unit 402 may be sent to the 5GC communication unit 412 or the EPC communication unit 401. This data may be, for example, uplink data sent to the 5GC communication unit 412 or the EPC communication unit 401 via another base station in a DC.

[0083] The series of processes performed by the base station 213 are controlled by the control unit 411. Therefore, although the control unit 411 is omitted in Figure 5, it is connected to each of the units 401, 402, 405-410, and 412.

[0084] Each part of the base station 213, for example, the control unit 411, the protocol processing unit 403, the 5GC communication unit 412, the EPC communication unit 401, the other base station communication unit 402, the encoder unit 405, and the decoder unit 410, are implemented as processing circuits including a processor and memory, similar to the mobile terminal 202 described above, or as dedicated processing circuits such as FPGA, ASIC, and DSP. In Figure 5, the number of antennas used by the base station 213 for transmission and the number of antennas used for reception may be the same or different.

[0085] As an example of the configuration of the CU215 shown in Figure 2, a configuration may be used in which a DU communication unit is added, excluding the encoder unit 405, modulation unit 406, frequency conversion unit 407, antennas 408-1 to 408-4, demodulation unit 409, and decoder unit 410 shown in Figure 5. The DU communication unit is connected to the protocol processing unit 403. The protocol processing unit 403 in the CU215 performs protocol processing such as PDCP and SDAP.

[0086] As an example of the DU216 configuration shown in Figure 2, a configuration may be used in which a CU communication unit is provided, excluding the EPC communication unit 401, other base station communication unit 402, and 5GC communication unit 412 shown in Figure 5. The CU communication unit is connected to the protocol processing unit 403. The protocol processing unit 403 in the DU216 performs protocol processing such as PHY, MAC, and RLC.

[0087] Figure 6 is a block diagram showing the configuration of the 5GC unit. Figure 6 shows the configuration of the 5GC unit 214 shown in Figure 2. Figure 6 shows the case where the 5GC unit 214 shown in Figure 2 includes the configuration of the AMF, SMF, and UPF. In the example shown in Figure 6, the AMF may have the function of the control plane control unit 525, the SMF may have the function of the session management unit 527, and the UPF may have the functions of the user plane communication unit 523 and the Data Network communication unit 521. The Data Network communication unit 521 transmits and receives data between the 5GC unit 214 and the Data Network. The base station communication unit 522 transmits and receives data between the 5GC unit 214 and the base station 213 via the NG interface. User data sent from the Data Network is passed from the Data Network communication unit 521 to the base station communication unit 522 via the user-plane communication unit 523, and then transmitted to one or more base stations 213. User data sent from the base station 213 is passed from the base station communication unit 522 to the Data Network communication unit 521 via the user-plane communication unit 523, and then transmitted to the Data Network.

[0088] Control data sent from base station 213 is passed from base station communication unit 522 to control plane control unit 525. The control plane control unit 525 may pass the control data to session management unit 527. Control data may also be sent from Data Network. Control data sent from Data Network may be sent from Data Network communication unit 521 to session management unit 527 via user plane communication unit 523. The session management unit 527 may also send the control data to control plane control unit 525.

[0089] The user-plane communication unit 523 includes the PDU processing unit 523-1 and the mobility anchoring unit 523-2, and performs all processing for the user-plane (hereinafter sometimes referred to as U-Plane). The PDU processing unit 523-1 processes data packets, for example, sending and receiving packets with the Data Network communication unit 521 and sending and receiving packets with the base station communication unit 522. The mobility anchoring unit 523-2 is responsible for maintaining the data path when the UE is in mobility.

[0090] The session management unit 527 manages the PDU session established between the UE and the UPF. The session management unit 527 includes the PDU session control unit 527-1 and the UE IP address assignment unit 527-2. The PDU session control unit 527-1 manages the PDU session between the mobile terminal 202 and the 5GC unit 214. The UE IP address assignment unit 527-2 assigns an IP address to the mobile terminal 202.

[0091] The control plane control unit 525 includes the NAS security unit 525-1 and the idle state mobility management unit 525-2, and performs all processing for the control plane (hereinafter sometimes referred to as C-Plane). The NAS security unit 525-1 performs security for NAS (Non-Access Stratum) messages, etc. The idle state mobility management unit 525-2 performs mobility management in the standby state (also referred to as RRC_IDLE state or simply idle), generation and control of paging signals in the standby state, addition, deletion, updating, searching, and tracking area management for one or more mobile terminals 202 under its umbrella.

[0092] The series of processes in the 5GC unit 214 are controlled by the control unit 526. Therefore, although the control unit 526 is omitted in Figure 6, it is connected to each of the units 521 to 523, 525, and 527. Each of the units in the 5GC unit 214 is implemented by a processing circuit consisting of, for example, a processor and memory, or by a dedicated processing circuit such as an FPGA, ASIC, or DSP, similar to the control unit 310 of the mobile terminal 202 described above.

[0093] Next, an example of a cell search method in a communication system is shown. Figure 7 is a flowchart illustrating the process from cell search to standby operation performed by a communication terminal (UE) in an NR-type communication system. When the communication terminal starts a cell search, in step ST601, it synchronizes the slot timing and frame timing using the first synchronization signal (P-SS) and the second synchronization signal (S-SS) transmitted from the surrounding base station.

[0094] P-SS and S-SS together are called the Synchronization Signal (SS). Each cell is assigned a synchronization code that corresponds one-to-one with its assigned PCI (Physical Cell Identifier). 1008 different PCI combinations are being considered. The communication terminal uses these 1008 PCI combinations to synchronize and also detects (identifies) the PCI of the synchronized cell.

[0095] The communication terminal then receives the PBCH in step ST602 for the next synchronized cell. The BCCH on the PBCH is mapped to the MIB (Master Information Block), which contains cell configuration information. Therefore, by receiving the PBCH and obtaining the BCCH, the MIB can be obtained. The information in the MIB includes, for example, the SFN (System Frame Number), scheduling information for SIB (System Information Block) 1, subcarrier spacing for SIB1 etc., and DM-RS position information.

[0096] Furthermore, the communication terminal obtains the SS block identifier from the PBCH. Part of the bit sequence of the SS block identifier is included in the MIB. The remaining bit sequence is included in the identifier used for sequence generation of the DM-RS associated with the PBCH. The communication terminal obtains the SS block identifier using the MIB included in the PBCH and the sequence of the DM-RS associated with the PBCH.

[0097] Next, in step ST603, the communication terminal measures the received power of the SS block.

[0098] Next, in step ST604, the communication terminal selects the cell with the best reception quality from among the one or more cells detected up to step ST603, for example, the cell with the highest received power, i.e., the best cell. The communication terminal also selects the beam with the best reception quality, for example, the beam with the highest received power of the SS block, i.e., the best beam. For example, the received power of the SS block for each SS block identifier is used to select the best beam.

[0099] Next, in step ST605, the communication terminal receives DL-SCH based on the scheduling information of SIB1 included in MIB and obtains SIB (System Information Block) 1 in the broadcast information BCCH. SIB1 contains information regarding access to the cell, cell configuration information, and scheduling information for other SIBs (SIBk: an integer k ≥ 2). SIB1 also contains the Tracking Area Code (TAC).

[0100] Next, in step ST606, the communication terminal compares the TAC of SIB1 received in step ST605 with the TAC portion of the Tracking Area Identity (TAI) in the Tracking Area List already held by the communication terminal. The Tracking Area List is also called the TAI list. TAI is identification information for identifying a tracking area and consists of MCC (Mobile Country Code), MNC (Mobile Network Code), and TAC (Tracking Area Code). MCC is the country code. MNC is the network code. TAC is the code number of the tracking area.

[0101] If, as a result of the comparison in step ST606, the TAC received in step ST605 is the same as a TAC included in the tracking area list, the communication terminal enters standby mode in that cell. If, after comparison, the TAC received in step ST605 is not included in the tracking area list, the communication terminal requests a change in the tracking area through that cell to the Core Network (CN) in order to perform a Tracking Area Update (TAU).

[0102] The devices constituting the core network (hereinafter sometimes referred to as "core network devices") update the tracking area list based on the identification number (UE-ID, etc.) of the communication terminal sent from the communication terminal along with the TAU request signal. The core network devices transmit the updated tracking area list to the communication terminal. The communication terminal rewrites (updates) its TAC list based on the received tracking area list. After that, the communication terminal enters a waiting state in that cell.

[0103] Next, we will show examples of random access methods in communication systems. In random access, four-step random access and two-step random access are used. Furthermore, for both four-step and two-step random access, there are contention-based random access, that is, random access where timing collisions with other mobile terminals may occur, and contention-free random access.

[0104] An example of a collision-based four-step random access method is shown. In the first step, the mobile terminal transmits a random access preamble to the base station. The random access preamble may be selected by the mobile terminal from a predetermined range, or it may be individually assigned to the mobile terminal and notified by the base station.

[0105] In the second step, the base station sends a random access response to the mobile terminal. The random access response includes uplink scheduling information used in the third step, and a terminal identifier used in the uplink transmission of the third step.

[0106] In the third step, the mobile terminal transmits data uplink to the base station. The mobile terminal uses the information obtained in the second step for the uplink transmission. In the fourth step, the base station notifies the mobile terminal whether or not a collision has been resolved. If the mobile terminal is notified that there is no collision, it terminates the random access process. If the mobile terminal is notified that there is a collision, it restarts the process from the first step.

[0107] The collision-free four-step random access method differs from the collision-based four-step random access method in the following ways: Prior to the first step, the base station pre-assigns a random access preamble and uplink scheduling to the mobile terminal. Furthermore, notification of collision resolution status is unnecessary in the fourth step.

[0108] An example of a collision-based two-step random access method is shown below. In the first step, the mobile terminal sends a random access preamble and an uplink transmission to the base station. In the second step, the base station notifies the mobile terminal whether a collision occurred. If the mobile terminal is notified that there was no collision, it terminates the random access process. If the mobile terminal is notified that there was a collision, it restarts the process from the first step.

[0109] The collision-free two-step random access method differs from the collision-based two-step random access method in the following ways: Prior to the first step, the base station pre-assigns a random access preamble and uplink scheduling to the mobile terminal. In the second step, the base station transmits a random access response to the mobile terminal.

[0110] Figure 8 shows an example of a cell configuration in NR. In an NR cell, a narrow beam is formed and transmitted by changing its direction. In the example shown in Figure 8, the base station 750 uses beam 751-1 to transmit and receive data with a mobile terminal at a certain time. At other times, the base station 750 uses beam 751-2 to transmit and receive data with a mobile terminal. Similarly, the base station 750 uses one or more of beams 751-3 to 751-8 to transmit and receive data with a mobile terminal. In this way, the base station 750 configures a wide-area cell 752.

[0111] Figure 8 shows an example where the base station 750 uses eight beams, but the number of beams may be different from eight. Also, in the example shown in Figure 8, the base station 750 uses one beam simultaneously, but it may use multiple beams.

[0112] The concept of Quasi-Colocation (QCL) is used for beam identification (see Non-Patent Document 14 (3GPP TS38.214)). That is, the beam is identified by information indicating which reference signal (e.g., SS block, CSI-RS) beam it can be considered identical to. This information may include information about the aspects of the beams that can be considered identical, such as Doppler shift, Doppler shift diffusion, mean delay, mean delay diffusion, and spatial Rx parameters (see Non-Patent Document 14 (3GPP TS38.214)).

[0113] In 3GPP, Side Link (SL) is supported for D2D (Device to Device) and V2V (Vehicle to Vehicle) communication (see Non-Patent Documents 1 and 16). SL is defined by the PC5 interface.

[0114] In SL communication, in addition to broadcast, support for unicast and groupcast is being considered, and therefore support for PC5-S signaling is being explored (see Non-Patent Document 27 (3GPP TS23.287)). For example, PC5-S signaling is implemented to establish a link for SL, i.e., PC5 communication. This link is implemented at the V2X layer and is also referred to as a Layer 2 link.

[0115] Furthermore, support for RRC signaling in SL communication is being considered (see Non-Patent Document 27 (3GPP TS23.287)). RRC signaling in SL communication is also referred to as PC5 RRC signaling. For example, it has been proposed to notify UEs of their capabilities between UEs performing PC5 communication, and to notify AS layer settings for performing V2X communication using PC5 communication.

[0116] Figure 9 shows an example of a mobile terminal connection configuration in SL communication. In the example shown in Figure 9, UE805 and UE806 are located within the coverage 803 of base station 801. UL / DL communication 807 is performed between base station 801 and UE805. UL / DL communication 808 is performed between base station 801 and UE806. SL communication 810 is performed between UE805 and UE806. UE811 and UE812 are located outside the coverage 803. SL communication 814 is performed between UE805 and UE811. Additionally, SL communication 816 is performed between UE811 and UE812.

[0117] As an example of communication between a UE and a NW via a relay in SL communication, UE805, shown in Figure 9, relays communication between UE811 and base station 801.

[0118] A UE performing relay may use a configuration similar to that shown in Figure 4. The relay processing in the UE will be explained using Figure 4. The relay processing by UE 805 in communication from UE 811 to base station 801 will be explained. The radio signal from UE 811 is received by antennas 307-1 to 307-4. The received signal is converted from the radio reception frequency to a baseband signal by the frequency conversion unit 306, and demodulation processing is performed in the demodulation unit 308. Weight calculation and multiplication processing may also be performed in the demodulation unit 308. The demodulated data is passed to the decoder unit 309, where decoding processing such as error correction is performed. The decoded data is passed to the protocol processing unit 301, where protocol processing such as MAC and RLC used for communication with UE 811 is performed, for example, operations such as header removal in each protocol. Protocol processing such as RLC and MAC used for communication with base station 801 is also performed, for example, operations such as header addition in each protocol. In the protocol processing unit 301 of the UE811, PDCP and SDAP protocol processing may also be performed. The processed data is passed to the encoder unit 304, where encoding processing such as error correction is performed. There may also be data that is output directly from the protocol processing unit 301 to the modulation unit 305 without encoding processing. The data encoded by the encoder unit 304 is then modulated by the modulation unit 305. Precoding in MIMO may be performed in the modulation unit 305. The modulated data is converted into a baseband signal and then output to the frequency conversion unit 306, where it is converted to a wireless transmission frequency. After that, the transmission signal is sent from antennas 307-1 to 307-4 to the base station 801.

[0119] As described above, an example of relaying by UE805 in communication from UE811 to base station 801 was shown, but the same process is used in relaying communication from base station 801 to UE811.

[0120] 5G base stations can support Integrated Access and Backhaul (IAB) (see Non-Patent Documents 2 and 20). A base station that supports IAB (hereinafter sometimes referred to as an IAB base station) consists of an IAB donor CU, which is a CU of the base station that operates as an IAB donor providing IAB functionality; an IAB donor DU, which is a DU of the base station that operates as an IAB donor; and an IAB node that is connected to the IAB donor DU and to the UE using a wireless interface. An F1 interface is provided between the IAB node and the IAB donor CU (see Non-Patent Document 2).

[0121] Figure 10 shows an example of IAB base station connections. IAB donor CU901 is connected to IAB donor DU902. IAB node 903 is connected to IAB donor DU902 using a wireless interface. IAB node 903 is connected to IAB node 904 using a wireless interface. In other words, multi-stage connections of IAB nodes may occur. UE905 is connected to IAB node 904 using a wireless interface. UE906 may be connected to IAB node 903 using a wireless interface, and UE907 may be connected to IAB donor DU902 using a wireless interface. Multiple IAB donor DU902 may be connected to IAB donor CU901, multiple IAB node 903 may be connected to IAB donor DU902, and multiple IAB node 904 may be connected to IAB node 903.

[0122] A BAP (Backhaul Adaptation Protocol) layer is provided in the connection between IAB donor DUs and IAB nodes, and in the connection between IAB nodes (see Non-Patent Document 29). The BAP layer performs operations such as routing received data to IAB donor DUs and / or IAB nodes, and mapping to RLC channels (see Non-Patent Document 29).

[0123] As an example of the configuration of an IAB donor CU, a configuration similar to that of CU215 is used.

[0124] As an example of the IAB donor DU configuration, a configuration similar to that of DU216 is used. In the protocol processing section of the IAB donor DU, BAP layer processing is performed, such as adding a BAP header to downlink data, routing to IAB nodes, and removing the BAP header from uplink data.

[0125] As an example of an IAB node configuration, a configuration excluding the EPC communication unit 401, other base station communication unit 402, and 5GC communication unit 412 shown in Figure 5 may be used.

[0126] The transmission and reception processing at the IAB node will be explained using Figures 5 and 10. The transmission and reception processing of the IAB node 903 in communication between the IAB donor CU 901 and UE 905 will be explained. In uplink communication from UE 905 to IAB donor CU 901, the radio signal from the IAB node 904 is received by the antenna 408 (part or all of antennas 408-1 to 408-4). The received signal is converted from the radio reception frequency to a baseband signal by the frequency conversion unit 407, and demodulation processing is performed by the demodulation unit 409. The demodulated data is passed to the decoder unit 410, where decoding processing such as error correction is performed. The decoded data is passed to the protocol processing unit 403, where protocol processing such as MAC and RLC used for communication with the IAB node 904 is performed, for example, operations such as header removal in each protocol. Furthermore, routing to the IAB donor DU902 is performed using a BAP header, and protocol processing such as RLC and MAC used for communication with the IAB donor DU902 is performed, such as adding headers for each protocol. The processed data is passed to the encoder unit 405, where encoding processing such as error correction is performed. There may be data that is output directly from the protocol processing unit 403 to the modulation unit 406 without encoding processing. The encoded data is then modulated in the modulation unit 406. Precoding in MIMO may be performed in the modulation unit 406. The modulated data is converted into a baseband signal, then output to the frequency conversion unit 407, where it is converted to a wireless transmission frequency. After that, a transmission signal is sent from antennas 408-1 to 408-4 to the IAB donor DU902. Similar processing is performed in downlink communication from the IAB donor CU901 to UE905.

[0127] The same transmission and reception processing is performed at IAB node 904 as at IAB node 903. In the protocol processing unit 403 of IAB node 903, as part of the BAP layer processing, for example, the addition of a BAP header and routing to IAB node 904 in uplink communication, and the removal of a BAP header in downlink communication are performed.

[0128] Hereafter, expressions using the symbol slash ( / ) mean that at least one of the two elements before and after the slash is included. The introduction of AI (Artificial Intelligence) / ML (Machine Learning) is being considered in mobile communication systems under 3GPP. For example, the use of AI / ML in radio link failure (RLF) processing / handover failure (HOF) processing is being discussed (see Non-Patent Documents 33, 34, and 35). These Non-Patent Documents propose that mobile communication systems reduce RLF / HOF by using AI / ML to predict RLF / HOF and taking early action. However, the above Non-Patent Documents do not disclose specific methods for how to introduce RLF / HOF processing using AI / ML into a communication system. For example, there is at least one of the following problems in RLF / HOF processing using AI / ML. (a1) What is to be predicted in RLF / HOF processing using AI / ML? (a2) How are the prediction results obtained using AI / ML notified? (a3) ​​What processing is performed in conjunction with the notification of the prediction results? As a result, RLF / HOF processing using AI / ML is not performed effectively, and problems such as not being able to improve communication quality occur.

[0129] This embodiment discloses a method for solving at least one of the above (a1) to (a3).

[0130] Figure 11 is a diagram illustrating the RLF processing (hereinafter sometimes simply referred to as "RLF processing") defined in the 3GPP standard (see Non-Patent Literature 1). The UE monitors whether the communication quality falls below a predetermined value (Qout). Such a decrease in communication quality is sometimes referred to as "out-of-sync". If the communication quality falls below the predetermined value (Qout) for a predetermined number of consecutive times (N310), the UE detects a Physical Layer Problem (hereinafter referred to as "PLP") and starts a timer (T310). The communication quality may be the BLER (BLock Error Ratio) or CQI of the PDCCH. After the timer (T310) starts, the UE monitors whether the communication quality exceeds a predetermined value (Qin). Such recovery of communication quality is sometimes referred to as "in-sync". The UE detects RLF if the communication quality does not exceed a predetermined value (Qin) for a predetermined number of consecutive times (N311) before the timer (T310) expires. When the UE detects RLF, it performs RRC connection re-establishment processing. If the RRC connection re-establishment processing is not successful before the timer (T311) expires, the UE transitions to RRC_Idle. Note that the expression "timer expires" can also be rephrased as "timer expires".

[0131] AI / ML can be used for RLF processing. Figure 12 is a schematic diagram of the case where an AI / ML model is used for RLF processing. The user e-engineer (UE) predicts RLF detection using the AI / ML model. Along with RLF detection, the UE may also predict the time when the RLF will be detected and the UE's position when the RLF is detected. The UE inputs input information to the AI / ML model. The prediction result is output from the AI / ML model. For example, the prediction result may include at least one of the following (b1) to (b2): (b1) The RLF is detected at time tx, where time tx is a future time. (b2) The RLF is detected when the UE is at position Lx, where position Lx is a future position that the UE is predicted to pass through. In this way, it becomes possible to predict when and where the RLF will be detected, and it can be applied to a wide variety of applications.

[0132] The UE may predict connectable cells along with RLF detection. The UE may predict one or more connectable cells. The prediction results may include information regarding the priority of one or more connectable cells. This allows for the early identification of cells that will undergo RRC connection re-establishment processing after RLF detection, enabling early connection to those cells.

[0133] RLF detection may be, for example, the expiration of a timer (T310) in SPCell, or the expiration of a predetermined period of time since the trigger of a measurement report in SPCell (expiration of a timer (T312)), or an indication of a random access problem (RA) from the MCG or SCG, or an indication that the maximum number of retransmissions has been reached from the MCG or SCG. When connected as an IAB node, RLF detection may be, for example, a backhaul (BH) RLF indication received via BAP from the MCG or SCG. RLF detection may be, for example, an LBT (Listen Before Talk) failure indication from the MCG or SCG. These may be predicted individually, or some or all of them may be predicted. At least one of these may be output as the cause of RLF detection. In this way, it is possible to obtain more detailed information about what kind of RLF detection was predicted in the RLF detection prediction. Control after deriving the RLF detection prediction result can be performed more effectively.

[0134] The output settings for what to output during the prediction process may be performed by the NW node. The definition of an NW node will be described later. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted by the prediction process.

[0135] The input information to the AI / ML model may include, for example, at least one of the following: RRM (Radio Resource Management) measurement results, RLM (Radio Link Monitoring) measurement results, and CSI measurement results. The input information may include, for example, at least one of the following: time, UE position, UE velocity, UE trajectory, etc. The input information may include, for example, at least one of the following: PLP detection conditions and RLF detection conditions. The input information may include, for example, transition conditions to RRC_Idle. The above detection conditions may include, for example, the conditions described above. The detection conditions may include parameters used for PLP or RLF detection. The parameters may include, for example, at least one of Qout, Qin, T310, N310, and N311. The above transition conditions may be, for example, the conditions described above. The transition conditions may be parameters used in the above conditions. The parameters may be, for example, T311, etc. By doing so, RLF detection can be directly predicted, and along with RLF detection, its time and location can also be predicted.

