Communication control method and user device

Abstract

A communication control method according to one aspect of the present invention is for a mobile communication system. The communication control method comprises a step in which a user device uses a first inference result from a first Artificial Intelligence (AI) / Machine Learning (ML) model as input data to a second AI / ML model, and infers a second inference result using the second AI / ML model. Here, the first AI / ML model is a Radio Resource Management (RRM) measurement prediction model that infers a measurement result of an entire beam, as the first inference result, from a measurement result of a partial beam. Furthermore, the second AI / ML model is any one of: a Radio Link Failure (RLF) prediction model that infers information indicating whether or not RLF has occurred, as the second inference result; a Handover Failure (HOF) prediction model that infers information indicating whether or not a HOF has occurred, as the second inference result; and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report, as the second inference result.

Description

Communication control method and user device

[0001] The present disclosure relates to a communication control method and a user device.

[0002] 3GPP (Third Generation Partnership Project) (registered trademark; hereinafter the same) is a standardization project for mobile communication systems. It is investigating the application of artificial intelligence (AI) technology, particularly machine learning (ML) technology, to wireless communication (air interface) of mobile communication systems. 3GPP has also begun discussions on applying AI technology to mobility (specifically, handover).

[0003] 3GPP TR 38.843 V18.0.0 (2023-12) RP-234055

[0004] A communication control method according to a first aspect is a communication control method in a mobile communication system. The communication control method includes a step of inferring a second inference result using a second AI / ML model, using a first inference result of a first AI / ML model as input data for the second AI / ML model. Here, the first AI / ML model is a Radio Resource Management (RRM) measurement prediction model that infers a measurement result of an overall beam as a first inference result from measurement results of partial beams. Also, the second AI / ML model is one of a Radio Link Failure (RLF) prediction model that infers information indicating whether or not a handover failure (HOF) has occurred as the second inference result, an HOF prediction model that infers information indicating whether or not a handover failure has occurred as the second inference result, and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report as the second inference result.

[0005] A user equipment according to a second aspect is a user equipment in a mobile communication system. The user equipment includes a controller configured to infer a second inference result using a second AI / ML model, with a first inference result of a first AI / ML model as input data to the second AI / ML model. Here, the first AI / ML model is a Radio Resource Management (RRM) measurement prediction model configured to infer a measurement result of an overall beam as a first inference result from measurement results of a partial beam, and the second AI / ML model is one of a Radio Link Failure (RLF) prediction model configured to infer information indicating whether or not a handover failure (HOF) has occurred as the second inference result, a Handover Failure (HOF) prediction model configured to infer information indicating whether or not a handover failure has occurred as the second inference result, and a measurement event prediction model configured to infer an event indicating a condition for transmitting a measurement report as the second inference result.

[0006] FIG. 1 is a diagram illustrating an example of the configuration of a mobile communication system according to the first embodiment. FIG. 2 is a diagram illustrating an example of the configuration of a UE (user equipment) according to the first embodiment. FIG. 3 is a diagram illustrating an example of the configuration of a network node (base station) according to the first embodiment. FIG. 4 is a diagram illustrating an example of the configuration of a protocol stack according to the first embodiment. FIG. 5 is a diagram illustrating an example of the configuration of a protocol stack according to the first embodiment. FIG. 6 is a diagram illustrating an example of the configuration of functional blocks of AI / ML technology according to the first embodiment. FIGS. 7(A) and 7(B) are diagrams illustrating an example of RRM measurement prediction in the time domain according to the first embodiment. FIGS. 8(A) and 8(B) are diagrams illustrating an example of RRM measurement prediction in the space domain according to the first embodiment. FIGS. 9(A) and 9(B) are diagrams illustrating an example of RRM measurement prediction in the frequency domain according to the first embodiment. FIGS. 10(A) and 10(B) are diagrams illustrating an example of RLF prediction according to the first embodiment. FIGS. 11(A) to 11(C) are diagrams illustrating an example of measurement event prediction according to the first embodiment. FIG. 12 is a diagram illustrating an example of operation according to the first embodiment.

[0007] The present disclosure aims to appropriately connect multiple AI / ML models.

[0008] [First embodiment] A mobile communication system according to a first embodiment will be described with reference to the drawings. In the description of the drawings, the same or similar parts are denoted by the same or similar reference numerals.

[0009] (Configuration of mobile communication system) The configuration of a mobile communication system according to the first embodiment will be described. FIG. 1 is a diagram showing an example of the configuration of a mobile communication system 1 according to the first embodiment. The mobile communication system 1 conforms to the 5th Generation System (5GS) of the 3GPP standard. Although 5GS will be described below as an example, the mobile communication system may also be at least partially applied to an LTE (Long Term Evolution) system. The mobile communication system may also be at least partially applied to a sixth generation (6G) system or later system.

[0010] The mobile communication system 1 includes a network (NW) 10 and a user equipment (UE) 100. The UE 100 is a mobile communication device that performs wireless communication with the NW 10. The UE 100 may be any device used by a user, and may be, for example, a mobile phone terminal (including a smartphone) and / or a tablet terminal, a notebook PC (Personal Computer), a communication module (including a communication card or chipset), a sensor or a device provided in a sensor, a vehicle or a device provided in a vehicle (Vehicle UE), or an aircraft or a device provided in an aircraft (Aerial UE).

[0011] The NW 10 includes a radio access network (RAN) 20 and a core network (CN) 30. When the mobile communication system is a 5th generation system (5GS), the RAN 20 is referred to as a Next Generation Radio Access Network (NG-RAN), and the CN 30 is referred to as a 5G Core Network (5GC).

[0012] The RAN 20 includes a plurality of network nodes 200 (network nodes 200a to 200c in the example of FIG. 1). The network nodes 200 are connected to each other via inter-network node interfaces. The network nodes 200 may be referred to as base stations in the RAN 20. When the network node 200 is a base station, the network node 200 may be configured (i.e., functionally divided) with a CU (Central Unit) and a DU (Distribution Unit), and the two units may be connected by a fronthaul interface. When the mobile communication system 1 is 5GS, the network node 200 is referred to as a gNB, the inter-network node interface is referred to as an Xn interface, and the fronthaul interface is referred to as an F1 interface.

[0013] When at least a part of the mobile communication system 1 is an LTE system, the network node 200 may be an evolved Node B (eNB) that is an LTE base station. When the mobile communication system 1 is a sixth-generation system or later, the network node 200 has a function of a base station and may be a device equivalent to a gNB or an eNB.

