Method performed by user equipment and method performed by radio access network

Abstract

A wireless communication system is described. In particular, means and methods for implementing AI / ML models in a wireless communication system, and user plane based data transfer for AI / ML model training, are described.

Description

METHOD PERFORMED BY USER EQUIPMENT AND METHOD PERFORMED BY RADIO ACCESS NETWORK

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

[0002] The disclosure has particular but not exclusive relevance to wireless communication systems and devices thereof operating according to the 3rd Generation Partnership Project (3GPP) standards or equivalents or derivatives thereof (including Long-Term Evolution (LTE)-Advanced, Next Generation or 5G / 6G networks, future generations, and beyond). The present disclosure in particular, but not exclusively, relates to data collection and data transfer for AI / ML models, and to related signalling and communications in a network.

[0003] Earlier developments of the 3GPP standards were referred to as the LTE of Evolved Packet Core (EPC) network and Evolved Universal Mobile Telecommunications System UMTS Terrestrial Radio Access Network (E-UTRAN), also commonly referred as '4G'. More recently, the term '5G' and 'new radio' (NR) has started to be used to refer to an evolving communication technology that is expected to support a variety of applications and services. Various details of 5G networks are described in, for example, the 'NGMN 5G White Paper' V1.0 by the Next Generation Mobile Networks (NGMN) Alliance, which document is available from https: / / www.ngmn.org / 5g-white-paper.html. 3GPP intends to support 5G by way of the so-called 3GPP Next Generation (NextGen) radio access network (RAN) and the 3GPP NextGen core network.

[0004] Under the 3GPP standards, a NodeB (or e.g., an eNB in LTE, and gNB in 5G) is the radio access network (RAN) node (or simply 'access node', 'access network node' or 'base station') via which communication devices (user equipments or 'UEs') connect to a core network and communicate with other communication devices or remote servers. For simplicity, the present application will use the term access network node, RAN node or base station to refer to any such access nodes.

[0005] For simplicity, the present application will use the term mobile device, user device, or UE to refer to any communication device that is able to connect to the core network via one or more RAN nodes. Although the present application may refer to mobile devices in the description, it will be appreciated that the technology described can be implemented on any communication devices (mobile and / or generally stationary) that can connect to a communication network for sending / receiving data, regardless of whether such communication devices are controlled by human input or software instructions stored in memory.

[0006] The RAN node structure may be, but need not necessarily be, split into two or more parts. In some RAN implementations there are two parts, known as the Central Unit (CU or sometimes gNB-CU) - sometimes referred to as a 'control unit' - and the Distributed Unit (DU or sometimes gNB-DU), connected by an F1 interface. This enables the use of a 'split' architecture in which the typically 'higher' CU layers (for example, but not necessarily or exclusively, Packet Data Convergence Protocol (PDCP) and Radio Resource Control (RRC) layers) and the, 'lower' DU layers (for example, but not necessarily or exclusively, Radio Link Control (RLC), Media (sometimes referred to as 'Medium') Access Control (MAC), and Physical (PHY) layers) are separated between a particular CU, and one or more DUs that are connected to and controlled by that CU via the F1 interface. Thus, for example, the higher layer CU functionality for a number of gNBs may be implemented centrally (for example, by a single processing unit, or in a cloud-based or virtualised system), whilst retaining the lower layer DU functionality locally separately for each gNB.

[0007] The choice of how to split functions in the architecture depends on, among other things, factors related to radio network deployment scenarios, constraints and intended supported use cases. Key considerations include: the need to support a specific quality of service for each service offered and for real / non-real time applications; support of specific user density and load demand in a given geographical area; and available transport networks with different performance levels.

[0008] Some recent developments in 3GPP relate to the use of artificial AI and ML, often abbreviated to AI / ML. Predictions or inferences generated using an AI / ML model can be used as part of various methods for improving the reliability or efficiency of communication in the network. For example, AI / ML models can be used to predict the path of a UE based on previous mobility of the UE, used for cell and / or beam management, or used in methods of encoding and transmitting information. An AI / ML model may be hosted at a RAN node, and the RAN node 5 may perform control of communication resources or control related to the status of a UE (e.g., control of UE mobility, or control of a radio resource control (RRC) state of the UE) based on an inference (e.g., determination or prediction) generated using the AI / ML model. The RAN node 5 may also transmit an inference generated using the model to another node in the network, for use at the other node.

[0009] Alternatively, an AI / ML model may be hosted at two nodes of the network, for example at a RAN node and at a UE. In this case, the RAN node and the UE may both make determinations or predictions using the model. For example, the UE may use the model as part of an encoding process for encoding (and / or compressing) channel state information (CSI) for transmission to the RAN node 5, and the RAN node 5 may use the same model as part of a corresponding decoding (and / or decompression) process for decoding the CSI received from the UE.

[0010] It will be appreciated that the use of AI / ML models may also be extended to other procedures and methods performed in a communication system to further improve the reliability or efficiency of communication in the network. For example, AI / ML models may be extended to generate inferences about the positions of UEs in a communication system using, for example, UE-assisted information and / or location management function (LMF)-based positioning assistance, and / or RAN node-based positioning assistance to facilitate life cycle management (LCM) operations specific to positioning accuracy enhancements in the communication system.

[0011] NPL 1: 'NGMN 5G White Paper' V1.0 by the Next Generation Mobile Networks (NGMN), available from https: / / www.ngmn.org / 5g-white-paper.html.

[0012] An AI / ML data collection entity in the network may provide AI / ML training data to an AI / ML model training function. The collected data may be, for example, data regarding mobility (e.g., handover of a UE 3, or a location of the UE 3), or any other suitable data related to the training of the AI / ML model. However, there is a problem that training data for the AI / ML model typically has a large data size, and so the collection and transmission of such data in the network must be carefully considered to avoid congestion of the network or inefficient use of the available communication resources. Two options for data collection for a UE-side model are to provide a training entity (e.g., an over-the top (OTT) server) or another server for data collection as the first termination entity in the network for transmission of the AI / ML training data, and in these cases the data could be transmitted using a user plane tunnel and the application layer. However, there is a problem that these options for data transmission for an AI / ML model lack control by the mobile network operator, which could cause issues in the network due to the large size of the data to be transmitted. Alternatively, the first termination entity in the network for transmission of the AI / ML data for a UE-side model may be provided in the core network, or inside an operations and maintenance (OAM) domain. In these options, a control plane tunnel could be used to transmit the AI / ML data, and either the NAS layer or RRC layer could be used. However, whilst this provides improved control of the AI / ML data transfer by the network operator, there is a problem that the control plane is used for control signalling and typically has limited communication resources available for data transfer, which can be problematic for transmission of the large amount of AI / ML data. It will be appreciated, therefore, that improved apparatus and methods for data transfer for AI / ML models (e.g. transmission of data for AI / ML model training) are needed.

[0013] For the case where the AI / ML data collection is terminated in the core network, if non-access stratum (NAS) signalling is used for the data transfer then the large data size of the AI / ML data can cause data transmission issues, as the NAS signalling is not typically designed for transmission of such a large amount of data. However, if the user plane (UP) is used for the AI / ML data transmission, there is a problem that it is not clear how the UE can be configured to efficiently collect and report the AI / ML data to the core network. This is because the UE is typically configured using RRC signalling for any type of data report, and the core network is not aware of which type of data the UE should collect and report for the AI / ML use case. Alternatively, for the case in which the AI / ML data transmission is terminated at the OAM, if RRC signalling is used for the data transfer then there is problem that the RRC signalling does not support transmission for the large data size of the AI / ML data, and the RRC signalling may be overwhelmed by the AI / ML data transmission. For the case of UP-based AI / ML data transfer, there is also a problem that it is unclear how the UE can be configured to collect the AI / ML data and establish a UP tunnel for transmission of the AI / ML data, since a UP tunnel is typically only established between the UE and a user plane function (UPF), under the control of an Access and Mobility Management Function (AMF) or Session Management Function (SMF).

[0014] The disclosure has a method performed by a User Equipment, UE, the method comprising receiving, from a Radio Access Network node, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and performing at least one of a periodic logging or an event-triggered logging based on the configuration information.

[0015] The disclosure has a method performed by a User Equipment, UE, the method comprising transmitting, to a Radio Access Network node, RAN, node, a first indication indicating that the UE supports a functionality of Channel State Information, CSI, compression based on an Artificial Intelligence / Machine Learning, AI / ML, dataset and / or an AI / ML model establishing a unicast Radio Bearer, RB, broadcast RB, or multicast RB; and receiving parameters for the AI / ML dataset and / or AI / ML model from the RAN node 5.