[0136] The input information may be set by the NW node. The NW node may transmit the input information to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling.

[0137] Hereafter, the prediction process for RLF detection using the AI / ML model may be referred to as the "RLF detection prediction process" to distinguish it from the actual RLF process (i.e., the process described in Figure 12). The RLF detection prediction process may be performed in parallel with the RLF process and the RLM process. The RLF detection prediction process may be performed independently of the RLF process and the RLM process, or it may be performed in correlation with at least one of the RLF process and the RLM process. The RLF detection prediction process may be started periodically or cyclically. Periodic startup may, for example, start the RLF detection prediction process at a predetermined date and time. Cyclical startup may, for example, repeatedly start the RLF detection prediction process at predetermined time intervals.

[0138] Periodic or cyclical startups may be configured by the NW node. For example, the NW node may send configuration information regarding the startup timing to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The configuration information may include at least one of the following: the startup start timing, the offset from the time the configuration information is received to the startup start timing, the period, the processing time for the RLF detection prediction process, and the end timing of the RLF detection prediction process.

[0139] The RLF detection and prediction process may be initiated in correlation with at least one of the RLF process and the RLM process. For example, the RLF detection and prediction process may be initiated by the first detection of CQI < Qout by the RLF process. For example, the RLF detection and prediction process may be initiated by PLP detection by the RLF process. For example, the RLF detection and prediction process may be initiated when the communication quality exceeds a predetermined value (Qin) for a predetermined number of consecutive times (N311). For example, the RLF detection and prediction process may be initiated when the timer (T310) stops. The RLF detection and prediction process may be initiated when an output from the RLF detection and prediction process is derived. The RLF detection and prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of the RLF detection and prediction process.

[0140] New activation conditions for the RLF detection and prediction process may be established. For example, the activation conditions may include at least one of the following (c1) to (c6): (c1) A UE is in a predetermined location. (c2) A UE has entered a predetermined area. (c3) A UE has entered the RRC_Connected state. (c4) A UE has connected to a cell. (c5) A UE has entered HO mode. (c6) A serving cell has been changed. In this way, the RLF detection and prediction process can be activated according to the status of the UE.

[0141] The activation conditions described above may be set by the NW node. For example, the NW node may send information regarding at least one of the above activation conditions to the UE. The UE may then activate the RLF detection and prediction process according to the activation condition indicated by that information.

[0142] RLF detection and prediction processing may be performed upon request from a network node. The RLF detection and prediction processing may be initiated by such request. The network node may be, for example, a base station, AMF, SMF, UPF, NWDAF (Network Data Analytics Function), or AS (Application Server). RLF detection and prediction processing may also be performed upon request from an application. A signaling system may be provided to request RLF detection and prediction processing. The requesting network node, AS, or application may notify the UE of this signaling system. This signaling system may be notified from the requesting network node to the UE via other network nodes. For example, this signaling system may be notified from the AS to the UE via an NEF (Network Exposure Function), AF (Application Function), AMF, or base station. The UE that receives the signaling system executes the RLF detection and prediction processing system. In this way, other nodes can control the RLF detection and prediction processing system on the UE. For example, it becomes possible to determine and control whether or not to perform RLF detection and prediction processing based on communication quality.

[0143] The stopping, resetting, or terminating of the RLF detection prediction process (hereinafter, unless otherwise specified, stopping, resetting, and terminating will be collectively referred to as "reset") may be performed independently of the RLF process and the RLM process. The reset of the RLF detection prediction process may be performed in correlation with at least one of the RLF process and the RLM process. After the RLF detection prediction process is started as described above, if an output is derived from the RLF detection prediction process, the UE may reset the RLF detection prediction process. The RLF detection prediction process may be reset after a predetermined time has elapsed since its start. The RLF detection prediction process may be reset if the communication quality exceeds a predetermined value (Qin) for a predetermined number of consecutive times (N311). For example, the RLF detection prediction process may be reset if it is predicted that the communication quality will exceed a predetermined value (Qin) for a predetermined number of consecutive times (N311). For example, the UE may reset the RLF detection prediction process by stopping the timer (T310). For example, the UE may reset the RLF detection prediction process if it predicts that the timer (T310) will stop. Alternatively, the UE may reset the RLF detection prediction process by stopping the recovery process described later. This allows for flexible operation of the RLF detection prediction process.

[0144] New reset conditions may be established for the RLF detection and prediction process. For example, the reset conditions may include at least one of the following (d1) to (d5): (d1) A UE is in a predetermined location. (d2) A UE has entered a predetermined area. (d3) A UE has entered the RRC_Idle state or the RRC_Inactive state. (d4) A UE has entered HO. (d5) A serving cell has been changed. In this way, the RLF detection and prediction process can be reset according to the status of the UE.

[0145] The reset conditions described above may be set by the NW node. For example, the NW node may send information regarding at least one of the reset conditions described above to the UE. The UE may reset the RLF detection and prediction process according to the conditions indicated by that information.

[0146] The RLF detection and prediction process may be reset upon request from a network node. The RLF detection and prediction process may be reset upon such request. A signaling mechanism for resetting the RLF detection and prediction process may be provided. This signaling mechanism may be notified to the user audience (UE) from a network node, AS, or application. The signaling mechanism may be notified to the UE from the requesting network node via other network nodes. For example, the signaling mechanism may be notified to the UE from the AS via NEF, AF, AMF, or base station. Upon receiving the signaling mechanism, the UE resets the RLF detection and prediction process. This allows other nodes to control the RLF detection and prediction process at the UE. For example, it becomes possible to determine and control whether or not to perform the RLF detection and prediction process based on communication quality.

[0147] While the startup timing and conditions for the RLF detection prediction process have been disclosed, the output derivation timing for the RLF detection prediction process may also be set. A processing time for the RLF detection prediction process may also be provided. This processing time may be set individually for each UE. For example, the processing time for the RLF detection prediction process may be set according to the processing capacity of the UE. The UE may notify the NW node of information regarding its processing capacity. The NW node may set the processing time for the RLF detection prediction process for the UE according to the information regarding the UE's processing capacity. The startup timing for the RLF detection prediction process may be the timing obtained by subtracting the processing time from the output derivation timing for the RLF detection prediction process. In this way, the startup timing for the RLF detection prediction process can be changed according to the processing capacity of the UE. The output derivation timing for the RLF detection prediction process can be set regardless of the processing capacity of the UE.

[0148] This document discloses the processing after RLF detection prediction by the RLF detection prediction process. If the UE predicts RLF detection by the RLF detection prediction process, the UE may determine that RLF has been detected. The UE may then perform post-RLF detection processing by the RLF process in accordance with the RLF detection. For example, the UE may start a timer (T311) and start the RRC connection re-establishment process. By doing so, the RRC connection re-establishment process can be performed earlier, improving the possibility of early connection, including with other cells.

[0149] The UE may notify the NW nodes of information regarding RLF detection predictions. The information regarding RLF detection predictions may include the RLF detection prediction result and prediction information predicted together with the RLF detection prediction result. This prediction information may include, for example, at least one of the time when the RLF is detected and the location of the UE when the RLF is detected. The NW nodes receive the RLF detection prediction result from the UE. The NW nodes become aware of the RLF detection prediction result at the UE.

[0150] A recovery process may be introduced. The network node derives a recovery method. For example, the network node may derive a recovery method based on the RLF detection prediction result received from the user. The recovery method may be, for example, a method to improve communication quality. The recovery method may include, for example, at least one of beam modification and increasing the maximum number of HARQ cycles. By performing such processing, it is possible to improve the communication quality between the user and the network.

[0151] Recovery processing may be performed after PLP detection in RLF processing. The UE may notify the NW node of PLP detection. The NW node becomes aware of PLP detection at the UE. The NW node may start recovery processing upon receiving PLP detection. The NW node derives a recovery method. The NW node may derive settings for applying the recovery method. The NW node may notify the UE of the recovery method. The NW node may notify the UE of recovery configuration information for applying the recovery method. As a method of notifying the recovery configuration information, for example, RRC signaling may be used. For example, the RRCReconfiguration message may be used. By using RRC signaling, a lot of configuration information can be notified to the UE. As another method, for example, MAC signaling may be used. For example, the NW node may include the recovery configuration information in the MAC CE (Control Element) and send it. The network node can notify the UE of the recovery method at an early stage. Alternatively, L1 / L2 signaling may be used. For example, the network node may include recovery configuration information in the DCI and transmit it. The DCI may be transmitted via PDCCH. The network node can notify the UE of the recovery method even earlier. The UE receives the recovery configuration information. The UE can recognize the recovery method.

[0152] Network nodes do not need to notify the UE of the recovery method. For example, a network node may apply a recovery method that does not require any configuration changes on the UE. A network node may apply a recovery method that can be processed only by that network node. This eliminates unnecessary signaling from the network node to the UE.

[0153] The UE may perform at least one of the RLF processing and the RLM processing in parallel with the recovery processing. For example, while the UE is performing the recovery processing, the UE may perform at least one of the following (e1) to (e4): (e1) The UE monitors whether CQI < Qout. (e2) The UE monitors whether CQI > Qin. (e3) The UE monitors whether CQI < Qout is detected N310 times consecutively. (e4) The UE monitors whether CQI > Qin is detected N311 times consecutively. For example, the UE may start post-RF detection processing upon RLF detection by the RLF processing. In this case, the UE may start a timer (T311) and start the RRC connection re-establishment processing. Since the UE can perform recovery processing until actual RLF detection occurs, communication quality can be improved and RLF detection can be suppressed.

[0154] A recovery processing time may be provided. The recovery processing time may be set by a timer. Hereafter, the timer for the recovery processing time will be referred to as timer (TX1), for example. The UE may start timer (TX1) at the timing of the start of the recovery process. When the communication quality improves, for example, when CQI>Qin is detected N311 times consecutively, the UE may reset timer (TX1). The UE may determine that RLF has been detected by the expiration of timer (TX1). The UE may perform post-RF detection processing upon RLF detection. The UE may notify the NW node of the recovery processing reset. The UE may notify the NW node that CQI>Qin has been detected N311 times consecutively. The NW node can recognize the recovery processing reset.

[0155] The value of Timer (TX1) may be 0. If Timer (TX1) is 0, there will be no recovery processing time. Setting Timer (TX1) to 0 allows you to set no recovery processing time.

[0156] Timer (TX1) may be set to any value. For example, Timer (TX1) may be set to be shorter than Timer (T310). If communication quality does not improve through recovery based on early RLF detection prediction, the system can transition to RLF processing earlier. RRC re-establishment processing can be started earlier. The possibility of connecting to a more suitable cell can be improved earlier.

[0157] After the recovery process is reset, the settings in the UE may be changed. For example, the settings in the UE may be changed to the settings used before the recovery process was initiated. After receiving notification of the recovery process reset, the NW node may change the settings for the UE and notify the UE of the changed settings. Regarding the notification method, it is advisable to apply a method similar to the method for notifying recovery setting information described above. The UE will be able to change its recovery settings. This can further improve communication quality.

[0158] Network nodes do not need to notify the UE of changes in their configuration. This configuration may be applied, for example, when no configuration changes are required on the UE. This configuration may also be applied, for example, when the process can be handled solely by the network nodes. This eliminates unnecessary signaling from the network nodes to the UE.

[0159] By doing so, network nodes can recognize RLF detection predictions at the UE early and perform recovery processing promptly. This improves communication quality.

[0160] Unlike actual RLF detection through RLF processing, RLF detection prediction does not necessarily mean that the UE is in an RLF state. The UE may still be able to communicate with the NW node. Performing recovery processing at this stage improves the communication quality between the UE and the NW node and suppresses actual RLF detection.

[0161] The UE may continue the RLF detection prediction process after performing RLF detection prediction by the RLF detection prediction process. The UE may use the RLF detection prediction result as input information to perform RLF detection prediction processing. The UE may perform a reset process after deriving the RLF detection prediction result by the RLF detection prediction process. The UE may perform the RLF detection prediction process again after the reset process.

[0162] A process for predicting RLF recovery may be provided. Hereafter, this process may be referred to as the "recovery prediction process." The recovery prediction process may, for example, be a process that predicts whether CQI > Qin will be detected N311 times consecutively. The output of the recovery prediction process may include, for example, at least one of the following: the RLF recovery prediction result, the prediction result of the time at which RLF recovery will be detected, and the prediction result of the location at which RLF recovery will be detected. The input information for the recovery prediction process may be appropriately adapted from the input information for the RLF detection prediction process. The output information for the recovery prediction process may be appropriately adapted from the output information for the RLF detection prediction process. The processing method for the recovery prediction process may be appropriately adapted from the processing method for the RLF detection prediction process. The recovery prediction process may be included in the RLF detection prediction process or performed separately. Hereafter, unless otherwise specified, the recovery prediction process will be included in the RLF detection prediction process. The UE may notify the NW nodes of the recovery prediction detection. The UE may include the recovery prediction detection in the information regarding the RLF detection prediction and notify the NW nodes. Network nodes will be able to recognize recovery prediction detection.

[0163] The UE may reset the recovery process based on the recovery prediction detection included in the RLF detection prediction process. For example, the UE may reset the recovery process at the time of detecting the recovery prediction. The UE may also reset the timer (TX1). In this way, the NW node can perform other processes earlier.

[0164] Figure 13 is a diagram illustrating an example of RLF detection prediction processing. Figure 13 shows an example using recovery processing. The UE starts RLF detection prediction processing at a predetermined timing, for example. The UE inputs the input information disclosed above into the AI / ML model and performs RLF detection prediction. The UE may also predict RLF recovery. The UE may perform RLF processing in parallel with the RLF detection prediction processing. The UE predicts RLF detection through the RLF detection prediction processing. The UE, having predicted RLF detection, sends information regarding the RLF detection prediction to the NW node. The NW node, upon receiving this information, performs recovery processing. The NW node may notify the UE of settings based on the recovery method (for example, the recovery setting information described above). The UE may perform recovery processing based on these settings.

[0165] If RLF is detected by RLF processing, the UE performs RRC connection re-establishment processing. If the RRC connection re-establishment processing is unsuccessful within the timer (T311), the UE transitions to RRC_Idle. If, during the recovery process, the communication quality exceeds a predetermined value (Qin) for a predetermined number of consecutive times (N311), the UE may reset the RLF detection prediction processing and perform the RLF detection prediction processing again. In this way, the UE can perform RLF detection prediction processing. By performing RLF detection prediction, recovery processing can be performed earlier, and actual RLF detection can be suppressed.

[0166] Figure 14 shows another example of RLF detection prediction processing. Figure 14 shows an example where a timer (TX1) is used for recovery processing. The UE determines that RLF has been detected when the timer (TX1) expires and performs RRC connection re-establishment processing. If the RRC connection re-establishment processing is not successful before the timer (T311) expires, the UE transitions to RRC_Idle. RLF processing does not have to be performed in parallel with RLF detection prediction processing. In this way, the time of recovery processing can be controlled. For example, it is possible to transition to RRC connection re-establishment processing due to RLF detection earlier.

[0167] As another example, the UE may transition to RRC_Idle upon RLF detection prediction. For example, the UE may set the value of timer (T311) to 0. When the RLF detection prediction result is obtained, the UE will transition to RRC_Idle due to the expiration of timer (T311), which is set to value 0. By transitioning to RRC_Idle early, a new cell search can be started, and a suitable cell for communication can be found.

[0168] Other methods using AI / ML for RLF processing are disclosed. The UE uses an AI / ML model to predict when the communication quality falls below a predetermined value, for example, when CQI < Qout is detected. The UE may also use an AI / ML model to predict when the communication quality exceeds a predetermined value, for example, when CQI > Qin is detected. Hereafter, these processes may be referred to as "communication quality detection prediction processing". Figure 15 is a schematic diagram of the case where CQI < Qout detection and CQI > Qin detection are predicted using an AI / ML model. The UE may predict the time and location in which CQI < Qout detection and CQI > Qin detection will occur. The prediction results are output from the AI / ML model. For example, the prediction results may include at least one of the following (f1) to (f4). (f1) CQI < Qout is detected at time tx. (f2) When a UE is located at position Lx, CQI < Qout is detected. (f3) At time tx, CQI < Qin is detected. (f4) When a UE is located at position Lx, CQI < Qin is detected. In this way, it becomes possible to predict when and where CQI < Qout and CQI > Qin will be detected, and it can be applied to a wide variety of applications.

[0169] Regarding the input information for the AI / ML model, it is advisable to appropriately apply the input information disclosed in the process for predicting RLF detection. This makes it possible to directly detect CQI<Qout> and CQI>Qin, and along with this detection, it becomes possible to predict their time and location.

[0170] Regarding the timing of the execution of the communication quality detection and prediction process, the method disclosed in the RLF detection and prediction process should be applied as appropriate. Similarly, the method disclosed in the RLF detection and prediction process should be applied as appropriate for resetting the communication quality detection and prediction process. The same effect can be obtained.

[0171] This document discloses the processing after communication quality detection prediction by the communication quality detection prediction process. The UE may notify the NW node of information regarding the communication quality detection prediction. The information regarding the communication quality detection prediction may include the communication quality detection prediction result and information predicted together with the communication quality detection prediction result. The NW node receives the information regarding the communication quality detection prediction from the UE. The NW node becomes aware of the communication quality detection prediction result at the UE.

[0172] A recovery process may be introduced. Regarding this recovery process, the method disclosed in the RLF detection and prediction process may be applied as appropriate. The NW node derives a recovery method. The recovery method may, for example, be a method to improve communication quality. The recovery method may include, for example, at least one of beam modification and increasing the maximum number of HARQ cycles. By performing such processing, it is possible to improve the communication quality between the UE and the NW.

[0173] The UE may perform at least one of the RLF processing and the RLM processing in parallel with the recovery processing. For example, while the UE is performing the recovery processing, the UE may perform at least one of the following (g1) to (g4): (g1) The UE monitors whether CQI < Qout. (g2) The UE monitors whether CQI > Qin. (g3) The UE monitors whether CQI < Qout is detected N310 times consecutively. (g4) The UE monitors whether CQI > Qin is detected N311 times consecutively. For example, the UE may start processing after RLF detection by PLP detection via RLF processing. In this case, the UE may start a timer (T310). After predicting CQI < Qout detection, recovery processing can be performed until actual PLP detection occurs. Therefore, communication quality can be improved and RLF detection can be suppressed.

[0174] Figure 16 is a diagram illustrating an example of communication quality detection and prediction processing. Figure 16 shows an example using recovery processing. The UE starts the communication quality detection and prediction processing at a predetermined timing, for example. The UE inputs the input information disclosed above into the AI / ML model to perform communication quality detection and prediction. The UE may also predict communication quality recovery. The UE may perform RLF processing in parallel with the communication quality detection and prediction processing. The UE predicts CQI < Qout detection through the communication quality detection and prediction processing. The UE, having predicted CQI < Qout detection, sends information regarding the CQI < Qout detection prediction to the NW node. The NW node, upon receiving this information, performs recovery processing. The NW node may notify the UE of settings based on the recovery method (for example, the recovery setting information described above). The UE may perform recovery processing based on these settings.

[0175] If a PLP is detected by RLF processing, the UE starts a timer (T310) and continues RLF processing. If CQI > Qin is not detected for N311 consecutive times before timer (T310) expires, the UE detects RLF and starts a timer (T311) to begin RRC connection re-establishment processing. If the RRC connection re-establishment processing is not successful before timer (T311) expires, the UE transitions to RRC_Idle. During the recovery process, for example, if the communication quality exceeds a predetermined value (Qin) or if the communication quality does not fall below a predetermined value (Qout) for N310 consecutive times, the UE may reset the communication quality detection prediction process and perform the communication quality detection prediction process again. In this way, the UE can perform the communication quality detection prediction process. Performing communication quality detection prediction allows for earlier execution of the recovery process and can suppress actual PLP detection and RLF detection.

[0176] After outputting the prediction results for communication quality detection, the UE may continue the communication quality detection prediction process. The UE may use an AI / ML model to predict, for example, CQI < Qout detection and CQI > Qin detection. The UE may also monitor whether the CQI < Qout detection prediction occurs N310 times consecutively. The UE may also monitor whether the CQI > Qin detection prediction occurs N311 times consecutively. If the CQI < Qout detection prediction occurs N310 times consecutively, the UE may determine that a PLP has been detected. If the CQI > Qin detection prediction occurs N311 times consecutively, the UE may reset the communication quality detection prediction process and perform the communication quality detection prediction process again. In this way, PLP detection can be predicted using the communication quality detection prediction process.

[0177] The UE may use an AI / ML model to predict whether the CQI < Qout detection prediction will occur N310 times consecutively. The UE may also predict whether the CQI > Qin detection prediction will occur N311 times consecutively. If the UE predicts that the CQI < Qout detection prediction will occur N310 times consecutively, it may determine that a PLP has been detected. If the UE predicts that the CQI > Qin detection prediction will occur N311 times consecutively, it may reset the communication quality detection prediction process and perform the communication quality detection prediction process again. In this way, it becomes possible to predict PLP detection using the communication quality detection prediction process.

[0178] While it is disclosed that PLP detection is determined by monitoring or predicting that the CQI < Qout detection prediction occurs N310 times consecutively, other methods may be applied. For example, the UE may determine PLP detection not by the condition that the CQI < Qout detection prediction occurs N310 times consecutively, but by the condition that the CQI < Qout detection prediction continues for a predetermined period of time. While it is disclosed that the reset of the communication quality detection prediction process is determined by monitoring or predicting that the CQI > Qin detection prediction occurs N311 times consecutively, other methods may be applied. The UE may determine the reset of the communication quality detection prediction process not by the condition that the CQI > Qin detection prediction occurs N311 times consecutively, but by the condition that the CQI > Qin detection prediction continues for a predetermined period of time. This allows for time setting and flexible control.

[0179] A recovery process may be provided. The UE may start the recovery process based on PLP detection prediction. The UE may terminate the recovery process based on PLP detection. The UE may start a timer (T310) when the recovery process is completed.

[0180] Figure 17 is a diagram illustrating another example of communication quality detection prediction processing. Figure 17 shows an example of PLP detection prediction. The UE predicts CQI < Qout detection through communication quality detection prediction processing. The UE, having predicted CQI < Qout detection, continues the communication quality detection prediction processing. The UE may determine that PLP has been detected if the CQI < Qout detection prediction occurs N310 times consecutively. That is, the UE may determine that PLP detection has been predicted. The UE may also determine that PLP has been detected if it is predicted that the CQI < Qout detection prediction will occur N310 times consecutively. When PLP detection is predicted in this way, the UE transmits information regarding the PLP detection prediction to the NW node. The NW node that receives this information performs recovery processing. The NW node may notify the UE of settings based on the recovery method (for example, the recovery setting information described above). The UE may perform recovery processing based on these settings. Regarding the notification method, it is advisable to apply the method disclosed in the aforementioned RLF detection and prediction process as appropriate.