[0014] Each network node 200 manages one or more cells. The network node 200 performs wireless communication with the UE 100 that has established a connection with the network node 200's cell. Each network node 200 has a radio resource management (RRM) function, a user data (also simply referred to as "data") routing function, a measurement control function for mobility control and scheduling, and the like. The term "cell" is used as a term indicating the smallest unit of a wireless communication area. The term "cell" is also used as a term indicating a function or resource for performing wireless communication with the UE 100. One cell belongs to one carrier frequency. One downlink component carrier and one uplink component carrier may be associated with one cell. The bandwidth (system bandwidth) corresponding to one cell may be divided into multiple band parts (BWP: Bandwidth Parts). In the following, a gNB may be used as an example of the network node 200.

[0015] The CN 30 includes a CN (Core Network) device 300. The CN device 300 may include a C-plane device corresponding to the control plane (C-plane) and a U-plane device corresponding to the user plane (U-plane). The C-plane device performs various mobility controls and paging for the UE 100. The C-plane device communicates with the UE 100 using NAS (Non-Access Stratum) signaling. The U-plane device controls data forwarding. When the mobile communication system is 5GS, the C-plane device is called an AMF (Access and Mobility Management Function), the U-plane device is called a UPF (User Plane Function), and the interface between the network node 200 and the CN device 300 is called an NG interface.

[0016] In the following description, the network node 200 and the CN device 300 may be referred to as a network device. The network device may be the network node 200. The network device may be the CN device 300.

[0017] 2 is a diagram showing an example of the configuration of a UE 100 (user equipment) according to the first embodiment. The UE 100 includes a receiving unit 110, a transmitting unit 120, and a control unit 130. The receiving unit 110 and the transmitting unit 120 configure a communication unit 140 that performs wireless communication with a network node 200. The UE 100 is an example of a communication device.

[0018] The receiving unit 110 performs various reception operations under the control of the control unit 130. The receiving unit 110 includes an antenna and a receiver. The receiver converts a radio signal received by the antenna into a baseband signal (received signal) and outputs the baseband signal to the control unit 130.

[0019] The transmitting unit 120 performs various transmissions under the control of the control unit 130. The transmitting unit 120 includes an antenna and a transmitter. The transmitter converts a baseband signal (transmission signal) output by the control unit 130 into a radio signal and transmits it from the antenna.

[0020] The control unit 130 performs various controls and processes in the UE 100. Such processes include processes of each layer described below. The control unit 130 includes at least one processor and at least one memory. The memory stores programs executed by the processor and information used in the processes by the processor. The processor may include a baseband processor and a CPU (Central Processing Unit). The baseband processor performs modulation / demodulation, encoding / decoding, etc. of baseband signals. The CPU executes programs stored in the memory to perform various processes. Note that the processes or operations performed in the UE 100 may be performed in the control unit 130.

[0021] 3 is a diagram showing an example of the configuration of the network node 200 according to the first embodiment. The network node 200 includes a transmitting unit 210, a receiving unit 220, a control unit 230, and a NW communication unit 240. The transmitting unit 210 and the receiving unit 220 constitute a communication unit 250 that performs wireless communication with the UE 100. The NW communication unit 240 constitutes a backhaul communication unit that communicates with the CN 30. The network node 200 is an example of a communication device.

[0022] The transmitting unit 210 performs various transmissions under the control of the control unit 230. The transmitting unit 210 includes an antenna and a transmitter. The transmitter converts a baseband signal (transmission signal) output by the control unit 230 into a radio signal and transmits it from the antenna.

[0023] The receiving unit 220 performs various types of reception under the control of the control unit 230. The receiving unit 220 includes an antenna and a receiver. The receiver converts a radio signal received by the antenna into a baseband signal (received signal) and outputs the baseband signal to the control unit 230.

[0024] The control unit 230 performs various controls and processes in the network node 200. Such processes include processes of each layer described below. The control unit 230 includes at least one processor and at least one memory. The memory stores programs executed by the processor and information used in the processes by the processor. The processor may include a baseband processor and a CPU. The baseband processor performs modulation / demodulation, encoding / decoding, etc. of baseband signals. The CPU executes programs stored in the memory to perform various processes. Note that the processes or operations performed in the network node 200 may be performed by the control unit 230.

[0025] The NW communication unit 240 is connected to adjacent network nodes via an Xn interface, which is an interface between network nodes. The NW communication unit 240 is connected to the CN device 300 via an NG interface, which is an interface between a network node and a core network. The network node 200 may be configured (i.e., functionally divided) with a central unit (CU) and a distributed unit (DU), and both units may be connected via an F1 interface, which is a fronthaul interface.

[0026] FIG. 4 is a diagram showing an example of the configuration of a protocol stack of a radio interface of a user plane that handles data.

[0027] The user plane air interface protocol includes a physical (PHY) layer, a medium access control (MAC) layer, a radio link control (RLC) layer, a packet data convergence protocol (PDCP) layer, and a service data adaptation protocol (SDAP) layer.

[0028] The PHY layer performs encoding / decoding, modulation / demodulation, antenna mapping / demapping, and resource mapping / demapping. Data and control information are transmitted between the PHY layer of the UE 100 and the PHY layer of the network node 200 via a physical channel. The PHY layer of the UE 100 receives downlink control information (DCI) transmitted on a physical downlink control channel (PDCCH) from the network node 200. Specifically, the UE 100 performs blind decoding of the PDCCH using a radio network temporary identifier (RNTI) and acquires successfully decoded DCI as DCI addressed to the UE. The DCI transmitted from the network node 200 has a CRC (Cyclic Redundancy Code) parity bit scrambled by the RNTI added.

[0029] In NR, the UE 100 can use a bandwidth narrower than the system bandwidth (i.e., the cell bandwidth). The network node 200 configures the UE 100 with a bandwidth portion (BWP) consisting of contiguous PRBs (Physical Resource Blocks). The UE 100 transmits and receives data and control signals in the active BWP. For example, up to four BWPs may be configurable for the UE 100. Each BWP may have a different subcarrier spacing. The BWPs may overlap in frequency. When multiple BWPs are configured for the UE 100, the network node 200 can specify which BWP to apply by controlling the downlink. This allows the network node 200 to dynamically adjust the UE bandwidth according to the amount of data traffic of the UE 100, etc., thereby reducing UE power consumption.

[0030] The network node 200 can configure, for example, up to three control resource sets (CORESETs) for each of up to four BWPs on the serving cell. A CORESET is a radio resource for control information to be received by the UE 100. Up to 12 or more CORESETs may be configured for the UE 100 on the serving cell. Each CORESET may have an index of 0 to 11 or more. A CORESET may consist of six resource blocks (PRBs) and one, two, or three consecutive Orthogonal Frequency Division Multiplex (OFDM) symbols in the time domain.

