Base station and communication system

Abstract

The present invention is a base station of a communication system that supports dual connectivity in which a communication terminal is simultaneously connected to two base stations, and when operating as a master base station that is a master node of dual connectivity, uses a model that has been trained using information acquired from a communication terminal connected to the base station itself and another base station adjacent thereto. And determines another base station that operates as a secondary base station that is a secondary node for dual connectivity.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to wireless communication technology. BACKGROUND

[0002] In the 3GPP (3rd Generation Partnership Project), which is a standardization organization of mobile communication systems, as a successor of LTE (Long Term Evolution) and LTE-A (Long Term Evolution Advanced) (see Non-Patent Literature 1), which is one of the 4th generation wireless access systems, the 5th generation (hereinafter sometimes referred to as "5G") wireless access system is under discussion (see Non-Patent Literature 2). The technology for the wireless section of 5G is called "New Radio Access Technology" (the "New Radio" is abbreviated as "NR"). The NR system is under discussion based on the LTE system and the LTE-A system.

[0003] For example, in Europe, METIS, which is an organization, is summarizing the requirements of 5G (see Non-Patent Literature 3). In the 5G wireless access system, as compared with the LTE system, the system capacity is 1000 times, the data transfer speed is 100 times, the data processing delay is 1 / 5, the number of simultaneous connections of communication terminals is 100 times, and the further reduction in power consumption and the reduction in cost of devices are cited as requirements (see Non-Patent Literature 3).

[0004] In order to meet such requirements, in the 3GPP, the discussion of the 5G standard is being advanced (see Non-Patent Literatures 4 to 23).

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

[0006] In NR, a higher frequency than LTE can be used to increase the transfer speed and reduce the processing delay.

[0007] In NR that sometimes uses a higher frequency than LTE, a transceiving range is formed in a narrow beam shape (beamforming) and the direction of the beam is changed (beam scanning), and thus the capability map ensures the cell coverage.

[0008] Using Figure 1 The decision matters in 3GPP regarding the frame structure of the NR system described in Non-Patent Literature 1 (Chapter 5) will be explained. Figure 1 is an explanatory diagram showing the structure of a radio frame used in a communication system of the NR system. Figure 1 In NR, one radio frame is 10 ms. The radio frame is divided into 10 subframes of equal size. In the frame structure of NR, one or a plurality of numerologies, that is, one or a plurality of subcarrier spacings (SCSs) are supported. In NR, one subframe is 1 ms, and one slot is constituted by 14 symbols, regardless of the subcarrier spacing. In addition, the number of slots included in one subframe is one in the case where the subcarrier spacing is 15 kHz, and the number of slots in other subcarrier spacings increases in proportion to the subcarrier spacing (refer to Non-Patent Literature 11 (3GPP TS 38.211)).

[0009] The decision matters in 3GPP regarding the channel structure in the NR system are described in Non-Patent Literature 2 (Chapter 5) and Non-Patent Literature 11.

[0010] The physical broadcast channel (Physical Broadcast Channel: PBCH) is a channel for downlink transmission from a base station device (hereinafter sometimes referred to simply as a "base station") to a mobile terminal device (hereinafter sometimes referred to simply as a "mobile terminal") and the like, a communication terminal device (hereinafter sometimes referred to simply as a "communication terminal" or a "terminal"). The PBCH is transmitted together with the downlink synchronization signal.

[0011] The downlink synchronization signal in NR has a primary synchronization signal (Primary Synchronization Signal: P-SS) and a secondary synchronization signal (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 an SS burst) in a prescribed period for a prescribed duration. The SS burst is constituted by a synchronization signal block (Synchronization Signal Block: hereinafter sometimes referred to as an SS block) of each beam of the base station.

[0012] The base station transmits SS blocks of each beam by changing the beam during the duration of the SS burst. The SS block is constituted by the P-SS, the S-SS, and the PBCH.

[0013] The physical downlink control channel (PDCCH) is a downlink transmission channel from the base station to the communication terminal. The PDCCH transmits downlink control information (DCI). In the DCI, there are included resource allocation information for a downlink shared channel (DL-SCH) which is one of the transport channels described later, resource allocation information for a paging channel (PCH) which is one of the transport channels described later, and HARQ (Hybrid Automatic Repeat reQuest) information related to the DL-SCH, and the like. In addition, the DCI sometimes includes uplink scheduling grant. In the DCI, there is sometimes included an Acknowledgement (Ack) / Negative Acknowledgement (Nack) which is a response signal to the uplink transmission. In addition, in order to flexibly perform DL / UL switching within a slot, the DCI sometimes includes slot format indication (SFI). The PDCCH or the DCI is also referred to as an L1 / L2 control signal.

[0014] In NR, there is provided a time domain · frequency domain which is a candidate including the PDCCH. This region is referred to as a control resource set (CORESET). The communication terminal monitors the CORESET to acquire the PDCCH.

[0015] The physical downlink shared channel (PDSCH) is a downlink transmission channel from the base station to the communication terminal. The PDSCH has mapped thereto a downlink shared channel (DL-SCH) which is a transport channel and a PCH which is a transport channel.

[0016] A physical uplink control channel (PUCCH) is an uplink transmission channel from a communication terminal to a base station. The PUCCH transmits uplink control information (UCI). Among the UCI, there are included response information (response signal) to downlink transmission, that is, Ack / Nack, CSI (Channel State Information), a scheduling request (SR), and the like. The CSI is constituted by RI (Rank Indicator), PMI (Precoding Matrix Indicator), and CQI (Channel Quality Indicator) reporting. The RI is rank information of a channel matrix in MIMO (Multiple Input Multiple Output). The PMI is information of a precoding matrix used in MIMO. The CQI is quality information indicating the quality of received data or the quality of a communication line. The UCI is sometimes transmitted by a PUSCH described later. The PUCCH or the UCI is also referred to as an L1 / L2 control signal.

[0017] A physical uplink shared channel (PUSCH) is an uplink transmission channel from a communication terminal to a base station. An uplink shared channel (UL-SCH) as one of transport channels is mapped in the PUSCH.

[0018] A physical random access channel (PRACH) is an uplink transmission channel from a communication terminal to a base station. The PRACH transmits a random access preamble.

[0019] A downlink reference signal (Reference Signal: RS) is a symbol known as a communication system of the NR system. Four kinds of downlink reference signals are defined. A UE-specific reference signal (UE-specific Reference Signal), that is, a data demodulation reference signal (Demodulation Reference Signal: DM-RS), a phase tracking reference signal (Phase Tracking Reference Signal: PT-RS), a positioning reference signal (Positioning Reference Signal: PRS), and a channel state information reference signal (Channel State Information Reference Signal: CSI-RS). As a measurement of the physical layer of the communication terminal, there are a reference signal received power (Reference Signal Received Power: RSRP) measurement and a reference signal received quality (Reference Signal Received Quality: RSRQ) measurement.

[0020] An uplink reference signal is also a symbol known as a communication system of the NR system. Three kinds of uplink reference signals are defined. A data demodulation reference signal (Demodulation Reference Signal: DM-RS), a phase tracking reference signal (Phase Tracking Reference Signal: PT-RS), and a sounding reference signal (Sounding Reference Signal: SRS).

[0021] A transport channel described in Non-Patent Literature 2 (Chapter 5) will be described. A broadcast channel (Broadcast Channel: BCH) among downlink transport channels is broadcast to the entire coverage of its base station (cell). The BCH is mapped to a physical broadcast channel (PBCH).

[0022] The HARQ-based retransmission control is applied to a downlink shared channel (DL-SCH). The DL-SCH can be broadcast to the entire coverage of a base station (cell). The DL-SCH supports dynamic or semi-static resource allocation. The semi-static resource allocation is also referred to as semi-persistent scheduling. The DL-SCH supports discontinuous reception (DRX) of a communication terminal for reducing power consumption of the communication terminal. The DL-SCH is mapped to a physical downlink shared channel (PDSCH).

[0023] A paging channel (PCH) supports DRX of a communication terminal for reducing power consumption of the communication terminal. The PCH is requested to be broadcast to the entire coverage of a base station (cell). The PCH is mapped to a physical resource such as a physical downlink shared channel (PDSCH) that can be dynamically used for traffic.

[0024] The HARQ-based retransmission control is applied to an uplink shared channel (UL-SCH) in an uplink transport channel. The UL-SCH supports dynamic or semi-static resource allocation. The semi-static resource allocation is also referred to as configured grant. The UL-SCH is mapped to a physical uplink shared channel (PUSCH).

[0025] A random access channel (RACH) is limited to control information. The RACH has a risk of collision. The RACH is mapped to a physical random access channel (PRACH).

[0026] The HARQ is described below. The HARQ refers to a technique for improving communication quality of a transmission line by combining an automatic repeat request (ARQ) and forward error correction (FEC). The HARQ has an advantage that the FEC can be effectively used by retransmission even for a transmission line in which communication quality is changed. In particular, when retransmission is performed, quality can be further improved by combining a reception result of initial transmission and a reception result of retransmission.

[0027] An example of the retransmission method is explained. When the reception side cannot correctly decode the received data, in other words, when a CRC (Cyclic Redundancy Check) error occurs (CRC = NG), "Nack" is transmitted from the reception side to the transmission side. The transmission side that has received "Nack" retransmits the data. When the reception side can correctly decode the received data, in other words, when no CRC error occurs (CRC = OK), "Ack" is transmitted from the reception side to the transmission side. The transmission side that has received "Ack" transmits the next data.

[0028] Other examples of the retransmission method are explained. In the case where a CRC error occurs in the reception side, a retransmission request is made from the reception side to the transmission side. The retransmission request is made by switching of NDI (New Data Indicator). The transmission side that has received the retransmission request retransmits the data. In the case where no CRC error occurs in the reception side, no retransmission request is made. In the case where the transmission side has not received the retransmission request within a prescribed time, it is considered that no CRC error is transmitted from the reception side.

[0029] A logical channel (Logical Channel) described in Non-Patent Literature 1 (Chapter 6) is explained. A broadcast control channel (Broadcast Control Channel: BCCH) is a downlink channel for broadcasting system control information. The BCCH as a logical channel is mapped to a broadcast channel (BCH) or a downlink shared channel (DL-SCH) as a transport channel.

[0030] A paging control channel (Paging Control Channel: PCCH) is a downlink channel for transmitting paging information (Paging Information) and a change of system information (System Information). The PCCH as a logical channel is mapped to a paging channel (PCH) as a transport channel.

[0031] A common control channel (Common Control Channel: CCCH) is a channel for transmitting control information between a communication terminal and a base station. The CCCH is used in the case where there is no RRC connection between a communication terminal and a network. In the downlink direction, the CCCH is mapped to a downlink shared channel (DL-SCH) as a transport channel. In the uplink direction, the CCCH is mapped to an uplink shared channel (UL-SCH) as a transport channel.

[0032] A dedicated control channel (DCCH) is a channel for transmitting dedicated control information between a communication terminal and a network in a point-to-point manner. The DCCH is used in a case where the communication terminal has an RRC connection with the network. The DCCH is mapped to an uplink shared channel (UL-SCH) in the uplink and to a downlink shared channel (DL-SCH) in the downlink.

[0033] A dedicated traffic channel (DTCH) is a channel for transmitting user information in point-to-point communication with a communication terminal. The DTCH exists in both the uplink and the downlink. The DTCH is mapped to an uplink shared channel (UL-SCH) in the uplink and to a downlink shared channel (DL-SCH) in the downlink.

[0034] Position tracking of a communication terminal is performed in units of an area composed of one or more cells. The position tracking is performed in order to track the position of the communication terminal even in a standby state, and thus to call the communication terminal, in other words, in order to enable calling of the communication terminal. An area for the position tracking of the communication terminal is referred to as a tracking area (TA).

[0035] In NR, calling of a communication terminal in a range of an area smaller than a tracking area is supported. The range is referred to as a RAN notification area (RNA). Paging of a communication terminal in the RRC_INACTIVE state described later is performed in the range.

[0036] In NR, in order to support wider transmission bandwidths, carrier aggregation (CA) in which two or more component carriers (CCs) are aggregated (also referred to as "aggregation") is studied. The CA is described in Non-Patent Literature 1.

[0037] In the case of CA, the UE as a communication terminal has an RRC connection unique to a network (NW). In the RRC connection, one serving cell provides NAS (Non-Access Stratum) mobility information and security input. This cell is referred to as a primary cell (PCell). Depending on the capability of the UE, a secondary serving cell (SCell) is constituted to form a group of serving cells together with the PCell. For one UE, a group of serving cells constituted by one PCell and one or more SCells is constituted.

[0038] In addition, in 3GPP, there is Dual Connectivity (DC) in which a UE is connected to two base stations to perform communication, and the like, to further increase the communication capacity. In relation to DC, it is described in Non-Patent Literature 1, 2.

[0039] One of the base stations in which dual connectivity (DC) is performed is sometimes referred to as a "master base station (MN)", and the other is sometimes referred to as a "secondary base station (SN)". The serving cell constituted by the master base station is sometimes collectively referred to as a master cell group (MCG), and the serving cell constituted by the secondary base station is sometimes collectively referred to as a secondary cell group (SCG). In DC, the primary cell in the MCG or the SCG is referred to as a special cell (SpCell or SPCell). The special cell in the MCG is referred to as a PCell, and the special cell in the SCG is referred to as a primary SCell (PSCell).

[0040] In addition, in NR, the base station sets a part of the carrier frequency band (hereinafter, sometimes referred to as a Bandwidth Part: BWP) in advance for the UE, and the UE transmits and receives between itself and the base station in the BWP, thereby reducing the power consumption in the UE.

[0041] In addition, in 3GPP, it is discussed that a service (application) using Side Link (SL) communication (also referred to as PC5 communication) is supported in the following-described EPS (Evolved Packet System) and 5G core system (refer to Non-Patent Documents 1, 2, 26 to 28). In SL communication, communication is performed between terminals. As a service using SL communication, for example, there are V2X (Vehicle-to-everything) services, proxy services, and the like. In SL communication, not only direct communication between terminals, but also communication between a UE via a relay and a NW is proposed (refer to Non-Patent Documents 26, 28).

[0042] A physical channel for SL is described (refer to Non-Patent Documents 2, 11). The physical sidelink broadcast channel (PSBCH) transmits information related to system synchronization and is transmitted from a UE.

[0043] The physical sidelink control channel (PSCCH) transmits control information from a UE for sidelink communication and V2X sidelink communication.

[0044] The physical sidelink shared channel (PSSCH) transmits data from a UE for sidelink communication and V2X sidelink communication.

[0045] The physical sidelink feedback channel (PSFCH) transmits HARQ feedback on the sidelink from a UE that received a PSSCH to a UE that transmitted the PSSCH.

[0046] A transport channel for SL is described (refer to Non-Patent Document 1). The sidelink broadcast channel (SL-BCH) has a predetermined transport channel format and is mapped to the PSBCH as a physical channel.

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

[0048] A logical channel for SL is described (refer to Non-Patent Literature 2). The sidelink broadcast control channel (SBCCH) is a sidelink channel for broadcasting sidelink system information from one UE to other UEs. The SBCCH is mapped to the SL-BCH, which is a transport channel.

[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 UEs with both sidelink communication capabilities is also implemented by the STCH. The STCH is mapped to the SL-SCH, which is a transport channel.

[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 SL-SCH, which is a transport channel.

[0051] In LTE, SL communication is only broadcast. In NR, in addition to broadcast, support of unicast and groupcast is also being studied as SL communication (refer to Non-Patent Literature 27 (3GPP TS 23.287)).

[0052] In unicast communication and groupcast communication of SL, feedback (Ack / Nack) of HARQ, CSI reporting, and the like are supported.

[0053] In addition, in 3GPP, Integrated Access and Backhaul (IAB) is being studied in which both an access link, which is a link between a UE and a base station, and a backhaul link, which is a link between base stations, are performed wirelessly (refer to Non-Patent Literatures 2, 20, and 29).

[0054] For mobile communication systems, some new technologies are required. For example, in order to improve communication performance, it is required to introduce Artificial Intelligence (AI) / Machine Learning (ML) for RAN. In 3GPP, such new technologies are beginning to be discussed (Non-Patent Literature 30, 31).

[0055] Prior Art Documents Non-Patent Literature Non-Patent Literature 1: 3GPP TS 36.300 V17.3.0 Non-Patent Literature 2: 3GPP TS 38.300 V17.3.0 Non-Patent Literature 3: “Scenarios, requirements and KPIs for 5G mobile and wireless system: Scenarios, requirements and KPIs for 5G mobile and wireless system, ICT-317669-METIS / D1.1 Non-Patent Literature 4: 3GPP TR 23.799 V14.0.0 Non-Patent Literature 5: 3GPP TR 38.801 V14.0.0 Non-Patent Literature 6: 3GPP TR 38.802 V14.2.0 Non-Patent Literature 7: 3GPP TR 38.804 V14.0.0 Non-Patent Literature 8: 3GPP TR 38.912 V16.0.0 Non-Patent Literature 9: 3GPP RP-172115 Non-Patent Literature 10: 3GPP TS 23.501 V18.0.0 Non-Patent Literature 11: 3GPP TS 38.211 V17.4.0 Non-Patent Literature 12: 3GPP TS 38.212 V17.4.0 Non-Patent Literature 13: 3GPP TS 38.213 V17.4.0 Non-Patent Literature 14: 3GPP TS 38.214 V17.4.0 Non-Patent Literature 15: 3GPP TS 38.321 V17.3.0 Non-Patent Literature 16: 3GPP TS 38.322 V17.2.0 Non-Patent Literature 17: 3GPP TS 38.323 V17.3.0 Non-Patent Literature 18: 3GPP TS 37.324 V17.0.0 Non-Patent Literature 19: 3GPP TS 38.331 V17.3.0 Non-Patent Literature 20: 3GPP TS 38.401 V17.3.0 Non-Patent Literature 21: 3GPP TS 38.413 V17.3.0 Non-Patent Literature 22: 3GPP TS 37.340 V17.3.0 Non-Patent Literature 23: 3GPP TS 38.423 V17.3.0 Non-Patent Literature 24: 3GPP TS 38.305 V17.3.0 Non-Patent Literature 25: 3GPP TS 23.273 V18.0.0 Non-Patent Literature 26: 3GPP TR 23.703 V12.0.0 Non-Patent Literature 27: 3GPP TS 23.287 V17.5.0 Non-Patent Literature 28: 3GPP TS 23.303 V17.0.0 Non-Patent Literature 29: 3GPP TS 38.340 V17.3.0 Non-Patent Literature 30: 3GPP TR 37.817 V17.0.0 Non-Patent Literature 31: 3GPP RP-213602 Non-Patent Literature 32: 3GPP R3-224427 Non-Patent Literature 33: 3GPP TS 38.314 V17.2.0 SUMMARY

[0056] PROBLEMS TO BE SOLVED BY THE INVENTION In a mobile communication system, the number of KPIs (Key Performance Indicator) required for delay, reliability, connection density, user experience, energy efficiency, and the like increases, and continuous optimization thereof is required. Therefore, AI / ML is introduced into a mobile communication system, and a technology for improving a plurality of KPIs by performing data collection and analysis has been studied (Non-Patent Literatures 30 and 31). However, the use cases in which AI / ML is introduced are limited to network energy saving, load balancing, and mobility optimization, and have not been studied in other use cases. Therefore, the following problem arises: in various communication conditions, a plurality of KPIs cannot be optimized.

[0057] In view of the above problems, an object of the present disclosure is to provide a base station that realizes a communication system capable of introducing AI / ML in various communication situations and implementing effective processing corresponding to the communication situations.

[0058] Technical solution to solve the technical problem The present disclosure is a base station of a communication system supporting dual connectivity in which a communication terminal is connected to two base stations, and in a case where the base station operates as a master node, i.e., a master base station, of dual connectivity, determines another base station operating as a secondary node, i.e., a secondary base station, of dual connectivity using a model on which training has been performed, the training using information obtained from a communication terminal connected to the base station and a neighboring other base station.

[0059] Effects of the invention According to the base station of the present disclosure, a communication system capable of implementing effective processing corresponding to a communication situation can be realized.

[0060] Objects, features, aspects and advantages of the present disclosure will become more apparent from the following detailed description and the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 is a diagram illustrating a structure of a radio frame used in a communication system of the NR system.

[0062] Figure 2 is a block diagram showing the overall structure of a communication system 210 of the NR system discussed in 3GPP.

[0063] Figure 3 is a block diagram showing the structure of DC based on a base station connected to an NG core.

[0064] Figure 4 is a block diagram showing the structure of a mobile terminal 202. Figure 2

[0065] Figure 5 is a block diagram showing the structure of a base station 213. Figure 2

[0066] Figure 6 is a block diagram showing the structure of a 5GC part.

[0067] Figure 7 is a flowchart showing an outline of cell search to standby operation by a communication terminal (UE) in a communication system of the NR system.

[0068] Figure 8 is a diagram showing one example of the structure of a cell in the NR system.

[0069] Figure 9 ​​is a connection structure diagram showing an example of a connection structure of a terminal in SL communication.

[0070] Figure 10 is a connection structure diagram showing an example of a connection structure of a base station supporting access backhaul integration.

[0071] Figure 11 is a diagram showing a structure example of a learning device relating to a communication system in Embodiment 1.

[0072] Figure 12 is a flowchart relating to a learning process of the learning device in Embodiment 1.

[0073] Figure 13 is a diagram showing a structure example of an inference device relating to a communication system in Embodiment 1.

[0074] Figure 14 is a flowchart relating to an inference process of the inference device in Embodiment 1.

[0075] Figure 15 is a sequence diagram showing an example of SN addition processing using DC AI / ML in Embodiment 1.

[0076] Figure 16 is a sequence diagram showing an example of SN change processing using DC AI / ML in Embodiment 1.

[0077] Figure 17 is a sequence diagram showing another example of SN change processing using DC AI / ML in Embodiment 1.

[0078] Figure 18 is a diagram showing a first half of a sequence showing MN HO processing with SN change using DC AI / ML in Embodiment 1.

[0079] Figure 19 is a diagram showing a second half of a sequence showing MN HO processing with SN change using DC AI / ML in Embodiment 1.

[0080] Figure 20 is a sequence diagram showing an example of CA processing using CA AI / ML in Embodiment 2.

[0081] Figure 21 is a sequence diagram showing an example of HO processing using CA consideration HO AI / ML in Embodiment 3.

[0082] Figure 22 is a sequence diagram showing an example of HO processing using beam consideration HO AI / ML in Embodiment 4.

[0083] Figure 23 This is a sequence diagram illustrating an example of group copying processing using AI / ML for group copying in Implementation 5. Detailed Implementation

[0084] Implementation method 1. Figure 2 This is a block diagram illustrating the overall structure of a communication system 210 using the NR method discussed in 3GPP. Figure 2 The following explanation is provided. The radio access network is referred to as NG-RAN (Next Generation Radio Access Network) 211. The communication terminal device, i.e., the mobile terminal device (hereinafter referred to as "user equipment" UE) 202, can wirelessly communicate with the base station device (hereinafter referred to as "NR base station (NG-RAN NodeB) gNB)" 213, and uses wireless communication to transmit and receive signals. NG-RAN 211 consists of one or more NR base stations 213.

[0085] Here, "communication terminal device" includes not only mobile terminal devices such as mobile phone terminals, but also stationary devices such as sensors. In the following description, "communication terminal device" will sometimes be abbreviated as "communication terminal".

[0086] Between UE202 and NG-RAN 211, the AS (Access Stratum) protocol is terminated. Protocols used as the AS include, for example, 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 for the control layer (hereinafter sometimes also called C-Plane, C-Plane, or CP), SDAP is used for the user layer (hereinafter sometimes also called U-Plane, U-Plane, or UP), and PDCP, MAC, RLC, and PHY are used for both C-Plane and U-Plane.

[0087] The control protocol RRC (Radio Resource Control) between the UE 202 and the NR base station 213 performs broadcast, paging, RRC connection management, and the like. As the state between the NR base station 213 and the UE 202 in the RRC, there are RRC_IDLE, RRC_CONNECTED, and RRC_INACTIVE.

[0088] In RRC_IDLE, PLMN (Public Land Mobile Network) selection, broadcast of system information (SI), paging, cell re-selection, mobility, and the like are performed. In RRC_CONNECTED, the mobile terminal has an RRC connection and can perform transmission and reception of data with the network. In RRC_CONNECTED, handover (HO), measurement of a neighbor cell, and the like are performed. In RRC_INACTIVE, the system information (SI) is broadcast, paging, cell re-selection, mobility, and the like are performed while maintaining the connection between the 5G core 214 and the NR base station 213.