[0016] The disclosure has a method performed by a Radio Access Network, RAN, node, the method comprising transmitting, to a User Equipment, UE, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and receiving, from the UE, logged information of the measurement performed based on the configuration information, wherein the logged information is based on at least one of a periodic logging or an event-triggered logging.

[0017] The present specification aims to disclose apparatus and methods that at least contribute to addressing one or more of the above needs and / or issues.

[0018] The various functional means described below that are part of the UE may be provided by a memory and one or more processors that execute instructions stored in the memory. Similarly, the various functional means described below that are part of the access network node may be provided by a memory and one or more processors that execute instructions stored in the memory.

[0019] Various example described below may be implemented by means of a computer program product comprising computer implementable instructions for causing a programmable computer to carry out the any of the methods described below. The computer implementable instructions may be provided as a signal or on a tangible computer readable medium.

[0020] Examples of apparatus and methods will now be described, by way of example, with reference to the accompanying drawings in which:

[0021] Fig. 1 schematically illustrates a mobile ('cellular' or 'wireless') communication system;Fig. 2 illustrates a functional framework for AI / ML models, and how various entities of the framework may interact with one another, that may be implemented in the communication system of Fig. 1;Fig. 3 schematically illustrates a method of training an AI / ML model, and of monitoring the performance of the AI / ML model, which may be implemented in the communication system of Fig. 1;Fig. 4 shows a simplified schematic illustration of various functions and entities that may be provided in the network;Fig. 5 illustrates an example of an improved method for configuration and control of data collection for an AI / ML model;Fig. 6 shows a modified version of the method of Fig. 5, in which there is an additional step of user plane tunnel establishment for reporting the AI / ML data;Fig. 7 illustrates an alternative improved method for method for configuration and control of data collection for an AI / ML model;Fig. 8 illustrates a method in which an AI / ML data report is triggered by the core network;Fig. 9 illustrates a method in which an AI / ML data reporting procedure is triggered by the UE;Fig. 10 illustrates a method in which an OAM transmits a request for the RAN node to configure a data type for data collection for an AI / ML model;Fig. 11 illustrates a further method in which the OAM transmits a request for the RAN node to configure a data type for data collection for an AI / ML model;Fig. 12 illustrates a further method in which AI / ML data is transmitted to the OAM;Fig. 13 illustrates a method in which a GTP-U tunnel is established for transmission of an AI / ML dataset and / or AI / ML model parameters;Fig. 14 illustrates a method of RAN node-based AI / ML dataset and / or AI / ML model parameter transfer;Fig. 15 is a simplified block schematic illustrating the main components of a UE for implementation in the communication system of Fig. 1; andFig. 16 is a simplified block schematic illustrating the main components of a RAN node for implementation in the communication system of Fig. 1.

[0022] <Overview> 　　An exemplary telecommunication system will now be described in general terms, by way of example only, with reference to Figs. 1 to 3.

[0023] Fig. 1 schematically illustrates a mobile ('cellular' or 'wireless') communication system (e.g., communication system 1) to which examples of the present disclosure are applicable.

[0024] In the communication system 1, user equipments (UEs) 3 (3-1, 3-2, 3-3) (e.g., mobile telephones and / or other mobile devices) can communicate with each other via a corresponding (radio) access network ((R)AN) node 5-1, 5-2 that operates according to one or more compatible radio access technologies (RATs). In the illustrated example, each RAN node 5 (5-1, 5-2) comprises a base station that respectively operates one or more associated cells. Communication via the RAN nodes 5 are typically routed through a core network 7 (e.g., a 5G / 6G or later generations core network or evolved packet core network (EPC)).

[0025] As those skilled in the art will appreciate, whilst three UEs 3 and two RAN nodes 5-1, 5-2 are shown in Fig. 1 for illustration purposes, the system, when implemented, will typically include other RAN nodes and UEs.

[0026] Each RAN node 5 controls one or more associated cells either directly, or indirectly via one or more other nodes (such as home base stations, relays, remote radio heads, distributed units, and / or the like). It will be appreciated that the RAN nodes 5 may be configured to support 4G, 5G, 6G and / or later generation, and / or any other 3GPP or non-3GPP communication protocols.

[0027] The UEs 3 are configured for communication with their serving RAN node 5 via an appropriate air interface (for example a so-called 'Uu' interface and / or the like). Neighbouring RAN nodes 5 may be connected to each other via an appropriate RAN node to RAN node interface (such as the so-called 'X2' interface, 'Xn' interface and / or the like - not shown in Fig. 1).

[0028] The core network 7 includes a number of logical nodes (or 'functions') for supporting communication in the communication system 1. In this example, the core network 7 comprises control plane functions (CPFs) 10 and one or more network node entities for the communication of user data (e.g. user plane functions (UPFs) 11). The CPFs 10 may include one or more network node entities for the communication of control signalling (e.g. Access and Mobility Management Functions (AMFs) 10-1), one or more network node entities for session management (e.g. Session Management Functions (SMFs) 10-2), a network data analytics function / data collection function (NWDAF / DCF) 10-7, and a number of other functions 10-n. Additional functions may include, for example: an Authentication Server Function (AUSF) which facilitates security processes; a Unified Data Management (UDM) entity for managing user specific data (e.g., for access authorisation, user registration, and data network profiles); a Policy Control Function (PCF); an Application Function (AF); a Security Anchor Function (SEAF) which is in a serving network and acts as a "middleman" during an authentication process between the UE 3 and its home network; an Authentication credential Repository and Processing Function (ARPF) which maintains the authentication credentials; and / or the like. It will be appreciated that the nodes or functions may have different names in different systems.

[0029] The communication system 1 also includes an Operations, Administration and Maintenance (OAM) 14 comprising one or more OAM functions for provisioning and managing network or elements within the wider communication system 1. The OAM 14 may be responsible for the storage and analysis of some radio-related measurements and may perform some data analytics functions including some RAN analytics. The OAM 14 may, for example, communicate with one or more of the core network CPFs 10 via a network data analytics function (NWDAF) or the like (not shown).

[0030] Each RAN node 5 is connected to the core network nodes via appropriate interfaces (or 'reference points') such as an N2 reference point between the RAN node 5 and the AMF 10-1 for the communication of control signalling, and an N3 reference point between the RAN node 5 and each UPF 11 for the communication of user data. The UEs 3 are each connected to the AMF 10-1 via a non-access stratum (NAS) connection over an appropriate reference point (e.g., N1 reference point (analogous to the S1 reference point in LTE)). It will be appreciated, that N1 communication is routed transparently via the RAN node 5.

[0031] One or more UPFs 11 are connected to an external data network 40 (e.g., an IP network such as the internet) via an appropriate reference point (e.g., N6 reference point) for communication of the user data.

[0032] The AMF 10-1 performs mobility management related functions, maintains the NAS connection with each UE 3 and manages UE registration. The AMF 10-1 is also responsible for managing paging. The AMF 10-1 receives user information sent through the network and forwards the information to the SMF 10-2.

[0033] The SMF 10-2 is connected to the AMF 10-1 via an appropriate reference point (e.g., N11 reference point). The SMF 10-2 provides session management functionality (that formed part of MME functionality in LTE) and additionally combines some control plane functions (provided by the serving gateway and packet data network gateway in LTE). The SMF 10-2 uses user information provided via the AMF 10-1 to determine what session manager would be best assigned to the user. The SMF 10-2 may be considered effectively to be a gateway from the user plane to the control plane of the network. The SMF 10-2 also allocates IP addresses to each UE 3.

[0034] < AI / ML> 　　The communication system 1 supports the use of artificial intelligence (AI) and machine learning (ML), often abbreviated to AI / ML in accordance with recent developments in cellular communication technology (e.g., as part of the work of the 3GPP) that those skilled in the art will be familiar with. These AI / ML features make use of trained AI / ML models to make one or more predictions or inferences, from a set of one or more input vectors, which can be used in the network (e.g., for improving the reliability or efficiency of communication in the network).