[0031] The MAC layer performs data priority control, retransmission processing using Hybrid Automatic Repeat reQuest (HARQ), random access procedures, etc. Data and control information are transmitted between the MAC layer of the UE 100 and the MAC layer of the network node 200 via a transport channel. The MAC layer of the network node 200 includes a scheduler. The scheduler determines the uplink and downlink transport format (transport block size, modulation and coding scheme (MCS)) and the resource blocks to be allocated to the UE 100.

[0032] The RLC layer transmits data to the RLC layer on the receiving side using the functions of the MAC layer and the PHY layer. Data and control information are transmitted between the RLC layer of the UE 100 and the RLC layer of the network node 200 via logical channels.

[0033] The PDCP layer performs header compression / decompression, encryption / decryption, and the like.

[0034] The SDAP layer maps IP flows, which are units for Quality of Service (QoS) control by the core network, to radio bearers, which are units for QoS control by the access stratum (AS). Note that if the RAN is connected to the EPC, SDAP may not be required.

[0035] FIG. 5 is a diagram showing the configuration of a protocol stack of a radio interface of a control plane that handles signaling (control signals).

[0036] The protocol stack of the radio interface of the control plane includes a Radio Resource Control (RRC) layer and a Non-Access Stratum (NAS) instead of the SDAP layer shown in FIG.

[0037] RRC signaling for various settings is transmitted between the RRC layer of the UE 100 and the RRC layer of the network node 200. The RRC layer controls logical channels, transport channels, and physical channels in accordance with the establishment, re-establishment, and release of radio bearers. When there is a connection (RRC connection) between the RRC of the UE 100 and the RRC of the network node 200, the UE 100 is in an RRC connected state. When there is no connection (RRC connection) between the RRC of the UE 100 and the RRC of the network node 200, the UE 100 is in an RRC idle state. When the connection between the RRC of the UE 100 and the RRC of the network node 200 is suspended, the UE 100 is in an RRC inactive state.

[0038] The NAS, which is located above the RRC layer, performs session management, mobility management, etc. NAS signaling is transmitted between the NAS of the UE 100 and the NAS of the AMF. Note that the UE 100 has an application layer in addition to the radio interface protocol. Also, the layer below the NAS is called an Access Stratum (AS).

[0039] (AI / ML Technology) Next, AI / ML (Artificial Intelligence / Machine Learning) technology according to an embodiment will be described. First, three points (functional blocks, use cases, and LCM) discussed in 3GPP as the AI / ML air interface (AI / ML for NR air interface) will be described below.

[0040] (Functional Blocks) FIG. 6 is a diagram showing an example of the configuration of functional blocks of the AI / ML technology in the mobile communication system 1 according to the first embodiment.

[0041] As shown in FIG. 6, the functional blocks include a data collection unit (Data Collection) A1, a model training unit (Model Training) A2, a model inference unit (Inference) A3, a management unit (Management) A5, and a model storage unit (Model Storage) A6.

[0042] The functional block configuration example shown in FIG. 6 represents a functional framework of a general AI / ML technology. Therefore, depending on the hypothetical use case, some of the functional block configuration example (e.g., model recording unit A6, etc.) may not be included in the functional block configuration example. The functional block configuration example shown in FIG. 6 may also be distributed between the UE 100 and the network device. Alternatively, some functions of the functional block configuration example (e.g., model learning unit A2 or model inference unit A3, etc.) may be located in both the UE 100 and the network device. Note that the network device refers to a node or entity included in the NW 10. The network device may be the CN device 300. The network device may also be the network node 200.

[0043] The data collection unit A1 provides input data to the model learning unit A2, the model inference unit A3, and the management unit A5. The input data includes training data for the model learning unit A2, inference data for the model inference unit A3, and monitoring data for the management unit A5.

[0044] The training data is data required as input when the AI / ML model is learning. The inference data is data required as input when the AI / ML model is inferring. The monitoring data is data required as input when the AI / ML model is managing.

[0045] In addition, data collection may refer to the process of collecting data at a network node, a management entity, or a UE 100, for example, to train an AI / ML model, manage an AI / ML model, and perform inference on an AI / ML model.

[0046] The model learning unit A2 performs AI / ML model training, AI / ML model validation, and AI / ML model testing.

[0047] AI / ML model learning is the process of learning an AI / ML model from input / output relationships to obtain a trained AI / ML model to be used for inference. For example, considering y = ax + b, AI / ML model learning may be the process of optimizing a (slope) and b (intercept) by providing input (x) and output (y) (i.e., providing learning data).

[0048] Generally, machine learning includes supervised learning, unsupervised learning, and reinforcement learning. Supervised learning is a method that uses correct answer data as training data. Unsupervised learning is a method that does not use correct answer data as training data. For example, unsupervised learning memorizes feature points from a large amount of training data and determines the correct answer (estimates the range). Reinforcement learning is a method that assigns a score to the output result and learns how to maximize the score. Although supervised learning will be described below, either unsupervised learning or reinforcement learning may be applied as machine learning.

[0049] The model learning unit A2 outputs a trained AI / ML model (Trained Model) obtained by AI / ML model learning to the model recording unit A6, and also outputs an updated AI / ML model (Updated Model) obtained by relearning the trained AI / ML model to the model recording unit A6.

[0050] In the following, AI / ML model learning may be referred to as "model learning" or "learning."

[0051] The model inference unit A3 performs AI / ML model inference. Specifically, the model inference unit A3 applies the inference data provided by the data collection unit A1 to the trained AI / ML model (or updated AI / ML model) to obtain inference output data. For example, in the equation y = ax + b, x corresponds to the inference data and y corresponds to the inference output data. Note that "y = ax + b" is an AI / ML model. A model with optimized slope and intercept, for example, "y = 5x + 3," is a trained AI / ML model. There are various model approaches, including linear regression analysis, neural networks, and decision tree analysis. The above "y = ax + b" can also be considered a type of linear regression analysis.

[0052] The model inference unit A3 outputs inference output data to the management unit A5. The model inference unit A3 also receives management instructions from the management unit A5. For example, the management instructions include selection of an AI / ML model, activation (deactivation) of an AI / ML model, switching of an AI / ML model, and fallback (performing inference without using an AI / ML model). The model inference unit A3 performs model inference in accordance with the management instructions.

[0053] Note that AI / ML model inference is, for example, a process of obtaining a set of outputs from a set of inputs using a trained AI / ML model (or an updated AI / ML model). Alternatively, model inference may be a process of obtaining inference output data from inference data using a trained AI / ML model (or an updated AI / ML model). Hereinafter, AI / ML model inference may be referred to as "model inference" or "inference."