[0089] The gNB 213 is connected to the 5G core (hereinafter, sometimes referred to as "5GC") 214 including an access and mobility management function (AMF), a session management function (SMF), or a user plane function (UPF). Communication of control information and / or user data is performed between the gNB 213 and the 5GC 214. The NG interface is the total of 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. One gNB 213 can be connected to a plurality of 5GCs 214. The gNBs 213 are connected by the Xn interface and communication of control information and / or user data is performed between the gNBs 213.

[0090] The 5GC section 214 is an upper device, specifically an upper node, and performs control of connection between the NR base station 213 and the mobile terminal (UE) 202, allocation of a paging signal to one or a plurality of NR base stations (gNB) 213 and / or LTE base stations (E-UTRAN NodeB: eNB), and the like. In addition, the 5GC section 214 performs mobility control of the idle state. The 5GC section 214 performs management of a tracking area list when the mobile terminal 202 is in the idle state and in the inactive state and the active state. The 5GC section 214 starts a paging protocol by transmitting a paging message to a cell belonging to a tracking area in which the mobile terminal 202 is registered.

[0091] The gNB 213 can constitute one or a plurality of cells. In a case where one gNB 213 constitutes a plurality of cells, each cell is configured to be able to communicate with the UE 202.

[0092] The gNB 213 can be divided into a central unit (Central Unit: hereinafter sometimes referred to as CU) 215 and a distributed unit (Distributed Unit: hereinafter sometimes referred to as DU) 216. The CU 215 is constituted as one in the gNB 213. The DU 216 is constituted as one or a plurality in the gNB 213. One DU 216 constitutes one or a plurality of cells. The CU 215 is connected to the DU 216 through an Fl interface, and communication of control information and / or user data is performed between the CU 215 and the DU 216. The Fl interface is constituted by an Fl-C interface and an Fl-U interface. The CU 215 assumes the functions of each protocol of RRC, SDAP, and PDCP, and the DU 216 assumes the functions of each protocol of RLC, MAC, and PHY. One or a plurality of TRPs (Transmission Reception Point) 219 are sometimes connected to the DCU 216. The TRP 219 performs transmission and reception of radio signals with the UE.

[0093] The CU 215 can be split into a C-layer CU (CU-C) 217 and a U-layer CU (CU-U) 218. The CU-C 217 is configured as one in the CU 215. The CU-U 218 is configured as one or more in the CU 215. The CU-C 217 is connected with the CU-U 218 through an E1 interface, and communication of control information is performed between the CU-C 217 and the CU-U 218. The CU-C 217 is connected with the DU 216 through an F1-C interface, and communication of control information is performed between the CU-C 217 and the DU 216. The CU-U 218 is connected with the DU 216 through an F1-U interface, and communication of user data is performed between the CU-U 218 and the DU 216.

[0094] The 5G system can include a Unified Data Management (UDM) function and a Policy Control Function (PCF) described in Non-Patent Literature 10 (3GPP TS 23.501). The UDM and / or the PCF can be included in the 5GC part 214. Figure 2

[0095] In the 5G system, a Location Management Function (LMF) described in Non-Patent Literature 24 (3GPP TS 38.305) can be provided. The LMF can be connected to the base station via the AMF as disclosed in Non-Patent Literature 25 (3GPP TS 23.273).

[0096] In the 5G system, a Non-3GPP Interworking Function (N3IWF) described in Non-Patent Literature 10 (3GPP TS 23.501) can be included. The N3IWF can be an access network (AN) between the UE in the non-3GPP access between the UE.

[0097] Figure 3 is a diagram illustrating a structure based on DC (dual connectivity) connected to an NG core. In Figure 3 , a solid line indicates a connection of a U-Plane, and a dashed line indicates a connection of a C-Plane. Figure 3 In the 5G system, the master base station 240-1 can be a gNB or an eNB. In addition, the secondary base station 240-2 can be a gNB or an eNB. For example, in Figure 3 , a DC structure in which the master base station 240-1 is a gNB and the secondary base station 240-2 is an eNB is sometimes referred to as NG-EN-DC. In Figure 3 ​The example shown illustrates a U-Plane connection between the 5GC unit 214 and the secondary base station 240-2 via the primary base station 240-1, but it can also be established directly between the 5GC unit 214 and the secondary base station 240-2. Additionally, Figure 3 In this configuration, the core network EPC (Evolved Packet Core) connected to the LTE and LTE-A systems can replace the 5GC unit 214 and connect to the main base station 240-1. The U-Plane connection between the EPC and the secondary base station 240-2 can be directly established.

[0098] Figure 4 It is shown Figure 2 The diagram shows the structure of the mobile terminal 202. Figure 4 The transmission processing of the mobile terminal 202 shown will be described. 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 is possible. This buffering can be set 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, and MAC, for example, determining the transmission target base station in DC and assigning headers to various protocols. The protocol-processed data is transmitted to the encoding unit 304 for error correction and other encoding processing. Alternatively, data may be output directly from the protocol processing unit 301 to the modulation unit 305 without encoding processing. The data encoded by the encoding unit 304 is modulated in the modulation unit 305. Precoding for MIMO may also be performed in the modulation unit 305. After the modulated data is converted into a baseband signal, it is output to the frequency conversion unit 306 and converted into a wireless transmission frequency. Subsequently, the transmitted signal was sent from antennas 307-1 to 307-4 to base station 213. Figure 4 The example shown has four antennas, but the number of antennas is not limited to four.

[0099] Furthermore, the receiving process of the mobile terminal 202 is performed as follows: Wireless signals from the base station 213 are received via antennas 307-1 to 307-4. The received signal is converted from the wireless receiving frequency to a baseband signal by the frequency conversion unit 306, and demodulation processing is performed in the demodulation unit 308. Waiting calculations and multiplication processes can be performed in the demodulation unit 308. The demodulated data is transmitted to the decoding unit 309 for error correction and other decoding processing. The decoded data is transmitted to the protocol processing unit 301, where protocol processing such as MAC, RLC, PDCP, and SDAP is performed, including actions such as header removal in each protocol. Of the data after protocol processing, control data is transmitted to the control unit 310, and user data is transmitted to the application unit 302.

[0100] The series of processes of the mobile terminal 202 are controlled by the control section 310. Thus, although the series of processes of the mobile terminal 202 are omitted in Figure 4 , the control section 310 is also connected to each of the sections 302, 304 to 309.

[0101] Each of the sections of the mobile terminal 202, such as the control section 310, the protocol processing section 301, the encoding section 304, and the decoding section 309, is realized by, for example, a processing circuit including a processor and a memory. For example, the control section 310 is realized by the processor executing a program describing the series of processes of the mobile terminal 202. The program describing the series of processes of the mobile terminal 202 is stored in the memory. Examples of the memory are nonvolatile or volatile semiconductor memories such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, and the like. Each of the sections of the mobile terminal 202, such as the control section 310, the protocol processing section 301, the encoding section 304, and the decoding section 309, can be realized by a dedicated processing circuit such as an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), a DSP (Digital Signal Processor), and the like. In Figure 4 , the number of antennas used for transmission of the mobile terminal 202 can be the same as the number of antennas used for reception, or can be different.

[0102] Figure 5 is a block diagram showing the structure of the base station 213 shown in Figure 2 . The transmission process of the base station 213 shown in Figure 5 will be described. The EPC communication section 401 performs transmission and reception of data between the base station 213 and the EPC. The 5GC communication section 412 performs transmission and reception of data between the base station 213 and the 5GC (the 5GC section 214 and the like). The other base station communication section 402 performs transmission and reception of data with other base stations. The EPC communication section 401, the 5GC communication section 412, and the other base station communication section 402 exchange information with the protocol processing section 403, respectively. Control data from the control section 411, and user data and control data from the EPC communication section 401, the 5GC communication section 412, and the other base station communication section 402 are transmitted to the protocol processing section 403. Buffering of the control data and the user data can be performed. The buffering of the control data and the user data can be provided in the control section 411, can be provided in the EPC communication section 401, can be provided in the 5GC communication section 412, or can be provided in the other base station communication section 402.

[0103] The protocol processing unit 403 performs protocol processing for SDAP, PDCP, RLC, MAC, etc., such as routing transmitted data in DC and assigning headers to various protocols. The protocol-processed data is transmitted to the encoding unit 405 for error correction and other encoding processing. Alternatively, data may be output directly from the protocol processing unit 403 to the modulation unit 406 without encoding processing. Furthermore, data can be transmitted from the protocol processing unit 403 to other base station communication units 402. For example, in DC, data transmitted from the 5GC communication unit 412 or the EPC communication unit 401 can be transmitted to other base stations, such as auxiliary base stations, via other base station communication units 402. The encoded data undergoes modulation processing in the modulation unit 406. Precoding for MIMO can also be performed in the modulation unit 406. After the modulated data is converted into a baseband signal, it is output to the frequency conversion unit 407 and converted into a wireless transmission frequency. Then, using antennas 408-1 to 408-4, the transmission signal is transmitted to one or more mobile terminals 202. Figure 5 The example shown has four antennas, but the number of antennas is not limited to four.

[0104] Furthermore, the reception processing of base station 213 is performed as follows: Wireless signals from one or more mobile terminals 202 are received by antennas 408-1 to 408-4. The received signals are converted from the wireless receiving frequency to a baseband signal by frequency conversion unit 407, and demodulated in demodulation unit 409. The demodulated data is transmitted to decoding unit 410 for error correction and other decoding processing. The decoded data is transmitted to protocol processing unit 403, where protocol processing such as MAC, RLC, PDCP, and SDAP is performed, including actions such as header removal in each protocol. Of the data after protocol processing, control data is transmitted to control unit 411, 5GC communication unit 412, EPC communication unit 401, or other base station communication unit 402, and user data is transmitted to 5GC communication unit 412, EPC communication unit 401, or other base station communication unit 402. Data sent from other base station communication units 402 can be transmitted to 5GC communication unit 412 or EPC communication unit 401. This data could be, for example, uplink data transmitted from the DC to the 5GC communication unit 412 or the EPC communication unit 401 via other base stations.

[0105] The series of processes of base station 213 are controlled by control unit 411. Therefore, although in Figure 5 The details have been omitted, but the control unit 411 is also connected to each of the parts 401, 402, 405 to 410, 412.

[0106] The various parts of base station 213, such as control unit 411, protocol processing unit 403, 5GC communication unit 412, EPC communication unit 401, other base station communication unit 402, encoding unit 405, and decoding unit 410, are implemented similarly to the mobile terminal 202, by a processing circuit consisting of a processor and memory, or by a dedicated processing circuit such as FPGA, ASIC, or DSP. Figure 5 In this system, the number of antennas used for transmitting and the number of antennas used for receiving in base station 213 can be the same or different.

[0107] As Figure 2 The example of the structure of CU 215 shown, except Figure 5 In addition to the encoding unit 405, modulation unit 406, frequency conversion unit 407, antennas 408-1 to 408-4, demodulation unit 409, and decoding unit 410 shown, a structure with a DU communication unit is sometimes used. The DU communication unit is connected to the protocol processing unit 403. The protocol processing unit 403 in CU215 performs protocol processing for PDCP, SDAP, etc.

[0108] As Figure 2 The example of the structure of DU216 shown, except Figure 5 In addition to the EPC communication unit 401, other base station communication units 402, and 5GC communication unit 412 shown, a structure with a CU communication unit is sometimes used. The CU communication unit is connected to the protocol processing unit 403. The protocol processing unit 403 in DU216 performs protocol processing for PHY, MAC, RLC, etc.

[0109] Figure 6 This is a block diagram showing the structure of the 5GC section. Figure 6 The above is shown in the figure. Figure 2 The structure of the 5GC section 214 shown. Figure 6 It shows in Figure 2 The 5GC section 214 shown includes the structures of AMF, SMF, and UPF. Figure 6In the example shown, the AMF can have the function of the control layer control section 525, the SMF can have the function of the session management section 527, and the UPF can have the functions of the user layer communication section 523 and the data network communication section 521. The data network communication section 521 performs data transmission and reception between the 5GC section 214 and a data network. The base station communication section 522 performs data transmission and reception between the 5GC section 214 and the base station 21 via an NG interface. User data transmitted from the data network is transferred from the data network communication section 521 to the base station communication section 522 via the user layer communication section 523, and is transmitted to one or more base stations 213. User data transmitted from the base station 213 is transferred from the base station communication section 522 to the data network communication section 521 via the user layer communication section 523, and is transmitted to the data network.

[0110] Control data transmitted from the base station 21 is transferred from the base station communication section 522 to the control layer control section 525. The control layer control section 525 can transfer the control data to the session management section 527. The control data can be transmitted from the data network. Control data transmitted from the data network can be transmitted from the data network communication section 521 to the session management section 527 via the user layer communication section 523. The session management section 527 can transmit the control data to the control layer control section 525.

[0111] The user layer communication section 523 includes a PDU processing section 523-1, a mobility anchor section 523-2, and the like, and performs overall processing for a user layer (hereinafter also referred to as a U-Plane). The PDU processing section 523-1 performs processing of data packets, for example, transmission and reception of packets between the data network communication section 521, and transmission and reception of packets between the base station communication section 522. The mobility anchor section 523-2 is responsible for connection of a data path at the time of UE mobility.

[0112] The session management section 527 performs management of a PDU session set between the UE and the UPF, and the like. The session management section 527 includes a PDU session control section 527-1, a UE IP address allocation section 527-2, and the like. The PDU session control section 527-1 performs management of a PDU session between the mobile terminal 202 and the 5GC section 214. The UE IP address allocation section 527-2 performs allocation of an IP address for the mobile terminal 202, and the like.

[0113] The control layer control section 525 includes a NAS security section 525-1, an idle state mobile management section 525-2, and the like, and performs overall processing for the control layer (hereinafter, also referred to as C-Plane). The NAS security section 525-1 performs security protection and the like of a NAS (Non-Access Stratum) message. The idle state mobile management section 525-2 performs mobile management of a standby state (idle state (RRC_IDLE state), or simply, idle), generation and control of a paging signal in the standby state, addition, deletion, update, search, and tracking area list management of a tracking area of one or a plurality of mobile terminals 202 in a coverage, and the like.

[0114] The series of processes of the 5GC section 214 are controlled by the control section 526. Thereby, although the 5GC section 214 is omitted in Figure 6 The control section 526 is connected to each of the sections 521 to 523, 525, and 527. Each part of the 5GC section 214 is realized by, for example, a processing circuit including a processor and a memory, or a dedicated processing circuit such as an FPGA, an ASIC, a DSP, and the like, similarly to the control section 310 of the mobile terminal 202 described above.

[0115] Next, one example of a cell search method in a communication system is shown. Figure 7 is a flowchart showing an outline of a procedure from cell search to standby movement by a communication terminal (UE) in a communication system of the NR system. If the communication terminal starts cell search, in step ST601, synchronization of a slot timing and a frame timing is acquired using a first synchronization signal (P-SS) and a second synchronization signal (S-SS) transmitted from a peripheral base station.

[0116] The P-SS and the S-SS are collectively referred to as a synchronization signal (SS). In the synchronization signal (SS), a synchronization code corresponding to each PCI (Physical Cell Identifier) allocated to each cell is allocated. It is discussed that the number of PCIs is set to 1008. The communication terminal acquires synchronization using the 1008 PCIs, and detects (determines) the PCI of a cell in which synchronization is acquired.

[0117] In step ST602, the communication terminal receives the PBCH of the next cell to acquire synchronization. The MIB (Master Information Block) including the cell configuration information is mapped in the BCCH on the PBCH. Therefore, by receiving the PBCH and acquiring the BCCH, the MIB can be acquired. As the information of the MIB, for example, there are the SFN (System Frame Number), the scheduling information of the SIB (System Information Block) 1, the subcarrier spacing of the SIB1 and the like, the information of the DM-RS position, and the like.

[0118] In addition, the communication terminal acquires the SS block identification through the PBCH. A part of the bit string of the SS block identification is included in the MIB. The remaining bit string is included in the identification of the sequence used for generating the DM-RS accompanying the PBCH. The communication terminal acquires the SS block identification using the MIB included in the PBCH and the sequence of the DM-RS accompanying the PBCH.

[0119] Next, in step ST603, the communication terminal measures the reception power of the SS block.

[0120] Next, in step ST604, the communication terminal selects the cell with the best reception quality, for example, the cell with the highest reception power, that is, the best cell, from the one or more cells detected until step ST603. In addition, the communication terminal selects the beam with the best reception quality, for example, the beam with the highest reception power of the SS block, that is, the best beam. The selection of the best beam uses, for example, the reception power of the SS block of each SS block identification.

[0121] Next, in step ST605, the communication terminal receives the DL-SCH based on the scheduling information of the SIB1 included in the MIB and acquires the SIB (System Information Block) 1 in the broadcast information BCCH. The SIB1 includes the information related to the access to the cell, the configuration information of the cell, the scheduling information of the other SIBs (SIBk: integer k ≥ 2). In addition, the SIB1 includes the TAC (Tracking Area Code).

[0122] Next, in step ST606, the communication terminal compares the TAC of the SIB1 received in step ST605 with the TAC part of the Tracking Area Identity (TAI) in the tracking area list that the communication terminal has held. The tracking area list is also called a TAI list. The TAI is identification information for identifying a tracking area, and is composed of an MCC (Mobile Country Code), an MNC (Mobile Network Code), and a TAC (Tracking Area Code). The MCC is a country code. The MNC is a network code. The TAC is a code number of a tracking area.

[0123] If the result of the comparison in step ST606 is that the TAC received in step ST605 is the same as the TAC included in the tracking area list, the communication terminal enters a standby operation in the cell. If the TAC received in step ST605 is not included in the tracking area list, the communication terminal passes through the cell and requests a change of tracking area to the core network (EPC) including an MME or the like to perform a TAU (Tracking Area Update).

[0124] The device that constitutes the core network (hereinafter sometimes referred to as a "core network side device") performs an update of the tracking area list based on the TAU request signal and the identification number (UE-ID or the like) of the communication terminal transmitted from the communication terminal. The core network side device transmits the updated tracking area list to the communication terminal. The communication terminal rewrites (updates) the TAC list held by the communication terminal based on the received tracking area list. Thereafter, the communication terminal enters a standby operation in the cell.

[0125] Next, an example of a random access method in a communication system is shown. In the random access, a 4-step random access and a 2-step random access are used. In addition, for the 4-step random access and the 2-step random access, there are a contention-based random access, that is, a random access in which timing contention can occur with other mobile terminals, and a contention-free random access.

[0126] An example of a 4-step random access method based on contention is shown. As a first step, the mobile terminal transmits a random access preamble to the base station. The random access preamble can be selected by the mobile terminal from a prescribed range, or can be individually allocated to the mobile terminal and notified by the base station.

[0127] As the second step, the base station transmits a random access response to the mobile terminal. The random access response contains uplink scheduling information used in the third step, a terminal identifier used in the uplink transmission of the third step, and the like.

[0128] As the third step, the mobile terminal performs uplink transmission to the base station. The mobile terminal uses the information acquired in the second step in the uplink transmission. As the fourth step, the base station notifies the mobile terminal of the presence or absence of collision resolution. The mobile terminal that is notified of the absence of collision ends the random access processing. The mobile terminal that is notified of the presence of collision re-performs the processing from the first step.

[0129] In the four-step random access method without collision, the following point is different from the four-step random access method based on collision. That is, before the first step, the base station pre-allocates a random access preamble and uplink scheduling to the mobile terminal. In addition, the notification of the presence or absence of collision resolution in the fourth step is not needed.

[0130] An example of the two-step random access method based on collision is shown. As the first step, the mobile terminal performs transmission of a random access preamble and uplink transmission to the base station. As the second step, the base station notifies the mobile terminal of the presence or absence of collision. The mobile terminal that is notified of the absence of collision ends the random access processing. The mobile terminal that is notified of the presence of collision re-performs the processing from the first step.

[0131] In the two-step random access method without collision, the following point is different from the two-step random access method based on collision. That is, before the first step, the base station pre-allocates a random access preamble and uplink scheduling to the mobile terminal. In addition, in the second step, the base station transmits a random access response to the mobile terminal.

[0132] Figure 8 One example of the structure of a cell in NR is shown. In a cell of NR, a narrower beam is formed, and its direction is changed to perform transmission. In Figure 8 In the example shown, the base station 750 uses the beam 751-1 at a certain time to perform transceiving with the mobile terminal. At other times, the base station 750 uses the beam 751-2 to perform transceiving with the mobile terminal. The same applies to the use of one or more of the beams 751-3 to 751-8 by the base station 750 to perform transceiving with the mobile terminal. Thus, the base station 750 constitutes a wide-range cell 752.

[0133] In Figure 8 , an example in which the number of beams used by the base station 750 is set to eight is shown, but the number of beams can also be different from eight. In addition, in Figure 8 the example shown, the number of beams simultaneously used by the base station 750 is set to one, but can also be plural.

[0134] The identification of the beam uses the concept of QCL (Quasi-CoLocation) (see Non-Patent Literature 14 (3GPP TS 38.214)). That is, it is identified by information indicating which reference signal (for example, SS block, CSI-RS) the beam can be considered the same as. In this information, there are cases where information on the kind of the viewpoint that the same beam can be considered, such as information on Doppler shift, Doppler shift spread, average delay, average delay spread, spatial Rx parameter (see Non-Patent Literature 14 (3GPP TS 38.214)).

[0135] In 3GPP, due to D2D (Device to Device) communication, V2V (Vehicle to Vehicle) communication, a SL (Side Link) is supported (see Non-Patent Literature 1, Non-Patent Literature 16). The SL is specified by a PC5 interface.

[0136] In the SL communication, in order to support unicast and groupcast in addition to broadcast, the support of PC5-S signaling is being studied (see Non-Patent Literature 27 (3GPP TS 23.287)). For example, PC5-S signaling is performed in order to establish a SL, that is, a link for implementing PC5 communication. This link is implemented in the V2X layer, and is also referred to as a layer 2 link.

[0137] In addition, in the SL communication, the support of RRC signaling is being studied (see Non-Patent Literature 27 (3GPP TS 23.287)). The RRC signaling in the SL communication is also referred to as PC5 RRC signaling. For example, it is proposed to notify the capabilities of the UE between the UEs performing PC5 communication, or to notify the settings of the AS layer for performing V2X communication using PC5 communication, and the like.

[0138] Figure 9 An example of a connection structure of a mobile terminal in the SL communication is shown in FIG. 8. Figure 9 In the example shown, there are a UE 805, a UE 806 within the coverage 803 of a base station 801. Between the base station 801 and the UE 806, UL / DL communication 805 is performed. Between the base station 801 and the UE 806, UL / DL communication 808 is performed. Between the UE 805 and the UE 806, SL communication 810 is performed. There are a UE 811, a UE 812 outside the coverage 803. Between the UE 805 and the UE 811, SL communication 814 is performed. In addition, between the UE 811 and the UE 812, SL communication 816 is performed.

[0139] As an example of communication between a UE and a NW via a relay in the SL communication,Figure 9 The UE 805 shown relays communications between the UE 811 and the base station 801.

[0140] The UE that relays sometimes uses the same structure as Figure 4 The processing of relaying in the UE is described using Figure 4 The relaying processing of the UE 805 in the communication from the UE 811 to the base station 801 is described. The wireless signal from the UE 811 is received by the antennas 307-1 to 307-4. The received signal is converted from the wireless reception frequency to the baseband signal by the frequency conversion section 306, and demodulation processing is performed in the demodulation section 308. In the demodulation section 308, waiting calculation and multiplication processing can be performed. The demodulated data is transferred to the decoding section 309, and decoding processing such as error correction is performed. The decoded data is transferred to the protocol processing section 301, and protocol processing such as MAC, RLC, and the like for communication with the UE 811, removal of the header in each protocol, and the like are performed. In addition, protocol processing such as RLC, MAC, and the like for communication with the base station 801, addition of the header in each protocol, and the like are performed. In the protocol processing section 301 of the UE 811, protocol processing of PDCP, SDAP, and the like is sometimes also performed. The data on which the protocol processing has been performed is transferred to the encoding section 304, and encoding processing such as error correction is performed. There can also be data that is output from the protocol processing section 301 to the modulation section 305 without being subjected to the encoding processing. The data that has been subjected to the encoding processing by the encoding section 304 is subjected to modulation processing in the modulation section 305. Pre-coding in MIMO can also be performed in the modulation section 305. The modulated data is converted to a baseband signal, and then output to the frequency conversion section 306, and converted to a wireless transmission frequency. Thereafter, the transmission signal is transmitted from the antennas 307-1 to 307-4 to the base station 801.