[0035] In respect of the communication system 1, for example, AI / ML models could potentially be trained and used for predicting the path of a UE 3 based on previous mobility of the UE 3, used for beam management, or used in methods of encoding and transmitting information. An AI / ML model may be hosted at a RAN node 5 (or any other suitable network node), and the RAN node 5 may perform control of communication resources for UEs 3 it serves, and / or perform control related to the status of a UE 3 (e.g. control of UE mobility, or control of a radio resource control, RRC, state of the UE 3) based on an inference (e.g. determination or prediction) generated using the AI / ML model. The RAN node 5 may also transmit an inference generated using the model to another node in the network, for use at the other node. An AI / ML model may also be hosted the UE 3, or at a plurality of locations within the network, for example at both the RAN node 5 and at the UE 3. For example, the RAN node 5 and the UE 3 may both make determinations and / or predictions using the same model or different models.

[0036] The support for such AI / ML features may involve different levels of collaboration between the network (RAN node 5 and / or core network 7) and a UE 3 served by the network when deploying and using such AI / ML features. For example, three possible 'network-UE collaboration levels' that may be supported are: Level x: Involving no collaboration between the network and the UE 3. Specifically, level x is an implementation-based AI / ML operation without any dedicated AI / ML-specific enhancement. Level y: Signalling-based collaboration without AI / ML model transfer. For example, this level is applicable when model training is performed offline, and models are registered to both a RAN node 5 and the UE 3. Here, the RAN node 5 and the UE 3 are aware of available models (before operation), and the RAN node 5 is only required to activate / deactivate the models residing at the UE 3 when needed. Level z: Signalling-based collaboration with AI / ML model transfer (e.g., where an AI / ML model is transferred to the UE 3 when needed.

[0037] The AI / ML model types that are supported in the communication system 1 may include, for example: - Single-sided model: A single-sided AI / ML model is an AI / ML model that is deployed (hosted) only at the UE side or at the network side. For example, an AI / ML model may be hosted (stored, for generating inferences) at a UE 3, RAN node 5 , or a central entity of the communication system 1, or an operations, administration, and maintenance (OAM) / over-the-top (OTT) server. When the AI / ML model is used at a UE 3, a RAN node 5, a central entity of the communication system 1, or an OAM / OTT server only, the AI / ML model may be referred to as a 'single-sided' model. An example of this type of 'single-sided' model is an AI / ML model for beam prediction in time, which can be deployed at the UE side. - However, even when the model is a single-sided model, it will be appreciated that the model need not necessarily be trained at the node at which it is deployed (e.g., a UE 3 or a RAN node 5). For example, the model could be trained at the RAN node 5 (or at another node in the network such as a core network node / function - e.g., a central entity of the communication system 1, or an OAM / OTT server - and then transferred to the UE 3 for use at the UE 3. - Two-sided model: A 'two-sided' model is an AI / ML model (or model pair) that has one AI / ML model hosted at one node (e.g., the UE 3), and a corresponding AI / ML model hosted at another node (e.g., a RAN node 5) - it will be appreciated that any pair of network nodes may be used. Such a two-sided model may also be referred to as a 'paired' AI / ML model. An inference using a two-sided model is performed jointly across the nodes at which the AI / ML models of the two-sided model are deployed. The joint inference may comprise, for example, a first part of the inference being performed at one node (e.g., the UE 3 or RAN node 5), and then the remaining part may be performed by the other (e.g., the RAN node 5 or UE 3). It will be appreciated that whilst the AI / ML model hosted at the different nodes may be the same AI / ML model, they need not necessarily be the same model. One example of this type of model is channel state information (CSI) compression, where the UE 3 performs CSI compression and network performs CSI decompression. As with the single-sided model case, the two-sided model (or models) may be trained at any suitable network node, and then transmitted to the UE 3 and the RAN node 5 (or other respective node or nodes).

[0038] A general discussion of how AI / ML may be implemented in the communication system 1 will now be provided, by way of example only, with reference to Figs. 2 and 3.

[0039] Fig. 2 illustrates a functional framework for AI / ML models, and how various entities of the framework may interact with one another, that may be implemented in the communication system 1.

[0040] The entities include a data collection entity 241, a model training function 243, a model inference function 245, an actor 247, a management function 249, and a model storage entity 251.

[0041] The model storage entity 251 may be a reference point for protocol terminations for model transfer and delivery. The AI / ML models could be stored at any suitable node in the network.

[0042] The data collection entity 241 provides training data to the model training function 243, inference data to the model inference function 245, and monitoring data to the management function 249. The collected data may be, for example, data regarding mobility (e.g., handover of a UE 3, or a location of the UE 3). The data may be obtained, for example, by a UE 3 or a RAN node 5 (e.g., by receiving a measurement report from a UE 3, or by receiving data from another RAN node 5 or a core network node / function) and transmitted to another RAN node 5 or core network node that generates the AI / ML model inference output (or alternatively, the same RAN node 5 that obtains the data may generate the AI / ML model output).

[0043] The model training function 243 performs the AI / ML model training, validation, and testing, and may generate model performance metrics as part of a model testing procedure. The model training function 243 may output a trained AI / ML model to the model storage entity 251 (though it will be appreciated that the output model may be stored at locations other than model storage entity 251).

[0044] The model inference function 245 provides AI / ML model inference output (e.g., predictions or decisions), and the actor 247 is a function or node that receives the output from the model inference function 245 and triggers or performs corresponding actions (e.g., a RAN node 5 that increases / reduces its transmit power or initiates a handover procedure for a UE 3). The AI / ML model inference output may be, for example, a prediction of mobility (e.g., expected path, route or trajectory, inter-cell, or inter-beam mobility, or expected handover) of the UE 3, or one or more parameters for use in encoding or decoding transmissions between the RAN node 5 and the UE 3. The model inference function 245 may receive an AI / ML model from the model storage entity 251, and inference data from the data collection entity 241 for use with the AI / ML model. The model inference function 245 may also output monitoring data for use at the management function 249 and receive information indicating an AI / ML to activate or deactivate from the management function 249.

[0045] The management function 249 receives monitoring data from the data collection entity 241 and may also receive monitoring data from the model inference function 245. The management function 249 may transmit to the model storage entity 251, an indication of an AI / ML model to be transmitted for use at the model inference function 245. The management function 249 may also transmit to the model training function 243, performance feedback or a retraining request for the AI / ML model.

[0046] The functions and entities illustrated in Fig. 2 may be co-located at a single node of the communication system 1 (e.g., at the RAN node 5 or core network node / function) or may be distributed amongst a plurality of network nodes (e.g., a plurality of the RAN nodes 5).

[0047] By way of example only, terms referred to by 3GPP in the context of this framework include: AI / ML model training: A process to train an AI / ML Model (e.g., by learning an input / output relationship) in a data driven manner and obtain the trained AI / ML Model for inference. Model training can be performed offline or online or combination of both. AI / ML model validation: A subprocess of training, to evaluate the quality of an AI / ML model using a dataset different from one used for model training, which helps selecting model parameters that generalize beyond the dataset used for model training. AI / ML model testing: A subprocess of training, to evaluate the performance of a final AI / ML model using a dataset different from one used for model training and validation. Differently from AI / ML model validation, testing does not assume subsequent tuning of the model. AI / ML model Inference: A process of using a trained AI / ML model to produce a set of outputs based on a set of inputs. Data collection: A process of collecting data by the network nodes, management entity, or UE for the purpose of AI / ML model training, data analytics and inference. Model monitoring: A procedure that monitors the inference performance of the AI / ML model. Model activation: Enable an AI / ML model for a specific function. Model deactivation: Disable an AI / ML model for a specific function. Model switching: Deactivating a currently active AI / ML model and activating a different AI / ML model for a specific function. Supervised learning: A process of training a model from input and its corresponding labels. Unsupervised leaning: A process of training a model without labelled data. Semi-supervised learning: A process of training a model with a mix of labelled data and unlabelled data. Reinforcement Learning (RL): A process of training an AI / ML model from input (also referred to as 'state') and a feedback signal (also referred to as 'reward') resulting from the model's output (also referred to as 'action') in an environment the model is interacting with.

[0048] The data collection by the data collection entity 241 may be performed at various nodes of the communication system 1 (e.g., at one or more RAN nodes 5 or UEs 3).

[0049] Fig. 3 schematically illustrates a method of training an AI / ML model, and of monitoring the performance of the AI / ML model. As illustrated in Fig. 3, stored data / features may first be extracted in a data extraction step. In the data validation step, a determination of whether to proceed with training or retraining the AI / ML model is made (e.g., based on the extracted data). In the data preparation stage, the data is prepared for use in training the AI / ML model. For example, the data may be cleaned (e.g., filtered), subject to a transformation, or modified in any other suitable manner. The data may also be divided in training data, validation data and test data sets in the data preparation stage.