[0181] If the recovery process detects CQI > Qin, or if CQI < Qout is not detected for N310 consecutive times, the UE may reset the communication quality detection and prediction process and perform the communication quality detection and prediction process again. The UE may also perform the RLF process in parallel with the recovery process. If the recovery process does not reset the communication quality detection and prediction process, and the RLF process detects PLP, the UE may start the timer (T310).

[0182] A recovery timer (TX2) may be provided. When PLP detection is predicted, the UE starts timer (TX2). If the timer (TX2) expires without the communication quality detection prediction process being reset by the recovery process, the UE may determine that PLP has been detected and start timer (T310). In this case, the UE does not need to execute RLF processing in parallel.

[0183] The UE may notify the NW nodes of at least one of the following (h1) to (h3): (h1) CQI>Qin was detected N311 times consecutively. (h2) CQI>Qin was predicted to occur N311 times consecutively. (h3) The prediction of CQI>Qin detection occurred N311 times consecutively. While the recovery process is running, the UE may notify the NW nodes of at least one of the following (h1) to (h3). The NW nodes may change the UE's settings in response to the notification from the UE. The NW nodes may also notify the UE of the changed settings. Regarding the notification method, it is advisable to apply a method similar to the method for notifying the recovery setting information described above as appropriate.

[0184] The value of Timer (TX2) may be 0. If Timer (TX2) is 0, there will be no recovery processing time. Setting Timer (TX2) to 0 makes it possible to set no recovery processing time.

[0185] Timer (TX2) may be set to any value. For example, Timer (TX2) may be set to be longer than Timer (TX1). For example, Timer (TX2) may be set to be longer than Timer (T310). By using Timer (TX2), measures to improve communication quality through recovery can be implemented, and an improvement in communication quality can be expected. By setting the time of Timer (TX2) to be longer, the possibility of avoiding actual PLP detection or RLF detection can be increased.

[0186] Different timer (T310) values ​​may be set depending on the recovery processing time. For example, if the value of timer (TX2) is set to be longer than 0, the value of timer (T310) may be set to be shorter than the value of timer (T310) when there is no recovery processing time. This can avoid a significant delay in the timing of RLF determination.

[0187] This approach enables PLP detection prediction and allows for early recovery processing. It also suppresses actual PLP and RLF detection.

[0188] Another method using AI / ML for RLF processing is disclosed. The UE predicts PLP detection using an AI / ML model. Figure 18 is a schematic diagram of the case where PLP detection is predicted using an AI / ML model. After the communication quality detection prediction process, the UE predicts PLP detection using an AI / ML model. The communication quality detection prediction process may use the method disclosed in Figure 15. The UE may predict at least one of CQI < Qout detection and CQI > Qin detection through the communication quality detection prediction process. The UE may use the prediction result of communication quality detection as input information for the AI / ML model for PLP detection prediction. In this way, the UE predicts PLP detection using the prediction result of communication quality detection. The information disclosed in the RLF detection prediction process may be appropriately applied as other input information to the AI / ML model for PLP detection prediction. In this way, the PLP detection prediction process can be executed.

[0189] Regarding the execution timing of the PLP detection prediction process, the method disclosed for the RLF detection prediction process may be applied as appropriate. Similar effects can be obtained. In addition, the UE may start the PLP detection prediction process upon derivation of the prediction result for communication quality detection. By doing so, the PLP detection prediction process can be avoided if the prediction result for communication quality detection is not derived, thereby improving processing efficiency. Similarly, the method disclosed for the RLF detection prediction process may be applied as appropriate for resetting the PLP detection prediction process. Similar effects can be obtained. In addition, the UE may start the communication quality detection prediction process upon resetting the PLP detection prediction process, or it may start from the PLP detection prediction process. The UE may be configured to choose whether to start with the communication quality detection prediction process or the PLP detection prediction process. This will improve processing efficiency.

[0190] This document discloses the processing after PLP detection prediction by the PLP detection prediction process. The UE may notify the NW node of information regarding the PLP detection prediction. The information regarding the PLP detection prediction may include the PLP detection prediction result and information predicted together with the PLP detection prediction result. The NW node receives the information regarding the PLP detection prediction from the UE. The NW node becomes aware of the PLP detection prediction result at the UE.

[0191] Recovery processing may be introduced. The NW node derives a recovery method. For example, the NW node may derive a recovery method based on the PLP detection prediction result received from the UE. The recovery method may be, for example, a method to improve communication quality. The recovery method may include, for example, at least one of beam modification and increasing the maximum number of HARQs. By performing such processing, it is possible to improve the communication quality between the UE and the NW.

[0192] Regarding the recovery method described above, the recovery methods disclosed in the RLF detection and prediction processing and communication quality detection and prediction processing described above may be applied as appropriate. The recovery method may also be a cell modification process. The recovery method may also be a PCell modification process. The recovery method may also be a PSCell modification process. For example, an NW node may initiate a PCell modification process on the UE. The recovery method may also be an HO process. For example, an NW node may initiate an HO process on the UE. By performing such processing, the communication quality between the UE and the NW can be improved.

[0193] After outputting the PLP detection prediction result, the UE may continue with the PLP detection prediction process. After outputting the PLP detection prediction result, the UE may continue with the communication quality detection prediction process. The UE may continue with both the communication quality detection prediction process and the PLP detection prediction process. For these processes, the method disclosed in the aforementioned communication quality detection prediction process may be applied as appropriate. PLP detection can be predicted using the communication quality detection prediction process.

[0194] The UE may start the timer (T310) based on the PLP detection prediction. The UE may monitor whether the state in which PLP detection is predicted will continue until the timer (T310) expires. The UE may monitor whether the CQI>Qin detection prediction occurs N311 times consecutively. The UE may reset the timer (T310) if the CQI>Qin detection prediction occurs N311 times consecutively. The UE may predict whether the state in which PLP detection is predicted will continue until the timer (T310) expires. The UE may predict whether the CQI>Qin detection prediction will occur N311 times consecutively. The UE may reset the timer (T310) if it is predicted that the CQI>Qin detection prediction will occur N311 times consecutively. In this case, the UE may perform the communication quality detection prediction process again, or the PLP detection prediction process. The UE may continue to perform communication quality detection prediction processing or PLP detection prediction processing. If the state in which PLP detection is predicted continues until the timer (T310) expires, the UE may determine that RLF has been detected. For processing after RLF, the methods disclosed above may be applied as appropriate. In this way, the timer (T310) can be started early, and recovery processing can be executed early. Communication quality can be improved early.

[0195] Figure 19 is a diagram illustrating an example of PLP detection prediction processing. Figure 19 shows an example of PLP detection prediction using communication quality detection prediction processing. The UE predicts CQI < Qout detection using communication quality detection prediction processing. The UE, having predicted CQI < Qout detection, continues the communication quality detection prediction processing. The UE may determine that a PLP has been detected if it is predicted that CQI < Qout detection will occur N310 times consecutively. In other words, the UE may determine that PLP detection has been predicted. When PLP detection is predicted in this way, the UE may send information regarding the PLP detection prediction to the NW node. The UE, having predicted PLP detection, starts a timer (T310).

[0196] After predicting CQI < Qout detection through communication quality detection prediction processing, the UE may determine whether the CQI < Qout detection prediction occurs N310 times consecutively. If the CQI < Qout detection prediction occurs N310 times consecutively, the UE may determine that a PLP has been detected. The UE, having determined that a PLP has been detected, may transmit information regarding the PLP detection prediction to the NW node. The UE, having determined that a PLP has been detected, starts a timer (T310).

[0197] If the CQI>Qin detection prediction occurs N311 times consecutively before timer (T310) expires, the UE may reset timer (T310). The UE then performs the PLP detection prediction process again, for example. If the state in which PLP detection is predicted continues until timer (T310) expires, the UE may determine that RLF has been detected. The UE starts timer (T311) and performs the RRC connection re-establishment process triggered by RLF. If the RRC connection re-establishment process is unsuccessful before timer (T311) expires, the UE transitions to the RRC_Idle state. In this way, timers (T310 and T311) can be started early, and the RRC connection re-establishment process due to RLF detection can be initiated early.

[0198] Another method using AI / ML for RLF processing is disclosed. The UE directly predicts PLP detection using an AI / ML model. Figure 20 is a schematic diagram of the method for directly predicting PLP detection using an AI / ML model. Regarding input information, the input information disclosed in the aforementioned communication quality detection prediction process and PLP detection prediction process may be applied as appropriate. In this way, the PLP detection prediction process can be executed. The UE may also notify the NW node of information regarding PLP detection prediction.

[0199] Regarding the execution timing of the PLP detection prediction process using this method, the execution timing of the PLP detection prediction process disclosed above should be applied as appropriate. For example, the PLP detection prediction process may be started by the first CQI < Qout detection by the RLF process. This can improve the accuracy of PLP detection prediction.

[0200] Regarding the processing after PLP detection and prediction, the processing disclosed above should be applied as appropriate. Similar effects can be obtained.

[0201] Recovery processing may be introduced. The NW node derives a recovery method. For example, the NW node may derive a recovery method based on the PLP detection prediction result received from the UE. The recovery method may be, for example, a method to improve communication quality. The recovery method may include, for example, at least one of beam modification and increasing the maximum number of HARQs. By performing such processing, it is possible to improve the communication quality between the UE and the NW.

[0202] Regarding the recovery method described above, the recovery methods disclosed in the RLF detection and prediction processing and communication quality detection and prediction processing described above may be applied as appropriate. The recovery method may also be a cell modification process. The recovery method may also be a PCell modification process. The recovery method may also be a PSCell modification process. For example, an NW node may initiate a PCell modification process on the UE. The recovery method may also be an HO process. For example, an NW node may initiate an HO process on the UE. By performing such processing, the communication quality between the UE and the NW can be improved.

[0203] Figure 21 is a diagram illustrating another example of PLP detection prediction processing. Figure 21 shows an example of direct PLP detection prediction. The UE monitors whether CQI < Qout and whether CQI > Qin using RLM processing. If the UE detects CQI < Qout, it performs PLP detection prediction processing. The UE predicts PLP detection through PLP detection prediction processing. If PLP detection is predicted, the UE may send information about the PLP detection prediction to the NW node.

[0204] Recovery processing may be performed. A network node that has received information regarding PLP detection prediction performs recovery processing. The network node derives a recovery method based on the information regarding PLP detection prediction. The network node may notify the UE of the settings based on the recovery method (for example, the recovery setting information described above). The UE may perform recovery processing based on these settings. Regarding the notification method, the method disclosed in the RLF detection prediction processing described above may be applied as appropriate.

[0205] If the recovery process detects CQI > Qin, or if CQI < Qout is not detected for N310 consecutive times, the UE may reset the PLP detection prediction process and perform the PLP detection prediction process again. The UE may also execute the RLF process in parallel with the recovery process. If the PLP detection prediction process is not reset by the recovery process and a PLP is detected by the RLF process, the UE may start the timer (T310).

[0206] This approach enables PLP detection prediction and allows for early recovery processing. It also suppresses actual PLP and RLF detection.

[0207] After predicting PLP detection through the PLP detection prediction process, the UE may determine whether the predicted state of PLP detection will continue until the timer (T310) expires. Figure 22 is a schematic diagram of the RLF detection prediction process using an AI / ML model. Figure 22 shows a method for determining whether the predicted state of PLP detection will continue until the timer (T310) expires after the PLP detection prediction process. Regarding the input information for the PLP detection prediction process, the input information disclosed above may be applied as appropriate. The UE uses the results of the PLP detection prediction process to determine whether the prediction result will continue until the timer (T310) expires. The input information for this determination may include, for example, determination conditions such as timer (T310), N310, and N311. In this way, the UE may use the results of the PLP detection prediction process and the determination conditions to determine or predict whether an RLF will be detected. By doing so, it becomes possible to execute the RLF detection prediction process using the results of the PLP detection prediction process.

[0208] Figure 23 is a diagram illustrating another example of PLP detection prediction processing. Figure 23 is a diagram illustrating an example of processing in which, after predicting PLP detection by PLP detection prediction processing, it is determined whether the predicted state of PLP detection will continue until the timer (T310) expires. The UE monitors whether CQI < Qout and whether CQI > Qin by RLM processing. If the UE detects CQI < Qout, it performs PLP detection prediction processing. The UE predicts PLP detection by PLP detection prediction processing. If PLP detection is predicted, the UE may send information regarding the PLP detection prediction to the NW node. If PLP detection is predicted, the UE starts the timer (T310) and determines whether the predicted state of PLP detection will continue until the timer (T310) expires. If the predicted state of PLP detection continues until the timer (T310) expires, the UE may determine that RLF has been detected. This enables RLF detection prediction. Upon RLF detection, the UE may start a timer (T311) and perform RRC connection re-establishment processing. If the PLP detection prediction state does not continue until the timer (T310) expires, the UE may reset the timer (T310). The UE may reset the PLP detection prediction processing. The UE may start the PLP detection prediction processing again.

[0209] After predicting PLP detection through the PLP detection prediction process, the UE may determine whether CQI > Qin occurs N311 times consecutively before the timer (T310) expires. If CQI > Qin does not occur N311 times consecutively before the timer (T310) expires, the UE may determine that RLF has been detected. If CQI > Qin occurs N311 times consecutively before the timer (T310) expires, the UE may reset the timer (T310). The UE may also reset the PLP detection prediction process. The UE may then restart the PLP detection prediction process.

[0210] By doing so, the timers (T310 and T311) can be started early, and the process of re-establishing the RRC connection due to RLF detection can be started early.

[0211] Another method using AI / ML for RLF processing is disclosed. The UE predicts the expiration of the timer (T310) using an AI / ML model. Figure 24 is a schematic diagram of the method for predicting T310 expiration using an AI / ML model. The UE directly predicts the expiration of the timer (T310) using an AI / ML model. Regarding input information, the input information disclosed in the RLF detection and prediction process described above may be applied as appropriate. In this way, the timer (T310) expiration prediction process can be executed. Hereafter, this prediction process may be referred to as the "T310 expiration prediction process".

[0212] Regarding the timing of the execution of the T310 expiration prediction process, the method disclosed in the RLF detection prediction process may be applied as appropriate. Similar effects can be obtained. In addition, the T310 expiration prediction process may be executed after the PLP detection prediction result is derived. By doing so, the T310 expiration prediction process can be avoided if the PLP detection prediction result is not derived, thereby improving processing efficiency. Similarly, the method disclosed in the RLF detection prediction process may be applied as appropriate for resetting the T310 expiration prediction process. Similar effects can be obtained. In addition, the UE may restart the T310 expiration prediction process after resetting it. This will improve processing efficiency.

[0213] This document discloses the processing after predicting the expiration of timer (T310) by the T310 expiration prediction processing. When the UE predicts the expiration of timer (T310), it may determine that an RLF has been detected. The processing after predicting the T310 expiration may be the same as the processing after detecting an RLF. After predicting the expiration of timer (T310), the UE performs RRC connection re-establishment processing. The UE may send an RRC connection re-establishment request. The NW node that receives the request re-establishes the RRC connection with the UE. After predicting the expiration of timer (T310), the UE may start timer (T311). The UE performs RRC connection re-establishment processing while timer (T311) is running. If the RRC connection re-establishment processing is successful before timer (T311) expires, the UE resets timer (T311). If the RRC connection re-establishment process is not successful before the timer (T311) expires, the UE transitions to RRC_Idle. By determining that RLF has been detected when the timer (T310) is expected to expire, the RRC connection re-establishment process can be executed earlier. This improves the likelihood of establishing an RRC connection with a more suitable cell.

[0214] Figure 25 is a diagram illustrating an example of T310 expiration prediction processing. Figure 25 shows an example of directly performing T310 expiration prediction. The UE monitors whether CQI < Qout and whether CQI > Qin using RLF processing. If the UE detects CQI < Qout N310 times consecutively, it detects a PLP. Upon detecting the PLP, the UE starts the T310 expiration prediction processing. If the UE predicts that CQI > Qin will not be detected N311 times consecutively before the timer (T310) expires, it outputs the timer (T310) expiration prediction.

[0215] If timer (T310) is expected to expire, the UE may determine that RLF has been detected. The UE starts timer (T311) and performs the RRC connection re-establishment process. The UE performs the RRC connection re-establishment process while timer (T311) is running. If the RRC connection re-establishment process is successful before timer (T311) expires, the UE resets timer (T311). If the RRC connection re-establishment process is unsuccessful before timer (T311) expires, the UE transitions to RRC_Idle.

[0216] If the UE predicts that CQI > Qin detection will occur N311 times consecutively before the timer (T310) expires, the UE may reset the T310 expiration prediction process. The UE may, for example, perform the T310 expiration prediction process again. The T310 expiration prediction can then be performed again.

[0217] After predicting the expiration of the timer (T310), the UE may notify the NW nodes of information regarding the T310 expiration prediction. The information regarding the T310 expiration prediction may include the T310 expiration prediction result and information predicted along with the T310 expiration prediction result. The NW nodes receive the information regarding the T310 expiration prediction from the UE. The NW nodes then become aware of the T310 expiration prediction result at the UE.

[0218] A recovery process may be introduced. The NW node derives a recovery method. For example, the NW node may derive a recovery method based on the T310 expiration prediction result received from the UE. The recovery method may be, for example, a method to improve communication quality. The recovery method may include, for example, beam modification and increasing the maximum number of HARQs. By performing such a process, it is possible to improve the communication quality between the UE and the NW. The recovery method may also be a cell modification process. The recovery method may also be a PCell modification process. The recovery method may also be a PSCell modification process. For example, the NW node may initiate a PCell modification process for the UE. The recovery method may also be an HO process. For example, the NW node may initiate an HO process for the UE. By performing such a process, it is possible to improve the communication quality between the UE and the NW.

[0219] If the recovery process detects CQI > Qin, or if CQI > Qin is detected N311 times consecutively, or if CQI < Qout is not detected N310 times consecutively, the UE may reset the T310 expiration prediction process and perform the T310 expiration prediction process again.

[0220] A recovery timer (TX3) may be provided. When the expiration of timer (T310) is predicted, the UE starts timer (TX3). If the communication quality is not restored by the recovery process, the UE may determine that RLF has been detected. For example, if the timer (TX3) expires without the T310 expiration prediction process being reset, the UE may determine that RLF has been detected. The UE may start timer (T311) upon RLF detection. While timer (T311) is running, the UE performs RRC connection re-establishment processing. If the RRC connection re-establishment processing is successful before timer (T311) expires, the UE resets timer (T311). If the RRC connection re-establishment processing is unsuccessful before timer (T311) expires, the UE transitions to RRC_Idle.

[0221] The value of Timer (TX3) may be 0. If Timer (TX3) is 0, there will be no recovery processing time. Setting Timer (TX3) to 0 makes it possible to set no recovery processing time.

[0222] The timer (TX3) may be set to any value. For example, the timer (TX3) may be set to be shorter than the timer (T310). If the communication quality does not improve due to recovery based on early prediction of timer (T310) expiration, the system can transition to RLF processing earlier. The RRC re-establishment process can be started earlier. The possibility of connecting to a more suitable cell can be improved earlier.

[0223] Different timer (T310) values ​​may be set depending on the recovery processing time. For example, if the value of timer (TX3) is set to be longer than 0, the value of timer (T310) may be set to be shorter than the value of timer (T310) when there is no recovery processing time. This can avoid a significant delay in the timing of RLF determination.

[0224] Unlike the actual timer (T310) failure due to RLF processing, in the prediction of timer (T310) failure, the UE is not necessarily in an RLF state. The UE may still be able to communicate with the NW node. Performing recovery processing at this stage improves the communication quality between the UE and the NW and suppresses actual RLF detection.

[0225] Figure 26 is a diagram illustrating another example of the T310 expiration prediction process. Figure 26 shows an example of directly performing T310 expiration prediction. Figure 26 shows an example in which recovery processing is introduced and a recovery timer (TX3) is provided. When the expiration of timer (T310) is predicted, the UE starts timer (TX3). The UE, having predicted the expiration of timer (T310), sends information regarding the T310 expiration prediction to the NW node. The NW node, having received this information, executes recovery processing. The NW node derives a recovery method. The NW node may notify the UE of settings based on the recovery method (for example, the recovery setting information described above). The UE may perform recovery processing based on these settings. Regarding the notification method, the method disclosed in the RLF detection and prediction processing described above may be applied as appropriate.

[0226] If the recovery process detects CQI > Qin, or if CQI < Qout is not detected for N310 consecutive times, the UE may reset the T310 expiration prediction process and perform the T310 expiration prediction process again. The UE may also execute the RLF process in parallel. If a PLP is detected by the RLF process after resetting the T310 expiration prediction process, the UE may start the T310 expiration prediction process again.

[0227] If the recovery process fails to restore communication quality and the timer (TX3) expires, the UE may determine that an RLF (Rapid Failure) has been detected. Upon detection of the RLF, the UE starts the timer (T311) and performs RRC connection re-establishment processing while the timer (T311) is running. If the RRC connection re-establishment processing is successful before the timer (T311) expires, the UE resets the timer (T311). If the RRC connection re-establishment processing is unsuccessful before the timer (T311) expires, the UE transitions to RRC_Idle.

[0228] This approach enables T310 expiration prediction and allows for earlier recovery processing. It also suppresses actual RLF detection.

[0229] The PLP detection prediction process may also be used in the T310 expiration prediction process. After the PLP detection prediction process, the UE predicts the expiration of the timer (T310) using an AI / ML model. The result of the PLP detection prediction process may be used as input information for the T310 expiration prediction process. The PLP detection prediction process may, for example, be a method that directly predicts PLP detection. Alternatively, the PLP detection prediction process may, for example, be a method that predicts PLP detection using the prediction result of CQI < Qout detection after the communication quality detection prediction process.

[0230] Figure 27 is another schematic diagram of predicting T310 expiration using an AI / ML model. After PLP detection and prediction processing, the UE predicts the expiration of the timer (T310) using an AI / ML model. The UE performs PLP detection and prediction using the method disclosed in Figure 20, and uses the prediction results to predict the expiration of the timer (T310). In this way, for example, when the T310 expiration prediction processing is reset, the UE can restart only the T310 expiration prediction processing part again. Therefore, the UE can restart from the appropriate processing.

[0231] Figure 28 is another schematic diagram of predicting T310 expiration using an AI / ML model. The UE performs communication quality detection and prediction processing using the method disclosed in Figure 18, and uses the results to perform PLP detection prediction. The UE further uses the results of the PLP detection prediction to predict timer (T310) expiration. In this way, the UE can restart from a more appropriate process depending on where and how it was reset. Therefore, efficient processing becomes possible.

[0232] Figure 29 is a diagram illustrating an example of T310 expiration prediction processing. Figure 29 shows an example of T310 expiration prediction processing using PLP detection prediction processing. The UE executes PLP detection prediction processing upon CQI < Qout detection. If PLP detection is predicted by the PLP detection prediction processing, the UE performs T310 expiration prediction processing. If timer (T310) expiration is predicted, the UE sends information regarding T310 expiration prediction to the NW node. Upon receiving this information, the NW node performs recovery processing. Recovery processing is executed based on T310 expiration prediction.