[0141] In the above, one example of the relaying of the UE 805 in the communication from the UE 811 to the base station 801 is shown, but the same processing is used in the relaying of the communication from the base station 801 to the UE 811.

[0142] A base station of the 5G system can support Integrated Access and Backhaul (IAB) (see Non-Patent Literature 2, 20). The base station that supports IAB (hereinafter sometimes referred to as an IAB base station) is constituted by an IAB donor CU that is a CU of a base station that operates as an IAB donor, an IAB donor DU that is a DU of a base station that operates as an IAB donor, and an IAB node that connects using a wireless interface between the IAB donor DU and between UEs. An F1 interface is provided between the IAB node and the IAB donor CU (see Non-Patent Literature 2).

[0143] Figure 10 An example of connection of an IAB base station is shown in FIG. 16. The IAB donor CU 901 is connected to the IAB donor DU 902. The IAB node 903 is connected to the IAB donor DU 902 using a wireless interface. The IAB node 903 is connected to the IAB node 904 using a wireless interface. That is, multi-stage connection of IAB nodes is sometimes performed. The UE 905 is connected to the IAB node 904 using a wireless interface. The UE 906 is sometimes connected to the IAB node 903 using a wireless interface, and the UE 907 is sometimes connected to the IAB donor 902 using a wireless interface. A plurality of IAB donor DUs 902 can be connected to the IAB donor CU 901, a plurality of IAB nodes 903 can be connected to the IAB donor DU 902, and a plurality of IAB nodes 904 can be connected to the IAB node 903.

[0144] In connection between the IAB donor DU and the IAB node and connection between the IAB nodes, a BAP (Backhaul Adaptation Protocol) layer is provided (see Non-Patent Literature 29). The BAP layer performs an action of routing received data to the IAB donor DU and / or the IAB node, mapping to an RLC channel, and the like (see Non-Patent Literature 29).

[0145] As an example of the structure of the IAB donor CU, the same structure as the CU 215 is used.

[0146] As an example of the structure of the IAB donor DU, the same structure as the DU 216 is used. In the protocol processing section of the IAB donor DU, processing of the BAP layer, for example, assignment of a BAP header in downlink data, routing to the IAB node, removal of the BAP header in uplink data, and the like are performed.

[0147] As an example of the structure of the IAB node, a structure other than the EPC communication section 401, the other base station communication section 402, and the 5GC communication section 412 shown in FIG. 17 is sometimes used. Figure 5

[0148] Figure 5 Figure 10 ​​​The reception and transmission processing in the IAB node will be described. The reception and transmission processing of the IAB node 903 in the communication between the IAB donor CU 901 and the UE 905 will be described. In the uplink communication from the UE 905 to the IAB donor CU 901, the radio signal from the IAB node 904 is received by the antenna 408 (part or all of the antennas 408-1 to 408-4). The received signal is converted from the radio reception frequency to the baseband signal by the frequency conversion section 407, and demodulation processing is performed in the demodulation section 409. The demodulated data is transferred to the decoding section 410, and decoding processing such as error correction is performed. The decoded data is transferred to the protocol processing section 403, and protocol processing such as MAC, RLC, and the like for communication with the IAB node 904, removal of the header in each protocol, and the like are performed. In addition, routing to the IAB donor DU 902 using the BAP header is performed, and protocol processing such as RLC, MAC, and the like for communication with the IAB donor DU 902, assignment of the header in each protocol, and the like are performed. The data on which the protocol processing has been performed is transferred to the encoding section 405, and encoding processing such as error correction is performed. There can also be data that is output from the protocol processing section 403 to the modulation section 406 without performing encoding processing. The encoded data is subjected to modulation processing in the modulation section 406. Pre-coding in MIMO can also be performed in the modulation section 406. The modulated data is converted to a baseband signal and then output to the frequency conversion section 407, which converts it to a radio transmission frequency. Thereafter, the transmission signal is transmitted to the IAB donor DU 902 using the antennas 408-1 to 408-4. The same processing is also performed in the downlink communication from the IAB donor CU 901 to the UE 905.

[0149] In the IAB node 904, the same reception and transmission processing as the IAB node 903 is also performed. In the protocol processing section 403 of the IAB node 903, as the processing of the BAP layer, for example, assignment of the BAP header in the uplink communication and routing to the IAB node 904, removal of the BAP header in the downlink communication, and the like are performed.

[0150] In the standards of the mobile communication system in 3GPP, there is a DC (Dual Connectivity) function in which a UE is connected to two base stations in order to achieve high-speed and large-capacity communication. In order to satisfy the requirements of various KPIs while achieving high-speed and large-capacity communication, it is desirable to perform DC more efficiently. In 3GPP, the introduction of AI / ML into the mobile communication system has been started. The introduction of AI / ML into DC has also been proposed (Non-Patent Literature 32). However, in this proposal, only prediction information of the movement direction of the UE and the traffic amount of the UE is notified between the MN-SN, and nothing else is disclosed. It is not sufficient to perform more efficient DC by only this information, and various KPIs cannot be optimized.

[0151] In the present embodiment, a method is disclosed that enables AI / ML to be introduced into a DC and enables more efficient DC.

[0152] Learning of the model (sometimes referred to as training in the present specification) is described. In the present specification, the model (also includes the learned model) is sometimes referred to as an AI / ML model.

[0153] In the communication system related to the present embodiment, a learning device 1100 shown in Fig. 11 is used. Figure 11 The structure of the learning device 1100 is shown. Figure 11 Fig. 11 is a diagram showing an example of the structure of the learning device 1100 related to the communication system. The learning device 1100 includes a data acquisition unit 1110 and a model generation unit 1120. The learned model generated by the model generation unit 1120 is stored by a learned model storage unit 1130. Figure 11 An example in which the learned model storage unit 1130 is present outside the learning device 1100 is shown in Fig. 11, but the learning device 1100 can also be configured to include the learned model storage unit 1130. Figure 11 The learning device 1100 shown in Fig. 11 can be applied to a base station.

[0154] The data acquisition unit 1110 acquires input information and feedback information as learning data.

[0155] The model generation unit 1120 learns output information based on the learning data including the input information and the feedback information. That is, a learned model that infers output information from input information of the communication system is generated.

[0156] The learning algorithm used by the model generation unit 1120 can use known algorithms such as supervised learning, unsupervised learning, reinforcement learning, and the like. As one example, a case in which reinforcement learning is applied is described. In reinforcement learning, an agent (action subject) within a certain environment observes the current state (parameters of the environment) and decides the action to be taken. According to the action of the agent, the environment dynamically changes, and the agent receives a reward according to the change in the environment. The agent repeats this process and learns the action policy that gives the most reward through a series of actions. As representative methods of reinforcement learning, Q-learning and TD-learning are known. For example, in the case of Q-learning, the general update formula of the action value function Q(s, a) is represented by formula (1).

[0157] [Mathematical formula 1]

[0158] In formula (1), s t represents the state of the environment at time t, and at represents an action at time t. By the action a t , the state becomes s t+1 t+1 represents a reward obtained according to a change in state, γ represents a discount rate, and a represents a learning coefficient. In addition, γ is set to a range of 0 < γ ≤ 1, and a is set to a range of 0 < a ≤ 1. Using input information is set to an action a t , a state s t , and the optimal action a t at the state s t at time t is learned.

[0159] If the action value Q of the action a of the highest Q value at time t + 1 is larger than the action value Q of the action a performed at time t, the update formula shown in Equation (1) makes the action value Q larger, and vice versa. In other words, the action value function Q(s, a) is updated so that the action value Q of the action a at time t approaches the optimal action value at time t + 1. Thus, the optimal action value under a certain environment is propagated to the action value under the previous environment in turn.

[0160] As described above, in a case where a learning completion model is generated by reinforcement learning, the model generation unit 1120 includes a reward calculation unit 1121 and a function update unit 1122.

[0161] The reward calculation unit 1121 calculates a reward on the basis of input information and feedback information. The reward calculation unit 1121 calculates a reward r on the basis of a reward criterion. For example, the reward r is increased (for example, a reward of "1" is given) in a case of a reward increase criterion, and on the other hand, the reward r is decreased (for example, a reward of "-1" is given) in a case of a reward decrease criterion.

[0162] The function update unit 1122 updates a function for determining an output in accordance with a reward calculated by the reward calculation unit 1121, and outputs to the learning completion model storage unit 1130. For example, in a case of Q-learning, a function for calculating the action value function Q(s t , a t ) shown in Equation (1) is used as output information.

[0163] The learning device 1100 repeatedly performs learning as described above. The learning completion model storage unit 1130 stores the action value function Q(s t , a t ) updated by the function update unit 1122, that is, a learning completion model.

[0164] Next, a process in which the learning device 1100 performs learning will be described using Figure 12 is a flowchart relating to the learning process of the learning device. Figure 12 is a flowchart relating to the learning process of the learning device.​

[0165] In step bl, the data acquisition unit 1110 acquires input information and feedback information as learning data.

[0166] In step b2, the model generation unit 1120 calculates a reward on the basis of the input information and the feedback information. Specifically, the reward calculation unit 1121 acquires the input information and the feedback information, and determines whether to increase the reward (step b3) or decrease the reward (step b4) on the basis of a predetermined reward criterion.

[0167] In a case where it is determined to increase the reward, the reward calculation unit 1121 increases the reward in step b3. On the other hand, in a case where it is determined to decrease the reward, the reward calculation unit 1121 decreases the reward in step b4.

[0168] In step b5, the function update unit 1122 updates the action value function Q(s t , a t ) represented by Expression (1) stored in the learning completed model storage unit 1130 on the basis of the reward calculated by the reward calculation unit 1121.

[0169] The learning device 1100 repeatedly executes the steps of steps bl to b5 above, and stores the generated action value function Q(s t , a t ) in the learning completed model storage unit 1130 as a learning completed model.

[0170] An explanation will be given of inference using the learning completed model.

[0171] Figure 13 is a block diagram of an inference device 1200 related to a communication system. The inference device 1200 includes a data acquisition unit 1210 and an inference unit 1220.

[0172] The data acquisition unit 1210 acquires input information.

[0173] The inference unit 1220 infers output information using the learning completed model. That is, by inputting the input information acquired by the data acquisition unit 1210 to the learning completed model, it is possible to infer output information appropriate for the input information.

[0174] In addition, in the present embodiment, an explanation is given of outputting output information using the learning completed model learned by the model generation unit 1120 of the communication system, but it is also possible to acquire the learning completed model from another communication system, and output output information on the basis of the learning completed model.

[0175] Figure 13The learning completion model storage section 1130 is shown as existing outside the inference device 1200, but the inference device 1200 can include the learning completion model storage section 1130. Figure 13 The inference device 1200 shown can be applied to a base station. The inference device 1200 can be applied to the same node as the learning device 1100, or to a different node.

[0176] Next, a process for obtaining output information using the inference device 1200 will be described. Figure 14

[0177] In step cl, the data acquisition section 1210 acquires input information.

[0178] In step c2, the inference section 1220 acquires a learning completion model from the learning completion model storage section 1130, inputs the input information to the acquired learning completion model, and obtains output information.

[0179] In step c3, the inference section 1220 outputs the output information obtained by the learning completion model to the communication system.

[0180] In step c4, the communication system performs a communication process using the output output information. Thus, a more efficient communication process using a learning completion model can be performed.

[0181] In addition, in the present embodiment, a case where reinforcement learning is applied to a learning algorithm used by the inference section 1220 is described, but the present application is not limited to this. With respect to the learning algorithm, in addition to reinforcement learning, supervised learning, unsupervised learning, or semi-supervised learning, or the like can be applied.

[0182] Further, as the learning algorithm used in the inference section 1220, deep learning that extracts a learning feature amount itself can be used, and machine learning can be performed according to other publicly known methods, such as a neural network, genetic programming, inductive logic programming, a support vector machine, or the like.

[0183] In addition, the learning device 1100 and the inference device 1200 can be connected to a communication system via a network, for example. Further, the learning device 1100 and the inference device 1200 can be built into the communication system. Further, the learning device 1100 and the inference device 1200 can exist on a cloud server.

[0184] ​Further, the model generation unit 1120 can learn the output information using the learning data acquired from the plurality of nodes. In addition, the model generation unit 1120 can acquire the learning data from the plurality of nodes used in the same area, or can learn the output information using the learning data collected from the plurality of nodes independently operating in different areas. Further, the nodes that collect the learning data can be added to the object in the middle, or can be removed from the object. Furthermore, the learning device 1100 that has learned the output information with respect to a certain node can be applied to a different node, or the output information can be relearned with respect to the other node and updated.

[0185] AI / ML is introduced for DC. In the present embodiment, a DC AI / ML model is provided to a base station. For example, the DC AI / ML model is used to perform selection of SN. The DC AI / ML model can be provided to a MN and / or a SN. For example, the DC AI / ML model is used to perform selection of an SN change target, selection of a PSCell change target.

[0186] The base station can transmit information on whether or not the DC AI / ML model is held (information on whether or not the model is held can also be transmitted) to a UE. The transmission can use RRC signaling. The base station can transmit information on whether or not the DC AI / ML model is held to a neighboring base station. The transmission can use Xn signaling. The base station can transmit information on whether or not the DC AI / ML model is held to a CN (Core Network) node (function can also be used). As the CN node, for example, a NWDAF (Network Data Analysis Function) can be used. The base station can transmit information on whether or not the DC AI / ML model is held to a management node. As the management node, for example, a MnS (Management Service) node or an OAM (Operation, Administration and Management) can be used. Thereby, the UE or node to which the information is notified can recognize whether or not the source base station of the notification holds the DC AI / ML model. It is possible to recognize whether or not AI / ML can be introduced in DC using the base station. The MN and / or the SN can perform the transmission. The UE or node to which the information is notified can recognize whether or not the MN and / or the SN for DC holds the DC AI / ML model.

[0187] Input information for the DC AI / ML model is provided. Twenty-six examples of the DC AI / ML model input information are disclosed below.

[0188] (1) Information on the position of the UE.

[0189] (2) Information on radio measurement of the UE.

[0190] (3) DC history information of the UE.

[0191] (4) DC history information of the UE from a neighboring base station.

[0192] (5) Information related to performance of the UE which has performed DC.

[0193] (6) Resource status of the current base station.

[0194] (7) Predicted resource status of the base station.

[0195] (8) Past DC failure information of the UE.

[0196] (9) Past DC failure information of the UE from a neighboring base station.

[0197] (10) Past communication interruption status of the UE.

[0198] (11) Past communication interruption status of the UE from a neighboring base station.

[0199] (12) Information related to trajectory of the UE.

[0200] (13) Information related to current traffic volume of the UE.

[0201] (14) Information related to predicted traffic volume of the UE.

[0202] (15) Information related to L2 (Layer 2) measurement of the UE.

[0203] (16) Predicted value of information related to L2 measurement of the UE.

[0204] (17) Information related to L2 measurement from a neighboring base station.

[0205] (18) Predicted value of information related to L2 measurement from a neighboring base station.

[0206] (19) Connected base station history information of the UE.

[0207] (20) Connected base station history information of the UE from a neighboring base station.

[0208] (21) Cell selection, cell reselection history information of the UE.

[0209] (22) Information related to MDT of the UE.

[0210] (23) Information related to QoE (Quality of Experience) measurement of the UE.

[0211] (24) Information related to the amount of data of the SN.

[0212] (25) Information related to UL data split.

[0213] (26) A combination of (1) to (25).

[0214] (1) For example, it can be coordinates, a serving cell identification, information identifying a DU to which the UE is connected, information identifying a TRP to which the UE is connected. In addition, it can be a moving speed. In addition, it can include information related to time. It is possible to associate the position of the UE and its time. It is possible to set the position or state of the UE as input information for training or inference.

[0215] (2) For example, it can be a wireless measurement result of the UE. For example, it can be RSRP, RSRQ, SINR (Signal to Interference Noise Ratio). It is possible to include information related to the position of the UE. It is possible to include information related to time. It is possible to associate the wireless measurement result with information about the position or time. It is possible to set the reception quality based on the position or state of the UE as input information for training or inference.

[0216] (3) For example, it can be information of a base station with which the UE has performed DC in the past. For example, it can be information identifying the MN, information identifying the SN. It can be setting information of DC. It can be setting information of MCG, SCG. It can be information identifying PCell, PSCell, SPCell. It can be setting information of PCell, PSCell, SPCell. For example, it can be information related to activation, deactivation of the SN. It can be information related to the terminal of DC. It can be information of the MN terminal or the SN terminal. It can be information related to the bearer of DC. It can be setting information of the bearer. It can be information of MCG bearer, SCG bearer, or split bearer. It can be information related to RLC bearer. It can be information related to path. For example, it can be information related to primary path or secondary path. It can be a list of information of a base station with which the UE has performed DC in the past. In addition, it is possible to include information related to the position of the UE. It is possible to include information related to time. It is possible to associate the information of the base station with which DC has been performed in the past with information about the position or time. It is possible to set the information about DC as input information for training or inference.

[0217] (4) is DC history information of the UE that the neighboring base station has. The DC history information of the UE can be the information disclosed in (3).

[0218] (5) For example, it can be the QoS of the UE performing DC. It can be the packet loss rate or the delay time. It can be information before performing DC. It can be information after performing DC. It can be information related to the performance of the UE of each MN or SN. Furthermore, it can include information related to the location of the UE. It can include information related to time. It can be possible to associate information on the performance of the UE performing DC with information on the location or time. Furthermore, it can be possible to associate with the history information of the DC of the UE of (3) or (4). It can be possible to set the communication performance of DC as input information for training or inference.

[0219] (6) For example, it can be the current resource usage amount of the base station for DC. For example, it can be the resource usage amount of the MN or the SN. It can be the resource usage amount of each cell, or the resource usage amount of each TRP. It can be the usage amount of the frequency resource. It can be the usage amount of the time resource. The usage amount of the frequency resource is the number of carriers, the number of RBs, the number of subcarriers, etc. It can be in the unit of carrier, in the unit of PRB (Physical Resource Block), or in the unit of subcarrier. The usage amount of the time resource is the number of subframes, the number of slots, the number of symbols, etc. It can be in the unit of subframe, in the unit of slot, or in the unit of symbol. It can be possible to set information on the load of the base station as input information for training or inference.

[0220] (7) For example, it can be information on the predicted resource of the base station for DC. The information on the resource of the base station can be the information disclosed in (6).

[0221] (8) is information about past DC processing failure of the UE. The DC processing is, for example, the DC processing described in Non Patent Literature 22. For example, SN addition processing, SN change processing, SN correction processing, SN release processing, and the like. The information about the DC processing failure can be, for example, past DC failure history information of the UE. For example, it can be information about a base station that the UE has failed in the past in the DC processing. It can be information that identifies an MN or an SN that the UE has failed in the past in the DC processing. It can be information that identifies a PCell, a PSCell, an SPCell that the UE has failed in the past in the DC processing. It can be information about a terminal of a DC that the UE has failed in the past in the DC processing. It can be information of an MN terminal or an SN terminal. It can be information about a bearer of a DC that the UE has failed in the past in the DC processing. It can be information of an MCG bearer, an SCG bearer, or a split bearer. It can be information about an RLC bearer. It can be information about a path. For example, it can be information about a primary path or a secondary path. It can be a list of information of a base station that the UE has failed in the past in the DC processing. Furthermore, it can include information about a location of the UE. It can include information about time. It can associate information about past DC failure with information about a location or time. It can set information about DC processing failure as input information for training or inference.

[0222] (9) is information about past DC processing failure of the UE by a neighboring base station. The information about past DC failure of the UE can be the information disclosed in (8).

[0223] (10) is information about past communication interruption of the UE in DC. For example, it can be historical information about past communication interruption of the UE. As the communication interruption, for example, it can be RLC (Radio Link Failure). For example, it can be information about communication interruption in the MN or the SN. For example, it can be information identifying the MN or the SN in which the communication interruption occurred. For example, it can be information identifying the PCell, the PSCell, the SPCell in which the communication interruption occurred. It can be information about the terminal of the DC in which the UE past communication interruption occurred. It can be information of the MN terminal or the SN terminal. It can be information about the bearer of the DC in which the UE past communication interruption occurred. It can be information of the MCG bearer, the SCG bearer, or the split bearer. It can be information about the RLC bearer. It can be information about the path. For example, it can be information about the primary path or the secondary path. It can be a list of information of the MN, the SN, the PCell, the PSCell, the SPCell, and the like in which the UE past communication interruption occurred. Furthermore, it can contain information about the location of the UE. It can contain information about the time. It can associate information about past communication interruption with information about the location or the time. It can set information about communication interruption as input information for training or inference.

[0224] (11) is information about past communication interruption of the UE in DC by the neighboring base station. The information about past communication interruption of the UE can be the information disclosed in (10).

[0225] (12) In (12), as the information about the trajectory of the UE, for example, the information about the location of the UE disclosed in (1) can be appropriately applied. For example, it can be the information of the base station in which DC is performed disclosed in (3). For example, it can be the information of the connected base station disclosed in (19). For example, it can be the information of the base station after cell selection, cell reselection disclosed in (21). It can be information about the past, current trajectory of the UE. The information about the past, current trajectory of the UE can be historical information. The information about the trajectory of the UE can be information about the future trajectory of the UE. The information about the future trajectory of the UE can be predicted information. These information can be listed. Furthermore, it can contain information about the time. It can associate information about the trajectory of the UE with information about the time. It can set information about the trajectory of the UE as input information for training or inference.

[0226] In (13), the traffic volume of the UE can be, for example, the number of connected UEs. For example, it can be the current traffic volume of the UE in the MN or the SN. For example, it can be the current traffic volume of the UE in the PCell, the PSCell, or the SPCell. Information on the load of the base station can be set as input information for training or inference.

[0227] (14) is, for example, information on the predicted traffic volume of the UE. The information on the traffic volume of the UE can be the information disclosed in (13).

[0228] (15) can be, for example, a measurement result of the UE on L2 measurement information. The L2 measurement information can be, for example, information measured by the UE as described in Non-Patent Literature 33. For example, it can be the packet delay amount. As the L2 measurement result of the UE, it can be the L2 measurement result of the UE in the DC. For example, it can be the L2 measurement result of the UE for the MN or the SN. For example, it can be the L2 measurement result of the UE for the PCell, the PSCell, or the SPCell. Furthermore, it can include information on the position of the UE. It can include information on the time. Information on the L2 measurement of the UE can be associated with information on the position or the time. Information on the communication performance with the base station can be set as input information for training or inference.

[0229] (16) is a predicted value of the L2 measurement information. is a predicted value of the L2 measurement information measured by the UE. It can be a predicted value of the L2 measurement information measured by the UE in the DC.

[0230] (17) can be, for example, an L2 measurement result of a base station. Can be an L2 measurement result that a neighboring base station has. The L2 measurement information of a base station can be, for example, information measured by a base station as described in Non-Patent Literature 33. For example, the number of received RA preambles, the amount of packet delay, the number of UEs in a connected state, the number of stored contexts of UEs in an inactive state, the packet loss rate, PRB information for MIMO (Multiple Input Multiple Output), the number of PDCP packets in a split DRB (Data Radio Bearer), the total RAN (Radio Access Network) delay amount in a split DRB, and the like. The L2 measurement result of a base station can be an L2 measurement result of a base station in DC. For example, can be an L2 measurement result in an MN or an SN. For example, can be an L2 measurement result in a PCell, a PSCell, or an SPCell. Can be an L2 measurement result of a base station that a neighboring base station has. Furthermore, for the L2 measurement result related to a UE, can include information related to the location of the UE. Can include information related to time. Can associate information on L2 measurement of a UE with information on location or time. Can set information on the communication performance with a UE as input information for training or inference.

[0231] (18) is a predicted value of L2 measurement information. is a predicted value of L2 measurement information measured by a base station. Can be a predicted value of L2 measurement information measured by a base station in DC. Can be a predicted value of L2 measurement information of a base station that a neighboring base station has.