[0050] In the model training step, the AI / ML model is trained (or retrained) using training data prepared in the data preparation step. It will be appreciated that any suitable training method can be used to train the AI / ML model (e.g., a method that comprises supervised learning or unsupervised learning). In the model evaluation step, the AI / ML model is evaluated (e.g., a prediction accuracy of the AI / ML model is evaluated) using a test data set (which may be generated in the data preparation step). In the model validation step, a determination of whether the AI / ML model is suitable for deployment in the communication system 1 is made (e.g., based on the results of the model evaluation step).

[0051] In the model serving step, the AI / ML model is deployed for use in the communication system 1. AI / ML model deployment may comprise compiling a trained AI / ML model, packaging the model into an executable format, and delivering the AI / ML model to a target device. For example, the AI / ML model may be transmitted to the RAN node 5 and / or the UE 3, for use at the RAN node 5 and / or the UE 3 to generate predictions or determinations using the AI / ML model as part of a prediction service. In the performance monitoring step, the performance of the deployed AI / ML model is monitored. The predictive performance of the AI / ML model may be monitored by comparing predictions generated using the model with one or more measurements. For example, when the AI / ML model is used to predict a location of a UE 3, the prediction accuracy of the AI / ML model may be assessed using a measurement of an actual location of the UE 3. If the AI / ML model is used for predicting future measurement results (e.g., the measured Reference Signal Received Power (RSRP) of reference signals) at some point in time, the prediction accuracy of the AI / ML model may be assessed using actual measurement results acquired by the UE 3 when that point in time is reached. If the AI / ML model is used for determining parameters for use in encoding and decoding data transmitted between a RAN node 5 and a UE 3, the model may be assessed based on the performance of the encoding and / or decoding processes. In the retraining trigger step, retraining of the AI / ML model is triggered (e.g., because the prediction accuracy of the AI / ML model has fallen below an acceptable threshold accuracy, or because a performance of a method that uses inferences from the AI / ML model has fallen below an acceptable threshold performance), and the method returns to the data extraction step.

[0052] Each step of the method of Fig. 3 may be executed at a single node of the communication system 1 (including at the RAN node 5), or alternatively steps of the method may be distributed between a plurality of different nodes (or indeed one or more of these steps may be performed online or offline).

[0053] As discussed above, information collected by nodes / functions in the communication system 1 (e.g., at a UE 3 and / or RAN nodes 5) can be used as training data for an AI / ML model and used as inference data for use in generating one or more model inferences using the AI / ML model. The information used as training data, monitoring data, and / or to generate the one or more model inferences may be referred to as 'AI / ML information' or 'AI / ML data.'

[0054] < Configuration information for AI / ML> 　　Configuration information for an AI / ML model (which may be referred to as "AI / ML configuration information") may be exchanged between nodes in the communication system. For example, a core network node may transmit AI / ML configuration information to a RAN node 5 (or any other entity of the network that supports an AI / ML-based prediction functionality) that hosts an AI / ML model. The AI / ML configuration information may include a list of supported use cases for the AI / ML model (the AI / ML model need not necessarily be for predicting UE mobility). The supported use cases may be, for example: energy saving; traffic steering; anomaly detection; quality of experience (QoE) optimisation; mobility robustness optimisation (MRO); RAN slice service level agreement (SLA) assurance; massive multiple-input multiple-output (MIMO) beamforming optimisation; network slice subnet instance (NSSI) resource allocation; optimisation coverage and capacity optimisation (CCO); mobility load balancing (MLB); RACH optimisation; or UE transmission power optimisation. The AI / ML configuration information may include an indication of a particular AI / ML model to use for a particular use case. The AI / ML configuration information may also include an indication of whether feedback is required (e.g., from another network node). The feedback may include, for example, communication performance feedback (e.g., indicating a communication performance for communication between a UE 3 and a RAN node 5).

[0055] A number of enhanced procedures and techniques that may be implemented in the communication system 1 will now be described, by way of example only, with reference to Figs. 4 to 14.

[0056] Fig. 4 shows a simplified schematic illustration of various functions and entities that may be provided in the system illustrated in Fig. 1. In this example, the system includes a Unified Data Management (UDM) entity 10-4, an SMF 10-2, a network exposure function (NEF) 10-5, an AF 10-6, and AMF 10-1, a UPF 11, and an NWDAF / Data Collection Function (DCF) 10-7 in addition to the RAN node 5 and the UE 3. The NEF 10-5 is a function that is provided between the core network and third-party applications outside of the network, and optionally between the core network and the AF 10-6 provided in the system illustrated in Fig. 4, to provide secure access to information from within the network. The NWDAF / DCF 10-7 is arranged in the core network domain, and may be used as a data collection function for AI / ML model training data. Alternatively, a specific DCF entity may be provided for data collection function, which is a new function that is provided for data collection for an AI / ML model (e.g. a UE-side AI / ML model). In a further alternative, the data collection function for AI / ML model training data could be provided by the AMF 10-1. Advantageously, a new functionality is defined between the UE 3 and the NWDAF / DCF 10-7 for the configuration and control of a data collection function for AI / ML models (e.g. UE-side AI / ML models). Alternatively, this functionality may be provided as part of the NAS functions between the UE 3 and AMF 10-1. Beneficially, the NWDAF / DCF 10-7 may be provided as the first termination point in the network for the AI / ML data collection, and may be connected with a specific server or AF 10-6 for AI / ML model training within the network.

[0057] Fig. 5 illustrates an example of an improved method for configuration and control of data collection for an AI / ML model. In step S501, the UE 3 is connected to the network, for communication via the RAN node 5 (e.g., in the 'RRC connected' state). The UE 3 may provide an indication to the network that UE 3 supports the operation of a particular AI / ML model (e.g., for the purpose of beam management).

[0058] In optional step S502, the UE 3 provides, to the AMF 10-1, an indication that the UE 3 requests data collection for UE-side AI / ML model training. The information provided by the UE 3 to the network in step S502 may include an identity of the AI / ML model (e.g., an identification number), an identity of the UE 3, an indication of model characteristics of the AI / ML model (e.g., the functionality supported by the AI / ML model), and / or a request that the AI / ML model be trained. The information provided by the UE 3 to the AMF 10-1 may be included in NAS signalling, or similar NAS-like signalling, between the UE 3 and the AMF 10-1, for example. In step S503, the AMF 10-1 provides an indication to the NWDAF / DCF 10-7 that an AI / ML model for the UE 3 is to be registered at the NWDAF / DCF 10-7. The AI / / ML model is registered at the NWDAF / DCF 10-7 for the purpose of training the AI / ML model. Advantageously, in this example, the AMF 10-1 registers the AI / ML model training at the NWDAF / DCF 10-7 using the identity of the UE 3 and the AI / ML model identity provided by the UE 3 in step S502. Therefore, the AMF 10-1 can efficiently and reliably register the training of the AI / ML model at the NWDAF / DCF 10-7. Alternatively, rather than the UE 3 transmitting the optional request for data collection in step S502, the NWDAF / DCF 10-7 may trigger the data collection procedure for training a particular AI / ML model for the UE 3. In a further alternative, the OAM 14 (not illustrated in Fig. 5) may trigger the data collection procedure.

[0059] Optionally, before the NWDAF / DCF 10-7 and / or AMF 10-1 configure the UE 3 for data collection for the AI / ML model training, the NWDAF / DCF 10-7 or AMF 10-1 may transmit a request to the RAN node 5, requesting the RAN node 5 to provide the data type for the data collection. For example, for an AI / ML model for beam management, the RAN node 5 may transmit an indication to the NWDAF / DCF 10-7 or AMF 10-1 that UE L1 measurements should be collected (e.g., for the top N beams). In a further example, for an AI / ML model for radio link failure (RLF) prediction, the RAN node 5 may transmit an indication to the NWDAF / DCF 10-7 or AMF 10-1 that information related to the occurrence of RLF events should be collected, and / or L3 measurements of a serving cell and neighbouring cells should be collected.

[0060] In step S504, the NWDAF / DCF 10-7 transmits a request to the AMF 10-1 for the UE 3 to be configured with the data type for data collection, and to be configured with report criteria, for the AI / ML model training. The information transmitted in step S504 may include the identity of the UE 3 (e.g., an identification number of the UE 3) and / or an identity of the AI / ML model (e.g., an identification number of the AI / ML model). Beneficially, therefore, the AMF 10-1 is able to efficiently and reliably configure the UE 3 for the AI / ML data collection and reporting, using the information provided by the NWDAF / DCF 10-7.