[0233] The UE may perform RLF processing in parallel with the PLP detection prediction processing and the T310 expiration prediction processing. If the communication quality is not restored by the recovery processing and RLF is detected without the recovery processing being reset, the UE starts the timer (T311) and performs the RRC reconnection establishment processing.

[0234] By predicting the timer (T310) expiration early in this way, recovery processing can be performed before RLF is actually detected. This allows for improvement of communication quality before RLF is detected, thereby suppressing actual RLF detection.

[0235] Figure 30 is a diagram illustrating another example of the T310 expiration prediction process. Figure 30 shows an example where the T310 expiration prediction process is performed using the PLP detection prediction process and a recovery timer (TX3) is provided. When the expiration of the timer (T310) is predicted, the UE transmits information regarding the T310 expiration prediction to the NW node. Upon receiving this information, the NW node performs the recovery process. The recovery process is executed based on the T310 expiration prediction. If the communication quality is not restored by the recovery process and the timer (TX3) expires without the recovery process being reset, the UE determines that an RLF has been detected. The UE starts the timer (T311) and performs the RRC reconnection establishment process.

[0236] This approach allows recovery processing to be performed before RLF is actually detected. In addition, if communication quality does not recover, processing can be moved to the post-RLF detection stage earlier. The UE can then perform the RRC connection establishment process with a more suitable cell earlier, enabling communication sooner.

[0237] A new timer may be provided in place of timer (T310). A different timer may be provided in addition to timer (T310). This timer may be used for T310 expiration prediction processing, or for other prediction processing. In addition to RLF processing, timers can be set individually for each prediction processing.

[0238] Figure 31 shows an example of prediction processing when a new timer (TX4) is installed in place of timer (T310). Hereafter, the process of predicting the expiration of timer (TX4) using an AI / ML model may be referred to as "TX4 expiration prediction processing". Timer (TX4) is set to a shorter time than timer (T310). After predicting PLP detection, if timer (TX4) remains unreset and expires, the UE determines that RLF has been detected. In this way, RLF detection can be performed early by the TX4 expiration prediction processing. The UE can then perform the RRC connection re-establishment process to a more suitable cell earlier.

[0239] Figure 32 shows another example of the TX4 expiration prediction process when a new timer (TX4) is provided in place of timer (T310). Timer (TX4) is set to a shorter time than timer (T310). After deriving the prediction result for PLP detection, the UE inputs this result as one of the input pieces to the AI / ML model for the TX4 expiration prediction process. The UE executes the TX4 expiration prediction process and predicts the expiration of timer (TX4). If the expiration of timer (TX4) is predicted, the UE sends information regarding the TX4 expiration prediction to the NW node. The NW node that receives this information performs recovery processing. When the expiration of timer (TX4) is predicted, the UE starts a recovery timer (TX3). If the communication quality does not recover through the recovery process and the timer (TX3) expires without the recovery process being reset, the UE determines that RLF has been detected. The UE starts a timer (T311) and performs the RRC reconnection establishment process. This allows recovery processing to be performed before RLF is actually detected. Furthermore, if communication quality does not recover, it can transition to post-RLF detection processing earlier. The UE can then perform RRC connection establishment processing with a more suitable cell earlier, enabling communication to resume sooner.

[0240] The timer (TX4) may be set to any value. For example, the timer (TX4) may be set to be shorter than the timer (T310). If the communication quality does not improve through recovery based on early PLP detection prediction, the system can transition to RLF processing earlier. The RRC re-establishment process can be started earlier. The possibility of connecting to a more suitable cell can be improved earlier.

[0241] The UE may determine that an RLF has been detected based on the PLP detection prediction. The value of the timer (TX4) may be 0. For example, the UE may set the timer (TX4) to 0, thereby executing a process that determines that an RLF has been detected based on the PLP detection prediction. The UE can start the RRC re-establishment process earlier. The possibility of connecting to a more suitable cell earlier can be improved.

[0242] Unlike actual timer (T310) failure due to RLF processing, in the prediction of timer (TX4) failure, the UE is not necessarily in an RLF state. The UE may still be able to communicate with the NW. Performing recovery processing at this stage improves the communication quality between the UE and the NW and suppresses actual RLF detection.

[0243] A new counter (NX1) may be provided instead of N310. A different counter may be provided in addition to N310. This counter may be used for prediction processing using the AI / ML model disclosed in this embodiment. This counter may be used as input information for prediction processing using the AI / ML model disclosed in this embodiment. This counter may be used, for example, for PLP detection prediction processing, or for other prediction processing. For example, the UE may determine that a PLP has been detected if, after predicting CQI < Qout detection, it is predicted that CQI < Qout detection prediction will occur NX1 times consecutively. Counters can be set individually for each prediction processing, separate from the RLF processing.

[0244] The counter (NX1) may be set to any value. For example, the counter (NX1) may be set to a value greater than N310. By setting the counter (NX1) to a larger value, for example, PLP detection prediction can be made more reliable, and the prediction probability by the prediction process can be improved.

[0245] A new counter (NX2) may be provided instead of N311. A different counter may be provided in addition to N311. This counter may be used for prediction processing using the AI / ML model disclosed in this embodiment. This counter may be used as input information for prediction processing using the AI / ML model disclosed in this embodiment. This counter may be used as a criterion for determining good communication quality in the prediction processing. For example, after predicting CQI < Qout detection, the UE starts a timer (T310) upon PLP detection. In this situation, if CQI > Qin detection prediction occurs NX2 times consecutively before the timer (T310) expires, the UE may reset the timer (T310). For example, if CQI > Qin detection prediction occurs NX2 times consecutively in the recovery process, the UE may reset the recovery process. For example, if CQI > Qin detection prediction occurs NX2 times consecutively in the prediction process, the UE may reset the prediction process. In addition to RLF processing, it becomes possible to set counters for each individual prediction process.

[0246] The counter (NX2) may be set to any value. For example, the counter (NX2) may be set to a value greater than N311. By setting the counter (NX2) to a larger value, the UE can reset the timer, or the recovery process or prediction process, while the communication quality is better.

[0247] The prediction process using the AI / ML model disclosed in this embodiment may be performed in parallel with the RLF process. The prediction process may be performed independently of the RLF process, or it may be performed in correlation with the RLF process. The prediction process may be initiated in correlation with the RLF process. For example, the RLF detection prediction process or the T310 expiration prediction process may be initiated by the actual detection of a PLP. For example, the UE may perform a recovery process after predicting PLP detection, and then stop the recovery process and start the timer (T310) for the RLF process upon actual PLP detection. The prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of the prediction process.

[0248] In the prediction process using the AI / ML model disclosed in this embodiment, a probability of the prediction result (hereinafter, "prediction probability") may be output. If the prediction probability in each prediction process falls below a predetermined value, the UE may consider (or determine) that no information regarding the prediction (e.g., the prediction result) was detected. If the prediction probability in each prediction process falls below a predetermined value, the UE may discard the information regarding the prediction. The UE may continue to execute each prediction process. In this way, malfunctions when the prediction probability is low can be avoided.

[0249] The UE may send prediction probabilities along with prediction information to the NW nodes. If the prediction probability for each prediction process received from the UE falls below a predetermined value, the NW nodes may consider (or determine) that no prediction information (e.g., prediction results) was detected. If the prediction probability for each prediction process received from the UE falls below a predetermined value, the NW nodes may discard the prediction information. This prevents incorrect control of the UE that may occur when the prediction probability is low.

[0250] As disclosed above, in the prediction process, the UE may notify the NW nodes of information regarding the prediction. For example, in the RLF detection prediction process, the UE may notify the NW nodes of information regarding the RLF detection prediction. The prediction process makes it possible to predict RLF detection earlier than actual RLF detection. By notifying the NW nodes of this information from the UE, the NW nodes can also recognize that RLF detection has been predicted early. The recovery process can improve communication quality and suppress actual RLF detection. In addition, the UE can perform RRC connection re-establishment processing to a more suitable cell at an earlier stage. The UE may also send information regarding other prediction processes described above to the NE nodes, similar to the RLF detection prediction process.

[0251] The UE may notify the NW nodes of the detection results in the RLF processing. For example, the UE may notify the N node W that it has detected CQI>Qout. For example, the UE may notify the NW nodes that PLP has been detected. For example, the UE may notify the NW nodes that CQI>Qin has been detected N311 times consecutively. The NW nodes may use these notifications received from the UE in combination with prediction processing. This enables recovery processing that combines actual detection by RLF processing with prediction processing. Therefore, more flexible control becomes possible.

[0252] In this embodiment and other embodiments described later, the NW node may be, for example, a base station. The base station may be a CU or a DU. The base station may be an MN or an SN. Notification to the MN may be made via the SN. Notification to the SN may be made via the MN. The NW node may be, for example, an IAB node. The NW node may be an IAB-DU or an IAB-CU. The NW node may be an IAB parent node. The NW node may be, for example, a NWDAF. The NW node may be, for example, an AMF, or an SMF, or an UPF. The NW node may be an AS. Notification from the UE to the NW node may be made via a base station. Notification from the UE to the NW node may be made via a base station connected to the UE. Different NW nodes may be notified for each piece of information. Several pieces of information regarding predictions may be notified to the same NW node. The network node controlling each prediction process may differ. The same network node may control several prediction processes. The network node to which information is sent may be determined according to the prediction process. This allows for flexible control of prediction processes.

[0253] This document discloses a method for notifying base stations of the above-mentioned forecast information and information regarding each detection in RLF processing. Hereafter, some or all of the forecast information and information regarding each detection in RLF processing will be referred to as "RLF-related information." The UE may transmit the RLF-related information using RRC signaling. For example, the UE may transmit a measurement report that includes the RLF-related information. When the RLF-related information is included in the measurement report, there is a problem that the timing of notification of the forecast information may differ from the timing of transmission of the measurement report. Alternatively, for example, the UE may transmit a UEAssistanceInformation message that includes the RLF-related information. UEAssistanceInformation is provided so that network nodes can use the information from the UE for control purposes. By including the RLF-related information in the message, the UE can transmit the RLF-related information at the same time as the notification of the forecast information. Network nodes can then recognize that the information is for control purposes. For example, UE may transmit RLF-related information in the UE information procedure. For example, UE may transmit an RLF report containing RLF-related information. For example, UE may transmit an RLF report containing information regarding RLF detection predictions.

[0254] For example, a UE may send Failure information containing RLF-related information. For example, a UE may send MCG Failure information containing RLF-related information, or SCG Failure information containing RLF-related information. For example, a UE may send MCG Failure information containing information about RLF detection predictions in the MCG, and SCG Failure information containing information about RLF detection predictions in the SCG. The Failure information may provide information indicating whether or not it is an RLF detection prediction. Information indicating whether or not it is an RLF detection prediction may be included in the Failure information. The NW node will be able to recognize whether it is an actual RLF detection or an RLF detection prediction.

[0255] For example, a UE may transmit RLF-related information using fast MCG link recovery.

[0256] A new RRC message may be created for notifying RLF-related information. By differentiating it from the conventional RRC message, the UE can send forecast-related information in various situations, without being constrained by the past.

[0257] Other notification methods are disclosed. For example, the UE may transmit a UCI containing RLF-related information. For example, the UE may transmit a CQI or CRI (Channel Rank Indicator) containing RLF-related information. The UE may transmit information regarding predictions and information regarding each detection in RLF processing as CQI or CRI information. New indicators may be provided separately from CQI and CRI. PUCCH may be used to transmit UCI, CQI, CRI, or new indicators. Control becomes more efficient because the measurement results at the PHY layer are used for information regarding CQI < Qout detections and PLP detection predictions.

[0258] RLF-related information and the settings of the PUCCH used to transmit said information may be associated. The settings of the PUCCH used to transmit RLF-related information may be notified to the UE by the base station (or other NW node). RRC signaling may be used to notify said settings. The UE can use the PUCCH to notify the base station of RLF-related information early.

[0259] The UE may notify RLF-related information using random access processing. For example, the UE may transmit RLF-related information using PRACH. The UE can notify RLF-related information earlier. For example, the UE may transmit RLF-related information using message 3 (Msg3). The UE can notify more information. The random access procedure may be a contention-based random access procedure. The UE can notify RLF-related information at any time. The random access procedure may be a non-contention-based random access procedure. The UE can notify early and reliably.

[0260] For example, a UE may notify RLF-related information using two-stage random access processing. A UE may transmit RLF-related information using message A (MsgA). A UE may transmit RLF-related information using the PRACH of MsgA. A UE may transmit RLF-related information using the PUSCH of MsgA. The PRACH settings available for notifying RLF-related information may be notified to the UE by a base station (or other NW node). RRC signaling may be used to notify the UE of these settings. The PUSCH settings used for notifying RLF-related information may be notified to the UE by a base station (or other NW node). These settings may be included in the SIB and broadcast, or they may be included in RRC signaling and notified to the UE. Using the PUSCH of MsgA for notifying RLF-related information allows for the notification of more information earlier.

[0261] The PRACH or PUSCH settings used for notifying RLF-related information may be shared with other applications. This allows for flexible notifications. Alternatively, separate PRACH or PUSCH settings may be provided for notifying information related to predictions or information about each detection in RLF processing. This avoids conflicts with other applications.

[0262] The UE may send a beam failer recovery request that includes RLF-related information. Alternatively, the UE may notify of RLF-related information using the same processing as a beam failer recovery request. The UE can use PRACH to send RLF-related information and notify of RLF-related information at an early stage. For example, the UE may send SPCell's RLF-related information using PRACH and SCell's RLF-related information using PUCCH. This method is compatible with conventional communication methods and improves control efficiency.

[0263] Other notification methods are disclosed. For example, the UE may transmit MAC signaling containing RLF-related information. For example, the UE may transmit MAC CE containing RLF-related information. For example, the UE may transmit MAC PDU containing RLF-related information. The RA problem, which is one of the RLF detection factors, is handled at the MAC layer. Therefore, control efficiency can be improved by transmitting notification of prediction information and information on each detection in RLF processing via MAC CE or MAC PDU.

[0264] The MAC entity of the UE may count the detection predictions for CQI<Qout and CQI>Qin. The MAC entity of the UE may also perform timer processing in the prediction process. In this way, for example, the UE can transmit the detection predictions for CQI<Qout, CQI>Qin, and PLP detection predictions via MAC signaling. This can improve the efficiency of control.

[0265] The UE may notify the MN of RLF-related information of the MCG. The UE may notify the PCell of RLF-related information of the MCG. The MCG's RLF-related information may be notified between the MN and the SN. Inter-base station signaling, such as Xn signaling, may be used to notify the MCG's RLF-related information. The UE may notify the MN of the MCG's RLF-related information via the SN. The UE may notify the MN of the MCG's RLF-related information via the PCell. The MN may provide control to the UE regarding the prediction. The MN may notify the UE of the control results. The control results may be notified between the MN and the SN. Inter-base station signaling, such as Xn signaling, may be used to notify the control results. The MN may notify the UE of the control results via the SN. The control results may include, for example, information regarding recovery processing (e.g., the recovery setting information described above). The UE can obtain the control results from the MN.

[0266] If at least one of the communication quality prediction, PLP detection prediction, and RLF detection prediction is performed on the SCG, the UE may notify the SN of the SCG's RLF-related information. The UE may notify the PSCell of the SCG's RLF-related information. The SCG's RLF-related information may be notified between the MN and the SN. Inter-base station signaling, such as Xn signaling, may be used to notify the SCG's RLF-related information. The UE may notify the SN of the SCG's RLF-related information via the MN. The UE may notify the PSCell of the SCG's RLF-related information via the MN. The SN may control the prediction for the UE. The SN may notify the UE of the control result. The control result may be notified between the MN and the SN. Inter-base station signaling, such as Xn signaling, may be used to notify the control result. The SN may notify the UE of the control result via the MN. The control result may include, for example, information related to the recovery process (e.g., the recovery configuration information mentioned above). The UE can obtain the control result from the SN.

[0267] Alternatively, the UE may notify the MN of the RLF-related information of the SCG. The UE may notify the PCell of the RLF-related information of the SN. The UE may notify the MN of the RLF-related information of the SCG via the SN. The UE may notify the PCell of the RLF-related information of the SCG via the SN. The MN may control the prediction for the UE. The MN may notify the UE of the control results. The MN may notify the UE of the control results via the SN. The control results may include, for example, information regarding the recovery process (e.g., the recovery setting information described above). The UE can obtain the control results from the MN.

[0268] When a network node receives RLF-related information from a user audience (UE), it may notify the UE of a reception response. The UE can then recognize that the RLF-related information has been successfully notified to the network node. This enables coordinated prediction processing between the UE and the network. For notifying the UE of the reception response from the base station, for example, RRC signaling, MAC signaling, or L1 / L2 signaling may be used. For example, if RRC signaling is used to notify the base station of the information from the UE, RRC signaling may also be used to notify the UE of the reception response from the base station. For example, if MAC signaling is used to notify the base station of the information from the UE, MAC signaling may also be used to notify the UE of the reception response from the base station. Processing at the same layer avoids complexity. Reception responses may also be notified between base stations. Inter-base station signaling, such as Xn signaling, may be used to notify the UE of the reception response between base stations. MN may notify UE of a reception response via SN. SN may notify UE of a reception response via MN.

[0269] A control processing time may be provided at the NW node. This control processing time may be the control processing time that the NW node performs for each prediction process or each RLF process detection. This control processing time may be provided individually for each prediction or for each RLF process detection, or it may be provided in common for multiple predictions or for multiple RLF processes detection. If the UE does not receive a notification from the NW node within the control processing time after sending RLF-related information to the NW node, the UE may consider it an RLF. This notification may be a control instruction from the NW node. This notification may be, for example, an RRC setting from the NW node. For example, a control processing time may be provided for information regarding PLP detection prediction. If the UE does not receive a notification from the NW node within the control processing time after sending PLP detection prediction information to the NW node, the UE may consider it an RLF. In this way, even if the communication status with the NW deteriorates rapidly, the RRC re-establishment process can be performed.

[0270] If the UE does not receive notification from the NW node within the control processing time after sending RLF-related information to the NW node, the UE may still perform RLF processing. For example, RLF processing can be performed even if the prediction processing does not work well and the prediction probability of information regarding each prediction decreases, preventing the NW node from performing control processing on the UE.

[0271] The NW node may set the control processing time and notify the UE. For example, RRC signaling may be used for this notification. The control processing time may be set as a timer.

[0272] A notification prohibition period may be set for each prediction. This prohibition period may be set individually for each prediction or may be set for multiple predictions in general. Once the UE has sent prediction information to the NW node, the UE should refrain from sending that prediction information to the NW node during the notification prohibition period. This prevents frequent notifications and reduces the signaling load. It also prevents malfunctions in control processing by the NW node.

[0273] The method disclosed in this embodiment may be appropriately applied to early RLF (including early-out-of-sync and early-in-sync) processing. Even in early RLF, for example, communication quality can be restored before actual RLF detection, and RRC connection re-establishment processing can be performed on a more suitable cell at an earlier stage.

[0274] The prediction processing method, such as which detection prediction to perform, the processing method after detection prediction, whether to perform recovery processing, and the configuration information used for processing, may be statically determined by standards, etc., or may be configured by the NW node. This configuration may be transmitted to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. For example, the setting of which detection prediction to perform may include at least one of RLF detection prediction, PLP detection prediction, communication quality detection prediction, and timer expiration detection prediction. By making it configurable by the NW node, it becomes possible to perform prediction processing appropriate to the radio wave propagation environment, for example.

[0275] As described above, a prediction probability may be output in each prediction process. The UE may reset the prediction process or execute the prediction process again depending on the prediction probability. For example, if the prediction probability is lower than a predetermined value, the UE may reset the prediction process and execute the prediction process again. If the prediction probability is lower than a predetermined value, the UE may consider (or determine) that the prediction has failed. If the UE detects a prediction failure, it may notify the NW node of information regarding the prediction failure. The NW node that receives this information may reset the prediction process. For example, if the NW node and / or UE are performing a recovery process, they may reset the recovery process. In this way, it is possible to prevent processing due to incorrect predictions and to suppress further deterioration of communication quality or communication interruption.

[0276] By using the method disclosed in this embodiment, RLF processing using AI / ML becomes possible. By introducing RLF processing using AI / ML into a communication system, it becomes possible to improve communication quality, for example, by reducing RLF through early recovery or by quickly returning to a communication-ready state after RLF.

[0277] Embodiment 2. The introduction of AI / ML in a mobile communication system under 3GPP is being considered. For example, the use of AI / ML in measurement processing is being discussed (see Non-Patent Documents 33, 36, and 37). These Non-Patent Documents propose that a mobile communication system can use AI / ML to predict measurement events and take early action, such as selecting an appropriate HO destination cell or performing low-latency HO processing, to reduce RLF / HOF or shorten communication downtime. However, the above Non-Patent Documents do not disclose specific methods for introducing measurement processing using AI / ML into a communication system. For example, in measurement processing using AI / ML, there is at least one of the following issues (i1) to (i3): (i1) What is predicted in measurement processing using AI / ML? (i2) How are the prediction results obtained using AI / ML notified? (i3) What processing is performed in conjunction with the notification of the prediction results? As a result, measurement processing using AI / ML is not performed effectively, leading to problems such as a failure to improve communication quality.

[0278] This embodiment discloses a method for solving at least one of the above (i1) to (i3).

[0279] Figure 33 is a diagram illustrating the measurement process defined in the 3GPP standard (see Non-Patent Literature 19). Figure 33 shows, as an example, the case in which HO processing is executed by measurement reporting. The UE performs a measurement, and if the measurement result satisfies the event entry condition, it starts a timer (TTT: Time To Trigger). If the event leaving condition is not met before the timer (TTT) expires, the measurement reporting process is activated due to the expiration of the timer (TTT). The activation of this measurement reporting process may hereafter be referred to as "event triggering". If the event leaving condition is met before the timer (TTT) expires, the timer (TTT) is reset. The event type, event entry condition, and event leaving condition are set from the NW node to the UE.

[0280] The UE sends a measurement report triggered by an event to the NW node. Upon receiving the report, the NW node initiates, for example, HO processing. Upon initiating HO processing, the NW node performs HO preparation processing. For example, negotiation takes place between the HO source cell and the HO target cell, and the HO source cell obtains the RRC settings of the HO target cell. Once the HO preparation processing is complete, the NW node sends an HO command to the UE. The HO command includes the RRC settings of the HO target cell. Upon receiving the HO command, HO execution processing takes place between the UE and the NW, and the HO is completed.

[0281] AI / ML can be used for measurement processing. Figure 34 is a schematic diagram of the case where an AI / ML model is used for measurement processing. The UE predicts event trigger detection using the AI / ML model. Along with event trigger detection, the UE may also predict the event type, the measurement object related to the event, the time at which the event trigger is detected, and the location of the UE where the event trigger is detected. The prediction results are output from the AI / ML model. For example, the prediction results may include at least one of the following (j1) to (j2): (j1) An event trigger is detected at time tx. (j2) An event trigger for a certain event is detected when the UE is at position Lx. In this way, it becomes possible to predict which events will be detected, when and where, and it can be applied to a wide variety of applications. Note that the following describes processing related to HO, but is not limited to this. The configuration of this embodiment may be applied to measurement processing related to other processing (e.g., beam directivity control).