[0232] (19) can be, for example, information of a base station to which a UE has connected in the past. For example, can be information identifying a base station. Can be information identifying a cell. Can be setting information of a cell or a cell group. Can be a list of information of a base station to which a UE has connected in the past. Furthermore, can include information related to the location of the UE. Can include information related to time. Can associate information on a base station to which a UE has connected in the past with information on location or time. The base station to which a UE has connected in the past can not be limited to a base station that performs DC. DC can be performed from which a more appropriate base station is selected from more base stations.

[0233] (20) is history information of a base station to which a UE has connected that a neighboring base station has. The connection base station history information of a UE can be information disclosed in (3).

[0234] (21) For example, it can be information of a base station where the UE has performed cell selection and / or cell reselection in the past. For example, it can be information identifying a base station. It can be information identifying a cell. It can be setting information of a cell or a cell group. It can be a list of information of a base station where the UE has performed cell selection and / or cell reselection in the past. Further, it can include information related to a location of the UE. It can include information related to time. It can be possible to associate information about a base station where cell selection and / or cell reselection has been performed in the past with information about a location or time. It can be possible to input information of the UE in an RRC_Idle or RRC_Inactive state. It can be possible to perform more suitable DC from more base stations.

[0235] (22) For example, it can be a measurement result of MDT (Minimization of Drive Tests). For example, it can be a measurement result of immediate MDT, or a measurement (Logged Measurement) result of logged MDT. It can be a measurement result of management-based MDT. It can be a measurement result of signaling-based MDT. It can be a measurement result of a trace function. It can be possible to associate a measurement result of MDT with information about a location or time of the UE. Information about a location or time of the UE can use information measured by MDT. By setting MDT for the UE, it can be possible to obtain a measurement result of MDT as input information for an AI / ML model for DC. It can be possible to input more states or more information of the UE.

[0236] (23) For example, it can be a QoE measurement result (sometimes also referred to as an application layer measurement result). It can be possible to associate a QoE measurement result with information about a location or time of the UE. By setting QoE measurement for the UE, it can be possible to obtain a measurement result of QoE as input information for an AI / ML model for DC. It can be possible to input measurement information in an application layer of the UE as input information.

[0237] (24) For example, it can be the amount of data for which transmission and / or reception is successful between the SN and the UE. The amount of data can be, for example, the number of packets. For example, in the DL, it can be the number of bytes of SDAP SDUs (Service Data Units) or PDCP SDUs that are successfully transmitted to the UE. In the UL, for example, it can be the number of bytes of SDAP SDUs or PDCP SDUs that are successfully received from the UE. Information on the amount of data for the SN can be associated with information on the location or time of the UE. This information can be for each PDU session, or it can be the sum of multiple PDU sessions. Alternatively, it can be the sum of all PDU sessions. This information can be for each QoS flow. This information can include the RAT (Radio Access Technology) type of the SN. By setting the QoE measurement for the UE, the measurement result of the QoE can be obtained as input information for the DC AI / ML model. Measurement information in the application layer of the UE can be input as input information.

[0238] (25) For example, it can be information on the setting of UL data splitting. For example, it can be UL data splitting threshold information. For example, it can be information on the number of times or the time at which the threshold is exceeded. For example, it can be information on which RAN node the UL data is transmitted to. For example, it can be information on the amount of transmitted UL data to each RAN node. Information on UL data splitting can be associated with information on the location or time of the UE. For example, in the case where a split bearer is set for UL, the threshold or the amount of transmitted UL data can be set as input information.

[0239] The base station with the DC AI / ML model inputs the above-disclosed input information for the DC AI / ML model into the DC AI / ML model for training or inference. As a result, more efficient DC can be performed using the DC AI / ML model.

[0240] The base station transmits the input information for the AI / ML model for DC to a neighboring base station. The base station can transmit the input information for the AI / ML model for DC to a base station having the AI / ML model for DC. The base station having the AI / ML model for DC can transmit a request for the input information for the AI / ML model for DC to a neighboring base station. The request can include information about the UE as a DC target. The information about the UE can be an identification of the UE. The base station receiving the request transmits the input information for the AI / ML model for DC to the base station of the request source. The transmission of the input information for the AI / ML model for DC between the base stations, the request for the information can use inter-base station signaling. For example, Xn signaling can be used. For example, an Xn message can be set for the predicted value. The MN and / or the SN for DC can make these transmissions. The transmission can use inter-base station signaling. For example, Xn signaling can be used. For example, an Xn message can be set for the predicted value. The transmission can be made in the DC process.

[0241] The DU transmits the input information for the AI / ML model for DC to a CU. The DU can transmit the input information for the AI / ML model for DC to a CU having the AI / ML model for DC. The CU having the AI / ML model for DC can transmit a request for the input information for the AI / ML model for DC to a DU. The request can include information about the UE as a DC target. The information about the UE can be an identification of the UE. The DU receiving the request transmits the input information for the AI / ML model for DC to the CU of the request source. The transmission of the input information for the AI / ML model for DC between the CU and the DU, the request for the information can use inter-CU and DU signaling. For example, F1 signaling can be used. For example, an F1 message can be set for the predicted value. These transmissions can be made between the CU and the DU constituting the MN and / or the SN for DC. The transmission can use inter-CU and DU signaling. For example, F1 signaling can be used. For example, an F1 message can be set for the predicted value. The transmission can be made in the DC process.

[0242] The UE transmits the input information for the AI / ML model for DC to the base station. The UE can transmit the input information for the AI / ML model for DC to the base station having the AI / ML model for DC. The base station having the AI / ML model for DC can transmit a request for the input information for the AI / ML model for DC to the UE. The request can include information about the UE as a target of DC. The information about the UE can be an identification of the UE. The UE receiving the request transmits the input information for the AI / ML model for DC to the base station as a source of the request. The transmission of the input information for the AI / ML model for DC between the UE and the base station and the request for the information can use signaling between the UE and the base station. For example, RRC signaling can be used. As another method, MAC signaling or L1 / L2 signaling can be used. The information and the request for the information can be transmitted as early as possible. The transmissions can be made between the UE and the MN and / or the SN for DC. The UE can transmit the input information for the AI / ML model for DC for each of the MN and the SN to the corresponding MN or SN, respectively. The MN or the SN can acquire the information for the node from the UE. As another method, the UE can transmit the input information for the AI / ML model for DC for each of the MN and the SN to either of the nodes. The node can acquire the information from the UE. The node receiving the input information for the AI / ML model for DC from the UE can transmit the information to the node having the AI / ML model for DC. The node having the AI / ML model for DC can acquire the information from the UE. The transmission of the input information for the AI / ML model for DC between the UE and the MN and / or the SN for DC and the request for the information can use signaling between the UE and the base station. For example, RRC signaling can be used. As UE-SN inter-signaling, SRB (Signaling Radio Bearer) 3 can be used, for example. As another method, MAC signaling or L1 / L2 signaling can be used. The transmission of the input information for the AI / ML model for DC from the UE or the request for the information for the UE between the MN and the SN can use inter-base station signaling. For example, Xn signaling can be used. For example, an Xn message can be set for the predicted value. The transmission can be made in DC processing.

[0243] The output information of the AI / ML model for DC is disclosed. Ten examples of the output information of the AI / ML model for DC are disclosed below.

[0244] (1) Information about a predicted trajectory of the UE.

[0245] (2) DC target information predicted.

[0246] (3) Priority order of a DC target predicted.

[0247] (4) Candidate PSCell information predicted.

[0248] (5) The probability of arrival of the predicted candidate PSCell.

[0249] (6) The predicted DC processing time.

[0250] (7) The predicted resource status of the base station.

[0251] (8) Information related to the predicted traffic amount of the UE.

[0252] (9) Information related to the predicted UL data split.

[0253] (10) A combination of (1) to (9).

[0254] (1) For example, it can be the predicted future trajectory information of the UE. The information related to the trajectory of the UE can be set as the information disclosed in input information example (12) for the DC AI / ML model. Information related to the past and current trajectory can be included. The history information of the past and current trajectory of the UE can be included together with the predicted future trajectory information of the UE. These information can be listed. In addition, information related to time can be included. The information on the trajectory of the UE can be associated with the information on time.

[0255] (2) is the DC target information predicted by the AI / ML model. The DC target information can be, for example, information related to the base station for DC in the DC processing described in Non-Patent Literature 22. For example, it can be information identifying the SN selected for DC. It can be information identifying the PCell, PSCell, or SPCell selected for DC. The DC target information can be, for example, setting information for DC. It can be MCG setting information. It can be SCG setting information. The DC target information can be, for example, information related to the terminal for DC. It can be information of the MN terminal or the SN terminal. The DC target information can be, for example, information related to the bearer for DC. It can be information of the MCG bearer, SCG bearer, or split bearer. It can be information related to the RLC bearer. In addition, information related to the predicted position of the UE can be included. Information related to the predicted time can be included. The predicted DC target information can be associated with the information on the position or time. The predicted DC target information is not limited to one, and can be plural. Information on the prediction accuracy of the DC target information predicted by the AI / ML model can be included.

[0256] (3) is information on the priority order of the predicted DC target shown by the information disclosed in (2). The predicted DC target and its priority order can be associated. For example, in the SN addition processing, a plurality of predicted SNs can be output. The plurality of output predicted SNs are respectively given a priority order.

[0257] (4) is prediction information of a candidate of a PSCell in conditional PSCell change (CPC) processing. The candidate of the PSCell is not limited to one, and can be plural. Information of prediction accuracy of the predicted PSCell candidate can be included.

[0258] (5) is prediction information of a reaching probability of a candidate of a PSCell in conditional PSCell change (CPC) processing by the UE. Information of a confidence interval of the reaching probability of the candidate PSCell predicted by the AI / ML model can be included.

[0259] (6) is information of a predicted DC processing time. The predicted DC processing time can be set for each of the DC target information, for example. The predicted DC processing time can be set for each of the PSCell candidate, for example. The predicted DC processing time can be associated with the predicted DC target information or the predicted PSCell candidate information.

[0260] (7) is information of a predicted resource status of a base station. The information of the resource status of the base station can be set as the information disclosed in the input information example (7) for the DC AI / ML model. The resource status of the base station can be the predicted resource status in the predicted DC target of (2) or the predicted PSCell candidate of (4).

[0261] (8) is information related to a predicted traffic amount of the UE. The information related to the traffic amount of the UE can be set as the information disclosed in the input information example (13) for the DC AI / ML model. The information related to the predicted traffic amount of the UE can be the predicted UE traffic status in the predicted DC target of (2) or the predicted PSCell candidate of (4).

[0262] (9) can be, for example, setting information of a predicted UL data split. It can be, for example, predicted UL data split threshold information. It can be set as the information disclosed in the input information example (25) for the DC AI / ML model.

[0263] The output information of the DC AI / ML model disclosed above is appropriately notified to the RAN node or the UE that performs the DC processing. Thus, more efficient DC can be performed using the DC AI / ML model.

[0264] The base station transmits the DC AI / ML model output information to a neighboring base station. The base station with the DC AI / ML model can transmit the DC AI / ML model output information to the base station related to the DC. The base station related to the DC can transmit a request for the DC AI / ML model output information to the base station with the DC AI / ML model. The base station receiving the request transmits the DC AI / ML model output information to the request source base station. The DC AI / ML model output information between the base stations, the transmission of the request for the information can use the inter-base station signaling. For example, the Xn signaling can be used. For example, the Xn message can be set for the prediction value. The MN and / or the SN for the DC can make these transmissions. The transmission can use the inter-base station signaling. For example, the Xn signaling can be used. For example, the Xn message can be set for the prediction value. The transmission can be made in the DC process.

[0265] The CU transmits the DC AI / ML model output information to the DU. The CU with the DC AI / ML model can transmit the DC AI / ML model output information to the DU. The DU can transmit a request for the DC AI / ML model output information to the CU with the DC AI / ML model. The CU receiving the request transmits the DC AI / ML model output information to the request source DU. The DC AI / ML model output information between the CU and the DU, the transmission of the request for the information can use the inter-CU-DU signaling. For example, the F1 signaling can be used. For example, the F1 message can be set for the prediction value. Between the CU-DU composed of the MN and / or the SN for the DC, these transmissions can be made. The transmission can use the inter-CU-DU signaling. For example, the F1 signaling can be used. For example, the F1 message can be set for the prediction value. The transmission can be made in the DC process.

[0266] The base station transmits the AI / ML model output information for DC to the UE. The base station with the AI / ML model for DC can transmit the AI / ML model output information for DC to the UE. The UE can transmit a request for the AI / ML model output information for DC to the base station with the AI / ML model for DC. The base station receiving the request transmits the AI / ML model output information for DC to the UE that is the source of the request. The transmission of the AI / ML model output information for DC between the UE and the base station and the transmission of the request for the information can use the signaling between the UE and the base station. For example, RRC signaling can be used. As another method, MAC signaling or L1 / L2 signaling can be used. The transmission of the information and the transmission of the request for the information can be performed as early as possible. The transmissions can be performed between the UE and the MN and / or the SN for DC. The AI / ML model output information for DC of each of the MN and the SN can be transmitted to the UE from the corresponding MN or SN, respectively. The UE can acquire the information from the MN or the SN. As another method, the AI / ML model output information for DC of each of the MN and the SN can be transmitted to the UE from the node of either party. The UE can acquire the information from the node. The transmission of the AI / ML model output information for DC between the UE and the MN and / or the SN for DC and the transmission of the request for the information can use the signaling between the UE and the base station. For example, RRC signaling can be used. As the inter-UE-SN signaling, for example, SRB3 can be used. As another method, MAC signaling or L1 / L2 signaling can be used. The transmission of the AI / ML model output information for DC between the MN and the SN from the UE or the request for the information for the UE can use the inter-base-station signaling. For example, Xn signaling can be used. For example, an Xn message can be provided for the predicted value. The transmission can be performed in the DC process.

[0267] Feedback information for the DC process using the AI / ML model for DC is disclosed. Five examples of the feedback information for the DC process using the AI / ML model for DC are disclosed below.

[0268] (1) Information related to the performance of the UE.

[0269] (2) Information related to the L2 measurement of the UE.

[0270] (3) Information related to the L2 measurement of the base station.

[0271] (4) Resource status of the base station.

[0272] (5) Combination of (1) to (4).

[0273] (1) For example, it can be (5) of the above example of the input information for the AI / ML model for DC.

[0274] (2) For example, it can be (15) of the above example of the input information for the AI / ML model for DC.

[0275] (3) For example, it can be (17) of the above-described input information example for the DC AI / ML model.

[0276] (4) For example, it can be (6) of the above-described input information example for the DC AI / ML model.

[0277] The DC-related node appropriately notifies the node having the DC AI / ML model of the DC processing feedback information using the DC AI / ML model disclosed above. Thereby, it is possible to evaluate the reward calculation of the DC AI / ML model and the like, and it is possible to update the model. Thereby, it is possible to perform more efficient DC.

[0278] The base station transmits DC processing feedback information using the DC AI / ML model to a neighboring base station. The DC-related base station can transmit DC processing feedback information using the DC AI / ML model to a base station having the DC AI / ML model. The base station having the DC AI / ML model can transmit a request for DC processing feedback information using the DC AI / ML model to the DC-related base station. The base station that receives the request transmits DC processing feedback information using the DC AI / ML model to the base station that is the source of the request. The transmission of DC processing feedback information using the DC AI / ML model between the base stations, and the request for the information can use inter-base station signaling. For example, Xn signaling can be used. The MN and / or the SN for DC can perform these transmissions. The transmission can use inter-base station signaling. For example, Xn signaling can be used. The transmission can be performed in the DC processing.

[0279] The DU transmits DC processing feedback information using the DC AI / ML model to the CU. The DC-related DU can transmit DC processing feedback information using the DC AI / ML model to the CU. The DC-related DU can transmit DC processing feedback information using the DC AI / ML model to the CU having the DC AI / ML model. The CU having the DC AI / ML model can transmit a request for DC processing feedback information using the DC AI / ML model to the DU. The DU that receives the request transmits DC processing feedback information using the DC AI / ML model to the CU that is the source of the request. The transmission of DC processing feedback information using the DC AI / ML model between the CU and the DU, and the request for the information can use inter-CU-DU signaling. For example, F1 signaling can be used. These transmissions can be performed between the CU and the DU that constitute the MN and / or the SN for DC. The transmission can use inter-CU-DU signaling. For example, F1 signaling can be used. The transmission can be performed in the DC processing.

[0280] The UE transmits DC processing feedback information using the DC AI / ML model to the base station. The UE can transmit DC processing feedback information using the DC AI / ML model to the base station having the DC AI / ML model. The base station having the DC AI / ML model can transmit a request for the UE to transmit DC processing feedback information using the DC AI / ML model. The UE receiving the request transmits DC processing feedback information using the DC AI / ML model to the base station of the request source. The transmission of DC processing feedback information using the DC AI / ML model and the request for the information between the UE and the base station can use UE-to-base station signaling. For example, RRC signaling can be used. As another method, MAC signaling or L1 / L2 signaling can be used. The information and the request for the information can be transmitted as early as possible. Between the UE and the MN and / or the SN for DC, these transmissions can be made. The UE can transmit DC processing feedback information using the DC AI / ML model for each of the MN and the SN to the corresponding MN or SN, respectively. The MN or the SN can acquire the information for the node from the UE. As another method, the UE can transmit DC processing feedback information using the DC AI / ML model for each of the MN and the SN to either node. The node can acquire the information from the UE. The node receiving DC processing feedback information using the DC AI / ML model from the UE can transmit the information to the node having the DC AI / ML model. The node having the DC AI / ML model can acquire the information from the UE. The transmission of DC processing feedback information using the DC AI / ML model and the request for the information between the UE and the MN and / or the SN for DC can use UE-to-base station signaling. For example, RRC signaling can be used. As UE-to-SN signaling, SRB3 can be used, for example. As another method, MAC signaling or L1 / L2 signaling can be used. The transmission of the input information for the DC AI / ML model from the UE or the request for the information for the UE between the MN and the SN can use base station-to-base station signaling. For example, Xn signaling can be used. The transmission can be made in DC processing.

[0281] Figure 15 FIG. 11 is a sequence diagram illustrating an example of SN addition processing using the DC AI / ML. An example in which a base station that is a MN for DC has a DC AI / ML model and performs training and inference is disclosed. In step ST1101, the MN is provided with a DC AI / ML model. In step ST1102, the MN performs measurement setting for the UE. For example, radio measurement setting is performed. The setting can use RRC signaling. An RRC Reconfiguration message can be used.

[0282] The UE performs measurement in step ST1103. The measurement performed by the UE can be, for example, radio measurement or L2 measurement. It can be measurement based on MDT (Minimization Drive Test). It can be measurement of immediate MDT (immediate MDT) or measurement of logged MDT (Logged Measurement). The UE can perform measurement using the measurement configuration received in step ST1102. The respective measurements can be performed at the same timing or at different timings.

[0283] In step ST1104, the MN transmits a request for the AI / ML model for DC use input information to the UE. The request can include information indicating the required information. The request can include information indicating the use of the AI / ML model for DC use input information. For example, the request can include information indicating that it is for model training. For example, it is effective in the case where the AI / ML model for DC use input information is provided with information for model training. For example, in the case where there is information for model training, information for inference, and information for feedback in the information held by the UE, the UE can quickly grasp the information to be notified to the MN, and as a result, the UE can quickly notify the MN. The transmission of the request can use, for example, RRC signaling. The request for the AI / ML model for DC use input information can be included in the above-described measurement configuration. In step ST1102, the MN can transmit the request for the AI / ML model for DC use input information to the UE by including it in the measurement configuration. The amount of signaling can be reduced. In step ST1106, the UE transmits the AI / ML model for DC use input information to the MN. As this information, the UE can transmit the result of measurement by the measurement result report. The measurement result can be a radio measurement result, an L2 measurement result, or a measurement result of MDT. The measurement result of logged MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. These measurement results can be transmitted by the same signaling or by different signaling. The transmission can use RRC signaling. For example, the transmission of the radio measurement result and the L2 measurement result can use the MeasurementReport message. For example, the transmission of the measurement result of MDT can use the UEInformationResponse message.

[0284] In step ST1105, the MN transmits a request for input information for the DC AI / ML model to the neighboring base station. The neighboring base station is not limited to one, and can be plural. The neighboring base station can include the SN as the DC target. In other words, the SN as the DC target can be selected within the neighboring base station. The request can include information indicating the required information. The request can include information indicating the use of the input information for the DC AI / ML model. For example, the request can include information indicating the use for model training. For example, this is effective in a case where the input information for the DC AI / ML model is provided with the information as the use for model training. For example, in a case where the information as the use for model training, the use for inference, and the use for feedback exist in the information held by the neighboring base station, the neighboring base station can quickly grasp the information to be notified to the MN, and as a result, the notification from the neighboring base station to the MN can be performed quickly. The transmission of the request can use, for example, the Xn signaling. In step ST1107, the neighboring base station transmits the input information for the DC AI / ML model (input information for model training) to the MN. The transmission of the information can use, for example, the Xn signaling.

[0285] In step ST1108, the MN performs the training of the DC AI / ML model using the input information for the DC AI / ML model held by the node and the input information for the DC AI / ML model acquired from the UE and the neighboring base station. The update of the DC AI / ML model is performed by the training. These processes can be appropriately performed. The processes can be periodically performed, or can be performed each time the DC is performed in the MN or the surrounding base station.

[0286] The MN determines to set the DC to the UE. In step ST1109, the MN transmits a request for the input information for the DC AI / ML model to the UE. The request can include information indicating the required information. The request can include information indicating the use of the input information for the DC AI / ML model. For example, the request can include information indicating the use for inference. For example, this is effective in a case where the input information for the DC AI / ML model is provided with the information as the use for inference. In step ST1111, the UE transmits the input information for the DC AI / ML model to the MN. As the information, the measurement result of the UE can be transmitted by the measurement result report. The measurement result can be a radio measurement result, can be an L2 measurement result, or can be a measurement result of MDT. The measurement result of MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. The signaling used in the transmission of step ST1109, step ST1111 can appropriately apply the method disclosed in step ST1104, step ST1106.

[0287] In step ST1110, the MN transmits a request for the DC AI / ML model input information to the neighboring base station. The request can include information indicating the required information. The request can include information indicating the use of the DC AI / ML model input information. For example, the request can include information indicating that it is for inference. For example, this is valid in a case where the DC AI / ML model input information is provided with information for inference. The transmission of the request can use, for example, Xn signaling. In step ST1112, the neighboring base station transmits the DC AI / ML model input information (inference input information) to the MN. The transmission of the information can use, for example, Xn signaling.

[0288] In step ST1113, the MN performs inference using the DC AI / ML model using the DC AI / ML model input information held by the node and the DC AI / ML model input information acquired from the UE and the neighboring base station. The MN derives the DC AI / ML model output information through the inference.

[0289] In step ST1114, the MN performs DC for the UE using the DC AI / ML model output information. In the example of the figure, SN addition processing is performed. The MN selects a DC target SN for the UE using the DC AI / ML model output information. Further, the MN performs settings for DC using the selected SN. The MN can transmit an SN addition request to the selected SN. SNaddition processing is performed among the MN, the UE, and the SN. As a result, the DC processing can be performed among the UE, the MN, and the SN using the output information derived by the DC AI / ML model.

[0290] In step ST1115, the MN transmits a request for DC processing feedback information using the DC AI / ML model to the UE. The request can include information indicating the required information. The transmission of the request can use, for example, RRC signaling. In step ST1117, the UE transmits DC processing feedback information using the DC AI / ML model to the MN. The transmission can use RRC signaling.

[0291] In step ST1116, the MN transmits a request for DC processing feedback information using the DC AI / ML model to the SN. The request can include information indicating the required information. The transmission of the request can use, for example, Xn signaling. In step ST1118, the SN transmits DC processing feedback information using the DC AI / ML model to the MN. The transmission of the information can use, for example, Xn signaling.

[0292] Thus, the MN can acquire the performance of the DC processing using the DC AI / ML model. The MN can evaluate the effect of the DC processing in the case where the DC AI / ML model is used, using the DC processing feedback information using the DC AI / ML model held by the node, the DC processing feedback information using the DC AI / ML model acquired from the UE and the SN.

[0293] Thus, the AI / ML is introduced to the SN addition processing, and more efficient SN addition processing can be performed.