[0061] In step S505, the AMF 10-1 configures the UE 3 for collection of the data for training the AI / ML model (in this example, a UE-side model). The AMF 10-1 configures the UE 3 using NAS signalling transmitted via the RAN node 5, but any other suitable signalling could alternatively be used. The information transmitted in step S505 may include the identity of the AI / ML model. In step S505 the AMF 10-1 may also provide the UE 3 (via the RAN node 5) with a start / stop condition or start / stop indication for starting or stopping the AI / ML data collection. In step S505 the AMF 10-1 may also provide the UE 3 (via the RAN node 5) with data collection time duration, specify a time period or duration for which the UE 3 is to collect the AI / ML data. The AMF 10-1 may also provide the UE 3 with a data collection time granularity(e.g., the data is logged every 100ms). The AMF 10-1 may also provide the UE 3 with a periodicity to be used by the UE 3 to collect the AI / ML data, or an event for which the AI / ML data is to be collected for. For example, for an RLF prediction model, the AMF 10-1 may provide an indication to the UE 3 that L3 measurements for a serving cell and a neighbouring cell are to be collected when RLF occurs. The AMF 10-1 may also provide an indication to the UE 3 of a periodicity to be used for reporting the collected AI / ML data to the network. Alternatively, or additionally, the AMF 10-1 may provide an indication to the UE 3 of an event for triggering the UE 3 to report the collected data to the network. For example, the AMF 10-1 may transmit an indication to the UE 3 that the UE 3 is to report collected AI / ML data to the network once the amount of AI / ML data (data size) stored at the UE 3 exceeds a threshold value.

[0062] Alternatively, rather than the AMF 10-1 explicitly specifying the configuration to be used for the UE 3 to collect and report the AI / ML data, the AMF 10-1 may transmit an indication to the RAN node 5 that the UE 3 is to be configured for data collection and / or data reporting for a particular AI / ML model, without the AMF 10-1 providing an explicit indication of the configuration to use for the data collection. In this case, the RAN node 5 may transmit an indication of a configuration for the UE 3 for AI / ML data collection and / or reporting using RRC signalling, or any other suitable signalling. For example, configuration information for AI / ML data collection and / or reporting may be stored at the RAN node 5, and the RAN node 5 may transmit the information for a particular AI / ML model to the UE 3 after receiving an indication from the AMF 10-1 that the UE 3 is to be configured for data collection for that AI / ML model. In step S505, the UE 3 may be configured in which state, or states, (e.g. RRC connected state) the UE should collect the data.

[0063] In step S506, the UE 3 collects (logs) the AI / ML data, using the configuration provided to the UE 3 from the AMF 10-1 via the RAN node 5 in step S505. In step S506 the UE 3 may also report the collected AI / ML data to the network using the configuration provided to the UE 3 from the AMF 10-1 via the RAN node 5 in step S505.

[0064] Fig. 6 shows a modified version of the method of Fig. 5, in which there is an additional step of user plane tunnel establishment for reporting the AI / ML data.

[0065] Steps S601 to S605 and S607 of Fig. 6 are the same as steps S501 to S506 of Fig. 5, except for the modifications set out below, and so will not be described again here.

[0066] In this example, a user plane tunnel corresponds to a GTP-U tunnel established between the RAN node 5 and the NWDAF / DCF 10-7, and also the user plane data transmission channel between the UE 3 and the RAN node 5.

[0067] In this example, the NWDAF / DCF 10-7 transmits a request to the AMF 10-1 for a user plane tunnel to be established for transmission of the AI / ML data by the UE 3. The request for the user plane tunnel to be established may be transmitted from the NWDAF / DCF 10-7 to the AMF 10-1 in step S604. The information provided to the AMF 10-1 by the NWDAF / DCF 10-7 may include any suitable information for establishing the user plane tunnel. The NWDAF / DCF 10-7 may transmit, to the AMF 10-1, an indication of a Tunnel ID to be used for the user plane tunnel. For example, the NWDAF / DCF 10-7 may transmit, to the AMF 10-1, an indication of a General Packet Radio Service (GPRS) Tunnelling Protocol for user data (GTP-U) Tunnel ID. The NWDAF / DCF 10-7 may additionally, or alternatively, transmit, to the AMF 10-1, an indication of one or more quality of service (QoS) parameters for a QoS flow to be carried by the tunnel. Accordingly, the AMF 10-1 may instruct the RAN node 5 with the GTP-U Tunnel ID (received from the NWDAF / DCF 10-7) to establish a GTP-U Tunnel with the NWDAF / DCF 10-7. The RAN node 5 responds by transmitting its GTP-U Tunnel ID to the AMF 10-1, and the AMF 10-1 forwards the GTP-U Tunnel ID (received from the RAN node 5) to the NWDAF / DCF 10-7.

[0068] It will be appreciated that the GTP-U Tunnel may be established between the RAN node 5 and the NWDAF / DCF 10-7 via a User Plane Functionality in the network. Accordingly, there may be user plane data transmission channel established between the UE 3 and the RAN node 5 before the corresponding AI / ML data is available for transmission by the UE 3, and the UE 3 may report the AI / ML data to the network using the tunnel according to a reporting configuration received in step S605 (e.g., from the AMF 10-1 via the RAN node 5). A further alternative for reporting the AI / ML data when a user plane tunnel has not been established will be described later.

[0069] Fig. 7 illustrates an alternative improved method for method for configuration and control of data collection for an AI / ML model. In this example, in step S701, the NWDAF / DCF 10-7 transmits, to the AMF 10-1, a request for AI / ML data for model training to be collected. The information transmitted from the NWDAF / DCF 10-7 to the AMF 10-1 in step S701 may include an identity of the AI / ML model for which the AI / ML data is to be collected (e.g., an identification number of the AI / ML model). The information transmitted from the NWDAF / DCF 10-7 to the AMF 10-1 in step S701 may include an indication of characteristics of the AI / ML model for which the AI / ML data is to be collected. The characteristics of the AI / ML model may comprise AI / ML model size, the functionality supported by AI / ML model, required data set, parameters to be configured to the AI / ML model, etc. The request of step S701 may include a request for the AMF 10-1 to configure the UE 3 with a type of data ('data type') for data collection. The request of step S701 may include a request for the AMF 10-1 to configure the UE 3 with one or more reporting criterion or reporting configurations for use by the UE 3 to transmit the AI / ML data to the network. For example, the request of step S701 may include a request for the AMF 10-1 to configure the UE 3 with a reporting periodicity or reporting trigger condition. More generally, the request of step S701 may include a request for the AMF 10-1 to configure the UE 3 with any suitable configuration for collecting and / or reporting the AI / ML data, for example any of the configuration information described above with reference to step S505 of Fig. 5.

[0070] The request of step S701 may additionally, or alternatively, include a request for the AMF 10-1 to establish a user plane tunnel for transmission of the AI / ML data by the UE 3. As described above with reference to Fig. 6, the information provided to the AMF 10-1 by the NWDAF / DCF 10-7 may include an indication of a Tunnel ID to be used for the user plane tunnel. For example, in step S701 the NWDAF / DCF 10-7 may transmit, to the AMF 10-1, an indication of a GTP-U Tunnel ID. Also as described above with reference to Fig. 6, the information provided to the AMF 10-1 by the NWDAF / DCF 10-7 may include an indication of one or more quality of service (QoS) parameters for a QoS flow to be carried by the user plane tunnel.

[0071] A UE 3 for collection of the AI / ML data may be selected by the AMF 10-1 based on a subscription of the UE 3, or a user consent associated with the UE 3. The user consent may be specific to AI / ML data collection.

[0072] In step S702, the UE 3 is provided with configuration information for the AI / ML data collection and reporting, based on the information provided to the AMF 10-1 by the NWDAF / DCF 10-7 in step S701. The configuration information may include any of the information described above with reference to step S505 of Fig. 5, for example.