[0282] The UE may predict the HO target cell along with event trigger detection. The UE may predict one or more HO target cells. The prediction result may include information regarding the priority of one or more HO target cells. The UE may predict candidate HO target cells. In this way, for example, the cell that will perform HO processing after event trigger detection can be derived early, and connections to the cell can be made early.

[0283] The output settings for what to output during the prediction process may be set by the NW node. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted by the prediction process.

[0284] The input information to the AI / ML model may include, for example, at least one of the following: RRM measurement results, RLM measurement results, and CSI measurement results. The input information may also include, for example, at least one of the following: time, UE position, UE velocity, UE trajectory, etc. The input information may also include information regarding decision conditions for performing prediction processing. For example, the input information may include at least one of the following: event type, event entry conditions, and event leaving conditions. For example, the input information may include at least one of the following used as event entry conditions and event leaving conditions: threshold, hysteresis parameter, and offset value. The input information may also include, for example, a timer (TTT). In this way, event trigger detection can be directly predicted, and along with event trigger detection, its time and location can also be predicted.

[0285] The input information may be set by the NW node. The NW node may transmit the input information to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling.

[0286] Hereafter, the process of predicting the detection of event triggers using an AI / ML model may be referred to as the "event trigger detection prediction process." The event trigger detection prediction process may be performed in parallel with the measurement process. The event trigger detection prediction process may be performed independently of the measurement process, or it may be performed in correlation with the measurement process. The event trigger detection prediction process may be started periodically or periodically. The event trigger detection prediction process may be started in correlation with the measurement process. For example, the prediction process may be started upon receiving a measurement event setting from a network node, or upon the start of the measurement process. For example, the prediction process may be started when the event entry condition is first met, or when the event leaving condition is first met. For example, the prediction process may be started upon the start of a timer (TTT). For example, the prediction process may be started upon the reset of a timer (TTT). For example, the prediction process may be started when an output from the event trigger detection prediction process is derived. For example, the prediction process may be started upon the transmission or reception of an HO instruction. For example, the prediction process may be triggered upon completion of the HO process. The prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of the event trigger detection and prediction process.

[0287] New activation conditions may be established for the event trigger detection and prediction process. For example, the activation conditions may include at least one of the following (k1) to (k6): (k1) The UE is in a predetermined location. (k2) The UE has entered a predetermined area. (k3) The UE has entered the RRC_Connected state. (k4) The UE has connected to a cell. (k5) The UE has entered HO. (k6) The serving cell has been changed. In this way, the event trigger detection and prediction process can be activated according to the status of the UE.

[0288] The activation conditions described above may be set by the network node. For example, the network node may send information about at least one of the above activation conditions to the user audience (UE). The UE may then initiate event trigger detection and prediction processing according to the activation condition indicated by that information.

[0289] The event trigger detection and prediction processing may be performed upon request from the network node. For this method, the method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the event trigger detection and prediction processing at the UE. For example, it becomes possible to determine and control whether or not to perform event trigger detection and prediction processing based on communication quality.

[0290] The event trigger detection prediction process may be reset independently of the measurement process, or it may be reset in correlation with the measurement process. After the event trigger detection prediction process is started as described above, the prediction process may be reset if an output is derived from the event trigger detection prediction process. The prediction process may be reset after a predetermined time has elapsed since its start. The prediction process may be reset if the event leaving conditions are met. The prediction process may be reset if it is predicted that the event leaving conditions will be met. The prediction process may be reset when the timer (TTT) expires. The prediction process may be reset when it is predicted that the timer (TTT) will expire. The prediction process may be reset when the recovery process described later is stopped. In this way, flexible operation of the event trigger detection prediction process becomes possible.

[0291] New reset conditions may be established for the event trigger detection and prediction process. For example, the reset conditions may include at least one of the following (l1) to (l5): (l1) A UE is in a predetermined location. (l2) A UE has entered a predetermined area. (l3) A UE has entered the RRC_Idle state or the RRC_Inactive state. (l4) A UE has entered HO. (l5) A serving cell has been changed. In this way, the event trigger detection and prediction process can be reset according to the status of the UE.

[0292] The reset conditions described above may be set by the network node. For example, the network node may send information about at least one of the reset conditions described above to the user audience (UE). The UE may then reset the event trigger detection and prediction process according to the condition indicated by that information.

[0293] The event trigger detection prediction process may be reset upon request from the network node. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the event trigger detection prediction process at the UE. For example, it becomes possible to determine and control whether or not to perform the event trigger detection prediction process based on communication quality.

[0294] While the startup timing and conditions for the event trigger detection prediction process have been disclosed, the output derivation timing for the event trigger detection prediction process may also be set. A processing time for the event trigger detection prediction process may also be provided. This processing time may be set individually for each UE. For example, the processing time may be set according to the processing capacity of the UE. The UE may notify the NW node of information regarding its processing capacity. The NW node may set the processing time for the event trigger detection prediction process for the UE according to the information regarding the UE's processing capacity. The startup timing for the event trigger detection prediction process may be the timing obtained by subtracting the processing time from the output derivation timing of the prediction process. Regarding the method for setting the processing time, the method disclosed in Embodiment 1 may be applied as appropriate. In this way, the startup timing of the prediction process can be changed according to the processing capacity of the UE. The output derivation timing of the prediction process can be set regardless of the processing capacity of the UE.

[0295] The processing after event trigger detection prediction is disclosed. The UE may notify the NW node of information regarding the event trigger detection prediction. The information regarding the event trigger detection prediction may include the event trigger detection prediction result and information predicted along with the prediction result. The NW node receives the event trigger detection prediction result from the UE. The NW node becomes aware of the event trigger detection prediction result at the UE.

[0296] The network node may, for example, perform HO processing upon receiving information regarding event trigger detection prediction. The network node may also perform HO preparation processing. This allows for early implementation of HO preparation processing.

[0297] Figure 35 shows an example of event trigger detection prediction processing. The UE starts event trigger detection prediction processing, for example, at the start of measurement processing. The UE performs event trigger detection prediction using the input information disclosed above. The UE predicts the detection of an event trigger through the event trigger detection prediction processing. Having predicted the detection of an event trigger, the UE sends information regarding the event trigger detection prediction to the NW node. The NW node, having received this information, performs HO preparation processing. After completing the HO preparation processing, the NW node notifies the UE of an HO instruction. The UE, having received the HO instruction, performs HO execution processing. In this way, the UE can predict event trigger detection and notify the NW node of information regarding the prediction, enabling early execution of HO processing. For example, it can suppress HOF caused by excessively late HO execution.

[0298] After receiving information regarding the prediction of event trigger detection, the NW node performs HO preparation processing. The NW node does not need to send an HO instruction to the UE during the period between the completion of HO preparation processing and the receipt of the actual event trigger measurement report. The NW node may wait during the period between the completion of HO preparation processing and the receipt of the actual event trigger measurement report. After the completion of HO preparation processing, if the NW node receives the actual event trigger measurement report, the NW node may send an HO instruction to the UE. Since HO preparation processing can be performed earlier than the actual measurement report, the processing time from the receipt of the actual measurement report to the transmission of the HO instruction can be shortened.

[0299] Figure 36 shows another example of the event trigger detection prediction process. The UE starts the event trigger detection prediction process, for example, at the same time as the start of the measurement process. The UE performs event trigger detection prediction using the input information disclosed above. The UE performs the measurement process in parallel with the event trigger detection prediction process. The UE predicts the detection of an event trigger through the event trigger detection prediction process. Having predicted the detection of an event trigger, the UE sends information regarding the event trigger detection prediction to the NW node. The NW node, upon receiving this information, executes the HO preparation process. The NW node waits to send an HO instruction to the UE during the period from the completion of the HO preparation process until it receives a measurement report due to an actual event trigger. After the completion of the HO preparation process, the NW node sends an HO instruction to the UE upon receiving a measurement report due to an actual event trigger. The UE, upon receiving the HO instruction, performs the HO execution process. In this way, since the NW node has executed the HO preparation process in advance, it is able to perform the HO execution process upon actual event trigger detection. Because network nodes can perform HO preparation processing earlier than the actual measurement report, the processing time from receiving the actual measurement report to sending the HO instruction can be shortened. Since the HO processing time can be shortened, HOF can be suppressed.

[0300] The UE may use an AI / ML model to predict that the event leaving condition will not be met before the timer (TTT) expires. Hereafter, the continuation (i.e., expiration) of the timer (TTT) without the event leaving condition being met may simply be referred to as "TTT continuation." The process of predicting TTT continuation may also be referred to as "TTT continuation prediction process." Figure 37 is a schematic diagram of the case where an AI / ML model is used in the measurement process to predict TTT continuation. The UE may execute this prediction process after the event entry condition has been met. After the event entry condition has been met, the UE predicts that the timer (TTT) will continue without meeting the event leaving condition. Along with predicting TTT continuation, the UE may also predict the event type, the duration of TTT continuation (i.e., TTT expiration), and the UE's position at the time of timer (TTT) expiration. The prediction results are output from the AI / ML model. For example, the prediction result might output something like, "An event matching the event entry conditions will be triggered after a period of ty has elapsed from the current time, and will continue for the duration of the timer (TTT)." This makes it possible to predict which events will occur, when, where, and for the duration of the timer (TTT), allowing it to be applied to a wide variety of applications.

[0301] The UE may predict the HO destination cell along with the prediction of TTT continuation. The UE may predict one or more HO destination cells. The prediction result may include information regarding the priority of one or more HO destination cells. The UE may predict candidate HO destination cells. In this way, for example, it becomes possible to derive the cell that will perform HO processing after predicting TTT continuation at an earlier stage, and to establish connections with the cell earlier.

[0302] Regarding the input information for the AI / ML model, it is advisable to apply the input information disclosed in the aforementioned event trigger detection and prediction process as appropriate.

[0303] Regarding the execution timing and reset of the TTT continuation prediction process using the AI / ML model, the method disclosed in the aforementioned event trigger detection and prediction process should be applied as appropriate.

[0304] For processing after TTT continuation is predicted, the methods disclosed in the event trigger detection prediction processing described above may be applied as appropriate. The UE may consider event trigger detection to have been predicted upon prediction of TTT continuation.

[0305] Figure 38 shows an example of TTT continuation prediction processing. The UE starts TTT continuation prediction processing, for example, when an event entry condition is met. The UE predicts TTT continuation using the input information disclosed above. The UE predicts TTT continuation through TTT continuation prediction processing. The UE, having predicted TTT continuation, sends information about the TTT continuation prediction to the NW node. The NW node, having received this information, performs HO preparation processing. After completing the HO preparation processing, the NW node notifies the UE of an HO instruction. The UE, having received the HO instruction, performs HO execution processing. In this way, the UE can predict TTT continuation and notify the NW node of information about the prediction, enabling it to execute HO processing early. For example, it can suppress HOF caused by HO execution being too late.

[0306] The NW node may wait during the period from the completion of the HO preparation process to the receipt of the measurement report triggered by the actual event. The method disclosed in the event trigger detection and prediction process described above may be applied as appropriate.

[0307] The UE may use an AI / ML model to predict when event entry conditions will be met. Hereafter, this prediction process will be referred to as the "event entry prediction process." For example, the UE may perform this prediction process at the start of measurement. The UE may predict not only when event entry conditions will be met, but also the type of event, the time when the event entry conditions will be met, and the location of the UE where the event entry conditions will be met. The prediction results are output from the AI / ML model. For example, the prediction result may be an output such as, "A certain event entry condition will be met after a period of ty has elapsed from the current time." In this way, it becomes possible to predict which events will meet the event conditions, when and where, and it can be applied to a wide variety of applications.

[0308] Regarding the input information for the AI / ML model, it is advisable to apply the input information disclosed in the aforementioned event trigger detection and prediction process as appropriate.

[0309] Regarding the execution timing and reset of the event entry prediction process using the AI / ML model, the method disclosed in the event trigger detection and prediction process described above should be applied as appropriate.

[0310] The processing after an event entry condition match is predicted is disclosed. The UE may notify the NW node of information regarding the event entry match prediction. The information regarding the event entry match prediction may include the event entry match prediction result and information predicted along with the prediction result. The NW node receives the event entry match prediction result from the UE. The NW node becomes aware of the event entry match prediction result at the UE.

[0311] The UE may perform TTT continuation prediction processing after event entry prediction processing. The UE may use the results of the event entry prediction processing for TTT continuation prediction processing. After it is predicted that the event entry conditions are met, the UE may perform TTT continuation prediction processing. Figure 39 is a schematic diagram of the case when event entry prediction processing and TTT continuation prediction processing are performed. The UE performs event entry prediction processing using an AI / ML model, and then performs TTT continuation prediction processing using an AI / ML model. The method disclosed in Figure 34 may be used as the event entry prediction processing. The UE may input the prediction results of the event entry prediction processing into the AI / ML model for TTT continuation prediction. The UE predicts TTT continuation using the prediction results of the event entry prediction processing. The information disclosed in the TTT continuation prediction processing may be applied as appropriate as other input information. In this way, the TTT continuation prediction processing can be executed.

[0312] Figure 40 shows another example of event entry prediction processing and TTT continuation prediction processing. The UE performs event entry prediction processing at the start of measurement. After the UE predicts that the event entry conditions are met, it starts TTT continuation prediction processing. The UE predicts TTT continuation through TTT continuation prediction processing. Having predicted TTT continuation, the UE sends information about the TTT continuation prediction to the NW node. The NW node that receives this information performs HO preparation processing. After completing the HO preparation processing, the NW node notifies the UE of an HO instruction. The UE that receives the HO instruction performs HO execution processing. In this way, the UE can predict TTT continuation and notify the NW node of information about the prediction, enabling it to execute HO processing early. For example, it can suppress HOF caused by HO execution being too late.

[0313] The NW node may wait during the period from the completion of the HO preparation process until it receives the measurement report triggered by the actual event. The method disclosed in the aforementioned event trigger detection and prediction process may be applied as appropriate.

[0314] The UE may, after executing the event entry prediction process, determine whether the timer (TTT) will continue (i.e., expire) or not. Figure 41 is a schematic diagram of the event trigger detection prediction process using an AI / ML model. Figure 41 shows a method for determining whether the TTT will continue or not after the event entry prediction process. Regarding the input information for the event entry prediction process, the input information disclosed above may be applied as appropriate. The UE uses the results of the event entry prediction process to determine whether the timer (TTT) will continue or not. For example, a determination condition such as the timer (TTT) may be used as input information for this determination. The UE may use the results of the event entry prediction process and the determination condition to determine or predict whether the timer (TTT) will continue (i.e., expire) or not. In this way, the event trigger detection prediction process can be executed using the event entry prediction process.

[0315] Figure 42 illustrates another example of event trigger detection prediction processing. After event entry prediction processing, the UE performs event trigger detection prediction by determining whether the timer (TTT) will continue. The UE performs event entry prediction processing at the start of measurement. After predicting that the event entry conditions are met, the UE determines whether the predicted state will continue until the timer (TTT) expires. If the predicted state will continue until the timer (TTT) expires, the UE determines that an event trigger has been detected. The UE sends information regarding the event trigger detection prediction to the NW node. The NW node that receives this information performs HO preparation processing. After completing the HO preparation processing, the NW node notifies the UE of an HO instruction. The UE that receives the HO instruction performs HO execution processing. In this way, by determining that the TTT will continue and notifying the NW node of the prediction, the UE can execute HO processing earlier. For example, it can suppress HOF caused by HO execution being too late.

[0316] The NW node may wait during the period from the completion of the HO preparation process to the receipt of the measurement report triggered by the actual event. The method disclosed in the event trigger detection and prediction process described above may be applied as appropriate.

[0317] Each prediction process may perform a detection to determine whether or not the event leaving conditions are met. For example, after an event entry condition is predicted to be met, the event trigger detection prediction process may perform a detection to determine whether or not the event leaving conditions are met. For example, after an event trigger is predicted to be detected, the detection process may be performed during the HO preparation process. For example, after an event trigger is predicted to be detected, the detection process may be performed between the time of the HO instruction being received and the time of the HO instruction being received.

[0318] In each prediction process, if a match is detected in the event leaving condition after an event entry condition is predicted to be met, the UE may reset the prediction process. If a match is detected in the event leaving condition, the UE may reset the HO process. For example, if an event leaving condition is met before the timer (TTT) expires, the UE may reset the timer (TTT). For example, if an event leaving condition is met while the TTT continuation prediction process is running, the UE may reset the TTT continuation prediction process. For example, if an event leaving condition is met during the HO preparation process, the UE may reset the HO preparation process. The UE may restart each prediction process by resetting. In this way, the conditions of the radio wave propagation environment and the UE, etc., resulting from the event leaving condition being met, can be reflected in each prediction process.

[0319] The UE may notify the network node of the detection of an event leaving condition match. The network node can then recognize the detection of the event leaving condition match. For example, if the network node receives the detection of an event leaving condition match during the HO preparation process, the network node can reset the HO preparation process.

[0320] For example, a base station may reset the HO preparation process. Requests for resetting the HO preparation process may be made between base stations. For example, during the HO preparation process, a base station that receives detection of an event leaving condition match may request the HO destination base station to reset the HO preparation process. The HO destination base station that receives the request may reset the HO preparation process. The HO destination base station that receives the HO preparation process reset request may notify the requesting base station that the HO preparation process reset is complete. The requesting base station can then recognize that the HO preparation process reset is complete at the HO destination base station. In this way, inconsistencies in the HO preparation process reset process can be suppressed between the UE and the NW node including the HO destination NW node.

[0321] Inter-base station signaling, such as Xn signaling, may be used to notify base stations of a request to reset the HO preparation process and to notify them of the completion of the HO preparation process reset. A new message may be provided for the notification of a request to reset the HO preparation process between base stations. Conventional messages may also be used for the notification of a request to reset the HO preparation process between base stations. For example, an HO cancel message may be used. The HO cancel message may include information indicating that an event leaving condition has been met. Using conventional messages can avoid complicating the signaling process.

[0322] The UE may use AI / ML to predict whether the event leaving conditions will be met. Hereafter, this prediction process may be referred to as the "event leaving prediction process." The UE may perform the event leaving prediction process in each of the prediction processes disclosed above. If it is predicted that the event leaving conditions will be met, the UE may reset each prediction process. Alternatively, even if it is predicted that the event leaving conditions will be met, the UE may continue each prediction process. The UE may continue each prediction process until the actual event leaving conditions are met. The UE may notify the NW node of the prediction that the event leaving conditions will be met. For example, the NW node will be able to decide whether to continue with the current communication settings.

[0323] The UE may predict that the timer (TTT) will not continue, that is, that the event leaving conditions will be met before the timer (TTT) expires. If the UE predicts that the timer (TTT) will not continue, it may reset each prediction process. Alternatively, if the UE predicts that the timer (TTT) will not continue, it may continue each prediction process. The UE may notify the NW node of the prediction that the timer (TTT) will not continue. The same effect as the configuration described above can be obtained.

[0324] Regarding the method of notification from the UE to the NW node, the method disclosed in Embodiment 1 may be applied as appropriate.

[0325] The prediction processing method, such as which predictions to perform, the post-prediction processing method, and the configuration information used for processing, may be statically determined by standards, etc., or may be configured by the network node. This configuration may be transmitted to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. For example, the setting of which predictions to perform may include at least one of event trigger detection prediction, TTT continuation prediction, event entry prediction, and event leaving prediction. By making it configurable by the network node, it becomes possible to perform prediction processing appropriate to the radio wave propagation environment, for example.

[0326] Similar to the embodiments described above, a prediction probability may be output in each prediction process. The UE may reset the prediction process or execute the prediction process again depending on the prediction probability. For example, if the prediction probability is lower than a predetermined value, the UE may reset the prediction process and execute the prediction process again. If the prediction probability is lower than a predetermined value, the UE may consider (or determine) that the prediction has failed. If the UE detects a prediction failure, it may notify the NW node of information regarding the prediction failure. The NW node that receives this information may reset the prediction process. For example, if the NW node and / or UE are performing HO preparation processing, they may reset the HO preparation processing. In this way, it is possible to prevent processing due to incorrect predictions and to suppress further deterioration of communication quality or communication interruption.

[0327] By using the method disclosed in this embodiment, it becomes possible to realize measurement processing using AI / ML. Measurement processing using AI / ML makes it possible to predict early event triggers, for example. This enables the selection of appropriate HO target cells and low-latency HO processing, thereby improving communication quality.

[0328] Embodiment 3. In a mobile communication system in 3GPP, support for Aerial UEs is being considered. For Aerial UE operation, the UE derives flight path information and notifies the NW node (see Non-Patent Documents 2 and 19). The NW node uses the flight path information received from the UE to analyze, for example, whether the UE is on the flight path and controls the UE to fly on the authorized flight path. However, the NW node's decision-making process and the control process for the UE take time. Therefore, simply using the UE's flight path information can lead to problems such as processing delays, for example, when the UE is moving at high speed or when the environment changes suddenly.

[0329] This embodiment discloses a method for solving these problems.

[0330] AI / ML is used to process flight path information. The UE predicts flight path information using AI / ML. Network nodes may also predict flight path information using AI / ML. Network nodes may be equipped with an AI / ML model and predict flight path information. The UE may predict flight path information using AI / ML. The UE may be equipped with an AI / ML model and predict flight path information. The UE may notify the network nodes of information regarding the flight path information predicted using AI / ML. By using AI / ML to process flight path information, it is possible to predict flight path information.

[0331] The UE predicts flight path information through a flight path information prediction process using AI / ML. Hereafter, this prediction process will be referred to as the "flight prediction process." Figure 43 is a schematic diagram of the flight prediction process using an AI / ML model. Flight path information may include, for example, at least one of the following: flight time, UE position, UE altitude, UE trajectory, and UE direction of travel. Position may be 3D position information. Flight path information may include, for example, waypoints. Flight path information may include, for example, the position of a waypoint and time associated with the waypoint. Time associated with a waypoint may include, for example, at least one of the following: the time the UE arrives at the waypoint, the time the UE departs from the waypoint, and the time the UE passes through the waypoint. The prediction result is output from the AI / ML model. The prediction result may include, for example, at least one of the following (m1) to (m4): (m1) The position of the UE at time tx. (m2) Waypoints that the UE will pass through at time tx. (m3) Waypoints that the UE will pass through after a period of time ty has elapsed from the current time. (m4) Time at which the UE will pass through position Lx. In this way, various types of flight path information can be predicted and applied to a wide variety of applications.

[0332] The output settings for what to output in the prediction process may be performed by the NW node. The definition of an NW node is as described above. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted in the prediction process.