[0294] Figure 16 is a diagram illustrating a sequence example of SN change processing using a DC AI / ML. An example in which the MN has a DC AI / ML model and performs training and inference is disclosed. The steps common to the SN change processing using the DC AI / ML are denoted by the same step number, and the common description is omitted. Figure 15

[0295] In steps ST1201 and ST1202, the MN transmits a request for input information for the DC AI / ML model to the SN (change source SN (S-SN)), the neighboring base station. The neighboring base station is not limited to one, and can be plural. The neighboring base station can include the change target SN (T-SN). In other words, the SN of the change target can be selected within the neighboring base station. The request can include information indicating the required information. The request can include information indicating that it is for model training. For example, it is effective in the case where the input information for the DC AI / ML model is provided as information for model training. The transmission of the request can use Xn signaling, for example. In steps ST1203 and ST1204, the S-SN, the neighboring base station transmits the input information for the DC AI / ML model (input information for model training) to the MN. The transmission of the information can use Xn signaling, for example.

[0296] In steps ST1205 and ST1206, the MN transmits a request for input information for the DC AI / ML model to the S-SN, the neighboring base station. The request can include information indicating the required information. The request can include information indicating that it is for inference. For example, it is effective in the case where the input information for the DC AI / ML model is provided as information for inference. The transmission of the request can use Xn signaling, for example. In steps ST1207 and ST1208, the SN, the neighboring base station transmits the input information for the DC AI / ML model (input information for inference) to the MN. The transmission of the information can use Xn signaling, for example.

[0297] ​In step ST1209, the MN performs DC for the UE using the DC AI / ML model output information derived in step ST1113. In the example of the figure, SN change processing is performed. For example, MN-initiated SN change processing can be performed. The MN selects a T-SN using the DC AI / ML model output information. Further, the MN performs setting for the DC using the selected T-SN. The MN can transmit an SN change request to the T-SN. SN addition processing is performed among the MN, the UE, the S-SN, and the T-SN. As a result, the DC processing among the UE, the MN, the S-SN, and the T-SN can be performed using the output information derived by the DC AI / ML model.

[0298] In steps ST1210 and ST1211, the MN transmits a request for DC processing feedback information using the DC AI / ML model to the S-SN and the T-SN. The request can include information indicating the required information. The transmission of the request can use, for example, Xn signaling. In steps ST1212 and ST1213, the S-SN and the T-SN transmit DC processing feedback information using the DC AI / ML model to the MN. The transmission of the information can use, for example, Xn signaling.

[0299] As a result, the MN can acquire the performance of the DC processing using the DC AI / ML model. The MN can evaluate the effect of the DC processing using the DC AI / ML model using the DC processing feedback information using the DC AI / ML model held by the node, and the DC processing feedback information using the DC AI / ML model acquired from the UE and the S-SN and the T-SN.

[0300] As a result, the AI / ML is introduced into the SN change processing, and more effective SN change processing can be performed.

[0301] Figure 17 is a figure showing another sequence example of SN change processing using a DC AI / ML. An example in which the SN has a DC AI / ML model and performs training and inference is disclosed. The steps common to Figure 15 , Figure 16 The same step number is given to the steps common to the above-described example, and the common description is omitted. In step ST1301, the S-SN is provided with a DC AI / ML model. In step ST1302, the S-SN performs measurement setting for the UE. For example, radio measurement setting is performed. The setting can use RRC signaling. An RRC reconfiguration message can be used.

[0302] The UE performs measurement in step ST1303. The measurement performed by the UE can be, for example, radio measurement or L2 measurement. It can be measurement based on MDT (Minimization Drive Test). It can be measurement of immediate MDT or measurement of Logged MDT (Logged Measurement). The UE can perform measurement using the measurement configuration received in step ST1302. The respective measurements can be performed at the same timing or at different timings.

[0303] In step ST1304, the S-SN transmits a request for the UE to transmit DC AI / ML model input information. The request can include information indicating the required information. The request can include information indicating that it is for model training. This is effective, for example, in a case where the DC AI / ML model input information is provided as information for model training. The transmission of the request can use, for example, RRC signaling. In step ST1307, the UE transmits DC AI / ML model input information to the S-SN. As this information, the measurement result of the UE can be transmitted by a measurement result report. The measurement result can be a radio measurement result, an L2 measurement result, or a measurement result of MDT. The measurement result of Logged MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. These measurement results can be transmitted by the same signaling or by different signaling. The transmission can use RRC signaling. For example, the transmission of the radio measurement result and the L2 measurement result can use a MeasurementReport message. For example, the transmission of the measurement result of MDT can use a UEInformationResponse message.

[0304] In steps ST1305 and ST1306, the S-SN transmits a request for the MN and the neighboring base station to transmit DC AI / ML model input information. The transmission from the S-SN to the neighboring base station can be performed directly by the S-SN or via the MN. Figure 17In the example of Fig. 13, a method via the MN is disclosed. The neighboring base station is not limited to one, but can be plural. The neighboring base station can include the T-SN. In other words, the T-SN can be selected in the neighboring base station. The request can include information indicating the required information. The request can include information indicating that it is for model training. For example, it is effective in the case where the input information for the AI / ML model for DC is provided as the information for model training. The transmission of the request can use, for example, Xn signaling. In steps ST1308, ST1309, the MN, the neighboring base station transmit the input information for the AI / ML model for DC to the S-SN. The transmission from the neighboring base station to the S-SN can be directly to the S-SN, or via the MN. Figure 17 In the example of Fig. 13, a method via the MN is disclosed. The transmission of the information can use, for example, Xn signaling.

[0305] In step ST1310, the S-SN performs training of the AI / ML model for DC using the input information for the AI / ML model for DC held by the node, and the input information for the AI / ML model for DC acquired from the UE, the MN, and the neighboring base station. The AI / ML model for DC is updated by the training. These processes can be appropriately performed. They can be periodically performed, or each time DC is performed in the MN, the S-SN, or the surrounding base station.

[0306] The S-SN determines to initiate SN change. In step ST1311, the S-SN transmits a request for the input information for the AI / ML model for DC to the UE. The request can include information indicating the required information. The request can include information indicating that it is for inference. For example, it is effective in the case where the input information for the AI / ML model for DC is provided as the information for inference. In step ST1314, the UE transmits the input information for the AI / ML model for DC to the S-SN. As the information, the measurement result of the UE can be transmitted by a measurement result report. The measurement result can be a radio measurement result, can be an L2 measurement result, or can be a measurement result of MDT. The measurement result of MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. The signaling used in the transmission of steps ST1311, ST1314 can appropriately apply the method disclosed in steps ST1304, ST1307.

[0307] In steps ST1312, ST1313, the S-SN transmits a request for the input information for the AI / ML model for DC to the MN, the neighboring base station. The transmission from the S-SN to the neighboring base station can be directly by the S-SN, or via the MN. Figure 17In the example of Fig. 13, a method via the MN is disclosed. The neighboring base station is not limited to one, but can be plural. The neighboring base station can include the T-SN. In other words, the T-SN can be selected in the neighboring base station. The request can include information representing the required information. The request can include information representing that it is for inference. This is effective, for example, in a case where the AI / ML model for DC is provided with the input information as the information for inference. The transmission of the request can use, for example, Xn signaling. In steps ST1315, ST1316, the MN, the neighboring base station transmit the input information for the AI / ML model for DC (inference input information) to the S-SN. The transmission from the neighboring base station to the S-SN can be directly to the S-SN, or via the MN. Figure 17 In the example of Fig. 13, a method via the MN is disclosed. The transmission of the information can use, for example, Xn signaling.

[0308] In step ST1317, the S-SN performs inference using the DC AI / ML model using the input information for the AI / ML model for DC held by the node, and the input information for the AI / ML model for DC acquired from the UE, the MN, and the neighboring base station. The S-SN derives the AI / ML model output information for DC by the inference.

[0309] In step ST1318, the S-SN performs DC processing using the AI / ML model output information for DC. In the example of the figure, SN change processing is performed. The S-SN can select the T-SN using the AI / ML model output information for DC. The S-SN can perform DC setting using the selected T-SN. The S-SN initiates SN change processing to the MN. The S-SN transmits a SN change request to the MN. The request can include information related to the selected T-SN. The request can include the DC setting. According to the SN change request initiated from the S-SN, SN change processing is performed among the MN, the UE, the S-SN, and the T-SN. As a result, the output information derived by the AI / ML model for DC can be used to perform DC processing among the UE, the MN, the S-SN, and the T-SN.

[0310] In step ST1319, the S-SN transmits a request for DC processing feedback information using the AI / ML model for DC to the UE. The request can include information representing the required information. The transmission of the request can use, for example, RRC signaling. In step ST1322, the UE transmits DC processing feedback information using the AI / ML model for DC to the S-SN. The transmission can use RRC signaling.

[0311] In step ST1320, step ST1321, the S-SN transmits a request for the MN, the T-SN to transmit DC processing feedback information using the DC AI / ML model to the S-SN. The transmission from the S-SN to the neighboring base station can be performed directly by the S-SN or via the MN. Figure 17 In the example of FIG. 13, a method performed via the MN is disclosed. The transmission of the request can use, for example, Xn signaling. In step ST1323, step ST1324, the MN, the T-SN transmits DC processing feedback information using the DC AI / ML model to the S-SN. The transmission from the T-SN to the S-SN can be performed directly to the S-SN or via the MN. Figure 17 In the example of FIG. 13, a method performed via the MN is disclosed. The transmission of the information can use, for example, Xn signaling.

[0312] Thus, the S-SN can acquire the performance of the DC processing using the DC AI / ML model. The S-SN can evaluate the effect of the DC processing using the DC AI / ML model using the DC processing feedback information using the DC AI / ML model held by the node, the DC processing feedback information using the DC AI / ML model acquired from the UE, the MN, and the T-SN.

[0313] Thus, by introducing AI / ML into the SN change processing, more efficient SN change processing can be performed.

[0314] The MN can have the DC AI / ML model. The MN can perform training, inference. In this case, the SN change processing initiated by the S-SN can be performed. For example, the training of the DC AI / ML model can appropriately apply the steps ST1101 to ST1104, ST1106, ST1201 to ST1204, ST1108 disclosed in FIG. 11. Figure 16 In the example of FIG. 13, a method performed via the MN is disclosed. The transmission of the request can use, for example, Xn signaling. In step ST1323, step ST1324, the MN, the T-SN transmits DC processing feedback information using the DC AI / ML model to the S-SN. The transmission from the T-SN to the S-SN can be performed directly to the S-SN or via the MN. Figure 16 In the example of FIG. 13, a method performed via the MN is disclosed. The transmission of the request can use, for example, Xn signaling. In step ST1323, step ST1324, the MN, the T-SN transmits DC processing feedback information using the DC AI / ML model to the S-SN. The transmission from the T-SN to the S-SN can be performed directly to the S-SN or via the MN. Figure 16The step ST1209 disclosed in the middle. The SN change target (T-SN) can be decided by the MN, for example, using the DC AI / ML model output information derived by the MN. The SN change processing can be performed using the information related to the T-SN decided by the MN. The request for the DC processing feedback information using the DC AI / ML model can be sent from the MN to the UE, the S-SN, the T-SN, and the DC processing feedback information using the DC AI / ML model can be sent from the UE, the S-SN, the T-SN to the MN. The MN can use the feedback information to perform the evaluation and update of the model.

[0315] Figure 18 and Figure 19 is a diagram showing a sequence example of the MN HO processing with SN change using DC AI / ML. Figure 18 shows the first half of the sequence example, Figure 19 shows the latter half of the sequence example. The following example is disclosed: the S-MN has an HO AI / ML model, the T-MN has a DC AI / ML model, and each performs training and inference. The HO AI / ML model can be applied, for example, to the AI / ML model for mobility disclosed in Chapter 5.3.2 of Non-Patent Literature 30. In step ST1401, the HO source MN (S-MN) is provided with an HO AI / ML model. In step ST1402, the base station that becomes the HO target (T-MN) is provided with a DC AI / ML model. In step ST1403, the S-MN performs measurement setting on the UE. For example, wireless measurement setting is performed. The setting can be performed using RRC signaling. The RRC reconfiguration (RRCReconfiguration) message can be used.

[0316] In step ST1404, the UE performs measurement. The measurement performed by the UE can be wireless measurement, or L2 measurement, for example. It can be measurement based on MDT (Minimization Drive Test). For example, it can be measurement of immediate MDT, or measurement of logged MDT (Logged Measurement). The UE can perform measurement using the measurement setting received in step ST1403. Each measurement can be performed at the same timing, or at different timings.

[0317] In step ST1411, the S-MN transmits to the UE a request for the input information for the AI / ML model for HO. The request can include information on the UE that is the object of HO. The information on the UE can be an identifier of the UE. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model for HO. The request can include information indicating that it is for model training. This is effective, for example, in a case where the input information for the AI / ML model for HO is provided as information for model training. The S-MN can transmit to the UE a request for the input information for the AI / ML model. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model. For example, it can include information indicating that it is for HO. For example, it can include information indicating that it is for training. The UE can promptly grasp the information to be transmitted to the S-MN, and as a result, prompt notification from the UE to the S-MN is possible. The transmission of the request can use, for example, RRC signaling. In step ST1413, the UE transmits to the S-MN the input information for the AI / ML model for HO. As the information, the measurement result of the UE can be transmitted by a measurement result report. The measurement result can be a radio measurement result, can be an L2 measurement result, or can be a measurement result of MDT. The measurement result of MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. These measurement results can be transmitted by the same signaling or by different signaling. The transmission can use RRC signaling. For example, the transmission of the radio measurement result and the L2 measurement result can use a MeasurementReport message. For example, the transmission of the measurement result of MDT can use a UEInformationResponse message.

[0318] In step ST1412, the S-MN transmits a request for input information for the AI / ML model for HO to the neighboring base station. The request can include information on the UE that is the object of HO. The information on the UE can be an identifier of the UE. The neighboring base station is not limited to one, and can be plural. The neighboring base station can include the T-MN. In other words, the T-MN can be selected in the neighboring base station. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model for HO. The request can include information indicating that it is for model training. This is effective, for example, in a case where the input information for the AI / ML model for HO is provided with the information that is for model training. The S-MN can transmit a request for the input information for the AI / ML model to the neighboring base station. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model. For example, it can include information indicating that it is for HO. For example, it can include information indicating that it is for training. The neighboring base station can quickly grasp the information to be transmitted to the S-MN, and as a result, the S-MN can be quickly notified from the neighboring base station. The transmission of the request can use Xn signaling, for example. In step ST1414, the neighboring base station transmits the input information for the AI / ML model for HO (input information for model training) to the S-MN. The transmission of the information can use Xn signaling, for example.

[0319] In step ST1415, the S-MN performs training of the AI / ML model for HO using the input information for the AI / ML model for HO held by the node and the input information for the AI / ML model for HO acquired from the UE and the neighboring base station. The AI / ML model for HO is updated by the training. These processes can be appropriately performed. They can be performed periodically, or each time HO is performed in the MN or the surrounding base station.

[0320] In steps ST1421, ST1423, and ST1424, the base station that has received the request for the input information for the AI / ML model for HO from the S-MN (the T-MN, the S-SN, or the like) transmits the input information for the AI / ML model for HO to the S-MN. Figure 18The T-MN can transmit a request for input information for the AI / ML model to the neighboring base station. The request can include information about the UE as the DC object. The information about the UE can be the identification of the UE. The UE that becomes the DC object can be the same as the UE that becomes the HO object. The neighboring base station is not limited to one, and can be plural. The neighboring base station can include the S-MN. For example, the source base station of the request for input information for the AI / ML model for HO received in step ST1412 can be set as the neighboring base station. The S-MN that receives the request can transmit a request for input information for the AI / ML model for DC to the UE in step ST1422. The request can include information about the UE as the DC object. The information about the UE can be the identification of the UE. The UE that becomes the DC object can be the same as the UE that becomes the HO object. The neighboring base station that transmits the request for input information for the AI / ML model for DC by the T-MN can include the T-SN. In other words, the T-SN can be selected within the neighboring base station. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model for DC. The request can include information indicating that it is for model training. For example, it is effective in the case where the input information for the AI / ML model for DC is provided as information for model training. The T-MN can transmit a request for input information for the AI / ML model to the neighboring base station. The S-MN can transmit a request for input information for the AI / ML model to the UE. The request can include information about the UE as the object. The information about the UE can be the identification of the UE. The request can include information indicating the required information. The request can include information indicating the use of the input information for the AI / ML model. For example, it can include information indicating that it is for DC. For example, it can include information indicating that it is for training. The neighboring base station or the UE can quickly grasp the information to be transmitted to the T-MN or the S-MN, and as a result, quick notification from the neighboring base station or the UE to the T-MN or the S-MN can be performed. The transmission of the request between the base stations can use, for example, Xn signaling. The transmission of the request from the base station to the UE can use RRC signaling.

[0321] In step ST1425, the UE transmits input information for the DC AI / ML model to the S-MN. As this information, a measurement result of the UE can be transmitted by a measurement result report. This measurement result can be a radio measurement result, can be an L2 measurement result, or can be a measurement result of MDT. The measurement result of MDT can be a result measured by the UE in the RRC_Inactive or RRC_Idle state. These measurement results can be transmitted by the same signaling or can be transmitted by different signaling. This transmission can use RRC signaling. For example, the transmission of the radio measurement result or the L2 measurement result can use a MeasurementReport message. For example, the transmission of the measurement result of MDT can use a UEInformationResponse message.

[0322] In steps ST1426, ST1427, and ST1428, the S-MN and the neighboring base station transmit input information for the DC AI / ML model to the T-MN. In step ST1426, the input information for the DC AI / ML model received from the UE in step ST1425 can be added to the input information for the DC AI / ML model transmitted from the S-MN to the T-MN. Thereby, the T-MN can acquire this information from the UE. The transmission of this information can use, for example, Xn signaling.

[0323] A request for the transmission of input information for the DC AI / ML model from the T-MN to the S-SN can be made. The T-MN can transmit this request to the S-SN via the S-MN. The S-SN can transmit input information for the DC AI / ML model to the T-MN. The S-SN can transmit this information to the T-MN via the S-MN. Thereby, the T-MN can reliably acquire this input information about the S-SN used in DC by the S-MN.

[0324] In step ST1429, the T-MN performs training of the DC AI / ML model using the input information for the DC AI / ML model held by the node and the input information for the DC AI / ML model acquired from the UE, the S-MN, and the neighboring base station. The update of the DC AI / ML model is performed by the training. These processes can be appropriately performed. They can be performed periodically or each time DC is performed in the T-MN, the S-SN, the T-SN, or the surrounding base station.

[0325] The process for training of the AI / ML model for HO and the process for training of the AI / ML model for DC can not be linked. They can be performed at different timings. In the example disclosed above, it is disclosed that the base station which received the input information request for the AI / ML model for HO from the S-MN in step ST1412 performs the process for training of the AI / ML model for DC in steps ST1421 to ST1429, but as another example, the base station can perform the process for training of the AI / ML model for DC without receiving the input information request for the AI / ML model for HO. The base station which has the AI / ML model for DC can perform the process for training of the AI / ML model for DC regardless of the reception of the request. Thereby, even if not targeted for HO, the AI / ML model can be used in the DC process.

[0326] The process for training of the AI / ML model for DC in steps ST1421 to ST1429 can be started after the T-MN receives the HO request message of step ST1436. In a case where the period until the training process of the AI / ML model for DC and the HO request becomes long, due to the movement of the UE, the change of the radio wave propagation environment, or other conditions, there are cases where the model becomes unsuitable. By adopting this method, the accuracy of the AI / ML model for DC can be improved. However, when the process for training of the AI / ML model for DC is performed after the HO request as such, the processing after the HO request increases, and the time until the HO process sometimes becomes long. In order to solve this problem, the process for training of the AI / ML model for DC can be performed before the HO request and after the HO request. For example, in the training of the AI / ML model for DC after the HO request, fine correction of the model can be performed. For example, in the training of the AI / ML model for DC after the HO request, the input information can be limited. For example, the input information which changes greatly in time can be limited. Thereby, the time until the HO process after the HO request can be shortened.

[0327] The S-MN determines to start HO. The S-MN can determine to start HO with SN change. The S-MN can determine to start HO without SN change. The S-MN can determine to start HO, and the T-MN can determine whether to perform SN change. Figure 18 And Figure 19 In the example of FIG. 17, the T-MN determines whether to perform SN change. The S-MN can use the AI / ML model for HO in the decision of the HO target.

[0328] In steps ST1431 to ST1434, transmission of a request for input information for the inference-use AI / ML model for HO, and transmission of the input information for the inference-use AI / ML model for HO are performed among the S-MN, the UE, and the neighboring base station. These processes can be appropriately applied to the processes of steps ST1411 to ST1414 described above. The model training can be replaced with inference. In step ST1435, the S-MN performs inference by the AI / ML model for HO and derives AI / ML model output information for HO. The S-MN uses the output information to decide the HO target base station (T-MN) and performs HO setting. In step ST1436, the S-MN transmits an HO request message to the T-MN.

[0329] The T-MN that receives the request message determines whether to perform SN change. The T-MN can select an SN for DC. In the case where the SN is the same as the S-SN, it can be determined that SN change is not performed, and in the case where the SN is different from the S-SN, it can be determined that SN change is performed on the selected SN. The T-MN can use the AI / ML model for DC for selection of the SN.

[0330] In steps ST1441 to ST1448, transmission of a request for input information for the inference-use AI / ML model for DC, and transmission of the input information for the inference-use AI / ML model for DC are performed among the T-MN, the S-MN, the UE, and the neighboring base station. These processes can be appropriately applied to the processes of steps ST1412 to ST1428 described above. The model training can be replaced with inference. In step ST1449, the T-MN performs inference by the AI / ML model for DC and derives AI / ML model output information for DC. The T-MN uses the output information to decide the DC target base station (T-SN) and performs DC setting. In step ST1450, the T-MN transmits an SN addition request message to the T-SN.

[0331] In step ST1460, the HO process of the MN with SN change is continued among the UE, the S-MN, the T-MN, the S-SN, and the T-SN. Thus, the process can be performed using the AI / ML model for HO and the AI / ML model for DC.

[0332] In step ST1471, step ST1472, the S-MN transmits to the UE and the T-MN a request for HO processing feedback information using the AI / ML model for HO. The S-MN can transmit to the T-MN a request for HO processing feedback information using the AI / ML model for HO for the UE. The T-MN can transmit to the UE a request for HO processing feedback information using the AI / ML model for HO. The S-MN can transmit to the UE the request via the T-MN. The request can contain information indicating the required information. The transmission of the request can use, for example, RRC signaling. In step ST1473, step ST1474, the UE and the T-MN transmit to the S-MN HO processing feedback information using the AI / ML model for HO. The UE can transmit to the T-MN HO processing feedback information using the AI / ML model for HO. The T-MN can transmit to the S-MN HO processing feedback information using the AI / ML model for HO. The UE can transmit to the S-MN the information via the T-MN. The transmission can use RRC signaling.

[0333] In step ST1481, step ST1482, step ST1483, the T-MN transmits to the S-SN, the UE, the T-SN a request for DC processing feedback information using the AI / ML model for DC. The transmission from the T-MN to the UE can be made directly by the T-MN. In step ST1484, step ST1485, step ST1486, the UE, the S-SN, the T-SN transmit to the T-MN DC processing feedback information using the AI / ML model for DC. The transmission from the UE to the T-MN can be made directly to the T-MN or via the S-MN. The transmission of the information between base stations can use, for example, Xn signaling. The transmission of the information between the UE and the base station can use, for example, RRC signaling.

[0334] Thereby, the T-MN can acquire the performance of the DC processing using the AI / ML model for DC. The T-MN can use the DC processing feedback information using the AI / ML model for DC held by the node, the DC processing feedback information using the AI / ML model for DC acquired from the UE and the T-SN to evaluate the effect of the DC processing with the AI / ML model for DC.

[0335] Thereby, the HO processing of the MN with SN change can be performed using the AI / ML model for HO and the AI / ML model for DC. In the case of combination of HO and DC, the AI / ML for HO and the AI / ML for DC can be executed, and more effective processing after the combination of HO and DC can be executed.

[0336] The HO AI / ML model can have a RAN node that becomes a HO object (may be a HO source or a HO target). The DC AI / ML model can have a RAN node that becomes a DC object (may be a MN or a SN). Further, a RAN node that becomes a HO object and becomes a DC object can have a HO AI / ML model and a DC AI / ML model. A model that has both functions of the HO AI / ML model and the DC AI / ML model can be provided. Thereby, the HO or the DC using the AI / ML model can be avoided from being limited to a specific RAN node.