[0054] In the following, an AI / ML model that is currently being trained (or updated) may be referred to as a training AI / ML model (or an updating AI / ML model). In the following, when there is no need to distinguish between a training (or updating) AI / ML model and a trained (or updated) AI / ML model, they may be simply referred to as an "AI / ML model."

[0055] The management unit A5 supervises operations on the AI / ML model (selection, activation, deactivation, switching, fallback, etc.). The management unit A5 also supervises monitoring of the AI / ML model. The management unit A5 can also perform operations to ensure appropriate inference operations based on monitoring data and inference output data. To this end, the management unit A5 outputs a model transfer and / or model delivery request (Model Transfer / Delivery Request) to the model recording unit A6, and causes the trained (or updated) AI / ML model recorded in the model recording unit A6 to be output to the model inference unit A3. The management unit A5 also outputs management instructions to the model inference unit A3 and supervises operations on the AI / ML model. Furthermore, the management unit A5 can output performance feedback and a re-learning request to the model learning unit A2, causing the model learning unit A2 to re-learn the AI / ML model (i.e., update the learned AI / ML model).

[0056] (Use Cases) Next, three use cases discussed in the AI / ML air interface will be described. There are three use cases as follows.

[0057] (A1) "CSI (Channel State Information) Feedback Enhancement"

[0058] (A2) "Beam management"

[0059] (A3) “Positioning accuracy enhancement”

[0060] (A1) CSI Feedback Enhancement CSI feedback enhancement represents a use case in which, for example, AI / ML techniques are applied to CSI fed back from the UE 100 to the network node 200. The CSI is information about a channel state in a downlink between the UE 100 and the network node 200. The CSI includes at least one of a Channel Quality Indicator (CQI), a Precoding Matrix Indicator (PMI), and a Rank Indicator (RI). The network node 200 performs, for example, downlink scheduling based on the CSI feedback from the UE 100.

[0061] The use case of improved CSI feedback has two sub-use cases: CSI compression in the frequency domain and CSI prediction in the time domain.

[0062] In the CSI compression, the CSI inferred in the UE 100 using the trained AI / ML model is compressed in the UE 100. Specifically, the compression is performed by quantizing the inferred CSI. The compressed (or quantized) CSI is transmitted from the UE 100 to the network node 200.

[0063] In CSI prediction, a trained AI / ML model is used to infer (predict) future CSI from the history of past CSI. The UE 100 has the trained AI / ML model, and uses the past history of CSI as input to infer future CSI (predicted CSI). The future CSI is transmitted to the network node 200 as CSI feedback.

[0064] (A2) Beam Management The beam management use case also has two sub-use cases: spatial-domain downlink beam prediction, which performs beam prediction in the spatial direction, and temporal downlink beam prediction, which performs beam prediction in the temporal direction. The spatial-domain downlink beam prediction is called "BM Case 1" (BM-Case 1), and the temporal downlink beam prediction is called "BM Case 2" (BM-Case 2).

[0065] In BM case 1, the input to the AI / ML model 103 is a measurement value for each beam included in beam set B. On the other hand, the output from the AI / ML model 103 is the probability that each (downlink) beam included in (predicted) beam set A will be the top beam. Beam set A and beam set B may be different. Alternatively, beam set B may be a subset of beam set A. In BM case 1, the inference may be performed on the UE 100 side or on the network equipment side.

[0066] In the case of BM case 2, the input to the AI / ML model 103 is the history of measurement values ​​for each beam included in beam set B. Measurement values ​​measured in the past for each beam are input to the AI / ML model 103. On the other hand, the output (inference output data) from the AI / ML model 103 is the probability that each (downlink) beam included in the (predicted) beam set A will be the top beam. Beam set A and beam set B may be different, or beam set B may be a subset of beam set A. Also, in the case of BM case 2, beam set A and beam set B may be the same. In the case of BM case 2, inference may be performed on the UE 100 side or on the network device side.

[0067] (A3) Location Accuracy Improvement The location accuracy improvement use case also has two sub-use cases: Direct AI / ML positioning, which uses a trained AI / ML model to directly infer the location of UE 100, and AI / ML assisted positioning, which infers intermediate position measurements.

[0068] In direct AI / ML positioning, a trained AI / ML model is used to directly infer the location of UE 100 from measurements at each measurement point (TRP: Transmission and / or Reception Point).

[0069] On the other hand, in AI / ML-assisted positioning, an intermediate location measurement value is inferred from measurements at each measurement point (TRP: Transmission and / or Reception Point) using a trained AI / ML model. The intermediate location measurement value is transmitted to a Location Management Function (LMF) and used by the LMF to measure (or infer) the location of the UE 100. The LMF may have a trained AI / ML model, and the location of the UE 100 is inferred using the trained AI / ML model.

[0070] (LCM) In the AI / ML wireless interface, LCM (Life Cycle Management) of the AI / ML model is being discussed.

[0071] Specifically, the LCM of the AI / ML model may include at least one of the following operations.

[0072] (L1) Data collection

[0073] (L2) Model training

[0074] (L3) Identification

[0075] (L4) Model delivery / transfer

[0076] (L5) Model inference operation

[0077] (L6) Selection, Activation, Deactivation, Switching, and Fallback Operations

[0078] (L7) Monitoring

[0079] (L8) Model updating

[0080] (L9) UE capability

[0081] On the network side, for example, by controlling each operation of the LCM, it becomes possible to appropriately manage the process from the generation of an AI / ML model to the deletion (or disposal) of the AI / ML model. Note that fallback means switching from an AI / ML model to a model that does not use an AI / ML model. A model that does not use an AI / ML model is sometimes called a "legacy model."

[0082] Regarding LCM, 3GPP defines functionality-based LCM and model ID-based LCM.

[0083] The function-based LCM may be an operation performed on a function of the AI / ML model. Specifically, the function-based LCM may be any of activation, deactivation, switching, and fallback operations performed on a function of the AI / ML model. The network side can instruct the LCM operation on the function of the AI / ML model using, for example, 3GPP signaling (RRC message, MAC CE, or DCI). In the function-based LCM, one AI / ML model may correspond to one function, or multiple AI / ML models may correspond to one function.

[0084] On the other hand, model ID-based LCM may be an operation performed on an individual AI / ML model using a model ID. The model ID is identification information for distinguishing an AI / ML model from other AI / ML models. Specifically, model ID-based LCM may be an operation of activating, deactivating, switching, or selecting an AI / ML model using the model ID.

[0085] (AI / ML Mobility) Next, AI / ML mobility (AI / ML for mobility) will be described.