[0333] The input information to the AI / ML model may include, for example, the location information of the UE. The input information may include the 3D location information of the UE. The input information may include, for example, at least one of the following: time, UE velocity, UE trajectory, UE direction of travel, UE altitude, etc. For example, the input information may include the surrounding environment of the UE. For example, the input information may include at least one of the following: atmospheric pressure, wind speed, turbulence, temperature, etc. For example, the input information may include information about surrounding objects. For example, the input information may include information related to other UEs. For example, the input information may include at least one of the following: position, velocity, trajectory, direction of travel, etc. of other UEs. For example, the input information may include information about boundaries. For example, the input information may include information about rules or regulations. In this way, flight path information can be predicted, and for example, the position of a flying UE and the time corresponding to that position can be predicted.

[0334] UE may predict deviations in flight prediction processing using AI / ML. Hereafter, the process of predicting deviations may be referred to as "deviation prediction processing". For example, UE may predict deviations from a predetermined flight path. Figure 44 is a schematic diagram of flight prediction processing that predicts deviations using an AI / ML model. The deviation prediction result is output from the AI / ML model. The prediction result may include information about the position and time of the UE in which the deviation is predicted, along with the deviation. In addition, the prediction result of flight path information may be output from the AI / ML model. For example, the prediction result may include at least one of the following (n1) to (n3): (n1) The UE deviates from a predetermined flight path after a period ty has elapsed from the present time. (n2) The UE deviates from a predetermined flight path at time tx, or does not deviate. (n3) The UE deviates from a predetermined flight path at position Ly. Here, position Ly may be a position on a plane or a 3D position. This makes it possible to predict deviations from a predetermined flight path, allowing it to be applied to a wide variety of applications.

[0335] The input information to the AI / ML model may include, in addition to the input information disclosed above, at least one of a predetermined flight path and a deviation condition. The deviation condition may include at least one of the following: the difference between the predicted position and the position of the predetermined flight path, and the difference between the predicted time and the time of the flight path. The AI / ML model may predict a deviation if the difference in position continues for a predetermined time, or if the difference in time continues for a predetermined time. Therefore, the deviation condition may include the predetermined time. The predetermined time may be information about a timer. The position information may be the position of a waypoint. The time information may be the time of a waypoint. For example, the prediction result may be an output such as "the difference between the position of the flight path at a predetermined time and the predicted position of the UE is greater than or equal to value D1," or "the difference between the time of the flight path at a predetermined position and the predicted time when the UE passes through that position is greater than or equal to value D2." One or more deviation conditions may be set as input information. Multiple settings and inputs make it possible to predict various types of deviations.

[0336] Configuration information, including predetermined flight paths and deviation conditions, may be set, for example, by a UE or by a NW node. An NW node may be, for example, a base station, AMF, SMF, UPF, NWDAF, LMF, PCF, or UDM. For example, an NW node may also be an AS. The above configuration information may be set by an application. The configuration information may be set by an AS and notified to a NW node via an NEF or AF, etc. The configuration information may be set by an application and notified to an AS. A NW node, AS, or application that has set the configuration information notifies a UE or NW node having an AI / ML model of the configuration information. The UE or NW node having an AI / ML model may use the configuration information as input information to the AI / ML model.

[0337] The UE may make a determination of deviation after the flight prediction processing. Figure 45 is a schematic diagram of a method for determining deviation after flight prediction processing using an AI / ML model. Regarding the input information for the flight prediction processing, the input information disclosed above may be applied as appropriate. The UE makes a determination of deviation using the results of the flight prediction processing. The input information for this determination may include, for example, a predetermined flight path and deviation conditions. The UE may determine or predict deviation using the results of the flight prediction processing, the predetermined flight path, and the deviation conditions. In this way, deviation determination can be performed using the flight prediction processing. This makes it possible to predict deviations.

[0338] Flight prediction and deviation prediction processing using AI / ML models may be performed in parallel with flight path information processing to obtain the current flight path information of the UE. The above prediction processing may be performed independently of the flight path information processing or in correlation with the flight path information processing. The prediction processing may be started periodically or periodically. The prediction processing may be started in correlation with the flight path information processing. For example, the prediction processing may be started when the UE departs from a predetermined waypoint or when the UE passes a predetermined waypoint. For example, the prediction processing may be started when an output from the flight prediction processing is derived. For example, the prediction processing may be started when an output from the deviation prediction processing is derived. For example, the prediction processing may be started when the predicted position from the flight path prediction is a predetermined distance away from the position from the previously obtained flight path information. For example, the prediction processing may be started when the predicted position from the current flight path prediction is a predetermined distance away from the predicted position from the previous flight path prediction. For example, the prediction process may be initiated when a predetermined time has elapsed since the time of the previously acquired flight path information. For example, the prediction process may be initiated when a predetermined time has passed between the predicted time from the previous flight path prediction and the predicted time from the current flight path prediction. The prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of flight prediction and deviation prediction processes.

[0339] New activation conditions for the prediction process may be established. For example, the activation conditions may include at least one of the following (o1) to (o6): (o1) The UE has reached a predetermined altitude. (o2) The UE has entered a predetermined area. (o3) The UE has entered the RRC_Connected state. (o4) The UE has connected to a cell. (o5) The UE has entered HO. (o6) The serving cell has been changed. In this way, the prediction process can be activated according to the status of the UE.

[0340] The activation conditions described above may be set by the NW node in the same manner as in Embodiment 1. For example, the NW node may transmit information regarding at least one of the above activation conditions to the UE. The UE may activate the flight prediction process and the deviation prediction process according to the activation condition indicated by the information.

[0341] The prediction process may be performed at the request of other nodes (including UEs, NW nodes, ASs, and applications) that do not have an AI / ML model. The method disclosed in Embodiment 1 may be appropriately applied to this method. This allows other nodes to control the prediction process. For example, it becomes possible to determine and control whether or not to perform the prediction process based on the location of the UE.

[0342] The reset of the prediction process may be performed independently of the flight path information processing, or it may be performed in correlation with the flight path information processing. The UE may reset the prediction process after it has started and has derived an output from the prediction process. The UE may reset the prediction process after a predetermined time has elapsed since it started. The UE may reset the prediction process when it reaches or passes a predetermined waypoint. The UE may also reset the prediction process by stopping the recovery process described later. This allows for flexible operation of flight path information prediction processing and deviation prediction processing.

[0343] A new reset condition may be established for the prediction process. For example, the reset condition may include at least one of the following (p1) to (p5): (p1) The UE has reached a predetermined altitude. (p2) The UE has entered a predetermined area. (p3) The UE has entered the RRC_Idle state or the RRC_Inactive state. (p4) The UE has entered HO. (p5) The serving cell has been changed. In this way, the prediction process can be reset according to the status of the UE.

[0344] The prediction process may be reset upon request from other nodes (including UE, NW nodes, AS, and applications) that do not have an AI / ML model. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the prediction process on the UE. For example, it becomes possible to determine and control whether or not to perform the prediction process based on communication quality.

[0345] While the startup timing and conditions for the prediction process have been disclosed, the output derivation timing for the prediction process may also be set. A processing time for the prediction process may also be provided. This processing time may be set individually for each UE. For example, the processing time for the prediction process may be set according to the processing capacity of the UE. The UE may notify the NW node of information regarding its processing capacity. The NW node may set the processing time for the prediction process for the UE according to the information regarding its processing capacity. The startup timing for the prediction process may be the timing obtained by subtracting the processing time from the output derivation timing for the prediction process. In this way, the startup timing for the prediction process can be changed according to the processing capacity of the UE. The output derivation timing for the prediction process can be set regardless of the processing capacity of the UE.

[0346] An NW node or UE having an AI / ML model may use input information collected or set by its own node. An NW node or UE having an AI / ML model may also use input information collected or set by other nodes. The input information is notified from other nodes to the NW node or UE having an AI / ML model. The node (NW node, UE, AS, and application, etc.) that collects or sets the input information may differ for each type of input information. For example, location information may be notified from the LMF. Surrounding environment information and surrounding object information may be notified from the sensing management function, etc. An NW node or UE having an AI / ML model may perform flight prediction processing and deviation prediction processing using input information collected or set by its own node. An NW node or UE having an AI / ML model may perform flight prediction processing and deviation prediction processing using input information received from other nodes.

[0347] The prediction results from the flight prediction processing and deviation prediction processing may be notified from an NW node or UE having an AI / ML model to nodes that use the prediction results (e.g., NW nodes, UEs, ASs, and applications). The prediction results may be notified to different nodes (e.g., NW nodes, UEs, ASs, and applications) depending on the content of the prediction results. For example, other nodes (e.g., NW nodes, UEs, ASs, and applications) may use the prediction results to perform recovery processing to recover flight path deviations.

[0348] Notification of prediction results may be made immediately after the prediction results are derived. Network nodes that receive the notification can perform recovery processing early. A method of reporting flight path information using flight path information processing may be used for notifying prediction results. The flight path information report may include information indicating that it is a notification of prediction results. This avoids complexity due to an increase in the number of notification messages.

[0349] Methods for notifying input and output information are disclosed. Such notification may be performed at the CP. For example, signaling may be used for such notification. Alternatively, such notification may be performed at the UP. For example, a PDU session may be used for such notification.

[0350] This document discloses a notification method between a UE and a base station. Regarding the notification method from the UE to the base station, the notification method for RLF-related information from the UE to the base station disclosed in Embodiment 1 may be applied as appropriate. Regarding the notification method from the base station to the UE, the notification method from the NW node to the UE of the recovery method disclosed in Embodiment 1 may be applied as appropriate. Wireless notification becomes possible between the UE and the base station.

[0351] After the prediction results are derived from the flight prediction processing and deviation prediction processing, recovery processing may be performed to recover from flight path deviations. For example, the application may set the recovery processing method. The application notifies the UE of the recovery processing method. The UE may then execute the recovery processing. The execution of the recovery processing becomes possible.

[0352] For example, a network node may set a recovery processing method. The network node notifies the user audience (UE) of the recovery processing method. The UE may then execute the recovery processing. For example, an application may notify a network node of the recovery processing method for the prediction results. The recovery processing method may be a control command to the UE. The network node, having obtained the prediction results, derives a control command to the UE using the recovery processing method previously obtained from the application, and sends the derived control command to the UE. The control command may be sent using a command post (CP) or a command upload (UP). The UE may then execute the recovery processing according to the received control command. This allows for earlier execution of the recovery processing.

[0353] Flight prediction and deviation prediction processes may be used to predict measurement events for Aerial UEs. For example, the flight prediction and deviation prediction processes may predict whether the UE meets event entry conditions related to altitude, or whether the UE meets event leaving conditions related to altitude. In this configuration, the input information to the AI / ML model may include, for example, at least one of the event type, event entry conditions, and event leaving conditions. For example, the input information may include at least one of the threshold, hysteresis parameter, and offset value used as event entry conditions and event leaving conditions.

[0354] The prediction processing method, such as which prediction to perform, the post-prediction processing method, whether to perform recovery processing, and the configuration information used for processing, may be statically determined by standards, etc., or may be configured by the NW node. This configuration may be transmitted to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. For example, the setting of which prediction to perform may include at least one of flight prediction processing and deviation prediction processing. By making it configurable by the NW node, it becomes possible to execute appropriate prediction processing according to the location of the UE, for example.

[0355] Similar to the embodiments described above, a prediction probability may be output in each prediction process. The UE may reset the prediction process or execute the prediction process again according to the prediction probability. For example, if the prediction probability is lower than a predetermined value, the UE may reset the prediction process and execute the prediction process again. If the prediction probability is lower than a predetermined value, the UE may consider (or determine) that the prediction has failed. If the UE detects a prediction failure, it may notify the NW node of information regarding the prediction failure. The NW node that receives this information may reset the prediction process. For example, if the NW node and / or UE are performing a recovery process, they may reset the recovery process. In this way, it is possible to prevent processing due to incorrect predictions and to suppress further deterioration of communication quality or communication interruption.

[0356] By using AI / ML to predict flight path information, decision-making processes at network nodes and control processes for the UE can be performed earlier. Therefore, processing delays can be reduced, for example, when the UE is moving at high speed or when the environment changes suddenly. This enables the safe, secure, and accurate operation of aerial UEs (payloads).

[0357] Embodiment 4. In a mobile communication system under 3GPP, measures to prevent overheating of the UE are supported. When the UE detects internal overheating, it notifies the NW node of the overheating detection and requests a reduction in CC or bandwidth. The NW node uses the information received from the UE to perform CC or bandwidth reduction processing on the UE. However, notification from the UE to the NW node or control by the NW node to the UE takes time. Therefore, relying solely on this overheating prevention process can lead to problems such as delays in processing in response to a rapid rise in temperature inside the UE due to, for example, a sudden change in the radio wave propagation environment or a sudden demand for increased communication capacity.

[0358] This embodiment discloses a method for solving these problems.

[0359] AI / ML is used for overheating prevention processing. The UE predicts overheating detection using AI / ML. Network nodes may also predict overheating detection using AI / ML. Network nodes may be equipped with an AI / ML model and predict overheating detection. The UE may also predict overheating detection using AI / ML. Network nodes may be equipped with an AI / ML model and predict overheating detection. The UE may notify the network nodes of information regarding the overheating detection prediction made using AI / ML. By using AI / ML for overheating prevention processing, it is possible to predict overheating detection. Hereafter, overheating prevention processing using AI / ML may be referred to as "overheating prediction processing".

[0360] The UE may predict its temperature using an AI / ML overheating prediction process. Figure 46 is a schematic diagram of the overheating prediction process using an AI / ML model. The UE temperature may be, for example, the internal temperature of the UE or the temperature of the UE enclosure. The UE temperature may also be the temperature at a predetermined location within the UE. The UE may also predict its thermal conductivity. The UE may predict the location and time corresponding to the UE temperature along with the UE temperature. The prediction results are output from the AI / ML model. The prediction results may include, for example, at least one of the following (p1) to (p2): (p1) The UE temperature at time tx. (p2) The UE temperature after a period ty has elapsed from the present time. In this way, the UE temperature can be predicted, and control using the predicted UE temperature becomes possible.

[0361] The output settings for what to output in the prediction process may be performed by the NW node. The definition of an NW node is as described above. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted in the prediction process.

[0362] The input information to the AI / ML model may include, for example, the current UE temperature or a history of UE temperatures from the past to the present. The input information may include, for example, the thermal conductivity of the UE. The input information may also include, for example, information about at least one of CC, BW, BWP, and the number of MIMO layers. The input information may include information about at least one of the maximum number of CCs, maximum BW, maximum BWP, and maximum number of MIMO layers. The input information may also include information about at least one of the number of CCs in use, BW, BWP, and MIMO layers in use. Furthermore, the input information may include at least one of the change in UE temperature relative to the difference in CCs, the change in UE temperature relative to the difference in BW, the change in UE temperature relative to the difference in BWP, and the change in UE temperature relative to the difference in the number of MIMO layers. For example, the change in UE temperature for a difference in CCs is information such as "increasing the number of CCs by 1 raises the UE temperature by Tm1," or "decreasing the number of CCs by 1 lowers the UE temperature by Tm2." Input information may include, for example, output power. Output power may be the output power of the transmitting amplifier or the output power of the UE. Input information may include, for example, power consumption. Power consumption may be, for example, the power consumption of individual devices inside the UE, or the power consumption of the entire UE. Input information may include, for example, ambient environment information. Ambient environment information may include, for example, atmospheric pressure, wind speed, turbulence, and at least one of temperature information. In this way, the UE temperature can be predicted.

[0363] The UE may predict overheating detection using an AI / ML overheating prediction process. Figure 47 is a schematic diagram of the overheating prediction process that predicts overheating detection using an AI / ML model. The overheating detection prediction result is output from the AI / ML model. Along with the overheating detection, the location and time at which overheating detection is predicted may also be output. In addition, the predicted UE temperature may be output from the AI / ML model. The prediction result may include, for example, at least one of the following (q1) to (q3): (q1) Overheating is detected after a period ty has elapsed from the present time. (q2) Overheating is detected at time tx. (q3) Overheating is detected when the UE is at location Lx. In this way, overheating detection can be predicted.

[0364] The UE may predict information about parameters to be reduced by using AI / ML-based overheating prediction processing. The predicted parameters may include at least one of the following: number of CCs, BW, BWP, and number of MIMO layers. The parameters may include at least one of the following: maximum number of CCs, maximum BW, maximum BWP, and maximum number of MIMO layers. The parameters may include at least one of the following: number of CCs to be used, BW to be used, BWP to be used, and number of MIMO layers to be used. Furthermore, the prediction results may include at least one of the following: difference in the number of CCs, difference in BW, difference in BWP, and difference in the number of MIMO layers. The prediction results may include, for example, at least one of the following: Reduced MaxCC, Reduced MaxBW, Reduced MaxBWP, and Reduced MaxMIMO-layers. For example, the prediction results may include the reduced CC value used, or the difference in the reduced CC. The prediction results may also include the difference and the resulting change in UE temperature. The UE may also use the overheating prediction process to predict what should be reduced, or what should be reduced to resolve overheating. The UE may also use the overheating prediction process to predict the priority of what should be reduced.

[0365] For example, the prediction result may include at least one of the following (r1) to (r4): (r1) Setting the maximum CC to a certain value eliminates overheating. (r2) Setting the maximum BW to a certain value causes overheating detection to be delayed until a period ty has elapsed from the current time. (r3) Changing the CC used from 2 CC to 3 CC causes the UE temperature to rise by Tm1 after a period ty has elapsed from the current time. (r4) Changing the BW used from 40 MHz to 20 MHz causes the UE temperature to fall by Tm2 after a period ty has elapsed from the current time. In this way, it becomes possible to predict effective methods for preventing overheating.

[0366] The input information to the AI / ML model may include, in addition to the input information disclosed above, overheating detection conditions, for example. The overheating conditions may be, for example, the internal temperature of the UE, or the temperature of individual devices within the UE. One or more overheating detection conditions may be set as input information. For example, overheating detection conditions for each individual device within the UE may be set and input as input information. For example, overheating detection conditions for multiple locations within the UE may be set and input as input information. Multiple settings and inputs make it possible to predict various types of overheating detection.

[0367] Regarding the method for setting predetermined overheat detection conditions, the method for setting predetermined flight paths or deviation conditions disclosed in Embodiment 3 may be applied as appropriate.

[0368] The UE may make a decision on overheating detection after predicting the UE temperature. Figure 48 is a schematic diagram of a method for determining overheating detection after predicting the UE temperature using an AI / ML model. Regarding the input information for predicting the UE temperature, the input information disclosed above may be applied as appropriate. The UE makes a decision on overheating detection using the results of the UE temperature prediction process. The input information for this decision may include, for example, overheating detection conditions. The UE may determine or predict overheating in the UE using the results of the UE temperature prediction process and the overheating detection conditions. In this way, the decision on overheating detection can be made using the UE temperature prediction process, and overheating detection can be predicted.

[0369] The overheating prediction process may be performed in parallel with the conventional overheating prevention process. The overheating prediction process may be performed independently of the conventional overheating prevention process, or it may be performed in correlation with the conventional overheating prevention process. The overheating prediction process may be started periodically or periodically. The overheating prediction process may be started in correlation with the conventional overheating prevention process. For example, the overheating prediction process may be started at the same time as the conventional overheating prevention process is started. For example, the overheating prevention process may be started when an output from the overheating prediction process is derived. For example, the overheating prediction process may be started when an output from the overheating prediction process is derived. The overheating prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of the overheating prediction process.

[0370] New activation conditions for the overheat prediction process may be established. For example, the activation conditions may include at least one of the following (s1) to (s10): (s1) The number of CCs used by the UE exceeds a predetermined number of CCs. (s2) The BW used by the UE exceeds a predetermined BW. (s3) The BWP used by the UE exceeds a predetermined BWP. (s4) The number of MIMO layers used by the UE exceeds a predetermined number of layers. (s5) The UE temperature exceeds a predetermined UE temperature. (s6) The UE enters a predetermined area. (s7) The UE enters the RRC_Connected state. (s8) The UE connects to a cell. (s9) The UE goes HO. (s10) The serving cell is changed. In this way, the overheat prediction process can be activated according to the status of the UE.

[0371] The activation conditions described above may be set by the NW node in the same manner as in Embodiment 1. For example, the NW node may send information regarding at least one of the above activation conditions to the UE. The UE may then activate the overheating prediction process according to the activation condition indicated by that information.

[0372] The overheating prediction process may be performed upon request from other nodes. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the overheating prediction process. For example, it becomes possible to determine and control whether or not to perform the overheating prediction process based on communication quality.

[0373] The overheat prediction process may be reset independently of the conventional overheat countermeasures process, or it may be performed in correlation with the conventional overheat countermeasures process. The UE may reset the overheat prediction process after it has started and has derived an output from the prediction process. The UE may reset the overheat prediction process after a predetermined time has elapsed since it started. The overheat prediction process may also be reset upon detection of overheating. The UE may also reset the overheat prediction process by stopping the recovery process described later. This allows for flexible operation of the overheat prediction process.

[0374] New reset conditions for the overheating prediction process may be established. For example, the reset conditions may include at least one of the following (t1) to (t9): (t1) The number of CCs used by the UE falls below a predetermined number of CCs. (t2) The BW used by the UE falls below a predetermined BW. (t3) The BWP used by the UE falls below a predetermined BWP. (t4) The number of MIMO layers used by the UE falls below a predetermined number of layers. (t5) The UE temperature falls below a predetermined UE temperature. (t6) The UE enters a predetermined area. (t7) The UE enters the RRC_Idle state or the RRC_Inactive state. (t8) The UE goes HO. (t9) The serving cell is changed. In this way, the prediction process can be reset according to the status of the UE.

[0375] The overheat prediction process may be reset upon request from another node. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the overheat prediction process. For example, it becomes possible to determine and control whether or not to perform the overheat prediction process based on communication quality.

[0376] While the startup timing and conditions for the overheating prediction process have been disclosed, the output derivation timing for the prediction process may also be set. A processing time for the prediction process may also be set. This processing time may be set individually for each UE. For example, the processing time for the prediction process may be set according to the processing capacity of the UE. The UE may notify the NW node of information regarding its processing capacity. The NW node may set the processing time for the prediction process for the UE according to the information regarding the UE's processing capacity. The startup timing for the prediction process may be the timing obtained by subtracting the processing time from the output derivation timing for the prediction process. In this way, the startup timing for the prediction process can be changed according to the processing capacity of the UE. The output derivation timing for the prediction process can be set regardless of the processing capacity of the UE.

[0377] An NW node or UE with an AI / ML model may use input information collected or set by its own node. An NW node or UE with an AI / ML model may also use input information collected or set by other nodes. The input information is notified from other nodes to the NW node or UE with the AI / ML model. The node (NW node, UE, AS, and application, etc.) that collects or sets the input information may differ for each piece of input information. An NW node or UE with an AI / ML model may perform overheat prediction processing using input information collected or set by its own node. An NW node or UE with an AI / ML model may perform overheat prediction processing using input information received from other nodes.

[0378] The prediction results from the overheating prediction process may be notified from an NW node or UE with an AI / ML model to nodes that use the prediction results (such as NW nodes, UEs, ASs, and applications). The prediction results may also be notified to different nodes (NW nodes, UEs, ASs, and applications) depending on the content of the prediction results. For example, other nodes (such as applications, other NW nodes, and UEs) may use the prediction results to implement overheating countermeasures.