[0337] The transmission of the request for the input information for the HO AI / ML model and the transmission of the request for the input information for the DC AI / ML model can be performed together. The transmission of the input information for the HO AI / ML model and the transmission of the input information for the DC AI / ML model can be performed together. For example, this can be performed in a case where a model that has both functions of the HO AI / ML model and the DC AI / ML model is used. Reduction of the amount of signaling can be attempted.

[0338] By adopting the method disclosed in the present embodiment, the DC AI / ML model can be used for training and inference in the DC processing. For example, by using the inference result using the DC AI / ML model, more efficient processing can be performed in the selection of the SN, the selection of the SN change target, the selection of the PSCell change target, and the like. Thereby, high-capacity communication at high speed can be achieved in the DC processing, and various KPIs can be optimized.

[0339] The method of performing the SN change processing using the DC AI / ML model is disclosed above. The disclosed method can also be appropriately applied to SN Modification processing. For example, it can be applied to PSCell change within an S-SN. It can be applied to SCell change within an S-SN. Information indicating SN change or SN Modification can be provided and included in the input information for the DC AI / ML model and the request therefor, the output information, and the feedback information. The RAN node or the UE that receives the information can identify whether it is SN change or SN Modification. Thereby, the DC AI / ML model can also be used in the SN Modification processing, and more efficient SN Modification processing can be performed.

[0340] Embodiment 2. In the standards of mobile communication systems in 3GPP, there is a CA (Carrier Aggregation) function for high-speed and large-capacity communication. In order to meet the requirements of various KPIs while achieving high-speed and large-capacity communication, it is desirable to more effectively perform CA. In the present embodiment, a method of achieving more effective CA in a mobile communication system is disclosed.

[0341] In the present embodiment, AI / ML is introduced for CA. A CA-use AI / ML model is provided in a base station. For example, the CA-use AI / ML model is used to perform setting of a CC (Component Carrier) for CA. By introducing AI / ML in CA, more effective CA can be achieved.

[0342] The base station can transmit information on whether or not the CA-use AI / ML model is held (information that the model is held can also be transmitted) to a UE. The transmission can use RRC signaling. The base station can transmit information on whether or not the CA-use AI / ML model is held to a neighboring base station. The notification can use Xn signaling. The base station can transmit information on whether or not the CA-use AI / ML model is held to a CN node (function can also be used). As the CN node, for example, a NWDAF (Network Data Analysis Function) can be used. The base station can transmit information on whether or not the CA-use AI / ML model is held to a management node. As the management node, for example, a MnS (Management Service) node or an OAM can be used. Thereby, the UE or node to which the information is notified can recognize whether or not the source base station of the notification holds the CA-use AI / ML model.

[0343] Input information for the CA-use AI / ML model is set. Twenty-one examples of CA-use AI / ML model input information are disclosed below.

[0344] (1) Information on the position of the UE.

[0345] (2) Information on radio measurement of the UE.

[0346] (3) CA history information of the UE.

[0347] (4) CA history information of the UE from a neighboring base station.

[0348] (5) Information on the performance of the UE performing CA.

[0349] (6) Resource status of the current base station.

[0350] (7) Predicted resource status of the base station.

[0351] (8) Past CA failure information of the UE.

[0352] (9) Past CA failure information of the UE from a neighboring base station.

[0353] (10) Past DC communication interruption status of the UE.

[0354] (11) Past communication interruption status of the UE from a neighboring base station.

[0355] (12) Information related to a trajectory of the UE.

[0356] (13) Information related to current traffic volume of the UE.

[0357] (14) Information related to predicted traffic volume of the UE.

[0358] (15) Information related to L2 measurement of the UE.

[0359] (16) Predicted value of information related to L2 measurement of the UE.

[0360] (17) Information related to L2 measurement from a neighboring base station.

[0361] (18) Predicted value of information related to L2 measurement from a neighboring base station.

[0362] (19) Carrier history information used in communication of the UE.

[0363] (20) Band history information used in communication of the UE.

[0364] (21) Combination of (1) to (20).

[0365] (1) and (2) can be applied to the input information examples (1) and (2) for the AI / ML model for DC disclosed in Embodiment 1, respectively.

[0366] (3) is a history of past CA information. The CA information can be, for example, information of an SCell (may also be a CC) used for CA. For example, it can be information identifying an SCell. For example, it can be setting information of an SCell. For example, it can be information of activation / deactivation of an SCell. For example, it can be information indicating which frequency band a CC belongs to. The information identifying the frequency band can be a frequency band number decided by 3GPP. For example, it can be information of a BWP used in CA. For example, it can be BWP setting information of an SCell used in CA. For example, it can be information indicating which BWP is used in CA. For example, it can be information on SUL. For example, it can be setting information of SUL used in CA. It can be information of a past base station of the UE and information of an SCell used for CA in each base station. It can be information of a change history of an SCell used in CA. These information can be listed. Further, information on a location of the UE can be included. Information on time can be included. For example, information of an SCell used in past CA can be associated with information on a location or time.

[0367] (4) is CA history information of a UE possessed by a neighboring base station. It can be information disclosed in (3) of the CA history information of the UE.

[0368] (5) can be, for example, QoS of a UE performing CA. It can be a packet loss rate or a delay time. It can be information before performing CA. It can be information after performing CA. It can be information each time an SCell used in CA is changed. Further, information on a location of the UE can be included. Information on time can be included. Information on performance of a UE performing CA can be associated with information on a location or time. Further, it can be associated with the CA history information of the UE of (3) or (4).

[0369] (6) can be, for example, current resource usage information of a base station used for CA. For example, it can be the number of SCells used in CA. It can be resource usage information of each SCell. For example, it can be usage information of a BWP used in CA. It can be usage information of a frequency resource. It can be usage information of a time resource. The usage information of the frequency resource can be in units of PRBs or in units of subcarriers.

[0370] (7) can be, for example, information on predicted resources of a base station used for CA. The information on resources of a base station used in CA can be information disclosed in (6).

[0371] (8) is information on past CA failure of the UE. For example, it can be information on deactivation of SCell based on communication quality. For example, it can be information on the number of times of exceeding the maximum number of HARQ in SCell. For example, it can be information on the number of times of ARQ retransmission failure. The information on CA failure can be, for example, past CA failure history information of the UE. It can be past base station history information of the UE and information on CA failure in each base station. These information can be listed. Further, information on the location of the UE can be included. Information on time can be included. Information on past CA failure can be associated with information on location or time.

[0372] (9) is information on past CA failure of the UE that the neighboring base station has. The information on past CA failure of the UE can be the information disclosed in (8).

[0373] (10) is information on past communication interruption of the UE in CA. For example, it can be history information on past communication interruption of the UE. As the communication interruption, for example, it can be RLF. For example, it can be information identifying the base station that has caused the communication interruption. For example, it can be information identifying the PCell, the PSCell, the SPCell that has caused the communication interruption. These information can be listed. Further, information on the location of the UE can be included. Information on time can be included. Information on past communication interruption can be associated with information on location or time.

[0374] (11) is information on past communication interruption of the UE in CA that the neighboring base station has. The information on past communication interruption of the UE can be the information disclosed in (10).

[0375] As the information on the trajectory of the UE of (12), for example, the information on the location of the UE disclosed in (1) can be appropriately applied. The information on the trajectory of the UE can be, for example, the information disclosed in (3), the information disclosed in (19), or the information disclosed in (20). It can be information on past, current trajectory of the UE. The information on past, current trajectory of the UE can be history information. The information on trajectory can be information on future trajectory of the UE. The information on future trajectory can be predicted information. These information can be listed. Further, information on time can be included. Information on the trajectory of the UE can be associated with information on time.

[0376] In (13), the traffic volume of the UE can be, for example, the number of connected UEs. For example, it can be the current traffic volume of UEs in the SCell. It can be the current traffic volume of UEs in the SCell for CA in the base station.

[0377] In (14), the information about the traffic volume of the UE can be the information disclosed in (13).

[0378] (15) can be, for example, a measurement result of L2 measurement information in the SCell of the UE. The L2 measurement information can be applied as appropriate to the input information example (15) for the DC AI / ML model disclosed in Embodiment 1. In addition, information about the position of the UE can be included. Information about the time can be included. Information about L2 measurement in the SCell of the UE can be associated with information about the position or the time.

[0379] (16) is a predicted value of L2 measurement information in the SCell. It is a predicted value of L2 measurement information in the SCell measured by the UE. It can be a predicted value of L2 measurement information in the SCell measured by the UE in CA.

[0380] (17) can be, for example, a measurement result of L2 measurement information in the SCell of the UE. It can be a measurement result of L2 measurement information in the SCell possessed by the neighboring base station. The L2 measurement information can be applied as appropriate to the input information example (17) for the DC AI / ML model disclosed in Embodiment 1. For example, it can be a L2 measurement result in the SCell of the base station when CA is performed. In addition, for the L2 measurement result related to the UE, information about the position of the UE can be included. Information about the time can be included. Information about L2 measurement of the UE can be associated with information about the position or the time.

[0381] (18) is a predicted value of L2 measurement information in the SCell. It is a predicted value of L2 measurement information in the SCell measured by the base station. It can be a predicted value of L2 measurement information in the SCell of the base station when CA is performed. It can be a predicted value of L2 measurement information in the SCell of the base station possessed by the neighboring base station.

[0382] (19) is information of carriers used by the UE in the past for communication. For example, it can be information about carrier frequencies with which the UE and the base station communicated. It can be information identifying carrier frequencies. It can be set per kind of link, e.g., downlink, uplink, sidelink, etc. It can be a list of information of carriers used by the UE in the past for communication. Further, it can contain information about a location of the UE. It can contain information about time. It can be that information about carriers used by the UE in the past for communication is associated with information about location or time. The information of carriers used by the UE in the past for communication can not be limited to carriers for CA. Better CA can be achieved from more carriers.

[0383] (20) is information of frequency bands used by the UE in the past for communication. For example, it can be information about frequency bands with which the UE and the base station communicated, it can be information identifying frequency bands. It can be set per kind of link, e.g., downlink, uplink, sidelink, etc. It can be a list of information of frequency bands used by the UE in the past for communication. Further, it can contain information about a location of the UE. It can contain information about time. It can be that information about frequency bands used by the UE in the past for communication is associated with information about location or time. The information of frequency bands used by the UE in the past for communication can not be limited to frequency bands for CA. Better CA can be achieved from more frequency bands.

[0384] The base station with the AI / ML model for CA inputs the above disclosed input information for the AI / ML model for CA into the AI / ML model for CA for training or inference. Thereby, more efficient CA can be performed using the AI / ML model for CA.

[0385] The base station transmits input information for the AI / ML model for CA, etc. to a neighboring base station. As for the transmission of input information for the AI / ML model for CA, etc. between base stations, the method of transmission of input information for the AI / ML model for DC disclosed in Embodiment 1 can be applied as appropriate. The input information for the AI / ML model for DC can be replaced with input information for the AI / ML model for CA to be applied as appropriate.

[0386] The DU transmits input information for the AI / ML model for CA, etc. to the CU. As for the transmission of input information for the AI / ML model for CA, etc. between CU and DU, the method of transmission of input information for the AI / ML model for DC disclosed in Embodiment 1 can be applied as appropriate. The input information for the AI / ML model for DC can be replaced with input information for the AI / ML model for CA to be applied as appropriate.

[0387] The UE transmits the input information for the CA AI / ML model to the base station. As for the transmission of the input information for the CA AI / ML model between the UE and the base station, the method of transmission of the input information for the DC AI / ML model between the CU and the DU disclosed in Embodiment 1 can be appropriately applied. The input information for the DC AI / ML model can be replaced with the input information for the CA AI / ML model to be appropriately applied.

[0388] The output information of the CA AI / ML model is disclosed. Nine examples of the output information of the CA AI / ML model are disclosed below.

[0389] (1) Information on the predicted trajectory of the UE.

[0390] (2) Predicted CA information.

[0391] (3) Priority order of the predicted CA.

[0392] (4) Predicted candidate SCell information.

[0393] (5) Arrival probability of the predicted candidate SCell.

[0394] (6) Predicted CA processing time.

[0395] (7) Resource status of the predicted base station.

[0396] (8) Information on the predicted traffic volume of the UE.

[0397] (9) Combination of (1) to (8).

[0398] (1) For example, it can be the predicted future trajectory information of the UE. The information on the trajectory of the UE can be set to the information disclosed in the input information for the CA AI / ML model example (12). Information on the past and current trajectory can be included. Historical information of the past and current trajectory of the UE can be included together with the predicted future trajectory information of the UE. These information can be listed. In addition, information on time can be included. The information on the trajectory of the UE can be associated with the information on time.

[0399] (2) is CA information predicted by the AI / ML model. The CA information can be information disclosed in (3) of the input information example (1). For example, it can be information related to the SCell used for CA. The SCell used for CA can be one or more. For example, it can be information identifying the SCell selected for CA. It can be setting information of the SCell selected for CA. For example, it can be information of whether or not to perform CA. Furthermore, it can include information related to the predicted position of the UE. It can include information related to the predicted time. The predicted CA information can be associated with information about the position or the time. The predicted CA information is not limited to one, and can be plural. It can include information of the prediction accuracy of the CA information predicted by the AI / ML model.

[0400] (3) is, for example, information of the predicted priority order of the SCell used for CA shown in the information disclosed in (2). The predicted SCell used for CA and the priority order thereof can be associated. As another example, it can be information of the priority order of the predicted CA information. The predicted CA information and the priority order thereof can be associated.

[0401] (4) is information of the SCell that becomes a candidate for CA. For example, it can be information identifying the SCell. It can be setting information of the SCell. The SCell that becomes a candidate for CA is not limited to one, and can be plural. It can include information of the prediction accuracy of the predicted SCell candidate.

[0402] (5) is prediction information of the probability that the SCell that becomes a candidate for CA can serve the UE. It can include information of the confidence interval of the arrival probability of the candidate SCell for CA predicted by the AI / ML model.

[0403] (6) is information of the predicted CA processing time. The CA processing time can be, for example, the set time of starting CA, the set time of performing CA, or the set time of ending CA. The CA processing time can be set for each CA, for example. It can be set for each SCell, for example. It can be, for example, the time of activating the SCell, the time of activating the SCell, the time of deactivating the SCell, or the time of deactivating the SCell. The predicted CA processing time can be associated with the predicted CA information or the SCell information, or the predicted SCell candidate information.

[0404] (7) is information on a predicted resource status of the base station. The information on the resource status of the base station can be set to the information disclosed in the input information example (7) for the CA AI / ML model. The resource status of the base station can be the predicted CA information of (2), or the predicted resource status in the predicted SCell candidate of (4).

[0405] (8) is information on a predicted traffic amount of the UE. The information on the traffic amount of the UE can be set to the information disclosed in the input information example (13) for the CA AI / ML model. The information on the predicted traffic amount of the UE can be the predicted CA information of (2), or the predicted traffic status of the UE in the predicted SCell candidate of (4).

[0406] The output information of the CA AI / ML model disclosed above is appropriately notified to the RAN node or the UE that performs the CA processing. Thereby, more efficient CA can be performed using the CA AI / ML model.

[0407] The base station transmits the CA AI / ML model output information and the like to the neighboring base station. As for the transmission of the CA AI / ML model output information and the like between the base stations, the transmission method of the DC AI / ML model output information and the like between the base stations disclosed in Embodiment 1 can be appropriately applied. The DC AI / ML model output information can be replaced with the CA AI / ML model output information to be appropriately applied.

[0408] The CU transmits the CA AI / ML model output information and the like to the DU. As for the transmission of the CA AI / ML model output information and the like between the CU and the DU, the transmission method of the DC AI / ML model output information and the like between the CU and the DU disclosed in Embodiment 1 can be appropriately applied. The DC AI / ML model output information can be replaced with the CA AI / ML model output information to be appropriately applied.

[0409] The base station transmits the CA AI / ML model output information and the like to the UE. As for the transmission of the CA AI / ML model output information and the like between the UE and the base station, the transmission method of the DC AI / ML model output information and the like between the UE and the base station disclosed in Embodiment 1 can be appropriately applied. The DC AI / ML model output information can be replaced with the CA AI / ML model output information to be appropriately applied.

[0410] Feedback information of CA using the CA AI / ML model is disclosed. Five examples of feedback information of CA processing using the CA AI / ML model are disclosed below.

[0411] (1) information on a performance of the UE.

[0412] (2) information on an L2 measurement of the UE.

[0413] (3) Information related to L2 measurement with the base station.

[0414] (4) Resource status of the base station.

[0415] (5) Combination of (1) to (4).

[0416] (1) For example, (5) can be the input information example for the CA AI / ML model described above.

[0417] (2) For example, (15) can be the input information example for the CA AI / ML model described above.

[0418] (3) For example, (17) can be the input information example for the CA AI / ML model described above.

[0419] (4) For example, (6) can be the input information example for the CA AI / ML model described above.

[0420] The CA-related node appropriately notifies the node having the CA AI / ML model of the CA processing feedback information using the CA AI / ML model disclosed above. Thereby, evaluation of the CA AI / ML model can be performed, and update of the model can be performed. Thereby, more efficient CA can be performed.

[0421] The base station transmits the CA processing feedback information using the CA AI / ML model and the like to the neighboring base station. With regard to the transmission of the CA processing feedback information using the CA AI / ML model and the like between the base stations, the method of transmission of the DC processing feedback information using the DC AI / ML model and the like between the base stations disclosed in Embodiment 1 can be appropriately applied. The DC processing feedback information using the DC AI / ML model can be replaced with the CA processing feedback information using the CA AI / ML model and appropriately applied.

[0422] The DU transmits the CA processing feedback information using the CA AI / ML model and the like to the CU. With regard to the transmission of the CA processing feedback information using the CA AI / ML model and the like between the CU and the DU, the method of transmission of the DC processing feedback information using the DC AI / ML model and the like between the CU and the DU disclosed in Embodiment 1 can be appropriately applied. The DC processing feedback information using the DC AI / ML model can be replaced with the CA processing feedback information using the CA AI / ML model and appropriately applied.

[0423] The base station sends CA processing feedback information using an AI / ML model for CA to the UE. Regarding the transmission of CA processing feedback information using an AI / ML model for CA between the UE and the base station, the method for transmitting DC processing feedback information using an AI / ML model for DC between the CU and DU disclosed in Implementation 1 can be appropriately applied. The DC processing feedback information using an AI / ML model for DC can be appropriately replaced with CA processing feedback information using an AI / ML model for CA.

[0424] Figure 20 This is a diagram illustrating an example of sequence processing using CA with AI / ML. It discloses an example of a CU with a CA with AI / ML model, and its training and inference. (For...) Figure 18 Common steps are labeled with the same step number, and common descriptions are omitted. In step ST1501, CU provides an AI / ML model for CA.

[0425] In step ST1511, the CU sends a request to the UE for input information for the AI / ML model used in CA. This request may contain information related to the UE as the CA object. The information related to the UE may be the UE's identifier. The request may contain information indicating the required information. The request may contain information indicating it is for model training. For example, this is valid if the input information for the AI / ML model used in CA includes information for model training. The transmission of this request may, for example, use RRC signaling. The CU may send this request to the DU to which the UE is connected. The DU sends the request to the UE. Transmission from the CU to the DU may use F1 signaling. In step ST1514, the UE sends the input information for the AI / ML model used in CA to the CU. This information may be transmitted via a measurement result report of the UE. This measurement result may be a wireless measurement result, an L2 measurement result, or an MDT measurement result. Recording the MDT measurement result may be the result measured by the UE in RRC_Inactive or RRC_Idle state. These measurement results may be transmitted via the same signaling or via different signaling. This transmission may use RRC signaling. For example, wireless measurement results and L2 measurement results can be transmitted using a MeasurementReport message. Similarly, MDT measurement results can be transmitted using a UE InformationResponse message. The UE can send this information to the DU it is connected to, and the DU then sends this information to the CU.

[0426] In step ST1512, step ST1513, the CU transmits a request for the CA AI / ML model input information to the DU. The request can include information related to the UE as a CA object. The information related to the UE can be an identifier of the UE. The DU is not limited to one, and can be plural. The DU can be provided to be connected to the CU. The request can include information indicating the required information. The request can include information indicating that it is for model training. For example, it is effective in a case where the CA AI / ML model input information is provided as information for model training. The transmission of the request can use, for example, F1 signaling. In step ST1515, step ST1516, the DU transmits the CA AI / ML model input information to the CU. The transmission of the information can use, for example, F1 signaling.

[0427] In step ST1517, the CU performs training of the CA AI / ML model using the CA AI / ML model input information held by the node and the CA AI / ML model input information acquired from the UE and the DU. The CA AI / ML model is updated by the training. These processes can be appropriately performed. The processes can be performed periodically, or can be performed by the base station each time HO is performed.

[0428] The CU determines to perform CA. In step ST1521 to step ST1526, transmission of a request for the inference AI / ML model input information for CA, and transmission of the inference AI / ML model input information for CA are performed among the UE, the DU, and the CU. These processes can appropriately apply the processes of step ST1511 to step ST1516 described above. The model training can be replaced with inference. In step ST1527, the CU performs inference by the CA AI / ML model, and derives the CA AI / ML model output information. The CU uses the output information to decide the SCell, the SCell candidate used in CA, and performs CA setting. In step ST1531, CA processing is performed among the CU, the DU, and the UE. Thereby, the output information derived by the CA AI / ML model can be used to perform CA processing among the CU, the DU, and the UE.

[0429] In step ST1541, step ST1542, step ST1543, the CU transmits a request for the UE and the DU to send CA processing feedback information using the AI / ML model for CA to the CU. The DU can be a DU related to CA. The request can include information indicating the required information. The transmission of the request from the CU to the UE can be performed via the DU. The transmission of the request between the UE and the DU can use, for example, RRC signaling. The transmission of the request between the DU and the CU can use, for example, F1 signaling. In step ST1544, step ST1545, step ST1546, the UE and the DU transmit CA processing feedback information using the AI / ML model for CA to the CU. The transmission of the information from the UE to the CU can be performed via the DU. The transmission of the information between the UE and the DU can use, for example, RRC signaling. The transmission of the information between the DU and the CU can use, for example, F1 signaling.

[0430] Thus, the CU can acquire the performance of CA processing using the AI / ML model for CA. The CU can use the CA processing feedback information using the AI / ML model for CA held by the node, the CA processing feedback information using the AI / ML model for CA acquired from the UE and the DU, to evaluate the effect of CA processing in the case of using the AI / ML model for CA.

[0431] Thus, CA processing can be performed using the AI / ML model for CA. More effective CA processing can be performed.

[0432] By adopting the method disclosed in the present embodiment, training and inference in CA processing can be performed using the AI / ML model for CA. For example, by using the inference result using the AI / ML model for CA, more effective processing can be performed in the selection of SCell, the selection of SCell candidates, and the like. Thus, high-speed and large-capacity communication can be achieved in CA processing, and various KPIs can be optimized.

[0433] Embodiment 3. AI / ML-based HO processing and DC processing are proposed. In order to optimize various KPIs, a method for more effectively performing AI / ML-based HO processing and DC processing is required. In the present embodiment, a method is disclosed, which can implement more effective AI / ML-based HO processing and DC processing.

[0434] In the present embodiment, an AI / ML model for HO that takes CA into account (hereinafter, sometimes referred to as an AI / ML model for HO that takes CA into account) is introduced. Further, an AI / ML model for DC that takes CA into account (hereinafter, sometimes referred to as an AI / ML model for DC that takes CA into account) is introduced. The AI / ML model for HO can apply, for example, the AI / ML model for mobility disclosed in Chapter 5.3.2 of Non-Patent Literature 30. The AI / ML model for DC can apply, for example, the AI / ML model for DC disclosed in Embodiment 1 as appropriate.

[0435] A method of taking CA into account in the AI / ML model for HO and the AI / ML model for DC is disclosed.

[0436] Input information for the AI / ML model for HO that takes CA into account and input information for the AI / ML model for DC that takes CA into account are set. As the input information for the AI / ML model for HO that takes CA into account, the input information for the AI / ML model for HO and the input information for the AI / ML model for CA disclosed in Embodiment 2 are combined. As the input information for the AI / ML model for DC that takes CA into account, the input information for the AI / ML model for DC disclosed in Embodiment 1 and the input information for the AI / ML model for CA disclosed in Embodiment 2 are combined. A part or all of the information can be combined. Thereby, the input information for the AI / ML model for HO that takes CA into account and the input information for the AI / ML model for DC that takes CA into account can be set.