[0073] In optional step S703, the AMF 10-1 establishes a user plane tunnel (e.g., a user plane tunnel requested by the NWDAF / DCF 10-7 in step S701) for transmission of the AI / ML data by the UE 3. The AMF 10-1 may transmit a request to the RAN node 5 to establish the tunnel. The tunnel may be, for example, a GTP-U tunnel having a NWDAF / DCF GTP-U Tunnel ID provided to the AMF 10-1 by the NWDAF / DCF 10-7 in step S701. As with the example described above with reference to Fig. 6, the user plane tunnel may be established between the UE 3 and the network before the corresponding AI / ML data is available for transmission by the UE 3, and the UE 3 may report the AI / ML data to the network using the tunnel according to a reporting configuration received in step S702.

[0074] Fig. 8 illustrates a method in which an AI / ML data report is triggered by the core network. In this example, in step S801, the UE 3 is initially in an idle state (e.g. in an RRC idle state or RRC inactive state) rather than being connected to the network (e.g. in an RRC connected state).

[0075] After the UE 3 collects AI / ML data, or whilst the UE 3 collects the AI / ML data, (e.g. according to a configuration received via the RAN node 5, in any of the examples described above), the UE 3 may be in the RRC idle or RRC inactive mode. Since the UE 3 may have collected a large amount of AI / ML data to report, it is advantageous for the AI / ML data reporting to be performed under the control of the core network. For example, the reporting of the AI / ML data may be controlled taking into account a load on the system, one or more network policies, or any other suitable criterion.

[0076] In step S802, the NWDAF / DCF 10-7 transmits, to the AMF 10-1, a request for the AMF 10-1 to page the UE 3 to report the collected AI / ML data for a particular AI / ML model, for training of the AI / ML model. The information transmitted from the NWDAF / DCF 10-7 to the AMF 10-1 may comprise an identity of the UE 3 to be paged. The information transmitted from the NWDAF / DCF 10-7 to the AMF 10-1 may comprise an identity of the AI / ML model for which AI / ML data is to be transmitted by the UE.

[0077] In step S803, the AMF 10-1 transmits, to the RAN node 5, are request for the RAN node 5 to page the UE 3 to report the collected AI / ML data for the AI / ML model. The information transmitted from the AMF 10-1 to the RAN node 5 may comprise an identity of the UE 3 to be paged. The information transmitted from the AMF 10-1 to the RAN node 5 may comprise an identity of the AI / ML model for which AI / ML data is to be transmitted by the UE 3. The information transmitted from the AMF 10-1 to the RAN node 5 may comprise a specific page cause (e.g., 'data report for AIML model training') to be used for paging the UE 3.

[0078] In step S804 the RAN node 5 pages the UE 3 to report the collected AI / ML data for the AI / ML model. The transmission of step S804 may comprise an indication of the identity of the UE 3, an indication of the identity of the AI / ML model, and / or the specific page cause. Alternatively, after receiving the request to page the UE 3 in step S803, the RAN node 5 may wait until the UE 3 returns to the RRC connected state without paging the UE 3, and then transmit the request for the UE 3 to report the AI / ML data.

[0079] In step S805 the UE 3 establishes a signalling connection with the network (e.g. with the AMF 10-1). The AMF 10-1 may transmit an indication to the NWDAF / DCF 10-7 that the UE 3 has established connection with the network.

[0080] In optional step S806, a user plane tunnel is established between the UE 3 and the NWDAF / DCF 10-7 for transmission of the AI / ML data in the AI / ML data report. The user plane tunnel may be established using the method described above with reference to step S606 of Fig. 6 or step S703 of Fig. 7, for example, with the NWDAF / DCF as the first termination entity. In step S807 the UE 3 transmits the AI / ML data to the NWDAF / DCF 10-7 via the user plane tunnel.

[0081] Fig. 9 illustrates a method in which an AI / ML data reporting procedure is triggered by the UE 3.

[0082] In step S901, the UE 3 is in the RRC idle state, and collects AI / ML data for training an AI / ML model.

[0083] In step S902, one or more reporting criteria are met, and the UE 3 therefore determines that the AI / ML data is to be reported. For example, the UE 3 may be configured to transmit an AI / ML data report after a predetermined time period or using a predetermined periodicity. Alternatively, for example, the criteria may be that the UE 3 is configured to transmit the AI / ML data report after the amount of AI / ML data obtained by the UE 3 (and stored at the UE 3) exceeds a threshold amount of data. In this example, the AI / ML data is to be reported to the NWDAF / DCF 10-7.

[0084] In step S903 the UE 3 transmits, to the AMF 10-1 via the RAN node 5, a service request for an AI / ML data report. Whilst in this example the service request is transmitted to the AMF 10-1, the service request could alternatively be transmitted to any other suitable entity in the core network. The service request indicates that the UE 3 is requesting the establishment of a packet data unit (PDU) session for the purpose of transmitting the AI / ML data in the AI / ML data collection report. The service request may also include an indication of the identity of the AI / ML model. After receiving the service request from the UE 3, the AMF 10-1 determines whether the PDU session is to be established. The AMF 10-1 may determine whether the PDU session is to be established based on, for example, a current load on the network, a network policy, or using any other suitable criteria. Alternatively, rather than transmitting the service request (or if the PDU session is not established), the UE 3 may simply continue to store the AI / ML data until the UE 3 enters the RRC connected state at a later time.

[0085] In step S904, the AMF 10-1 has determined that the PDU session is to be established and the UE 3 is therefore instructed, via the RAN node 5, to establish a signalling connection with the network.

[0086] In optional step S905, the AMF 10-1 transmits, to the NWDAF / DCF 10-7, an indication that a PDU session is to be established for transmission of the AI / ML data. The AMF 10-1 may include an indication of the identity of the UE 3 corresponding to the collected AI / ML data. The AMF 10-1 may also include an indication of the identity of the AI / ML model to which the AI / ML data corresponds.

[0087] In step S906, the PDU session and a user plane tunnel between the UE 3 and the NWDAF / DCF 10-7 via UPF is established, for transmission of the AI / ML data report from the UE 3 to the NWDAF / DCF 10-7, with the DCF-NWDAF 10-7 as the first termination entity. In step S907, the UE 3 transmits the AI / ML data report to the NWDAF / DCF 10-7 using the established user plane tunnel.

[0088] Figure 10 illustrates a method in which an OAM 14 transmits a request for the RAN node 5 to configure a data type for data collection for an AI / ML model.

[0089] Is this example, in step S1001 the UE 3 is connected to the network, for communication via the RAN node 5 (e.g., in the 'RRC connected' state). The UE 3 may provide an indication to the network that UE 3 supports a particular AI / ML model.

[0090] In step S1002, the UE 3 transmits, to the OAM 14 via the RAN node 5, a request for data collection for training of a UE-side AI / ML model. The information provided by the UE 3 to the OAM 14 in step S1001 may include an identity of the AI / ML model (e.g., an identification number), an identity of the UE 3, an indication of model characteristics of the AI / ML model (e.g., the functionality supported by the AI / ML model), and / or a request that the AI / ML model be trained. The transmission of step S1002 is sent between the OAM client at the UE 3 and the OAM 14 (e.g., an OAM server), and may be transmitted using the Uu interface and an interface between the RAN node 5 and the OAM 14. Over the Uu interface, the information transmitted in step S1001 may be transmitted using an RRC message, and may be transmitted (or 'carried') using a transparent container. The information transmitted by the UE 3 in step S1002 may also include an indication of a data type to be collected, for training the AI / ML model at the network side.

[0091] In step S1003, the OAM 14 transmits, to the RAN node 5, a request for the RAN node 5 to configure the data type for data collection for the AI / ML model training, and in step S1004 the RAN node 5 transmits configuration information for use by the UE 3 to collect the AI / ML data. For example, for beam management, the RAN node 5 may indicate to the UE 3 that UE L1 measurements (e.g., for the top N beams) are to be collected. Alternatively, for example, for RLF prediction the RAN node 5 may indicate to the UE 3 that RLF events and L3 measurements of a serving cell and one or more neighbouring cells are to be collected. More generally, the RAN node 5 transmits configuration information to the UE 3 so that the UE 3 becomes configured to collect suitable AI / ML data for training the AI / ML model. The information transmitted to the UE 3 in step S1004 may comprise a start / stop indication, indicating when the UE 3 is to start obtaining the requested AI / ML data and when the UE 3 is to stop obtaining the requested AI / ML data. The information transmitted to the UE 3 in step S1004 may comprise a time duration, indicating a time duration for which the UE 3 is to obtain the requested AI / ML data. The information transmitted to the UE 3 in step S1004 may comprise a data collection time granularity, indicating a time granularity to use for collecting the AI / ML data. The information transmitted to the UE 3 in step S1004 may comprise a periodicity, indicating a periodicity at which the requested AI / ML data is to be obtained (e.g., a measurement periodicity). The information transmitted to the UE 3 in step S1004 may comprise an event, indicating an event for triggering the UE 3 to perform a measurement to obtain the requested AI / ML data (e.g, for RLF prediction, the event may be an RLF occurrence). The information transmitted to the UE 3 in step S1004 may comprise a reporting periodicity, indicating a periodicity at which the requested AI / ML data is to be reported (e.g., transmitted to the network in an AI / ML measurement report). The information transmitted to the UE 3 in step S1004 may comprise one or more reporting criteria, indicating a criteria or trigger event for triggering reporting of the AI / ML data. The reporting criteria may comprise, for example, a condition that the UE 3 is to transmit the AI / ML measurement report if the amount of AI / ML data stored at the UE 3 exceeds a threshold amount of data.