[0379] Notification of the prediction result may be given immediately after the prediction result is derived. Network nodes that receive the notification can take overheating countermeasures early. Conventional overheating detection notification methods may be used for notifying the prediction result. Conventional overheating detection notifications may include information indicating that they are notifications of prediction results. This avoids complexity due to an increase in the number of notification messages.

[0380] Methods for notifying input and output information are disclosed. Such notification may be performed at the CP. For example, signaling may be used for such notification. Alternatively, such notification may be performed at the UP. For example, a PDU session may be used for such notification.

[0381] This document discloses a notification method between the UE and the base station. Regarding the notification method from the UE to the base station, the notification method for RLF-related information from the UE to the base station disclosed in Embodiment 1 may be applied as appropriate. Regarding the notification method from the base station to the UE, the notification method from the NW node to the UE of the recovery method disclosed in Embodiment 1 may be applied as appropriate. Wireless notification becomes possible between the UE and the base station.

[0382] After the prediction results are derived from the overheating prediction process, recovery processing may be performed to resolve the overheating. The NW node may set the recovery processing method. The NW node notifies the UE of the recovery processing settings. For example, the NW node may notify the UE of the recovery processing settings using RRC signaling. The NW node receives information regarding overheating prediction from the UE. The information regarding overheating prediction may include the maximum number of reduced CCs (Reduced MaxCC), the maximum reduced BW (Reduced MaxBW), the maximum reduced BWP (Reduced MaxBWP), and the maximum number of MIMO layers (Reduced MaxMIMO-layers). In the recovery processing, the NW node, for example, tells the UE to set the number of CCs, BW, BWP, and MIMO layers to be used to be less than or equal to Reduced MaxCC, Reduced MaxBW, Reduced MaxBWP, and Reduced MaxMIMO-layers, respectively. For example, the NW node sets the number of CCs, the value of BW, the value of BWP, and the number of MIMO layers to be used based on the predicted value of the UE temperature, the predicted value of the change in UE temperature for the difference in CC, the predicted value of the change in UE temperature for the difference in BW, and the predicted value of the change in UE temperature for the difference in the number of MIMO layers. For example, the maximum output of the transmitting amplifier may be reduced during the recovery process. For example, the maximum number of HARQ retransmissions may be increased during the recovery process. The UE may then perform the recovery process. This makes it possible to perform such a recovery process. Furthermore, since the recovery process can be started based on the overheat prediction result, the recovery process can be performed earlier.

[0383] The UE may reset the recovery process based on the actual UE temperature. For example, the UE resets the recovery process when the actual UE temperature falls below a predetermined temperature. If the UE detects that the actual UE temperature has fallen below a predetermined temperature, or if it detects that the actual UE temperature has risen above a predetermined temperature, the UE may notify the NW node of the detection result. Alternatively, the UE may notify the NW node of the UE temperature periodically or periodically. The UE may include information in the detection result such as the UE temperature, the time corresponding to the UE temperature, and the UE location corresponding to the UE temperature. The NW node can recognize the detection result. The NW node can use the detection result to reset the recovery process.

[0384] By resetting the recovery process, for example, the UE may remove the limits on the maximum number of CCs, maximum BW, maximum BWP, and maximum number of MIMO layers, and communicate according to the UE capabilities. This makes it possible to perform higher-capacity communication.

[0385] The prediction processing method, such as whether or not to perform a prediction, the post-prediction processing method, whether or not to perform recovery processing, and the configuration information used for processing, may be statically determined by standards, etc., or may be configured by the network node. This configuration may be transmitted to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. By making it configurable by the network node, it becomes possible to perform prediction processing appropriate to the radio wave propagation environment, for example.

[0386] Similar to the embodiments described above, the prediction probability may be output during the prediction process. The UE may reset the prediction process or execute the prediction process again depending on the prediction probability. For example, if the prediction probability is lower than a predetermined value, the UE resets the prediction process and executes the prediction process again. If the prediction probability is lower than a predetermined value, the UE may consider (or determine) that the prediction has failed. If the UE detects a prediction failure, it may notify the node of information regarding the prediction failure. The NW node that receives this information may reset the prediction process. For example, if the NW node and / or UE are performing a recovery process, they may reset the recovery process. In this way, processing due to incorrect predictions can be prevented, and further deterioration of communication quality or communication interruption can be suppressed.

[0387] By using AI / ML to derive prediction results for UE temperature, overheating countermeasures can be implemented early. Therefore, it is possible to suppress rapid temperature increases in the UE caused by, for example, sudden changes in the radio wave propagation environment or sudden demands for increased communication capacity. This helps to suppress deterioration of communication quality, communication interruptions, and UE damage caused by rising UE temperatures.

[0388] Embodiment 5. In a mobile communication system under 3GPP, a function for countermeasures against interference within the device is supported. This function is sometimes referred to as "IDC function (in-device coexistence functionality)". When the UE detects interference within the device, it notifies the NW node of the detection of internal device interference. The UE also analyzes the interference factors along with the detection of internal device interference and notifies the NW node of, for example, the affected carrier frequency and the desired activation time interval. The NW node uses the information received from the UE to schedule the UE to reduce internal device interference. However, this internal device interference countermeasures process suffers from a problem in that the processing delay becomes large due to the UE's detection of internal device interference, analysis of interference factors, notification of these to the NW node, and the NW node's control of the UE.

[0389] This embodiment discloses a method for solving these problems.

[0390] Figure 49 is a diagram representing the IDC processing defined in the 3GPP standard (see Non-Patent Literature 1). The UE detects IDC interference. This causes the UE to begin analyzing the interference cause. In Phase 1, the UE analyzes the interference cause and derives the analysis results. Having derived the analysis results, the UE notifies the NW node of IDC support information. This notification to the NW node is sometimes referred to as "IDC indication." The IDC support information includes IDC detection, affected carrier frequencies, and desired activation time intervals. Upon receiving the IDC support information from the UE, the NW node begins analyzing countermeasures for the IDC problem. In Phase 2, the NW node analyzes countermeasures for the IDC problem and derives countermeasures. Having derived countermeasures, the NW node notifies the UE of these countermeasures. Upon receiving the countermeasures from the NW node, the UE applies these countermeasures. In Phase 3, countermeasures will be applied to both the UE and NW to address the IDC problem.

[0391] AI / ML is used in IDC processing. The UE predicts IDC interference using AI / ML. Network nodes may also predict IDC interference using AI / ML. Network nodes may be equipped with an AI / ML model and predict IDC interference. The UE may predict IDC interference using AI / ML. The UE may be equipped with an AI / ML model and predict IDC interference. The UE may notify the network nodes of information regarding the IDC interference prediction made using AI / ML. By using AI / ML in IDC processing, it is possible to predict IDC interference. Hereafter, IDC processing using AI / ML may be referred to as "IDC prediction processing".

[0392] The UE may predict IDC interference using AI / ML-based IDC processing. Figure 50 is a schematic diagram of IDC prediction processing using an AI / ML model. IDC interference may be interference with other systems mounted in the same device as the UE. The direction of interference may be the giving direction of interference or the receiving direction of interference. The direction of interference may be one or both. The UE may predict the location and time in which the IDC interference will be detected, along with the IDC interference itself. The prediction results are output from the AI / ML model. The prediction results may be, for example, "IDC interference will be detected when the UE is at position Lx" or "IDC interference with system B will be detected after a period ty has elapsed from the present time." In this way, IDC interference can be predicted, and control using the IDC interference prediction results becomes possible.

[0393] The output settings for what to output in the prediction process may be performed by the NW node. The definition of an NW node is as described above. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted in the prediction process.

[0394] The input information to the AI / ML model may include, for example, at least one of the following: communication bandwidth information, communication time information, transmitted power information, and received power information. The communication bandwidth information may include at least one of the following: band, CC, BW, BWP, RB (Resource Block) number, and subcarrier number. The communication time information may include at least one of the following: SFN, subframe number, slot number, symbol number, and DRX time information. For example, the input information may include at least one of the following: RRM measurement results, RLM measurement results, and CSI measurement results. For example, the input information may include measurement results of other systems within the device. Regarding the measurement indicator, it is good practice to apply an appropriate indicator from the RRM measurement, for example, RSSI (Received Signal Strength Indicator) or SNR (Signal-to-Noise Ratio). For example, the input information may include at least one of the following: time and date. For example, the input information may include at least one of the following: the position of the UE, the velocity of the UE, and the trajectory of the UE. For example, the input information may include system information of other systems within the device. The system information may include at least one of the following: system type, operational / non-operational information, operational / non-operational time information, operating frequency information, transmitted power information, and received power information.

[0395] The UE may perform interference factor analysis after IDC prediction processing. Figure 51 is a schematic diagram of a method for performing interference factor analysis after IDC prediction processing using an AI / ML model. Regarding the input information for IDC prediction processing, the input information disclosed above may be applied as appropriate. The UE predicts IDC interference through IDC prediction processing. The UE performs interference factor analysis using the prediction results of IDC device internal interference. The input information for this analysis may include, for example, at least one of wireless conditions and communication quality. Wireless conditions and communication quality may include, for example, at least one of RRM measurement results, RLM measurement results, and CSI measurement results. For example, the input information may include communication settings. Communication settings may include, for example, communication settings set in RRC signaling. The input information may be information input to IDC prediction processing, or information used when performing analysis. The input information may be, for example, information related to the analysis of interference factors. Information regarding the analysis of interference factors may include, for example, at least one of the methods, conditions, and limitations used in the analysis. This makes it possible to predict the interference factors using the IDC interference prediction results. The interference factor prediction results may include, for example, at least one of the IDC interference prediction results, the affected carrier frequencies, and the desired activation time interval. The interference factor prediction results may also include prediction results for information to be included in the IDC support information.

[0396] Figure 52 is a diagram illustrating an example of IDC prediction processing. Figure 52 discloses a case where interference factors are analyzed after IDC prediction processing. The UE starts IDC prediction processing when it transitions to the RRC_Connected state, for example. The UE predicts IDC interference using the input information disclosed above. Having obtained the IDC interference prediction result, the UE starts analyzing the interference factors. The UE may use the analysis input information disclosed above for the analysis of the interference factors. In Phase 1, the UE analyzes the interference factors and predicts them. When these processes are performed within the UE, the UE that has derived the analysis results may notify the NW node of the interference factor prediction results. In this way, the NW node can recognize the IDC interference prediction results and interference factor prediction results from the UE. The NW node starts analyzing countermeasures for the IDC interference problem. In Phase 2, the NW node analyzes countermeasures and derives countermeasures. The NW node that derives the countermeasure notifies the UE of the countermeasure. The countermeasure may include communication settings. The UE, upon receiving the countermeasure from the NW node, applies the countermeasure. In Phase 3, the countermeasure is applied by the UE and the NW node, and countermeasures against IDC interference are taken. In this way, IDC interference countermeasures are implemented and IDC interference is suppressed. This improves the communication quality not only of the 3GPP mobile communication system but also of other systems.

[0397] After deriving the analysis results for interference factors, the UE may wait to notify the NW node of the predicted interference factors until actual internal device interference is detected. In other words, the UE may wait to proceed to Phase 2. Figure 53 shows an example of IDC prediction processing when the notification of the predicted interference factors is waited for. The UE, which has predicted IDC interference through IDC prediction processing, analyzes the interference factors in Phase 1 and predicts the interference factors. The UE, having derived the analysis results, does not immediately notify the NW node of the predicted interference factors, but waits until actual IDC interference is detected. When actual IDC interference is detected, the UE notifies the NW node of the predicted interference factors that were previously analyzed in Phase 1. The NW node that receives the notification analyzes countermeasures in Phase 2 and derives countermeasures. The NW node that has derived countermeasures notifies the UE of these countermeasures. The UE, having received a countermeasure from the NW node, applies the countermeasure. In Phase 3, the countermeasure is applied to both the UE and the NW node, and the IDC interference problem is addressed.

[0398] Another method is disclosed. A UE that has predicted IDC interference through IDC prediction processing analyzes the interference factors in Phase 1 and predicts the interference factors. The UE that has derived the analysis results may immediately notify the NW node of the predicted interference factors. Upon receiving the notification, the NW node may not immediately proceed to Phase 2, but may wait until actual IDC interference is detected.

[0399] The UE may notify the network nodes of the detection of actual IDC interference. The network nodes can then recognize the detection of actual IDC interference by the UE.

[0400] Upon receiving the detection of actual IDC interference at the UE, the NW node proceeds to Phase 2 to analyze mitigation methods and derives a solution. The NW node that has derived a solution notifies the UE of the solution. The UE, having received the solution from the NW node, applies the solution. In Phase 3, the solution is applied by both the UE and the NW node, and the IDC interference problem is addressed.

[0401] This approach ensures that countermeasures are applied only when IDC interference is actually detected. For example, if a countermeasure involves restricting certain communication settings, this prevents the unnecessary application of such restrictions. Furthermore, because the transition to Phase 1 through IDC prediction processing can be initiated earlier, the results of the interference factor analysis can be derived sooner. Consequently, the time from actual IDC interference detection to notification of the detection to network nodes can be shortened. This allows for the early initiation of countermeasure analysis, leading to the early derivation, notification, and application of countermeasures.

[0402] The UE may predict IDC interference and interference factors using AI / ML-based IDC prediction processing. Figure 54 is a schematic diagram showing another example of IDC prediction processing using an AI / ML model. The UE predicts IDC interference and interference factors using the AI / ML model. Regarding the input information to the AI / ML model, the input information for IDC interference prediction and the input information for factor analysis disclosed above may be applied as appropriate. The prediction results for IDC interference and interference factors are output from the AI / ML model. The prediction results may be, for example, "When the UE is at position Lx, IDC interference is detected, and the interference factor is factor Cx" or "After a period ty has elapsed from the present time, IDC interference with system B is detected, and the interference factor is factor Cy". In this way, IDC interference and interference factors can be directly predicted. Direct prediction can improve, for example, the prediction probability.

[0403] Figure 55 is a diagram illustrating another example of IDC prediction processing. Figure 55 discloses a case where IDC interference and interference factors are predicted using AI / ML-based IDC prediction processing. The UE starts IDC prediction processing, for example, when it transitions to the RRC_Connected state. Regarding input information, the input information for IDC interference prediction and the input information for factor analysis disclosed above should be applied as appropriate. The UE obtains prediction results for IDC interference and interference factors through IDC prediction processing. When this prediction processing is performed within the UE, the UE may notify the NW node of the derived prediction results. In this way, the NW node can recognize the prediction results for IDC interference and interference factors in the UE. For subsequent processing, the processing in Figure 52 should be applied as appropriate. In this way, countermeasures against IDC interference are implemented and IDC interference is suppressed. This can improve the communication quality not only of the 3GPP mobile communication system but also of other systems.

[0404] NW nodes may wait to notify the mitigation method for IDC interference during the period from the derivation of the mitigation method to the detection of actual IDC interference. They may also wait to proceed to Phase 3. Figure 56 shows an example of IDC prediction processing when the notification of IDC interference mitigation method is waited for. The UE, which has predicted IDC interference and interference factors through IDC prediction processing, notifies the NW node of the prediction results. The NW node, having received the prediction results of IDC interference and interference factors from the UE, proceeds to Phase 2 and analyzes the interference mitigation method. The NW node, having derived the analysis results of the interference mitigation method, does not immediately notify the UE of the mitigation method, but waits until actual IDC interference is detected. When actual IDC interference is detected (for example, an IDC indication is received), the NW node notifies the UE of the analysis results of the mitigation method that it had previously analyzed in Phase 2. The UE, upon receiving the notification, applies the mitigation method. In Phase 3, mitigation methods are applied to the UE and NW nodes to address IDC interference. This ensures that mitigation methods are applied only when IDC interference is actually detected. For example, if a mitigation method involves restricting certain communication settings, this prevents the unnecessary application of such restrictions. Furthermore, because the transition to Phase 2 through IDC prediction processing can be started earlier, interference mitigation methods can be derived sooner. Consequently, the time from actual IDC interference detection to notification of mitigation methods to the UE can be shortened. Analysis of mitigation methods can be started earlier, and the derivation, notification, and application of mitigation methods can be carried out more quickly.

[0405] The UE may wait after Phase 1 and then perform further Phase 1 processing (for example, Phase 1x). If the actual IDC interference differs from the predicted result, the UE may further analyze the interference factors. The period for further analysis of interference factors may be designated as Phase 1x. Similarly, the NW node may wait after Phase 2 and then perform further Phase 2 processing (for example, Phase 2x). If the actual IDC interference differs from the prediction, the NW node may further analyze the interference countermeasures. The period for further analysis of interference countermeasures may be designated as Phase 2x.

[0406] Figure 57 shows an example of IDC prediction processing when Phase 2x processing is introduced. In Phase 2, the NW node that has derived the analysis results for interference mitigation methods does not immediately notify the UE of the mitigation methods, but waits until actual IDC interference is detected. When actual IDC interference is detected (i.e., an IDC indication is received), the NW node analyzes the difference between the predicted IDC interference and the actual IDC interference. If the actual IDC interference differs from the predicted result, the NW node performs further analysis of the interference mitigation in Phase 2x. In Phase 2x, the NW node derives the analysis results for interference mitigation and notifies the UE of the interference mitigation methods. The UE that receives the notification applies the mitigation methods. In this way, even if the predicted IDC interference differs from the actual IDC interference, it is possible to implement interference mitigation processing that is appropriate for the actual IDC interference.

[0407] The IDC prediction process may be performed in parallel with the conventional IDC process. The IDC prediction process may be performed independently of the conventional IDC process, or it may be performed in correlation with the conventional IDC process. The IDC prediction process may be started periodically or periodically. The IDC prediction process may be started in correlation with the conventional IDC process. For example, the IDC prediction process may be started at the same time as the conventional IDC process. For example, the IDC prediction process may be started when an output from the IDC prediction process is derived. The IDC prediction process may be performed continuously for a predetermined number of times. This allows for flexible operation of the IDC prediction process.

[0408] New activation conditions for IDC prediction processing may be established. For example, the activation conditions may include at least one of the following (u1) to (u7): (u1) The communication quality at the UE falls below a predetermined communication quality. (u2) The device on which the UE is installed runs another system. (u3) The UE enters a predetermined area. (u4) The UE enters the RRC_Connected state. (u5) The UE connects to a cell. (u6) The UE goes HO. (u7) The serving cell is changed. In this way, IDC prediction processing can be activated according to the status of the UE.

[0409] The activation conditions described above may be set by the NW node in the same manner as in Embodiment 1. For example, the NW node may send information regarding at least one of the above activation conditions to the UE. The UE may then activate the IDC prediction process according to the activation condition indicated by that information.

[0410] IDC prediction processing may be performed upon request from other nodes. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the IDC prediction processing. For example, it becomes possible to determine and control whether or not to perform IDC prediction processing based on communication quality.

[0411] The IDC prediction process may be reset independently of the IDC process, or it may be reset in correlation with the IDC process. For example, the UE may reset the IDC prediction process after it has started and has produced an output. For example, the UE may reset the IDC prediction process after a predetermined time has elapsed since it started. For example, the UE may reset the IDC prediction process when interference countermeasures are applied. For example, the IDC prediction process may be reset by notification of IDC support information from the actual IDC process. This allows for flexible operation of the IDC prediction process.

[0412] New reset conditions for the IDC prediction process may be established. For example, the reset conditions may include at least one of the following (v1) to (v6): (v1) The communication quality at the UE exceeds a predetermined communication quality. (v2) Another system terminates on the device on which the UE is installed. (v3) The UE enters a predetermined area. (v4) The UE enters the RRC_Idle state or the RRC_Inactive state. (v5) The UE goes HO. (v6) The serving cell changes. In this way, the prediction process can be reset according to the status of the UE.

[0413] The IDC prediction process may be reset upon request from another node. The method disclosed in Embodiment 1 may be applied as appropriate. This allows other nodes to control the IDC prediction process. For example, it becomes possible to determine and control whether or not to perform the IDC prediction process based on communication quality.

[0414] While the startup timing and conditions for the IDC prediction process have been disclosed, the output derivation timing for the IDC prediction process may also be set. For example, a processing time for the IDC prediction process may be set. A processing time for the IDC interference and interference factor prediction process may also be set. The processing time for the IDC prediction process may be set individually for each UE. For example, it may be set according to the processing capacity of the UE. The UE may notify the NW node of information regarding its processing capacity. The NW node may set the processing time for the prediction process for the UE according to the information regarding the UE's processing capacity. The startup timing for the prediction process may be the timing obtained by subtracting the processing time from the output derivation timing of the prediction process. In this way, the startup timing for the prediction process can be changed according to the processing capacity of the UE. The output derivation timing for the prediction process can be set regardless of the processing capacity of the UE.

[0415] Furthermore, for example, a processing time for interference mitigation analysis may be set. For instance, the IDC processing time may be defined as the time from the start of IDC prediction processing to the application of the interference mitigation method. This allows the timing of the application of the interference mitigation method to be set.

[0416] The UE may predict that IDC interference will disappear during the IDC prediction process. If the UE predicts that IDC interference will disappear, it may notify the NW node of information regarding this prediction. If the UE predicts that IDC interference will disappear, it may reset Phase 1. If the NW node receives information regarding the prediction that IDC interference will disappear, it may reset Phase 2. If the UE predicts that IDC interference will disappear, it may reset the IDC prediction process. If the UE predicts that IDC interference will disappear, it may start the IDC prediction process. This can suppress the continuation of unnecessary IDC interference prediction processes.

[0417] An NW node or UE with an AI / ML model may use input information collected or set by its own node. An NW node or UE with an AI / ML model may also use input information collected or set by other nodes. The input information is notified from other nodes to the NW node or UE with the AI / ML model. The nodes (NW nodes, UEs, ASs, and applications, etc.) that collect or set the input information may differ for each piece of input information. An NW node or UE with an AI / ML model may perform IDC prediction processing using input information collected or set by its own node. An NW node or UE with an AI / ML model may perform IDC prediction processing using input information received from other nodes.

[0418] The prediction results from the IDC prediction process may be notified from an NW node or UE with an AI / ML model to nodes that use the prediction results (such as NW nodes, UEs, ASs, and applications). The prediction results may also be notified to different nodes (NW nodes, UEs, ASs, and applications) depending on the content of the prediction results. For example, other nodes (such as applications, other NW nodes, and UEs) may use the prediction results to implement countermeasures against IDC interference.

[0419] Conventional IDC detection notification methods may be used to notify prediction results. The IDC detection notification may include information indicating that it is a notification of prediction results. This avoids complexity due to an increase in the number of notification messages.

[0420] Methods for notifying input and output information are disclosed. Such notification may be performed at the CP. For example, signaling may be used for such notification. Alternatively, such notification may be performed at the UP. For example, a PDU session may be used for such notification.