[0437] The base station transmits the input information for the AI / ML model for HO that takes CA into account and the input information for the AI / ML model for DC that takes CA into account (hereinafter, sometimes collectively referred to as “input information for the AI / ML model for HO / DC that takes CA into account”) and the like to the neighboring base station. As for the transmission of the input information for the AI / ML model for HO / DC that takes CA into account and the like between the base stations, the transmission method of the input information for the AI / ML model for DC disclosed in Embodiment 1 can be applied as appropriate. The input information for the AI / ML model for DC can be replaced with the input information for the AI / ML model for HO / DC that takes CA into account as appropriate.

[0438] The DU transmits the input information for the AI / ML model for HO / DC that takes CA into account and the like to the CU. As for the transmission of the input information for the AI / ML model for HO / DC that takes CA into account and the like between the CU and the DU, the transmission method of the input information for the AI / ML model for DC disclosed in Embodiment 1 can be applied as appropriate. The input information for the AI / ML model for DC can be replaced with the input information for the AI / ML model for HO / DC that takes CA into account as appropriate.

[0439] The UE transmits to the base station input information for the AI / ML model for CA-considered HO / DC. With regard to the transmission of the input information for the AI / ML model for CA-considered HO / DC between the UE and the base station, the method of transmission of the input information for the AI / ML model for DC disclosed in Embodiment 1 can be appropriately applied. The input information for the AI / ML model for DC can be replaced with the input information for the AI / ML model for CA-considered HO / DC to be appropriately applied.

[0440] The AI / ML model output information for HO considering CA (AI / ML model output information for CA-considered HO) and the AI / ML model output information for DC considering CA (AI / ML model output information for CA-considered DC) are set. As the AI / ML model output information for CA-considered HO, the AI / ML model output information for HO and the AI / ML model output information for CA disclosed in Embodiment 2 are combined. As the AI / ML model output information for CA-considered DC, the AI / ML model output information for DC disclosed in Embodiment 1 and the AI / ML model output information for CA disclosed in Embodiment 2 are combined. A part or all of the information can be combined. As the method of combination, for example, information on CA can be added to the information on the HO target, the DC target. For example, information on the SCell performing CA in the base station or the PCell of the HO target can be added to the information on the base station or the PCell. For example, information on the SCell performing CA in the base station or the PSCell of the DC target can be added to the information on the base station or the PSCell. For example, information on CA can be added to the information on the candidate PSCell of CHO (Conditional Handover). Information on the candidate SCell can be added. For example, information on the candidate SCell performing CA in the PCell can be added to the information on each candidate PCell. For example, information on CA can be added to the information on the candidate PSCell of CPC. Information on the candidate SCell can be added. For example, information on the candidate SCell performing CA in the PSCell can be added to the information on each candidate PSCell.

[0441] As another example of the method of combination, for example, information on CA can be added to the traffic amount information of the UE of the base station of the HO target, the DC target. For example, information on the traffic amount of the UE in the SCell predicted for CA in the base station of the predicted HO target, the DC target can be added.

[0442] As other examples of the combination method, for example, information on resource status of the base station of the HO target, the DC target can be appended with information on CA. For example, information on resource status of an SCell predicted for CA in the base station of the predicted HO target, the DC target can be appended.

[0443] Thereby, it is possible to set the AI / ML model output information for HO / DC with CA taken into account.

[0444] The base station transmits the AI / ML model output information for HO / DC with CA taken into account to the neighboring base station or the like. As for the transmission of the AI / ML model output information for HO / DC with CA taken into account between the base stations, the transmission method of the AI / ML model output information for DC between the base stations disclosed in Embodiment 1 can be appropriately applied. The AI / ML model output information for DC can be replaced with the AI / ML model output information for HO / DC with CA taken into account to be appropriately applied.

[0445] The CU transmits the AI / ML model output information for HO / DC with CA taken into account to the DU or the like. As for the transmission of the AI / ML model output information for HO / DC with CA taken into account between the CU and the DU, the transmission method of the AI / ML model output information for DC between the CU and the DU disclosed in Embodiment 1 can be appropriately applied. The AI / ML model output information for DC can be replaced with the AI / ML model output information for HO / DC with CA taken into account to be appropriately applied.

[0446] The base station transmits the AI / ML model output information for HO / DC with CA taken into account to the UE or the like. As for the transmission of the AI / ML model output information for HO / DC with CA taken into account between the UE and the base station, the transmission method of the AI / ML model output information for DC between the UE and the base station disclosed in Embodiment 1 can be appropriately applied. The AI / ML model output information for DC can be replaced with the AI / ML model output information for HO / DC with CA taken into account to be appropriately applied.

[0447] The system sets up HO processing feedback information using an AI / ML model that takes CA into account (HO processing feedback information using an AI / ML model that takes CA into account) and DC processing feedback information using an AI / ML model that takes CA into account (DC processing feedback information using an AI / ML model that takes CA into account). As HO processing feedback information using an AI / ML model that takes CA into account, the system combines the HO processing feedback information using an AI / ML model with the CA processing feedback information using an AI / ML model disclosed in Embodiment 2. As DC processing feedback information using an AI / ML model that takes CA into account, the system combines the DC processing feedback information using an AI / ML model disclosed in Embodiment 1 and the CA processing feedback information using an AI / ML model that takes CA into account disclosed in Embodiment 2. Part or all of the information can be combined. Therefore, it is possible to set up HO processing feedback information using an AI / ML model that takes CA into account and DC processing feedback information using an AI / ML model that takes CA into account.

[0448] The base station transmits HO processing feedback information using an AI / ML model that considers CA to neighboring base stations and DC processing feedback information using an AI / ML model that considers CA to neighboring base stations (hereinafter, they are sometimes collectively referred to as "HO / DC processing feedback information using an AI / ML model that considers CA"). Regarding the transmission of HO / DC processing feedback information using an AI / ML model that considers CA between base stations, the method for transmitting DC processing feedback information using an AI / ML model that uses a DC model between base stations disclosed in Implementation 1 can be appropriately applied. The DC processing feedback information using an AI / ML model can be appropriately replaced with HO / DC processing feedback information using an AI / ML model that considers CA.

[0449] The DU sends HO / DC processing feedback information using an AI / ML model that considers CA to the CU. Regarding the transmission of HO / DC processing feedback information using an AI / ML model that considers CA between CU and DU, the transmission method for DC processing feedback information using an AI / ML model that uses DC between CU and DU, as disclosed in Implementation 1, can be appropriately applied. The DC processing feedback information using an AI / ML model can be replaced with HO / DC processing feedback information using an AI / ML model that considers CA.

[0450] The UE transmits feedback information on the HO / DC processing using the AI / ML model for HO / DC considering CA, and the like, to the base station. With regard to the transmission of the feedback information on the HO / DC processing using the AI / ML model for HO / DC considering CA, and the like, between the UE and the base station, the transmission method of the feedback information on the DC processing using the AI / ML model for DC disclosed in Embodiment 1 can be appropriately applied. The feedback information on the DC processing using the AI / ML model for DC can be replaced with the feedback information on the HO / DC processing using the AI / ML model for HO / DC considering CA to be appropriately applied.

[0451] Figure 21 is a sequence example illustrating the HO processing using the AI / ML for HO considering CA. The steps common to Figure 18 The same step number is given to the common steps, and the common description is omitted. In step ST1601, the RAN node (S-RAN node) that becomes the HO source is provided with the AI / ML model for HO considering CA.

[0452] In step ST1604, the S-RAN node transmits a request for input information for the AI / ML model for HO considering CA to the UE. The request can include information on the UE that is the object of the HO. The information on the UE can be an identifier of the UE. The request can include information indicating that the information is for model training. For example, it is effective in the case where the input information for the AI / ML model for HO considering CA is provided with information for model training. The transmission of the request can use, for example, RRC signaling. The transmission from the S-RAN node to the UE can use RRC signaling. In step ST1606, the UE transmits the input information for the AI / ML model for HO considering CA to the S-RAN node. As the information, the measurement results of the UE can be transmitted by a measurement result report. The measurement results can be radio measurement results, can be L2 measurement results, or can be measurement results of MDT. The measurement results of MDT can be results measured by the UE in the RRC_Inactive or RRC_Idle state. These measurement results can be transmitted by the same signaling or can be transmitted by different signaling. The transmission can use RRC signaling. For example, the transmission of the radio measurement results and the L2 measurement results can use a MeasurementReport message. For example, the transmission of the measurement results of MDT can use a UEInformationResponse message.

[0453] In step ST1605, the S-RAN node transmits a request for input information for the AI / ML model for CA-considered HO to the neighboring base station. The request can include information about the UE that is the object of HO. The information about the UE can be an identifier of the UE. The neighboring base station is not limited to one, and can be plural. The neighboring base station can include the T-RAN node. In other words, the T-RAN node can be selected within the neighboring base station. The request can include information indicating that the information is for model training. For example, this is effective in a case where the input information for the AI / ML model for CA-considered HO is provided with information for model training. The transmission of the request can use, for example, Xn signaling. In step ST1607, the neighboring base station transmits the input information for the AI / ML model for CA-considered HO to the S-RAN node. The transmission of the information can use, for example, Xn signaling.

[0454] In step ST1608, the S-RAN node performs training of the AI / ML model for CA-considered HO using the input information for the AI / ML model for CA-considered HO held by the node and the input information for the AI / ML model for CA-considered HO acquired from the UE and the neighboring base station. The AI / ML model for CA-considered HO is updated by the training. These processes can be performed as appropriate. They can be performed periodically, or each time HO is performed in the base station or the surrounding base stations.

[0455] The S-RAN node determines to perform CA-considered HO. In steps ST1609 to ST1612, transmission of a request for input information for the AI / ML model for CA-considered HO for inference and transmission of the input information for the AI / ML model for CA-considered HO for inference are performed among the UE, the S-RAN node, and the T-RAN node. These processes can apply the processes of steps ST1604 to ST1607 described above as appropriate. The model training can be replaced with inference. In step ST1613, the S-RAN node performs inference by the AI / ML model for CA-considered HO and derives output information for the AI / ML model for CA-considered HO. The S-RAN node decides the T-RAn node of the HO target using the output information while taking CA into consideration in the HO target, and performs setting of HO. In step ST1614, HO processing is performed among the S-RAN node, the UE, and the T-RAN node.

[0456] In step ST1615, the S-RAN node can transmit AI / ML model output information for CA-considered HO to the T-RAN node. For example, the S-RAN node uses the AI / ML model for CA-considered HO to select an SCell or an SCell candidate for CA in the T-RAN node, and transmits information indicating the result of the selection as the AI / ML model output information for CA-considered HO. The S-RAN node transmits the information to the T-RAN node. In step ST1621, the T-RAN node uses the information to decide the CC. An SCell for CA can be decided. In step ST1622, CA processing is performed between the UE and the T-RAN node. The processing of step ST1615 and / or step ST1621 and / or step ST1622 can be performed in the HO processing of step ST1614. For example, the AI / ML model output information for CA-considered HO can be transmitted by being included in an HO request message transmitted from the S-RAN node to the T-RAN node. For example, the setting of the SCell decided by the T-RAN node can be transmitted by being included in an HO request response message transmitted from the T-RAN node to the S-RAN node. The S-RAN node can transmit the setting of the SCell decided by the T-RAN node to the UE in an RRC Reconfiguration for HO.

[0457] Thus, it is possible to perform CA-considered HO processing between the UE, the S-RAN node, and the T-RAN node using the output information derived by the AI / ML model for CA-considered HO.

[0458] In steps ST1631 and ST1632, the S-RAN node transmits a request for HO processing feedback information using the AI / ML model for CA-considered HO to the UE and the T-RAN node. The request can be transmitted from the S-RAN node to the UE via the T-RAN node. The request can include information indicating the required information. In steps ST1633 and ST1634, the UE and the T-RAN node transmit the HO processing feedback information using the AI / ML model for CA-considered HO to the S-RAN node. The information can be transmitted from the UE to the S-RAN node via the T-RAN node. The transmission of the information between the UE and the base station can use, for example, RRC signaling. The transmission of the information between the base stations can use, for example, F1 signaling.

[0459] Thus, the S-RAN node can acquire the performance of the CA-considered HO processing using the AI / ML model for CA-considered HO. The S-RAN node can evaluate the effect of the CA-considered HO processing in the case where the AI / ML model for CA-considered HO is used, using the HO processing feedback information using the AI / ML model for CA-considered HO held by the node, and the HO processing feedback information using the AI / ML model for CA-considered HO acquired from the UE and the T-RAN node.

[0460] Thus, the CA-considered HO processing can be performed using the AI / ML model for CA-considered HO. More effective HO processing can be performed.

[0461] By adopting the method disclosed in the present embodiment, information on CA can be transmitted between base stations, between CUs and DUs, and between UEs and base stations. By combining the information on CA, such as information on SCells, with input information, output information, and feedback information of the AI / ML model for HO / DC, an AI / ML model for HO / DC that takes CA into consideration can be set. By using the AI / ML model for HO / DC that takes CA into consideration, a HO target or a DC target that can perform more effective CA can be selected. By performing HO or DC processing that takes CA into consideration, high-speed and large-capacity communication can be achieved, and various KPIs can be optimized.

[0462] Embodiment 4. In the present embodiment, other methods by which more effective HO processing and DC processing using AI / ML can be implemented are disclosed.

[0463] In the present embodiment, an AI / ML model for HO that takes beams into consideration (hereinafter, sometimes referred to as an AI / ML model for HO that takes beams into consideration) is introduced. In addition, an AI / ML model for DC that takes beams into consideration (hereinafter, sometimes referred to as an AI / ML model for DC that takes beams into consideration) is introduced. Methods by which beams are taken into consideration in the AI / ML model for HO and the AI / ML model for DC are disclosed.

[0464] Input information for the AI / ML model for HO that takes beams into consideration and input information for the AI / ML model for DC that takes beams into consideration (hereinafter, sometimes collectively referred to as “input information for the AI / ML model for HO / DC that takes beams into consideration”) are set. Eight examples of the input information for the AI / ML model for HO / DC that takes beams into consideration are disclosed below.

[0465] (1) Information related to radio measurement of the UE.

[0466] (2) Information related to radio measurement of the base station.

[0467] (3) Beam history information.

[0468] (4) Beam history information from a neighboring base station.

[0469] (5) Communication interruption status in past beams.

[0470] (6) Communication interruption status in past beams from a neighboring base station.

[0471] (7) Information related to a trajectory of the UE.

[0472] (8) Combination of (1) to (7).

[0473] (1) For example, information related to wireless measurement of the UE. Wireless measurement results from each beam of the base station can be added. Furthermore, information related to a position of the UE can be included. Information related to time can be included. Wireless measurement results of each beam can be associated with information about the position or time.

[0474] (2) For example, information related to wireless measurement of the base station. Wireless measurement results from each beam of the UE can be added. Furthermore, information related to a position of the UE can be included. Information related to time can be included. Wireless measurement results of each beam can be associated with information about the position or time.

[0475] (3) For example, history of beams received by the UE in the past. The beam information can be, for example, information related to SSB (Synchronization Signal Block). It can be information related to CSI-RS. It can be information related to SSBRI (SS / PBCH Resource Indicator). It can be information related to CRI (CSI-RS Resource Indicator). For example, it can be information related to a direction of the beam. For example, it can be information related to a structure of the beam. It can be history information of past base stations or cells (PCell, PSCell, SPCell, SCell) of the UE and information of beams used in each base station or cell. These information can be listed. As other examples, it can be information related to beams received by the base station from the UE in the past. Furthermore, information related to a position of the UE can be included. Information related to time can be included. Information of past beams can be associated with information about the position or time.

[0476] (4) Is history information of beams that a neighboring base station has. The beam information can be information disclosed in (3).

[0477] (5) is information on communication interruption in a past beam of the UE. For example, it can be historical information on communication interruption in a past beam of the UE. As the communication interruption, for example, it can be a beam failure. It can be information on a beam failure request. For example, it can be information identifying a beam in which the communication interruption occurred. It can be historical information of past base stations or cells (PCell, PSCell, SPCell, SCell) of the UE and information on communication interruption in beams used in each base station or cell. These information can be listed. As other examples, it can be information on communication interruption in a past beam received by the base station from the UE. Further, it can include information on a location of the UE. It can include information on time. It can associate information on communication interruption in a past beam with information on a location or time.

[0478] (6) is information on communication interruption in a past beam of a neighboring base station. The information on communication interruption in a past beam can be information disclosed in (5).

[0479] (7) is information on a trajectory of the UE. For example, it can be the beam history information disclosed in (3). It can be information on a past, current trajectory of the UE. The information on a past, current trajectory of the UE can be historical information. As the information on a trajectory, it can be information on a future trajectory of the UE. The information on a future trajectory can be predicted information. These information can be listed. Further, it can include information on time. It can associate the information on a trajectory of the UE with information on time.

[0480] The AI / ML model for HO / DC considering a beam disclosed above can be combined with the AI / ML model for HO / DC disclosed in Embodiment 3 using input information to be AI / ML model for HO / DC considering a beam. A part or all of each information can be combined. The AI / ML model for HO can apply, for example, the AI / ML model for mobility disclosed in Chapter 5.3.2 of Non-Patent Literature 30. The AI / ML model for DC can apply, for example, the AI / ML model for DC disclosed in Embodiment 1 as appropriate.

[0481] The AI / ML model for HO / DC considering a beam is input to the AI / ML model for HO considering a beam and the AI / ML model for DC considering a beam (hereinafter, collectively referred to as “AI / ML model for HO / DC considering a beam”) for training or inference. Thereby, more efficient HO or DC can be performed using the AI / ML model for HO / DC considering a beam.

[0482] The AI / ML model output information for HO that takes into account beams (AI / ML model output information for HO that takes into account beams) and the AI / ML model output information for DC that takes into account beams (AI / ML model output information for DC that takes into account beams) are set. Hereinafter, these output information are collectively referred to as “AI / ML model output information for HO / DC that takes into account beams”. As the AI / ML model output information for HO / DC that takes into account beams, the AI / ML model output information for HO / DC disclosed in Embodiment 1 (AI / ML model output information for HO and AI / ML model output information for DC) can be supplemented with information about beams. For example, information about beams can be supplemented to the information of the HO target, the information of the DC target. For example, information about beams used in the base station or the PCell of the HO target can be supplemented to the information of the base station or the PCell. For example, information about beams used in the base station or the PSCell of the DC target can be supplemented to the information of the base station or the PSCell. For example, information about beams used in the candidate PCell of the CHO can be supplemented to the information of the candidate PCell. Information about candidate beams can be supplemented. For example, information about beams can be supplemented to the information of the candidate PSCell of the CPC. Information about candidate beams can be supplemented.

[0483] Thus, the AI / ML model output information for HO / DC that takes into account beams can be set.

[0484] The HO processing feedback information using the AI / ML model for HO that takes into account beams (HO processing feedback information using the AI / ML model for HO that takes into account beams) and the DC processing feedback information using the AI / ML model for DC that takes into account beams (DC processing feedback information using the AI / ML model for DC that takes into account beams) are set. Hereinafter, these feedback information are collectively referred to as “HO / DC processing feedback information using the AI / ML model for HO / DC that takes into account beams”. As the HO / DC processing feedback information using the AI / ML model for HO / DC that takes into account beams, the HO / DC processing feedback information using the AI / ML model for HO / DC disclosed in Embodiment 1 (HO processing feedback information using the AI / ML model for HO and DC processing feedback information using the AI / ML model for DC) can be supplemented with information about beams. Thus, the HO / DC processing feedback information using the AI / ML model for HO / DC that takes into account beams can be set.

[0485] The transmission method of the input information, the output information, and the feedback information for the AI / ML model for DC between the base stations, between the CU and the DU, and between the UE and the base station can be appropriately applied to the transmission method of the input information, the output information, and the feedback information for the AI / ML model for DC disclosed in Embodiment 1.

[0486] Figure 22 is a sequence example illustrating the HO processing using the AI / ML for HO considering the beam. The steps common to Figure 18 Embodiment 1 are denoted by the same step numbers, and the common explanations are omitted. In step ST1701, the RAN node (S-RAN node) that becomes the HO source is provided with the AI / ML model for HO considering the beam.

[0487] In step ST1704, the S-RAN node transmits a request for the input information for the AI / ML model for HO considering the beam to the UE. The request can include information about the UE that is the object of HO. The information about the UE can be the identification of the UE. In step ST1706, the UE transmits the input information for the AI / ML model for HO considering the beam to the S-RAN node. The transmission method thereof can be appropriately applied to the method disclosed in steps ST1604 and ST1606 of Figure 21 Embodiment 1.

[0488] In step ST1705, the S-RAN node transmits a request for the input information for the AI / ML model for HO considering the beam to the neighboring base station. The request can include information about the UE that is the object of HO. The information about the UE can be the identification of the UE. In step ST1707, the neighboring base station transmits the input information for the AI / ML model for HO considering the CA to the S-RAN node. The transmission method thereof can be appropriately applied to the method disclosed in steps ST1605 and ST1607 of Figure 21 Embodiment 1.

[0489] In step ST1708, the S-RAN node performs the training of the AI / ML model for HO considering the beam using the input information for the AI / ML model for HO considering the beam held by the node and the input information for the AI / ML model for HO considering the beam acquired from the UE and the neighboring base station. The update of the AI / ML model for HO considering the beam is performed by the training. These processes can be appropriately performed. They can be performed periodically, or each time the HO is performed in the base station or the surrounding base stations.

[0490] The S-RAN node judges the beam consideration HO. In steps ST1709 to ST1712, transmission of an input information request for the beam consideration HO AI / ML model for inference, and transmission of input information for the beam consideration HO AI / ML model for inference are performed among the UE, the S-RAN node, and the T-RAN node. These processes can appropriately apply the processes of steps ST1704 to ST1707 described above. The model training can be replaced with inference. In step ST1713, the S-RAN node performs inference by the beam consideration HO AI / ML model, and derives beam consideration HO AI / ML model output information. The S-RAN node takes into account the beam in the HO target, uses the output information to decide the T-RAN node of the HO target, and performs setting of the HO. In step ST1714, HO processing is performed among the S-RAN node, the DU, and the T-RAN node.

[0491] Thus, the beam consideration HO processing can be performed among the UE, the S-RAN node, and the T-RAN node using the output information derived by the beam consideration HO AI / ML model.

[0492] In steps ST1731 and ST1732, the S-RAN node transmits a request for the UE and the T-RAN node to transmit HO processing feedback information using the beam consideration HO AI / ML model. The transmission can be performed from the S-RAN node to the UE via the T-RAN node. The request can include information indicating the required information. In steps ST1733 and ST1734, the UE and the T-RAN node transmit the HO processing feedback information using the beam consideration HO AI / ML model to the S-RAN node. The transmission can be performed from the UE to the S-RAN node via the T-RAN node. The transmission of the information between the UE and the base station can use, for example, RRC signaling. The transmission of the information between the base stations can use, for example, F1 signaling.

[0493] Thus, the S-RAN node can acquire the performance of the beam consideration HO processing using the beam consideration HO AI / ML model. The S-RAN node can use the HO processing feedback information using the beam consideration HO AI / ML model held by the node, and the HO processing feedback information using the beam consideration HO AI / ML model acquired from the UE and the T-RAN node to evaluate the effect of the beam consideration HO processing in the case of using the beam consideration HO AI / ML model.

[0494] Thus, the beam consideration HO processing can be performed using the beam consideration HO AI / ML model. More effective HO processing can be performed.

[0495] By employing the method disclosed in this embodiment, information about beams can be transmitted between base stations, between CUs and DUs, and between UEs and base stations. By adding information about beams to the input information, output information, and feedback information of the AI / ML model for HO / DC, an AI / ML model for HO / DC that takes beams into account can be set. By using an AI / ML model for HO / DC that takes beams into account, a more effective beam of the HO target or DC target can be selected. For example, a RACH corresponding to a more suitable beam of the HO target or DC target can be set, and reduction in failure of HO processing or DC processing and reduction in access time to the HO target or DC target can be achieved. By performing HO or DC processing that takes beams into account, high-speed and high-capacity communication can be achieved, and various KPIs can be optimized.