[0092] Fig. 11 illustrates a further method in which the OAM 14 transmits a request for the RAN node 5 to configure a data type for data collection for an AI / ML model. In contrast to the method of Fig. 10, in this example the UE 3 begins in the RRC idle or RRC inactive state in step S1101, rather than the RRC connected state.

[0093] After the UE 3 collects AI / ML data, or whilst the UE 3 collects the AI / ML data, (e.g. according to a configuration received via the RAN node 5, in any of the examples described above), the UE 3 may be in the RRC idle or RRC inactive mode. Since the UE 3 may have collected a large amount of AI / ML data to report, it is advantageous for the AI / ML data reporting to be performed under the control of the core network. For example, the reporting of the AI / ML data may be controlled taking into account a load on the system (e.g, a real-time system load), one or more network policies, or any other suitable criterion.

[0094] In step S1102, the OAM 14 transmits, to the RAN node 5, a request for the AI / ML data to be collected from a particular UE 3 and for a particular AI / ML model. The information transmitted in step S1002 may comprise an identity of the UE 3 and an identity of the AI / ML model.

[0095] In step S1103, the RAN node 5 pages the UE 3 to report the collected AI / ML data for the AI / ML model. The information transmitted in step S1103 may comprise an indication of the identity of the UE 3, an indication of the identity of the AI / ML model, and / or a specific page cause (e.g., 'data report for AIML model training').

[0096] In step S1104, the UE 3 establishes a signalling connection with the RAN node 5. In step S1105, in this example the RAN node 5 reconfigures the UE 3 to establish a specific RAN Data Radio Bearer (DRB) for AI / ML data reporting. The parameters of the QoS (e.g., priority) for the RAN DRB may be preconfigured at the RAN node 5 (e.g., provided by the OAM 14), or predefined at the RAN node 5 for the specific type of AI / ML data transfer, for training the AI / ML model. Alternatively, the RAN node 5 may simply determine the QoS parameters (e.g. independently). The QoS parameters may comprise data priority, packet delay budget, reliability requirements, or any other suitable QoS parameters. The DRB may be a new type of radio bearer since it is an end-to-end radio bearer between the UE 3 and the RAN node 5 that need not involve the core network. In comparison with legacy signalling radio bearers (SRB) and legacy DRBs, the priority of the radio bearer for transmitting the AI / ML data may be lower (but need not necessarily be lower). The priority and other parameters of such radio bearer for transmission of the AI / ML data may be configured by the RAN node 5 using an RRC message (e.g., 'RRCReconfiguration') transmitted from the RAN node 5 to the UE 3.

[0097] In step S1106, the UE 3 transmits AI / ML data report to the OAM 14 via the RAN node 5. The data is first transmitted from the UE 3 to the RAN node 5 using the RAN DRB established in step S1105, with the RAN node 5 as the first termination entity. The RAN node 5 then forwards the AI / ML data report to the OAM 14 (e.g., an OAM server).

[0098] In step S1203, the UE 3 transmits an RRC connection request to the RAN node 5, for the purpose of transmitting the AI / ML data report. The RRC connection request (or RRC connection resume request) may include an indication that the request is for transmitting the AI / ML data. The RRC connection request may optionally include an indication of the identity of the corresponding AI / ML model (e.g., and identification number of the AI / ML model). The UE 3 may include an establishment cause in the RRC connection request (or RRC connection resume request) that indicates that the connection requestion is for the transmission of AI / ML data. Advantageously, the RAN node 5 is therefore able to determine that the RRC connection request is for the transmission of AI / ML data, and can determine whether to allow the data transfer, for example based on the current load on the network.

[0099] In this example the RAN node 5 determines to allow the AI / ML data transmission, and therefore in step S1204 the RRC connection is established.

[0100] In step S1205, a RAN DRB is established for transmission of the AI / ML data report. The RAN node 5 may reconfigure the UE 3 to establish a specific RAN DRB for AI / ML data reports. The parameters of the QoS (e.g., priority) for the RAN DRB may preconfigured at the RAN node 5 (e.g., provided by the OAM 14), or predefined at the RAN node 5 for the specific type of AI / ML data transfer, for training the AI / ML model. Alternatively, the RAN node 5 may simply determine the QoS parameters (e.g. independently). The QoS parameters may comprise data priority, packet delay budget, reliability requirements, or any other suitable QoS parameters. The DRB may be a new type of radio bearer since it is an end-to-end radio bearer between the UE 3 and the RAN node 5 that need not involve the core network. In comparison with legacy signalling radio bearers (SRB) and legacy DRBs, the priority of the radio bearer for transmitting the AI / ML data may be lower (but need not necessarily be lower). Advantageously, the priority and other parameters of the radio bearer for transmission of the AI / ML data may be configured by the RAN node 5 using an RRC message (e.g., 'RRCReconfiguration') transmitted from the RAN node 5 to the UE 3.

[0101] In step S1206, the UE 3 transmits AI / ML data report to the OAM 14 via the RAN node 5. The data is first transmitted from the UE 3 to the RAN node 5 using the RAN DRB established in step S1205, with the RAN node 5 as the first termination entity. The RAN node 5 then forwards the AI / ML data report to the OAM 14 (e.g., OAM server).

[0102] < Dataset Transfer and Model Parameters Transfer> 　　As described above, the UE 3 may use an AI / ML model as part of an encoding process for encoding (and / or compressing) channel state information (CSI) for transmission to the RAN node 5, and the RAN node 5 may use the same model as part of a corresponding decoding (and / or decompression) process for decoding the CSI received from the UE 3. It will be appreciated that various other types of 'two-sided' model in which one AI / ML model hosted at one node (e.g., the UE 3), and a corresponding AI / ML model hosted at another node (e.g., a RAN node 5) may also be provided. For such two-sided models, there can be a need to share model parameters, or a dataset for the model, between the two nodes that host the AI / ML models. For example, for the case of an AI / ML model for CSI encoding, there can be a need to provide model parameters to the UE 3. The UE 3 may receive the model parameters or a dataset from the network, and the model parameters or dataset may then be subject to UE-side offline engineering (e.g., using a UE-side OTT server), for example for training or re-training the model at the UE 3. The dataset transmitted to the UE 3 may comprise, for example, a training data set of sample data comprising a number of target CSI, and corresponding CSI feedback, in the form of '{target CSI, CSI feedback}' pairs. Alternatively, parameters transmitted to the UE 3 may comprise parameters corresponding to a nominal CSI encoder. In a further alternative, both the data comprising t]he {target CSI, CSI feedback} pairs and the parameters corresponding to a nominal CSI encoder may be provided to the UE 3. However, the data comprising the {target CSI, CSI feedback} pairs typically has a large data size (e.g., 200 to 300 MB), and so transmission of such a large data set in the network should be considered. In contrast, for most cases the RRC messages sent from the RAN node 5 to the UE 3 are typically less than 9000 bytes. Extremely large RRC reconfiguration messages (e.g., 'RRCReconfiguration') or RRC resume messages (e.g., 'RRCResume') may require downlink RRC message segmentation in order to fit into a single PDCP service data unit (SDU). Whilst the data comprising the {target CSI, CSI feedback} pairs (or another suitable dataset) may be transmitted to the UE 3 via a user plane data transmission tunnel, traditionally the user plane data tunnel is established between the UE 3 and the UPF 11. Therefore, there is a need for improved apparatus and methods for transmitting data sets and model parameters for AI / ML models between entities in the network, in particular for two-sided AI / ML models.