[0421] This document discloses a notification method between the UE and the base station. Regarding the notification method from the UE to the base station, the notification method for RLF-related information disclosed in Embodiment 1 may be applied as appropriate. Regarding the notification method from the base station to the UE, the notification method from the NW node to the UE of the recovery method disclosed in Embodiment 1 may be applied as appropriate. Wireless notification becomes possible between the UE and the base station.

[0422] The prediction processing method, such as whether or not to perform prediction, the post-prediction processing method, whether or not to implement IDC interference countermeasures, and the configuration information used for processing, may be statically determined by standards, etc., or may be configured by the network node. This configuration may be transmitted to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. By making it configurable by the network node, it becomes possible to perform prediction processing appropriate to the system used in the device, for example.

[0423] Similar to the embodiments described above, a prediction probability may be output in each prediction process. The UE may reset the prediction process or execute the prediction process again depending on the prediction probability. For example, if the prediction probability is lower than a predetermined value, the UE may reset the prediction process and execute the prediction process again. If the prediction probability is lower than a predetermined value, the UE may consider (or determine) that the prediction has failed. If the UE detects a prediction failure, it may notify the NW node of information regarding the prediction failure. The NW node that receives this information may reset the prediction process. For example, if the NW node and / or UE are implementing IDC interference countermeasures, they may reset the IDC interference countermeasures process. In this way, it is possible to prevent processing based on incorrect predictions and to suppress further deterioration of communication quality or communication interruptions.

[0424] By using AI / ML to derive prediction results for IDC interference and interference factors, countermeasures against IDC interference can be implemented earlier. Therefore, for example, delays in IDC interference countermeasures can be reduced, and deterioration of communication quality and communication interruptions due to IDC interference can be suppressed.

[0425] Embodiment 6. In a mobile communication system under 3GPP, a delay budget report function is supported to improve voice communication quality (see Non-Patent Documents 1, 19, and 38). If a UE has good radio conditions but, for example, poor radio conditions on the opposing UE, resulting in high jitter or delay and poor voice communication quality, it can request an increase or decrease in the DRX period as a delay budget report. This enables the opposing UE to take corrective action to improve communication quality. However, this delay budget processing presents a problem in that the processing delay for corrective action to improve communication quality becomes large due to the UE's evaluation of radio conditions and communication quality, resetting of the DRX period, or corrective action on the opposing UE.

[0426] This embodiment discloses a method for solving these problems.

[0427] AI / ML is used for delay budget processing. The UE predicts the radio condition using AI / ML. The UE may also predict voice communication quality using AI / ML. Network nodes may perform the prediction using AI / ML. Network nodes may be equipped with an AI / ML model and perform the prediction. The UE may perform the prediction using AI / ML. The UE may be equipped with an AI / ML model and perform the prediction. The UE may notify the network nodes of the prediction results using AI / ML. By using AI / ML for delay budget processing, it is possible to predict the radio condition and voice communication quality. Hereafter, delay budget processing using AI / ML may be referred to as "delay budget prediction processing".

[0428] Figure 58 is a schematic diagram of the delay budget prediction process. The UE predicts the radio condition and voice communication quality using AI / ML. The radio condition may include, for example, at least one of CQI and BLER. The voice communication quality may include, for example, at least one of QoS (Quality of Service) and QoS (Quality of Experience) of voice communication. For example, the voice communication quality may include at least one of jitter, delay, and R value (see Non-Patent Document 39). The voice communication quality may also include E2E (End-to-End) delay. The UE may use AI / ML to predict each of the above radio condition and voice communication quality, or to predict two or more of the above radio condition and voice communication quality together. Prediction processing may be performed by providing an AI / ML model that predicts each of the above radio condition and voice communication quality individually. Prediction processing may be performed by providing an AI / ML model that predicts two or more of the above radio condition and voice communication quality together. The UE may predict its location and time, along with the wireless state and voice communication quality. The prediction results are output from the AI / ML model. The prediction results may include, for example, "predicted wireless state and voice communication E2E delay after a period ty has elapsed from the present time," and "predicted voice communication quality after a period ty has elapsed from the present time, assuming the UE is at location Lx." In this way, the wireless state and voice communication quality can be predicted, and control using the prediction results becomes possible.

[0429] The output settings for what to output in the prediction process may be performed by the NW node. The definition of an NW node is as described above. The NW node may transmit the output settings to the UE using at least one of RRC signaling, MAC signaling, and L1 / L2 signaling. The output settings may include at least one of the pieces of information predicted in the prediction process.

[0430] The input information to the AI / ML model may include, for example, information about voice communication quality and at least one of information about radio conditions. The information about radio conditions may include, for example, at least one of RSRP, RSRQ, RSSI, SNR, CQI, CSI, and BLER. For example, the input information may include RRM measurement results. The RRM measurement results may include, for example, at least one of RSRP, RSRQ, RSSI, and SNR. For example, the input information may include RLM measurement results. The RLM measurement results may include at least one of RS for RLM, SS block, CSI-RS, and CRS (Cell-specific Reference Signal). For example, the input information may include CSI measurement results. The input information may include, for example, at least one of time and date. The input information may include, for example, at least one of UE position, UE velocity, and UE trajectory. The input information may include, for example, DRX settings. The DRX settings may include, for example, at least one of the DRX period, on period, and inactive period. The DRX settings may include, for example, at least one of the DRX settings when RRC_Connected and the DRX settings when RRC_Idle. The input information may include, for example, at least one of the HARQ retransmission count and the maximum HARQ retransmission count. By using various types of input information in this way, the prediction probability can be improved.

[0431] The UE may determine bad radio condition and bad voice communication quality after predicting the radio condition and voice communication quality. Bad voice communication quality may also be bad voice communication E2E delay. Figure 59 is a schematic diagram of a method for determining bad radio condition and bad voice communication quality after predicting the radio condition and voice communication quality using an AI / ML model. Regarding the input information for predicting the radio condition and voice communication quality, the input information disclosed above may be applied as appropriate. The UE obtains the prediction results for the radio condition, the prediction results for voice communication quality (e.g., VoLTE / VoNR quality), and the prediction results for voice communication delay (e.g., VoLTE / VoNR E2E delay) using an AI / ML model. The UE uses the above prediction results for radio condition and voice communication quality to determine bad radio condition and bad voice communication quality. Other input information for this determination may include, for example, determination conditions. For example, the judgment condition may include at least one of the following: information relating to a predetermined radio state and information relating to a predetermined voice communication quality. For example, the judgment condition may include a predetermined duration. Other input information may include one or more indicators of a predetermined radio state and one or more indicators of voice communication quality. The UE may output a prediction result for a bad radio state if the radio state falls below a predetermined radio state. For example, the UE may output a prediction result for a bad radio state if the radio state falls below a predetermined radio state for a predetermined duration. The UE may output a prediction result for a bad voice communication quality if the voice communication quality falls below a predetermined voice communication quality for a predetermined duration.

[0432] Alternatively, the UE may determine good radio condition and good voice communication quality after predicting radio condition and voice communication quality. Good voice communication quality may be good voice communication E2E delay. Good voice communication quality may be a drop in voice communication E2E delay. The determination conditions may include, for example, at least one of information regarding a predetermined radio condition and information regarding a predetermined voice communication quality. The determination conditions may also include a predetermined duration. Other input information may include one or more predetermined radio condition indicators and one or more voice communication quality indicators. The UE may output a predicted good radio condition result if the radio condition exceeds a predetermined radio condition. The UE may output a predicted good radio condition result if the radio condition exceeds a predetermined radio condition for a predetermined duration. The UE may output a predicted good voice communication quality result if the voice communication quality exceeds a predetermined voice communication quality. The UE may output a prediction result for good voice communication quality if the voice communication quality exceeds a predetermined voice communication quality for a predetermined duration.

[0433] The criteria for determining poor radio conditions and poor voice communication quality may be the same as, or different from, the criteria for determining good radio conditions and good voice communication quality. For example, the criteria for determining good radio conditions and good voice communication quality may be set higher than the criteria for determining poor radio conditions and poor voice communication quality. A transition region can be provided, allowing the decision processing to be performed stably.

[0434] The UE may directly predict poor radio conditions and poor voice communication quality in the delay budget prediction process. Alternatively, the UE may directly predict good radio conditions and good voice communication quality through the delay budget prediction process. The UE may make multiple predictions individually or together. Predictions may be made using AI / ML models that make predictions individually. Predictions may be made using AI / ML models that make multiple predictions together. Figure 60 is a schematic diagram showing another example of the delay budget prediction process. The UE directly predicts poor radio conditions and poor voice communication quality using an AI / ML model. Regarding the input information to the AI / ML model, it is advisable to apply the input information disclosed above and the input information for the decision process as appropriate. The UE may predict good radio conditions and good voice communication quality using an AI / ML model. The prediction results may include outputs such as, for example, "The wireless condition will be poor after a period of time ty has elapsed from the current time," and "If the UE is located at position Lx, the voice communication quality will be poor after a period of time ty has elapsed from the current time." This allows for direct prediction of poor wireless conditions and poor voice communication quality. Direct prediction can, for example, improve the prediction probability.

[0435] The UE may predict the delay adjustment amount in the delay budget prediction process. The delay adjustment amount may be, for example, an acceptable delay adjustment amount in good radio conditions, or a desired delay adjustment amount in poor radio conditions. Depending on the acceptable or desired delay adjustment amount, control can be performed, for example, by increasing or decreasing the maximum number of HARQ retransmissions.

[0436] The UE may predict the DRX setting in the delay budget prediction process. The DRX setting may be, for example, the DRX setting value required in the case of poor voice communication quality, or the DRX setting value that is acceptable in the case of good voice communication quality. The UE may also predict the CDRX (Connected Mode DRX) setting. The CDRX setting may be, for example, the amount of change for each setting. For example, the DRX setting value required in the case of poor voice communication quality can reduce voice communication jitter and delay.

[0437] The UE may perform each of the prediction processes disclosed above together. Using these prediction results, for example, it may predict good wireless conditions and poor voice communication quality. In this way, for example, it becomes possible to predict that the UE's own wireless condition is good, but the voice communication quality deteriorates due to the deterioration of the opposing UE's wireless condition.

[0438] The UE may perform voice communication quality prediction processing, good voice communication quality prediction processing, and / or poor voice communication quality prediction processing after the wireless state prediction processing and / or poor wireless state prediction processing disclosed above. Figure 61 is a schematic diagram showing another example of delay budget prediction processing. Figure 61 discloses a case where poor voice communication quality prediction processing is performed after good wireless state prediction processing. Regarding the input information for good wireless state prediction processing, the input information disclosed above may be applied as appropriate. The UE predicts that the wireless state will be better than a predetermined value by good wireless state prediction processing using the first AI / ML model and outputs information regarding the prediction of good wireless state. The UE inputs this output to a second AI / ML model for poor voice communication quality prediction processing and executes poor voice communication quality prediction processing. Other input information may be applied as appropriate, as disclosed above. The UE predicts that the voice communication quality will fall below a predetermined value through a prediction process for poor voice communication quality, and outputs information regarding the prediction of poor voice communication quality. The UE may also predict the DRX setting using a second AI / ML model. For example, the UE may predict the desired DRX setting. In this way, it becomes possible to predict poor voice communication quality using the prediction result for good wireless conditions. For example, the prediction of poor voice communication quality may be started based on the prediction result for good wireless conditions, enabling efficient prediction processing.

[0439] An NW node may perform delay budget prediction processing for multiple UEs performing voice communication. For example, an NW node may have an AI / ML model for performing delay budget prediction processing for multiple UEs performing voice communication. Figure 62 is a schematic diagram showing an example of delay budget prediction processing for multiple UEs performing voice communication. The NW node uses the AI / ML model to predict the radio state and voice communication quality of UE1 and UE2 performing voice communication, and predicts the delay amount and DRX setting for each UE. Regarding the input information to the AI / ML model, input information set for each UE performing voice communication may be used. Regarding the input information for each UE, the input information disclosed above and the input information for the decision processing may be applied as appropriate. The prediction results of the delay adjustment amount and DRX setting for each UE performing voice communication are output from the AI / ML model. The prediction results may include outputs such as, "Set the delay adjustment amount in UE1 to a certain value and set the DRX setting to the first setting," and "Set the delay adjustment amount in UE2 to a certain value and set the DRX setting to the second setting." In this way, the delay budgets of multiple UEs performing voice communication can be predicted, and coordinated measures to improve wireless conditions and communication quality can be implemented by multiple UEs.

[0440] The processing after delay budget prediction is disclosed. When the UE performs delay budget prediction processing, the UE may notify the NW nodes of information regarding the delay budget prediction. The information regarding the delay budget prediction may include the delay budget prediction result and information predicted along with the prediction result. The NW nodes receive the delay budget prediction result from the UE. The NW nodes become aware of the delay budget prediction result at the UE.

[0441] An NW node that obtains information regarding a UE's delay budget prediction may notify the opposing UE of this information. If a UE has performed delay budget prediction processing, it may notify other opposing UEs of this information. Notification from one UE to other opposing UEs may be done via an NW node or through direct communication between UEs. Direct communication between UEs may be done using an SL (Service Line). For notifying opposing UEs of information regarding delay budget predictions, for example, RTP (Real Time Transport Protocol) / RTP (RTP Control Protocol) may be used. CP (Control Line) or UP (Upline) may be used to notify an opposing UE of information regarding delay budget predictions from an NW node. For notifying an opposing UE of information regarding delay budget predictions from an NW node, for example, signaling may be used. For signaling between UEs and NW nodes, RRC signaling, MAC signaling, or L1 / L2 signaling may be used. In this way, the UE can recognize information regarding the opposing UE's delay budget prediction.

[0442] A network node that has obtained information regarding a UE's delay budget prediction may request other UEs facing that UE to perform delay budget prediction processing. A network node that has obtained information regarding a UE's delay budget prediction may request other UEs facing that UE to perform countermeasures to improve the wireless state or countermeasures to improve voice communication quality. The other UEs facing the UE will then be able to perform delay budget prediction processing, countermeasures to improve the wireless state, or countermeasures to improve voice communication quality.

[0443] After delay budget prediction processing, the NW node may derive countermeasures to improve the wireless state and countermeasures to improve voice communication quality. For example, an NW node that has obtained a prediction result for poor voice communication quality derives countermeasures to improve voice communication quality for a UE performing voice communication. The countermeasures may include communication settings for the UE. The NW node may use the delay adjustment amount and DRX settings obtained from the delay budget prediction as countermeasures. The communication settings for the UE may include, for example, DRX settings. For example, the NW node may use the DRX settings obtained from the delay budget prediction as countermeasures. For example, the NW node may use the requested DRX settings as countermeasures. For example, the NW node may set the DRX settings within the range of acceptable DRX settings as countermeasures. The communication settings for the UE may include, for example, the maximum number of HARQ retransmissions. For example, the NW node may use the delay adjustment amount obtained from the delay budget prediction as countermeasures. The network node may set the maximum number of HARQ retransmissions according to the desired or acceptable delay adjustment amount. In this way, it becomes possible to derive methods for improving the wireless condition and methods for improving voice communication quality.

[0444] A network node that has derived measures to improve wireless conditions and measures to improve voice communication quality may notify the UE of these measures. The measures may also be communication settings for the UE. For example, the network node may notify the UE of these communication settings using RRC signaling. The network node may also use an RRC Reconfiguration message. The UE, upon receiving these communication settings, implements them. This allows for measures to improve wireless conditions and voice communication quality, thereby suppressing deterioration of wireless conditions and voice communication quality at the UE.

[0445] A network node that has derived countermeasures to improve wireless conditions and countermeasures to improve voice communication quality does not need to notify the UE of these countermeasures. For example, this is the case when the countermeasures do not require any configuration on the UE. Not notifying the UE of these countermeasures can reduce the signaling load.

[0446] Figure 63 is a diagram illustrating an example of delay budget prediction processing. For example, the UE starts delay budget prediction processing at the same time as voice communication begins. The UE performs delay budget prediction using the input information disclosed above. Through delay budget prediction processing, the UE predicts, for example, good wireless conditions and poor voice communication quality. When the UE performs delay budget prediction processing, it transmits information about the prediction to the NW node. The NW node that has received the information about the prediction derives, for example, a method to improve voice communication quality. In the diagram, this derivation period is referred to as Phase Y. The method disclosed above may be applied as appropriate to derive the method to improve voice communication quality. The NW node that has derived a method to improve voice communication quality in Phase Y notifies the UE of the method. The UE applies the received method to improve voice communication quality. In this way, delay budget prediction processing makes it possible to apply methods to improve wireless conditions and methods to improve voice communication quality earlier compared to conventional delay budget processing. This allows for early improvement of wireless connectivity and voice communication quality.

[0447] After the UE derives the delay budget prediction result, the transition to Phase Y may be waited until the UE submits a delay budget report based on the actual delay budget. Figure 64 shows an example of the delay budget prediction process when a waiting process is performed in Phase Y. For example, even if the UE obtains prediction results for good wireless conditions and poor voice communication quality through the delay budget prediction process, the NW node does not immediately start Phase Y. For example, even if the UE notifies the NW node of the prediction results for good wireless conditions and poor voice communication quality, the NW node does not immediately start Phase Y. The NW node waits to derive countermeasures to improve voice communication quality until the UE submits a delay budget report based on the actual delay budget. When the NW node receives the actual delay budget report from the UE, it starts Phase Y and derives countermeasures to improve voice communication quality. A network node that has derived a method to improve voice communication quality using PhaseY notifies the UE of the method. The UE then applies the received method. In this way, delay budget prediction processing makes it possible to apply methods to improve wireless conditions and voice communication quality earlier compared to conventional delay budget processing. This allows for early improvement of both wireless conditions and voice communication quality.

[0448] In this way, the countermeasures are applied only when a report due to delayed budget processing is actually made. For example, if the countermeasures involve some kind of restriction on communication settings, it is possible to prevent those restrictions from being applied unnecessarily.

[0449] Another method is disclosed. For example, when the UE obtains the prediction results of a good radio state and poor voice communication quality, the UE notifies the NW node of the prediction results, and the NW node may immediately start PhaseY to derive a countermeasure method for improving the voice communication quality. After the NW node derives the countermeasure method in PhsaeY, instead of immediately notifying the UE of the countermeasure method, the NW node waits for the notification to the UE until a delay budget report based on the actual delay budget is made. When the NW node receives the actual delay budget report from the UE, the NW node notifies the UE of the countermeasure method derived in PhaseY for improving the voice communication quality.

[0450] By doing so, the transition to PhaseY due to the delay budget prediction process can be started early, so that a countermeasure method for improving the radio state and a countermeasure method for improving the voice communication quality can be derived early. Therefore, it is possible to shorten the period from the delay budget report based on the actual delay budget to the notification of the countermeasure method to the UE. It becomes possible to notify and apply the countermeasure method early. Also, when the report by the delay budget process is actually made, the countermeasure method is applied. For example, when the countermeasure method involves some communication setting restrictions, it is possible to prevent the restrictions from being applied uselessly.

[0451] When the content of the delay budget report based on the actual delay budget is different from the prediction result of the delay budget, the NW node may derive a further countermeasure method. After deriving the further countermeasure method, the NW node may notify the UE of the countermeasure method. Since the NW node derives the countermeasure method using the information related to the delay budget prediction in advance, the derivation of the further countermeasure method can be carried out in a short period. It is effective, for example, when the prediction probability is high.

[0452] The above method may also be applied to communication between a UE and another UE facing the UE. For example, the NW node may derive a countermeasure method for improving the radio state and voice communication quality in another UE facing the UE. The NW node that has derived the countermeasure method in another UE facing it may notify the countermeasure method to the other UE facing it. The NW node may also notify a request for applying the countermeasure method to the other UE facing it. The other UE facing it may apply the countermeasure method for improving the radio state and the countermeasure method for improving the voice communication quality received from the NW node. It becomes possible to take countermeasures to improve the radio state and voi...

Claims

1. A communication system comprising: a communication terminal device; and a network node configured to communicate with the communication terminal device, wherein the communication terminal device is configured to perform predictive processing related to at least one of the following: radio link failure (RLF), measurement processing, flight of the communication terminal device, overheating in the communication terminal device, in-device coexistence (IDC) interference in the communication terminal device, and delay budget, using at least one of AI (Artificial Intelligence) and ML (Machine Learning).

2. The communication terminal device is configured to perform predictive processing related to the RLF, the predictive processing comprising at least one of the following: processing to predict a deterioration or recovery of communication quality; processing to predict the expiration of a timer related to a PLP (Physical Layer Problem); processing to predict the detection of the PLP; processing to predict the expiration of a timer related to the RLF; and processing to predict the detection of the RLF, the communication system according to claim 1.

3. The communication terminal device is configured to perform a prediction process related to the measurement process, the prediction process includes at least one of: a process to predict the detection of an event trigger; a process to predict the expiration of a timer related to the event trigger; a process to predict the matching of an event entry condition; and a process to predict the matching of an event leaving condition, according to claim 1.

4. The communication system according to claim 1, wherein the communication terminal device is configured to perform prediction processing related to the flight of the communication terminal device, and the prediction processing includes at least one of: a process for predicting flight path information of the communication terminal device; a process for predicting deviation of the communication terminal device from a predetermined flight path; and a process for predicting the detection of an event trigger.

5. The communication system according to claim 1, wherein the communication terminal device is configured to perform a prediction process related to overheating in the communication terminal device, and the prediction process includes at least one of: a process for predicting the temperature of the communication terminal device; a process for predicting overheating of the communication terminal device; a process for predicting the amount of change in the temperature of the communication terminal device with respect to the difference in parameters; and a process for predicting the parameters to be reduced.

6. The communication terminal device is configured to perform prediction processing related to the IDC interference in the communication terminal device, and the prediction processing includes at least one of the following: processing to predict the IDC interference and processing to predict the factors of the IDC interference, according to claim 1.

7. The communication system according to claim 1, wherein the communication terminal device is configured to perform prediction processing related to the delay budget, and the prediction processing includes at least one of the following: processing to predict delay; processing to predict the wireless state in the communication terminal device; processing to predict the voice communication quality in the communication terminal device; and processing to predict the delay adjustment amount or communication settings.

8. The communication system according to any one of claims 1 to 7, wherein the communication terminal device is configured to transmit information related to the results of the prediction process to the network node.

9. The communication system according to claim 8, wherein the network node is configured to derive a recovery method and a countermeasure method using the information related to the results of the prediction process.

10. The communication system according to claim 9, wherein the network node is configured to transmit information relating to the recovery method or the countermeasure method to the communication terminal device.

11. The communication system according to any one of claims 1 to 10, wherein the communication terminal device is configured to receive input information for at least one of the AI ​​and the ML from the network node.

12. The communication system according to any one of claims 1 to 11, wherein the communication terminal device is configured to receive information related to the decision conditions for performing the prediction processing from the network node.

13. The communication system according to any one of claims 1 to 12, wherein the communication terminal device is configured to receive information related to the activation conditions of the prediction process from the network node.

14. The communication system according to any one of claims 1 to 13, wherein the communication terminal device is configured to receive information related to the reset conditions of the prediction process from the network node.

Images