[0496] Information about beams can be added to the input information, output information, and feedback information of the AI / ML model for CA disclosed in Embodiment 2 and the AI / ML model for HO / DC that takes CA into account disclosed in Embodiment 3. CA or HO / DC that takes CA into account using an AI / ML model that takes beams into account can be performed. More effective CA or HO / DC that takes CA into account can be performed, and high-speed and high-capacity communication can be achieved, and various KPIs can be optimized.

[0497] The method disclosed in Embodiment 2 can be applied to mobility at the beam level. It can also be applied to inter-cell beam management. An AI / ML model that takes beams into account can be introduced to these processes. More effective mobility at the beam level and inter-cell beam management become possible.

[0498] Embodiment 5. In mobile communication systems standardized by 3GPP, in order to improve the reliability of communication, a packet duplication function using DC or CA is provided (Non-Patent Literature 2). In packet duplication, a method that enables more effective packet duplication processing is also required so that not only the reliability is improved but also various KPIs are optimized. In this embodiment, a method that enables more effective packet duplication processing is disclosed.

[0499] In this embodiment, AI / ML is introduced for packet duplication. A packet duplication AI / ML model is provided in a base station. The packet duplication AI / ML model is used to perform, for example, setting of packet duplication such as whether packet duplication is needed, the number of packet duplications, and the like. By introducing AI / ML in packet duplication, more effective packet duplication can be achieved.

[0500] Input information for the packet duplication AI / ML model is provided. Thirteen examples of input information for the packet duplication AI / ML model are disclosed below.

[0501] (1) Information about packet duplication.

[0502] (2) History information of packet duplication of the UE.

[0503] (3) History information of packet duplication of the UE from a neighboring base station.

[0504] (4) Information related to performance of the UE which has performed packet duplication.

[0505] (5) Resource status of the current base station.

[0506] (6) Resource status of the predicted base station.

[0507] (7) Information related to traffic amount of the current UE.

[0508] (8) Information related to traffic amount of the predicted UE.

[0509] (9) Information related to L2 measurement of the UE.

[0510] (10) Predicted value of information related to L2 measurement of the UE.

[0511] (11) Information related to L2 measurement from a neighboring base station.

[0512] (12) Predicted value of information related to L2 measurement from a neighboring base station.

[0513] (13) Combination of (1) to (12).

[0514] (1) For example, it can be setting information of packet duplication. For example, it can be information of packet duplication status, CA-based packet duplication, and / or DC-based packet duplication, information related to a base station used in packet duplication, information related to a cell used in packet duplication, information related to PDCP which has performed packet duplication, number of packet duplications, information related to RLC bearer of packet duplication, number of the RLC bearers, information related to logical channel of packet duplication, and the like. Further, it can include information related to location of the UE. It can include information related to time. It can be associated with information related to location or time.

[0515] (2) For example, it can be a history of packet duplication in the past. As a history of packet duplication, for example, it can be information about a base station and / or a cell used in packet duplication of the UE. It can be a history of the information about packet duplication of the above (1). It can be a history of packet duplication set. It can be a history of packet duplication activated. It is possible to reduce the amount of signaling. Furthermore, it is possible to perform inference with high accuracy. It is possible to list such information. Furthermore, it is possible to include information about a position of the UE. It is possible to include information about time. It is possible to associate the history information of packet duplication of the UE with information about a position or time.

[0516] (3) is history information of packet duplication of the UE possessed by the neighboring base station. The history information of packet duplication can be the information disclosed in (2).

[0517] (4) For example, it can be QoS of the UE which performed packet duplication. It can be packet throughput, packet loss rate, delay time. It can be packet repetition amount, packet repetition rate. The packet repetition amount is the amount of repeated packets. As the amount of packets, it can be the number of packets, or it can be the number of bytes. The packet repetition rate, for example, can be a ratio of the amount of repeated packets to the amount of received packets. It can be the packet repetition amount or the packet repetition rate in a prescribed period. It can be packet discard amount, packet discard rate. The packet discard amount is the amount of discarded packets. The packet discard rate, for example, can be a ratio of the amount of discarded packets to the amount of received packets. It can be the packet discard amount or the packet discard rate in a prescribed period. The information about the performance of the UE can be information before packet duplication is performed. It can also be information after packet duplication is performed. Such information can be for each RLC bearer. It can be for each base station. It can be for each SCell. Furthermore, it is possible to include information about a position of the UE. It is possible to include information about time. It is possible to associate the information about the performance of the UE which performed packet duplication with information about a position or time.

[0518] (5) For example, it can be the current resource usage amount of the base station or the cell for packet duplication. It can be the usage amount of frequency resources. It can be the usage amount of time resources.

[0519] (6) For example, it can be information about predicted resources of the base station or the cell for packet duplication. The information about resources of the base station or the cell used in packet duplication can be the information disclosed in (5).

[0520] (7) In (7), the traffic amount of the UE, for example, can be the number of connected UEs. For example, it can be the number of UEs constituting an RLC bearer. For example, it can be the number of UEs using resources. For example, it can be the current traffic amount of the UE of the base station or the cell for packet duplication.

[0521] (8) For example, information related to predicted traffic volume of the UE. The information related to the traffic volume of the UE can be the information disclosed in (7).

[0522] (9) For example, can be a measurement result of the UE on the L2 measurement information. As the measurement result of the UE on the L2 measurement information, can be a measurement result of the UE on the L2 measurement information in packet duplication. Information related to the L2 measurement of the UE for the input information for AI / ML disclosed in Embodiment 1 or Embodiment 2 can be appropriately applied.

[0523] (10) is a predicted value of the L2 measurement information. is a predicted value of the L2 measurement information measured by the UE. can be a predicted value of the L2 measurement information measured by the UE in packet duplication.

[0524] (11) For example, can be a measurement result of the base station. can be a measurement result of the L2 measurement information possessed by the neighboring base station. As the measurement result of the base station, can be a measurement result of the base station in packet duplication. Information related to the L2 measurement from the neighboring base station for the input information for AI / ML disclosed in Embodiment 1 or Embodiment 2 can be appropriately applied.

[0525] (12) is a predicted value of the L2 measurement information. is a predicted value of the L2 measurement information measured by the base station. can be a predicted value of the L2 measurement information measured by the base station in packet duplication. can be a predicted value of the L2 measurement information of the base station possessed by the neighboring base station.

[0526] The input information for the AI / ML model for packet duplication disclosed above can be information related to packet duplication of the DL or information related to packet duplication of the UL.

[0527] The input information for the AI / ML model for packet duplication disclosed above, the input information for the AI / ML model for DC, and the input information for the AI / ML model for CA can be combined as the input information for the AI / ML model for packet duplication. A part or all of each information can be combined.

[0528] The input information for the AI / ML model for packet duplication is input to the AI / ML model for packet duplication for training or inference. As a result, more efficient packet duplication can be performed using the AI / ML model for packet duplication.

[0529] Output information of the AI / ML model for packet duplication is disclosed. Six examples of the output information of the AI / ML model for packet duplication are disclosed below.

[0530] (1) Information related to predicted packet duplication.

[0531] (2) Priority order of predicted packet duplication.

[0532] (3) predicted packet duplication processing time.

[0533] (4) predicted resource status of base station.

[0534] (5) information related to predicted traffic of UE.

[0535] (6) combination of (1) to (5).

[0536] (1) is information related to predicted packet duplication by AI / ML model. The information related to packet duplication can be set to the information disclosed in the input information example (1) for AI / ML model for packet duplication. For example, it can be information related to base station for packet duplication. It can be information related to cell for packet duplication. The base station or cell for packet duplication can be one or more. In addition, it can include information related to predicted location of UE. It can include information related to predicted time. It can be associated with information on predicted packet duplication and information on location or time. The information related to predicted packet duplication is not limited to one, and can be plural. It can include information on prediction accuracy of information related to predicted packet duplication by AI / ML model.

[0537] (2) is, for example, information on priority order of predicted base station or cell for packet duplication disclosed in (1). It can be that predicted base station or cell for packet duplication is associated with its priority order. As another example, it can be information on priority order of information related to predicted packet duplication. The information related to predicted packet duplication and its priority order can be associated.

[0538] (3) is predicted packet duplication processing time. As the packet duplication processing time, for example, it can be a set time to start packet duplication, a set time to perform packet duplication, a set time to end packet duplication. The packet duplication processing time can be set, for example, for each packet duplication. For example, it can be set for each base station for packet duplication or each cell. For example, it can be a time to activate a cell for packet duplication, a time when the cell is being activated, a time to deactivate the cell, a time when the cell is being deactivated. For example, it can be set for each RLC bearer or each logical channel. The predicted packet duplication processing time can be associated with information on predicted packet duplication.

[0539] (4) is information related to predicted resource status of base station. The information related to resource status of base station can be set to the information disclosed in the input information example (6) for AI / ML model for packet duplication. The resource status of base station can be, for example, predicted resource status in predicted base station or cell for packet duplication.

[0540] (5) is information related to the predicted traffic volume of the UE. The information related to the traffic volume of the UE can be set to the information disclosed in the input information example (8) for the AI / ML model for packet duplication. The information related to the predicted traffic volume of the UE may, for example, be information related to the predicted traffic volume of the UE in the base station or cell for packet duplication.

[0541] The AI / ML model output information for packet duplication disclosed above can be information related to packet duplication of DL or information related to packet duplication of UL.

[0542] The AI / ML model output information for packet duplication disclosed above, the AI / ML model output information for DC, and the AI / ML model output information for CA can be combined as the AI / ML model output information for packet duplication. Part or all of each information can be combined.

[0543] The AI / ML model output information for packet duplication disclosed above is appropriately notified to the RAN node or the UE that performs packet duplication. Thereby, more efficient packet duplication can be performed using the AI / ML model for packet duplication.

[0544] Feedback information of packet duplication using the AI / ML model for packet duplication is disclosed. Six feedback information examples of packet duplication processing using the AI / ML model for packet duplication are disclosed below.

[0545] (1) Information related to the performance of the UE that performs packet duplication.

[0546] (2) Information related to L2 measurement of the UE.

[0547] (3) Information related to L2 measurement of the base station.

[0548] (4) Resource status of the base station.

[0549] (5) Information related to the traffic volume of the UE.

[0550] (6) Combination of (1) to (5).

[0551] (1) may, for example, be (4) of the input information example for the AI / ML model for packet duplication.

[0552] (2) may, for example, be (9) of the input information example for the AI / ML model for packet duplication.

[0553] (3) may, for example, be (11) of the input information example for the AI / ML model for packet duplication.

[0554] (4) may, for example, be (5) of the input information example for the AI / ML model for packet duplication.

[0555] (5) can be (7) for example, an example of input information for an AI / ML model for packet duplication.

[0556] The packet duplication processing feedback information disclosed above using an AI / ML model for packet duplication is appropriately notified to a node having an AI / ML model for packet duplication. Thereby, evaluation of the AI / ML model for packet duplication can be performed, and updating of the model can be performed. Thereby, more efficient packet duplication can be performed.

[0557] The transmission method of input information, output information, and feedback information for an AI / ML model for packet duplication between base stations, between CUs and DUs, and between UEs and base stations can be appropriately applied to the transmission method of input information, output information, and feedback information for an AI / ML model for DC disclosed in Embodiment 1.

[0558] Figure 23 is a diagram showing an example of a sequence of packet duplication processing using an AI / ML for packet duplication. The steps common to Figure 15 The same step number is noted for the steps common to the steps, and common explanations are omitted. In step ST1801, the MN is provided with an AI / ML model for packet duplication. Figure 23 In the example of, a case where DC using MN and SN is performed is disclosed.

[0559] In step ST1804, the MN transmits a request for input information for an AI / ML model for packet duplication to the UE. The request can include information about the UE as a packet duplication target. The information about the UE can be an identifier of the UE. In step ST1806, the UE transmits input information for an AI / ML model for packet duplication to the MN. The transmission method thereof can be appropriately applied to the method disclosed in steps ST1104, ST1106 of. Figure 15

[0560] In step ST1805, the MN transmits a request for input information for an AI / ML model for packet duplication to the SN. The request can include information about the UE as a packet duplication target. The information about the UE can be an identifier of the UE. In step ST1807, the SN transmits input information for an AI / ML model for packet duplication to the MN. The transmission method thereof can be appropriately applied to the method disclosed in steps ST1105, ST1107 of. Figure 15

[0561] ​​In step ST1808, the MN performs training of the packet duplication AI / ML model using the packet duplication AI / ML model input information held by the node and the packet duplication AI / ML model input information acquired from the UE and the SN. The packet duplication AI / ML model is updated by the training. These processes can be appropriately performed. They can be performed periodically or each time packet duplication is performed in the MN or the SN.

[0562] The MN determines to perform packet duplication. In steps ST1809 to ST1812, transmission of the inference AI / ML model input information request and transmission of the inference AI / ML model input information are performed among the UE, the MN, and the SN. These processes can be appropriately applied to the processes of steps ST1804 to ST1807 described above. The model training can be replaced with inference. In step ST1813, the MN performs inference by the packet duplication AI / ML model and derives AI / ML model output information. The MN uses the output information to determine the SN and the SCell for packet duplication and performs setting of packet duplication. In step ST1814, packet duplication processing is performed among the MN, the SN, and the UE.

[0563] Thus, the output information derived by the packet duplication AI / ML model can be used to perform packet duplication processing among the MN, the SN, and the UE.

[0564] In steps ST1831 and ST1832, the MN transmits a request for packet duplication processing feedback information using the packet duplication AI / ML model to the UE and the SN. The request can include information indicating the required information. In steps ST1833 and ST1834, the UE and the SN transmit packet duplication processing feedback information using the packet duplication AI / ML model to the MN. The transmission of the information between the UE and the base station can use, for example, RRC signaling. The transmission of the information between the base stations can use, for example, F1 signaling.

[0565] Thus, the MN can acquire the performance of the packet duplication processing using the packet duplication AI / ML model. The MN can evaluate the effect of the packet duplication processing in the case where the packet duplication AI / ML model is used using the packet duplication processing feedback information held by the node and the packet duplication processing feedback information acquired from the UE and the SN.

[0566] Thus, the packet duplication processing can be performed using the packet duplication AI / ML model. More effective packet duplication processing can be performed.

[0567] By employing the method disclosed in the present embodiment, it is possible to perform training and inference in the packet duplication process using the packet duplication AI / ML model. By using the inference result using the packet duplication AI / ML model, it is possible to perform more efficient setting of packet duplication, such as selection of base stations, cells for packet duplication, setting of RLC bearers, and the like. Thus, it is possible to achieve higher reliability in the packet duplication process, and to optimize various KPIs.

[0568] The method disclosed in the present embodiment can be combined with the DC process using the DC AI / ML model disclosed in Embodiment 1, the CA process using the CA AI / ML model disclosed in Embodiment 2. The packet duplication AI / ML model input information, output information can be combined with the DC AI / ML model input information, CA AI / ML model input information, output information. Part or all of each information can be combined. Thus, it is possible to perform more efficient DC or CA processing that takes into account packet duplication.

[0569] The method disclosed in the present embodiment can be combined with the HO process using the HO AI / ML model. The packet duplication AI / ML model input information, output information can be combined with the HO AI / ML model input information, output information. Part or all of each information can be combined. Thus, it is possible to perform more efficient HO processing that takes into account packet duplication.

[0570] Each AI / ML model, each AI / ML model learning device, and inference device using each AI / ML model disclosed in the present specification can be applied to a base station, can be applied to a UE, can be applied to a CN node, such as an AMF, can be applied to an SMF, can be applied to a UPF, can be applied to an LMF, and can be applied to a NWDAF (Network Data Analytics Function). As other examples, it can be applied to a MnS (Management Service), and can be applied to an OAM. As a MnS, it can be applied to a MaS (Management System), for example. They can be applied to the same node, or can be applied to different nodes. For example, the AI / ML model learning device is applied to a node of a MnS, a CN node, or an OAM to perform training that requires time in advance, the inference device using the AI / ML model is applied to a RAN node to perform inference, and output information required for performing a dynamic communication process is obtained. Thus, it is possible to provide a communication system that can perform a communication process with lower latency.

[0571] Each AI / ML model disclosed in this specification can be transmitted to a node provided with each AI / ML model, each AI / ML model learning device, and an inference device using each AI / ML model with each AI / ML model learning device using input information and feedback information. Thus, each AI / ML model can be generated and updated on each node. For example, various communication processes more suitable for the situation of the communication system, such as causing a node with less load to perform AI / ML execution, and the like, can be performed.

[0572] Each AI / ML model input information, output information, and feedback information disclosed in this specification is not limited to the above-described disclosed processing, and can be appropriately applied to other uses. Each AI / ML model input information, output information, and feedback information disclosed in this specification can be transmitted to a base station, can be transmitted to a UE, can be transmitted to a CN node, such as an AMF, can be transmitted to an SMF, can be transmitted to an UPF, can be transmitted to an LMF, and can be transmitted to a NWDAF. As other examples, it can be transmitted to a MnS, and can be transmitted to an OAM. As a MnS, it can be transmitted to a MaS, for example. Further, for example, these information can be transmitted to an AF (Application Function) via a NEF (Network Exposure Function). Various processes can be performed using these information.

[0573] In this specification, a node can be a function.

[0574] In the communication system related to the present disclosure, one gNB constitutes one or more cells. In the present disclosure, it is described as a gNB or a cell, but unless otherwise specified, it can be a gNB or a cell.

[0575] In the present disclosure, a gNB can be a MCG or a SCG.

[0576] Each of the above-described embodiments and modified examples thereof is merely an example, and each of the embodiments and modified examples thereof can be freely combined. In addition, any structural element of each of the embodiments and modified examples thereof can be appropriately changed or omitted.

[0577] For example, in each of the above-described embodiments and modified examples thereof, a subframe is one example of a time unit of communication in a 5th generation communication system. A slot can be a scheduling unit. In each of the above-described embodiments and modified examples thereof, the processing described in the time unit of a slot can be performed in the TTI unit, the subframe unit, the sub-slot unit, or the mini-slot unit.

[0578] For example, the method disclosed in each of the above-described embodiments and modified examples thereof can be applied to IAB. It can be applied to communication between an IAB donor and an IAB node. It can also be applied to processing using Uu in IAB.

[0579] Hereinafter, the modes of the present disclosure are collectively described as an appendix.

[0580] 1. A base station of a communication system supporting dual connectivity in which a communication terminal is connected to two base stations simultaneously, characterized by comprising: when acting as a master node, i.e., a master base station, of the dual connectivity, using a model on which training has been performed, the training using information obtained from a communication terminal connected to the base station and a neighboring other base station, to determine an other base station to act as a secondary node, i.e., a secondary base station, of the dual connectivity.

[0581] (appendix 2) The communication system according to the appendix 1, characterized by comprising: when acting as the master base station, in a case where it is determined that a change of the secondary base station is required, using the model to determine an other base station to act as the changed secondary base station.

[0582] (appendix 3) The communication system according to the appendix 1, characterized by comprising: when acting as the secondary base station, in a case where it is determined that a change of the secondary base station is required, using the model to determine an other base station to act as the changed secondary base station.

[0583] (appendix 4) The communication system according to any one of the appendices 1 to 3, characterized by comprising: when acting as the master base station, in a case where it is determined that a change of the master base station is required, using the model to determine an other base station to act as the changed master base station.

[0584] (appendix 5) The communication system according to the appendix 4, characterized by comprising: the base station supports carrier aggregation in which a plurality of component carriers are used centrally, the information used for the training of the model includes information associated with a decision of the component carriers to be the target of the carrier aggregation.

[0585] (appendix 6) The communication system according to the appendix 4, characterized by comprising: the information used for the training of the model includes information associated with a decision of a beam to be used in communication with the communication terminal.

[0586] (appendix 7) The communication system according to the appendix 4, characterized by comprising: the information used for the training of the model includes information associated with a setting of a packet duplication function to be used in communication with the communication terminal.

[0587] (appendix 8) The communication system according to any one of the appendices 1 to 7, characterized by comprising: In a case where the base station starts operating as the changed main base station in association with the change processing of the main base station, another base station operating as the secondary base station is decided using the model.

[0588] 9. A base station of a communication system composed of a central unit and one or more distributed units and supporting carrier aggregation using a plurality of component carriers, characterized by the central unit deciding component carriers as targets of the carrier aggregation using a model subjected to training using information acquired from a communication terminal connected to the base station and the distributed units.

[0589] (Addendum 10) A communication system supporting dual connectivity in which a communication terminal is connected to two base stations at the same time, characterized by a base station operating as a master node of the dual connectivity deciding another base station operating as a secondary node of the dual connectivity using a model subjected to training using information acquired from a communication terminal connected to the base station and an adjacent other base station.

[0590] (Addendum 11) A communication system including a base station supporting carrier aggregation using a plurality of component carriers, characterized by the base station supporting the carrier aggregation being composed of a central unit and one or more distributed units, the central unit deciding component carriers as targets of the carrier aggregation using a model subjected to training using information acquired from a communication terminal connected to the base station and the distributed units.

[0591] Explanation of Reference Signs 202 Communication terminal device (mobile terminal) 210 Communication system 213, 240-1, 240-2, 750 Base station device (NR base station, base station) 214 5G core 215 Central unit 216 Distributed unit 217 Central unit for control layer 218 Central unit for user layer 219 TRP 301, 403 Protocol processing section 302 Application section 304, 405 Encoding section 305, 406 Modulation section 306, 407 Frequency conversion section 307-1 - 307-4, 408-1 - 408-4 Antenna 308, 409 Demodulation section 309, 410 Decoding section 310, 411, 526 Control section 401 EPC communication section 402 Other base station communication section 412 5GC communication section 521 Data network communication section 522 Base station communication section 523 User layer communication section 523-1 PDU processing section 523-2 Mobility anchor section 525 Control layer control section 525-1 NAS security section 525-2 Idle state mobility management section 527 Session management section 527-1 PDU session control section 527-2 UE IP address allocation section 751-1 - 751-8 Beam 752 Cell 1100 Learning device 1110, 1210 Data acquisition section 1120 Model generation section 1121 Reward calculation section 1122 Function update section 1130 Learning completed model storage section 1200 Inference device 1220 Inference section

Claims

1. A base station, which is a base station of a dual-connection communication system supporting simultaneous connection of a communication terminal to two base stations, characterized in that, When operating as the master node, i.e., the master base station, in the dual-connection scenario. The trained model is used to determine which base stations will act as auxiliary nodes, i.e., auxiliary base stations, in the dual-connection system. The training uses information obtained from communication terminals connected to this base station and other adjacent base stations.

2. The base station as described in claim 1, characterized in that, If, when acting as the primary base station, it is determined that a change to the secondary base station is required, the model is used to determine which other base stations will act as the changed secondary base station.

3. The base station as described in claim 1, characterized in that, If it is determined that a change to the secondary base station is required when the secondary base station is operating as such, the model is used to determine which other base stations will operate as the changed secondary base station.

4. The base station as described in any one of claims 1 to 3, characterized in that, If it is determined that a change of the main base station is required when the main base station is acting as the main base station, the model is used to determine other base stations that will act as the changed main base station.

5. The base station as described in claim 4, characterized in that, The base station supports carrier aggregation that centrally uses multiple component carriers. The information used for training the model includes information associated with the determination of the component carriers that are the objects of the carrier aggregation.

6. The base station as described in claim 4, characterized in that, The information used for training the model includes information associated with the determination of the beam used in communication with the communication terminal.

7. The base station as described in claim 4, characterized in that, The information used for training the model includes information associated with the settings of the packet copy function used in communication between the communication terminal and the communication terminal.

8. The base station as described in any one of claims 1 to 7, characterized in that, When this base station begins to operate as the changed primary base station along with the change processing of the primary base station, the model is used to determine other base stations that will operate as the secondary base station.

9. A base station, comprising a central unit and one or more distributed units, and supporting carrier aggregation of multiple component carriers in a communication system, characterized in that, The central unit uses a trained model to determine the component carriers to be aggregated, the training using information obtained from communication terminals connected to the base station and the distributed unit.

10. A communication system that supports simultaneous connection of a communication terminal to two base stations, characterized in that, The base station acting as the master node of the dual connection uses a trained model to determine which other base stations will act as the auxiliary nodes of the dual connection. The training uses information obtained from communication terminals connected to this base station and other adjacent base stations.

11. A communication system comprising a base station supporting carrier aggregation using multiple component carriers, characterized in that, The base station supporting the carrier aggregation consists of a central unit and one or more distributed units. The central unit uses a trained model to determine the component carriers to be aggregated, the training using information obtained from communication terminals connected to the base station and the distributed unit.

Images