[0086] Here, the existing Layer 3 (L3) handover will be considered as follows. That is, in the existing Layer 3 (L3) handover, it is triggered and executed based on past measurement results and event conditions reported from the UE 100. In that handover is executed using past results, the existing L3 handover can be said to be a kind of reactive scheme. When the mobility of the UE 100 is below a certain level, such an L3 handover may function well. However, when the mobility of the UE 100 is higher than a certain level or when cells are deployed more densely than a certain level, various problems may occur. Specifically, unintended problems such as radio link failure (RLF), premature handover, premature handover, handover failure (HOF), ping-pong handover, etc. may occur.

[0087] To address such problems, it is possible to avoid the above-mentioned problems by predicting future beam measurement results or future cell measurement results using, for example, an AI / ML model. That is, by establishing a proactive scheme for handover using the AI / ML model, it is possible to predict the occurrence of unintended problems in advance and take preventative measures, thereby improving handover performance. The shift from a reactive scheme to a proactive scheme is one of the motivations for research on AI / ML mobility in 3GPP.

[0088] Next, three prediction models for AI / ML mobility that are being discussed in 3GPP will be described: RRM (Radio Resource Management) measurement prediction, measurement event prediction, and RLF / HOF prediction.

[0089] (B1) RRM Measurement Prediction First, RRM measurement prediction will be described. In RRM measurement prediction, an AI / ML model is used to finally output L3 cell-level measurements. Specifically, the L3 cell-level measurements correspond to the output after L1 filtering is performed on layer 1 (L1) beam-level measurements, cell quality is derived by solidation, and L3 filtering is performed on the cell quality. The L3 cell-level measurements are included in a measurement report and transmitted from the UE 100 to the network node 200.

[0090] RRM measurement prediction has three subcases: measurement prediction in the time domain (intra-cell temporal domain prediction), measurement prediction in the spatial domain (intra-cell spatial domain prediction), and measurement prediction in the frequency domain (inter-frequency inter-cell prediction).

[0091] (B1-1) RRM Measurement Prediction in the Time Domain In the RRM measurement prediction in the time domain, basically, an AI / ML model is used to predict unmeasured measurement results from measured measurement results in the time direction.

[0092] 7(A) and 7(B) are diagrams illustrating examples of RRM measurement prediction in the time domain according to the first embodiment. In the RRM measurement prediction in the time domain, as shown in FIG. 7(A), there is Case A in which future measurement results in a prediction window are predicted based on measurement results actually measured by the UE 100 in an observation window. The observation window represents a period during which the UE 100 actually performs measurements. For example, the UE 100 reports future measurement results to the network node 200, which enables the network node 200 to trigger handover at an appropriate timing. Furthermore, in the RRM measurement prediction in the time domain, there is Case B in which a subset of measurement results is predicted based on actually measured measurement results, as shown in FIG. 7(B).

[0093] (B1-2) RRM Measurement Prediction in the Spatial Domain In the RRM measurement prediction in the spatial domain, basically, an AI / ML model is used to output cell level measurement results for L3 from partial beam level measurement results for L1.

[0094] 8A and 8B are diagrams illustrating an example of RRM measurement prediction in the spatial domain according to the first embodiment. In the spatial domain RRM measurement prediction, as shown in FIG. 8A, an AI / ML model is used to predict overall L1 beam level measurement results from partial L1 beam level measurement results, and legacy consolidation and L3 filtering are applied to the predicted results to output L3 cell level measurement results. In the case shown in FIG. 8A, the AI / ML model is used to predict L1 level beam level measurement results. Also, as shown in FIG. 8B, an AI / ML model is used to directly infer L3 cell level measurement results from partial L1 beam level measurement results.

[0095] (B1-3) RRM Measurement Prediction in the Frequency Domain Figures 9(A) and 9(B) are diagrams showing examples of RRM measurement prediction in the frequency domain according to the first embodiment. As shown in Figure 9(A), in the RRM measurement prediction in the frequency domain, an AI / ML model is used to predict the measurement results in a cell B, which is different from cell A, from the measurement results in cell A. However, as shown in Figure 9(B), it is assumed that cell A and cell B exist in the same sector.

[0096] An example of RRM measurement prediction has been described above, and in 3GPP, RRM measurement prediction is currently given the highest priority compared to other prediction models.

[0097] (B2) Measurement Event Prediction In 3GPP, it has been agreed that there is at least a method using the output of the RRM measurement prediction described above for measurement event prediction. However, it has also been agreed in 3GPP that discussion on measurement event prediction will be suspended until progress in the discussion on RRM measurement prediction is made.

[0098] (B3) RLF / HOF Prediction FIGS. 10(A) and 10(B) are diagrams showing examples of RLF prediction according to the first embodiment.

[0099] In 3GPP, it has been agreed that there are two methods for predicting RLF / HOF: an indirect method (FIG. 10(A)) that uses the output of RRM measurement prediction, and a direct method (FIG. 10(B)) that does not use RRM measurement prediction. It has also been agreed that in RLF / HOF prediction, RLF prediction takes priority over HOF prediction.

[0100] (Communication Control Method According to First Embodiment) In the first embodiment, of the three prediction models of AI mobility being discussed in 3GPP, the following description will mainly focus on RLF prediction.

[0101] As described above, regarding RLF prediction, 3GPP has agreed on two methods: an indirect method that predicts RLF based on the inference results of time-domain RRM measurement prediction (AI / ML model), and a direct method that directly predicts whether or not RLF will occur using the AI / ML model.

[0102] Fig. 11(A) is a diagram showing an example of RLF prediction using the indirect method, and Fig. 11(B) is a diagram showing an example of RLF prediction using the direct method, Fig. 11(A) and Fig. 11(B) are diagrams showing more details of the examples of RLF prediction shown in Fig. 10(A) and Fig. 10(B).

[0103] As shown in FIG. 11A , the indirect method uses an RRM measurement prediction (AI / ML model) to infer the measurement results of the overall beam from the measurement results of partial beams, for example. The RLF monitoring unit predicts RLF using legacy criteria without using the AI / ML model. The legacy criteria are implemented, for example, as follows: That is, the RLF monitoring unit focuses on the measurement results of each beam in the overall beam measurement results, and detects out-of-sync when the measurement results fall below a certain value. When out-of-sync is detected N310 times consecutively, the RLF monitoring unit starts counting timer N310. Then, when the measurement results exceed a certain value and in-sync is not detected N311 times consecutively within the timer N310 period, the RLF monitoring unit detects RLF. If the RLF monitoring unit detects In-Sync N311 times in succession within the timer N310 period, it stops the timer N310 and does not detect RLF. The RLF monitoring unit outputs whether or not an RLF has occurred using the legacy criteria.

[0104] Also, as shown in FIG. 11B, in the direct method, the presence or absence of RLF is directly inferred from input data using RLF prediction, which is an AI / ML model.

[0105] On the other hand, a method of linking multiple AI / ML models to obtain a final inference result is also possible. Hereinafter, the method of linking multiple AI / ML models to obtain an inference result is referred to as a hybrid method.

[0106] 11(C) is a diagram showing an example of RLF prediction using a hybrid method. In the example shown in FIG. 11(C), an inference result (e.g., a first inference result) of the RRM measurement prediction, which is an AI / ML model (e.g., a first AI / ML model), is used as input data for the RLF prediction, which is an AI / ML model (e.g., a second AI / ML model). In the RLF prediction, the presence or absence of RLF occurrence (e.g., a second inference result) is predicted from the inference result of the RRM measurement prediction.

[0107] In this way, the hybrid method makes it possible to predict the presence or absence of RLF using, for example, RRM prediction that is already used as an AI / ML model, and thus enables the AI / ML model to be flexibly configured. However, it is currently not clear what settings are required to configure such a hybrid method.

[0108] Therefore, the first embodiment aims to enable a plurality of AI / ML models to be appropriately linked.

[0109] Therefore, in the first embodiment, the RRM measurement prediction model and the RLF prediction model are configured so that the inference result of the RRM measurement prediction model is input to the RLF prediction model.

[0110] Specifically, a user equipment (e.g., UE 100) uses a first inference result of a first AI / ML model (e.g., an RRM measurement prediction model) as input data to a second AI / ML model (e.g., an RLF prediction model) and infers a second inference result (e.g., an RLF prediction) using the second AI / ML model. Here, the first AI / ML model is an RRM measurement prediction model that infers a measurement result of an entire beam as a first inference result from measurement results of partial beams. Also, the second AI / ML model is an RLF prediction model that infers information indicating the presence or absence of a radio link failure (RLF) as a second inference result.

[0111] In this way, the inference result of the RRM measurement prediction model is configured to be input to the RLF prediction model, so that the RRM measurement prediction model and the RLF prediction model can be appropriately linked.

[0112] 11(C), the latter stage AI / ML model may be, instead of the RLF prediction model, a HOF prediction model that infers information indicating the presence or absence of a handover failure (HOF) (e.g., a second inference result). Furthermore, the latter stage AI / ML model may be, instead of the RLF prediction model, a measurement event prediction model that infers an event (e.g., event A1) that indicates a measurement report transmission condition. Both the HOF prediction model and the measurement event prediction model are inferred using the inference result of the RRM measurement prediction model.

[0113] 11C, in the hybrid method, inference of the AI / ML model may be performed using not only the inference result but also actual measurement data. In FIG. 11C, an example is shown in which actual measurement data other than the inference result is used as input data in the RLF prediction model. Hereinafter, the AI / ML model configured by the hybrid method may be referred to as a hybrid model.

[0114] (Example of Operation According to First Embodiment) Next, an example of operation according to the first embodiment will be described.

[0115] Fig. 12 is a diagram illustrating an example of operation according to the first embodiment. Before the example of operation illustrated in Fig. 12 is started, it is assumed that the AI / ML model is stored in advance in the UE 100. Furthermore, in Fig. 12, it is assumed that the UE 100 is in an RRC connected state with the network node 200.

[0116] As shown in FIG. 12 , in step S10, the transmitter 120 of the UE 100 may transmit a message including information indicating that the UE 100 is capable of executing the hybrid model to the network node 200. The message may be an RRC message such as a UE Capability Information (UECapabilityInformation) message. The message may include information indicating that the UE 100 is capable of executing two or more AI / ML models for one function. Alternatively, the message may include information indicating the number of AI / ML models that can be executed for one function. Alternatively, the message may include information indicating that the UE 100 is capable of connecting two or more AI / ML models for one function.

[0117] In step S11, the receiver 110 of the UE 100 receives hybrid model setting information transmitted from the transmitter 210 of the network node 200. The hybrid model setting information indicates, for example, setting information for linking multiple AI / ML models. The hybrid model setting information includes configuration information indicating that multiple AI / ML models are linked. The UE 100 can link multiple AI / ML models by using the hybrid model setting information.

[0118] First, in the case of functionality ID-based configuration information, the configuration information may be specified by a functionality ID that defines the configuration of the linked model.

[0119] Specifically, the configuration information may be specified by a function ID indicating the function of the AI / ML model in the first stage and a function ID indicating the function of the AI / ML model in the second stage. For example, the first stage may be specified by function ID = 1 (= RRM measurement prediction model), and the second stage may be specified by function ID = 2 (= RLF prediction model). In the case of three or more stages, the configuration information may be indicated by specifying the function IDs of each stage in order (in list form). Note that which model corresponds to which function ID may be defined in the specifications.

[0120] The hybrid model setting information may include information specifying a target function (e.g., RLF prediction, HOF prediction, or event prediction) as well as information indicating whether inference data of another model may be used as input data for the target function. For example, in the example of Fig. 11(C), the hybrid model setting information may include information indicating that the target function is RLF prediction as well as information indicating that inference data of another model may be used as input data.

[0121] Second, in the case of model ID-based configuration, the configuration information may be specified by the number of stages of AI / ML models to be linked and a model ID specifying the AI / ML model of each stage. Specifically, the configuration information may include a list of model IDs. Each entry in the list may represent an AI / ML model of each stage. For example, the first entry in the list represents the AI / ML model of the first stage, the second entry represents the AI / ML model of the second stage, ..., and the nth entry represents the model of the nth stage. In the example of FIG. 11(C), the configuration information includes a list of model ID=1 (=RRM measurement prediction model) and model ID=2 (RLF prediction model) in this order. The hybrid configuration information may include information indicating to which function the linked models are ultimately applied. For example, in the case of FIG. 11(C), since the final function is RLF prediction, the hybrid model configuration information includes information (which may be indicated by a model ID) indicating that RLF prediction represents the final function.

[0122] Third, whether function ID-based or model ID-based, when inference by the AI / ML model is performed using input data other than inference result data, information specifying the type of input data other than the inference result data may be included in the hybrid model configuration information. For example, when location information of UE 100 is used as input data, information indicating that the input data is location information of UE 100 may be included in the hybrid model configuration information. Alternatively, the hybrid model configuration information may include information indicating whether inference result data of another model or actual measurement data is used as input data to the AI / ML model. Alternatively, the hybrid model configuration information may include information specifying another AI / ML model that generates the input data. In the example of FIG. 11(C), information specifying an RRM measurement prediction model may be included in the hybrid model configuration information as information specifying another AI / ML model.

[0123] The linked model is executed when inference is performed in the AI / ML model. When learning is performed in the AI / ML model, learning is performed individually using actual measurement data without being linked. That is, for each linked AI / ML model, actual measurement data is used when learning is performed, and the inference results of other AI / ML models are used when inference is performed. In the example of FIG. 11(C), the RLF prediction model is learned using actual measurement data (e.g., beam measurement values ​​for each beam, etc.) when learning, and the inference results of the RRM measurement prediction model are used when inference is performed. In the linked AI / ML models, different data sets (actual measurement data or inference result data) are used when learning and when inference is performed.

[0124] The hybrid model setting information may be transmitted in a state that it is included in an RRC message.The hybrid model setting information may be transmitted in a state that it is included in a message of a layer newly defined for AI / ML (e.g., an AI / ML layer).

[0125] In step S12, the control unit 130 of the UE 100 generates a hybrid model by a hybrid method by linking the AI / ML models based on the hybrid model setting information. For example, a hybrid model in a linked state shown in Fig. 11(C) is generated. As shown in Fig. 11(C), a hybrid model in which multiple AI / ML models are linked can ultimately obtain a desired inference result (RLF prediction).

[0126] In step S13, the receiving unit 110 of the UE 100 may receive a notification of activation or deactivation for a hybrid model. In the notification, a function ID may be specified to instruct activation or deactivation. Alternatively, in the notification, a model ID may be specified to instruct activation or deactivation. Alternatively, in the notification, activation or deactivation may be instructed for all AI / ML models linked to the AI / ML model by instructing activation or deactivation for a model ID for an AI / ML model of an arbitrary stage. Alternatively, the notification may instruct activation or deactivation for the model ID of the final stage AI / ML model (the AI / ML model that performs the desired inference), thereby instructing activation or deactivation for all AI / ML models linked to that AI / ML model.

[0127] In step S14, the control unit 130 of the UE 100 performs inference using the hybrid model. As described above, the AI / ML model at each stage of the linked model may use not only the inference result of the AI / ML model at the previous stage but also actual measurement data. For example, as shown in FIG. 11(C), location information (such as latitude, longitude, and / or altitude) of the UE 100 may be used as input data. Alternatively, movement information (such as speed and / or movement direction) of the UE 100 may be used as input data. Alternatively, information regarding the Quality of Experience (QoE) for an application executed on the UE 100 may be used as input data. The control unit 130 of the UE 100 performs inference using the hybrid model and outputs inference result data. In the example of FIG. 11(C), information regarding the occurrence or non-occurrence of RLF is output from the hybrid model as inference result data. The RLF prediction model may infer a predicted time at which RLF will occur. Alternatively, the RLF prediction model may infer the probability of RLF occurring. The inference result data may include a predicted time and / or probability of occurrence.

[0128] In step S15, the transmitter 120 of the UE 100 may transmit the inference result data to the network node 200. The inference result data may be transmitted by being included in an RRC message. The inference result data may be transmitted by being included in a message of a new AI / ML layer. The inference result data may be transmitted as U-plane data.

[0129] (Another Operation Example According to First Embodiment) In the first embodiment, as shown in FIG. 11C , an RLF prediction model has been described as an example of the final-stage AI / ML model. For example, instead of the RLF prediction model, a HOF prediction model that predicts whether or not an HOF will occur may be used. Even in this case, the inference result of the RRM measurement prediction model is used as an input to the HOF prediction model. In this case, the inference purpose of the hybrid model can be to predict whether or not an HOF will occur. Alternatively, instead of the RLF prediction model, a measurement event prediction model that predicts an event used in a transmission condition for a measurement report may be used. Even in this case, the inference result of the RRM measurement prediction model is used as an input to the measurement event prediction model. In this case, the inference purpose of the hybrid model can be to predict an event.

[0130] (Operation Example 2 According to First Embodiment) In the first embodiment, the description has been given on the assumption that the AI / ML model used in the hybrid model in the UE 100 is already held in the UE 100. For example, the AI / ML model used in the hybrid model may be included in hybrid model setting information and transmitted from the network node 200 to the UE 100 (step S11).

[0131] [Other Embodiments] In the first embodiment described above, supervised learning has been mainly described, but the present invention is not limited to this. For example, unsupervised learning or reinforcement learning may be applied to the first embodiment.

[0132] The above-described operational flows are not limited to being implemented independently, but can be implemented by combining two or more operational flows. For example, some steps of one operational flow may be added to another operational flow, or some steps of one operational flow may be replaced with some steps of another operational flow. In each flow, it is not necessary to execute all steps, and only some steps may be executed. Furthermore, the order of steps in each flow may be changed as appropriate.

[0133] In the above embodiment, an example in which the base station is an NR base station (gNB) has been described, but the base station may be an LTE base station (eNB) or a 6G base station. The base station may also be a relay node such as an IAB (Integrated Access and Backhaul) node. The base station may also be a DU of the IAB node. The UE 100 may also be an MT (Mobile Termination) of the IAB node. That is, the UE 100 may be a terminal function unit (a type of communication module) for the base station to control a relay that relays signals. Such a terminal function unit is referred to as an MT. Examples of MTs include, in addition to IAB-MT, NCR (Network Controlled Repeater)-MT and RIS (Reconfigurable Intelligent Surface)-MT.

[0134] The term "network node" primarily refers to a base station, but may also refer to a core network device or a part of a base station (CU, DU, or RU). A network node may also be configured by a combination of at least a part of a core network device and at least a part of a base station.

[0135] A program may be provided that causes a computer to execute each process according to the above-described embodiments. The program may be recorded on a computer-readable medium. The computer-readable medium can be used to install the program on a computer. Here, the computer-readable medium on which the program is recorded may be a non-transitory storage medium. The non-transitory storage medium is not particularly limited, and may be, for example, a storage medium such as a CD-ROM and / or a DVD-ROM. Furthermore, circuits that execute each process performed by the device according to the above-described embodiments may be integrated, and at least a part of the device may be configured as a semiconductor integrated circuit (chipset, SoC: System on a Chip).

[0136] The functions performed by the apparatus according to the above-described embodiments may be implemented in circuitry or processing circuitry, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), a central processing unit (CPU), conventional circuits, and / or combinations thereof, programmed to perform the described functions. A processor includes transistors and / or other circuits and is considered to be circuitry or processing circuitry. A processor may also be a programmed processor that executes a program stored in a memory. In this disclosure, circuitry, units, and means refer to hardware that is programmed to perform or executes the described functions. The hardware may be any hardware disclosed herein or any hardware known to be programmed or capable of performing the described functions. If the hardware is a processor, the circuitry, means, or unit is a combination of the hardware and software used to configure the hardware and / or processor.

[0137] As used in this disclosure, the terms "based on" and "depending on / in response to" do not mean "based only on" or "depending only on," unless expressly stated otherwise. The term "based on" means both "based only on" and "based at least in part on." Similarly, the term "depending on" means both "depending only on" and "depending at least in part on." The terms "include," "comprise," and variations thereof do not mean including only the listed items, but may mean including only the listed items or may include additional items in addition to the listed items. Additionally, the term "or," as used in this disclosure, is not intended to mean an exclusive or. Furthermore, any reference to elements using designations such as "first," "second," etc., as used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used herein as a convenient method of distinguishing between two or more elements. Thus, a reference to a first and a second element does not imply that only two elements may be employed therein or that the first element must precede the second element in some way. In this disclosure, where articles are added by translation, such as a, an, and the in English, these articles shall include the plural unless the context clearly indicates otherwise.

[0138] Although the embodiments have been described in detail above with reference to the drawings, the specific configuration is not limited to the above, and various design changes can be made within the scope of the gist. Furthermore, the embodiments, operation examples, and processes can be combined as appropriate within the scope of not causing any contradiction.

[0139] This application claims priority to U.S. Provisional Application No. 63 / 679,729 (filed August 6, 2024), the entire contents of which are incorporated herein by reference.

[0140] (Additional Notes) The above can be summarized as in the additional notes, but the additional notes do not limit the embodiments.

[0141] (Supplementary Note 1) A communication control method in a mobile communication system, comprising: a step in which a user device infers a second inference result using a second AI / ML model, using a first inference result of a first AI / ML model as input data to the second AI / ML model; the first AI / ML model is an RRM (Radio Resource Management) measurement prediction model that infers a measurement result of an overall beam as the first inference result from measurement results of partial beams; and the second AI / ML model is any one of an RLF prediction model that infers information indicating whether or not a radio link failure (RLF) has occurred as the second inference result, an HOF prediction model that infers information indicating whether or not a handover failure (HOF) has occurred as the second inference result, and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report as the second inference result.

[0142] (Supplementary Note 2) The communication control method according to Supplementary Note 1, further comprising the steps of: the user equipment receiving hybrid model setting information from a network node; and the user equipment linking the first AI / ML model and the second AI / ML model based on the hybrid model setting information.

[0143] (Supplementary Note 3) The communication control method according to Supplementary Note 1 or Supplementary Note 2, wherein the hybrid model setting information includes configuration information indicating that the first AI / ML model and the second AI / ML model are linked.

[0144] (Appendix 4) A communication control method according to any one of Appendices 1 to 3, wherein the configuration information is specified using first function identification information indicating a function of the first AI / ML model and second function identification information indicating a function of model identification information of the second AI / ML model.

[0145] (Supplementary Note 5) The communication control method according to any one of Supplementary Note 1 to Supplementary Note 4, wherein the configuration information is specified using model identification information of the first AI / ML model and model identification information of the second AI / ML model.

[0146] (Supplementary Note 6) The communication control method according to any one of Supplementary Notes 1 to 5, wherein the hybrid model setting information includes information indicating whether actual measurement data is used as input data, and / or information indicating whether inference results of other models are used as the input data.

[0147] (Supplementary Note 7) The communication control method according to any one of Supplementary Note 1 to Supplementary Note 6, further comprising a step in which the user device trains the first AI / ML model and the second AI / ML model using different training data for the first AI / ML model and the second AI / ML model.

[0148] (Supplementary Note 8) A user device in a mobile communication system, comprising: a control unit that uses a first inference result of a first AI / ML model as input data for a second AI / ML model to infer a second inference result using the second AI / ML model; the first AI / ML model is an RRM (Radio Resource Management) measurement prediction model that infers a measurement result of an overall beam as the first inference result from measurement results of a partial beam; and the second AI / ML model is any one of an RLF prediction model that infers information indicating whether or not a radio link failure (RLF) has occurred as the second inference result, an HOF prediction model that infers information indicating whether or not a handover failure (HOF) has occurred as the second inference result, and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report as the second inference result.

[0149] 1: Mobile communication system 20: RAN 30: CN 100: UE 110: Receiving unit 120: Transmitting unit 130: Control unit 200: Network node 210: Transmitting unit 220: Receiving unit 230: Control unit 240: NW communication unit 300: CN device

Claims

1. A communication control method in a mobile communication system, comprising: a user device inferring a second inference result using a first AI (Artificial Intelligence) / ML (Machine Learning) model, using a first inference result of the first AI / ML model as input data to the second AI / ML model; the first AI / ML model is an RRM (Radio Resource Management) measurement prediction model that infers a measurement result of an overall beam as the first inference result from measurement results of partial beams; and the second AI / ML model is one of an RLF prediction model that infers information indicating whether or not a radio link failure (RLF) has occurred as the second inference result, an HOF prediction model that infers information indicating whether or not a handover failure (HOF) has occurred as the second inference result, and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report as the second inference result.

2. The communication control method according to claim 1, further comprising: the user equipment receiving hybrid model setting information from a network node; and the user equipment linking the first AI / ML model and the second AI / ML model based on the hybrid model setting information.

3. A communication control method according to claim 2, wherein the hybrid model setting information includes configuration information indicating that the first AI / ML model and the second AI / ML model are linked.

4. A communication control method as described in claim 3, wherein the configuration information is specified using first function identification information indicating the function of the first AI / ML model and second function identification information indicating the function of the model identification information of the second AI / ML model.

5. A communication control method according to claim 3, wherein said configuration information is specified using model identification information of said first AI / ML model and model identification information of said second AI / ML model.

6. A communication control method as described in claim 2, wherein the hybrid model setting information includes information indicating whether actual measurement data is to be used as input data, and / or information indicating whether the inference results of another model are to be used as the input data.

7. The communication control method according to claim 1, further comprising the user device training the first AI / ML model and the second AI / ML model using different training data for the first AI / ML model and the second AI / ML model.

8. A user device in a mobile communication system, comprising: a control unit that uses a first inference result of a first AI (Artificial Intelligence) / ML (Machine Learning) model as input data to a second AI / ML model and infers a second inference result using the second AI / ML model; the first AI / ML model is an RRM (Radio Resource Management) measurement prediction model that infers a measurement result of an overall beam as the first inference result from measurement results of partial beams; and the second AI / ML model is any one of an RLF prediction model that infers information indicating whether or not a radio link failure (RLF) has occurred as the second inference result, an HOF prediction model that infers information indicating whether or not a handover failure (HOF) has occurred as the second inference result, and a measurement event prediction model that infers an event indicating a condition for transmitting a measurement report as the second inference result.

Images