[0103] Fig. 13 illustrates a method in which a GTP-U tunnel is established for transmission of an AI / ML dataset and / or AI / ML model parameters. The method of Fig. 13 is an example of a core network-based AI / ML dataset and / or AI / ML parameters transfer.

[0104] In this example, the core network is used to distribute the AI / ML dataset and / or AI / ML parameters from the network to the UE 3, for example for training an AI / ML model for CSI compression at the UE 3. In this example, in step S1301 the UE 3 is connected with the network (e.g., in the RRC connected state). The UE 3 may transmit, to the RAN node 5 (or to an entity in the core network, via the RAN node 5) an indication that the UE 3 supports AI / ML-based CSI compression functionality. The UE 3 may include an indication that the UE 3 supports a particular AI / ML model for CSI compression, for example by provided an indication of a model identity or an indication of a model structure, or any other suitable indication. The UE 3 may also provide, in step S1301, an indication of a need for training of the model based on a shared dataset, and / or a need for parameters for the AI / ML model. The information transmitted by the UE 3 in step S1301 may be transmitted using an RRC message, for example using a UE Assistance Information (UAI) message, or using any other suitable message. Beneficially, therefore, the RAN node 5 is able to determine an AI / ML dataset and / or AI / ML model parameters to be transmitted to the UE 3 for training of the AI / ML model.

[0105] In step S1302, a GTP-U tunnel is established between the RAN node 5 and the NWDAF / DCF 10-7. The RAN node 5 may transmit, to the AMF 10-1, a request to establish a GTP-U tunnel to transfer the AI / ML dataset and / or AI / ML model parameters to the NWDAF / DCF 10-7, for example for the purpose of CSI compression AI / ML training. In the request transmitted to the AMF 10-1 by the RAN node 5, the RAN node 5 may include an indication of a suggested RAN node-side GTP-U Tunnel ID, an indication of the UE ID(s), and / or an indication of the identity of the AI / ML model. After receiving the request from the RAN node 5, the AMF 10-1 transmits a request to the NWDAF / DCF 10-7 to establish a GTP Tunnel for the transfer of the AI / ML dataset and / or AI / ML model parameters. The request ] from the AMF 10-1 to the NWDAF / DCF 10-7 optionally includes a RAN node-side GTP-U tunnel ID, and indication of one or more UE identities, and an indication of the identity of the AI / ML model. After receiving the request from the AMF 10-1, the NWDAF / DCF 10-7 transmits the NWDAF / DCF-side GTP-U Tunnel ID to the AFM 10-1, and the AMF 10-1 forwards the NWDAF / DCF-side GTP-U Tunnel ID to the RAN node 5. The RAN node 5 then establishes the GTP-U tunnel with the NWDAF / DCF 10-7. In step S1303, the RAN node 5 transmits the AI / ML data set and / or AI / ML model parameters to the NWDAF / DCF 10-7 using the established GTP-U tunnel.

[0106] In step S1304, a request to provide a unicast service, broadcast service or multicast service for transmission of the AI / ML dataset and / or AI / ML model parameters to UEs 3 within the network is transmitted from the NWDAF / DCF 10-7 to the AMF 10-1.

[0107] In step S1305, if the GTP-U tunnel between the NWDAF / DCF 10-7 and the RAN node 5 is not available, the AMF 10-1 may command the establishment of a GTP-U tunnel between the NWDAF / DCF 10-7 and the RAN node 5 for the purpose of the transfer the AI / ML dataset and / or AI / ML model parameters for the one or more UEs 3. In addition, the AMF 10-1 may instruct the establishment of the unicast RB, broadcast RB or multicast RB between the UE 3 and the network. The establishment of the unicast RB, broadcast RB or multicast RB is performed, for transmission of the AI / ML dataset and / or AI / ML model parameters to the UE 3. As part of S1305, the AMF 10-1 may instruct the UEs 3 to establish a specific unicast PDU session and user plane tunnel with the NWDAF / DCF 10-7 in order to transfer the AI / ML dataset and / or AI / ML model parameters to the UEs 3 that support the corresponding AI / ML model (e.g. for CSI compression). Alternatively, a broadcast session or multicast session may be used to distribute the AI / ML dataset and / or AI / ML model parameters to all of the UEs 3, or to multiple UEs 3. Alternatively, during this process, the NWDAF / DCF 10-7 may transmit a request to the RAN node 5 for the RAN node 5 to provide the broadcast or multicast service for the AI / ML dataset and / or the AI / ML parameters. The RAN node 5 configures the UE 3 to establish the unicast DRB, broadcast RB or multicast RB, so that the UE 3 can receive the AI / ML dataset and / or the AI / ML parameters. The UE 3 then receives the AI / ML dataset and / or the AI / ML parameters transmitted via the unicast, broadcast or multicast.

[0108] In optional step S1306, the UE 3 may switch to RRC idle or RRC inactive mode, but advantageously in step S1307 the UE 3 continues to receive the AI / ML dataset and / or the AI / ML parameters even whilst in the RRC idle mode or RRC inactive mode. The established unicast DRB, broadcast RB or multicast RB is kept (or 'maintained') by the UE 3 during the RRC state transition. When the UE 3 is in the RRC idle mode or the RRC inactive mode and changes cells, the UE 3 may also continue to receive the AI / ML dataset and / or the AI / ML parameters without resuming the RRC connection if the RB configuration of the new cell can be acquired by the UE 3 after cell reselection. If the RB configuration is not provided in the new cell via system information or another common channel based transmission (e.g., multicast control channel (MCCH)), then the UE 3 may resume the RRC connection, to enter the RRC connected state, in order to obtain the RB configuration. The RAN node 5 may also be configured to provide, via system information broadcast in the cell or any other suitable common channel, a list of neighbour cells that can provide the same AI / ML dataset and / or AI / ML model parameters to the UE 3.

[0109] Fig. 14 illustrates a method of RAN node-based AI / ML dataset and / or AI / ML model parameter transfer.

[0110] In this example, the RAN node 5 is used to distribute the AI / ML dataset and / or AI / ML model parameters from the network to the UE 3 (e.g. for CSI compression).

[0111] In step S1401, the UE 3 is in the RRC connected state. The UE 3 may transmit, to the RAN node 5 (e.g. using an RRC message, or any other suitable message), an indication that the UE 3 supports AI / ML-based CSI compression functionality, for example. The UE 3 may include an indication that the UE 3 supports a particular AI / ML model for CSI compression, for example by provided an indication of a model identity or an indication of a model structure, or any other suitable indication.

[0112] In step S1402, a unicast RB, broadcast RB or multicast RB is established, so that the UE 3 can receive the AI / ML dataset and / or AI / ML model parameters from the RAN node 5. The UE 3 may be instructed by the RAN node 5 to establish a specific RAN level radio bearer (e.g., similar to DRB) that is specific to AI / ML data transfer. The parameters of the QoS (e.g., priority) for the RAN DRB may be preconfigured at the RAN node (e.g., provided by an OAM 14), or predefined at the RAN node 5 for the specific type of AI / ML data transfer, for training the AI / ML model. Alternatively, the RAN node 5 may simply determine the QoS parameters (e.g. independently). The QoS parameters may comprise data priority, packet delay budget, reliability requirements, or any other suitable QoS parameters. The DRB may be a new type of radio bearer since it is an end-to-end radio bearer between the UE 3 and the RAN node 5 that need not involve the core network. The priority and other parameters of the radio bearer for transmission of the AI / ML dataset and / or AI / ML model parameters may be configured by the RAN node 5 using an RRC message (e.g., 'RRCReconfiguration') transmitted from the RAN node 5 to the UE 3. The RAN-level radio bearer may be established in the manner of a unicast specific to one UE 3, or in the manner of a multicast specific to multiple UEs 3, or in the manner of a broadcast applicable to all of the UEs within a cell. For the case of multicast or broadcast, a RAN broadcast RB or RAN multicast RB may be used to distribute the AI / ML dataset or AI / ML model parameters to all of the UEs in the cell or multiple UEs 3 that support the AI / ML model (e.g. for CSI compression). The UE(s) 3 then receive the AI / ML dataset and / or AI / ML model parameters.

[0113] In optional step S1403, the UE 3 may switch to RRC idle or RRC inactive mode, but advantageously in step S1404 the UE 3 continues to receive the AI / ML dataset and / or the AI / ML parameters even whilst in the RRC idle mode or RRC inactive mode. The established unicast DRB, broadcast RB or multicast RB is kept (or 'maintained') by the UE 3 during the RRC state transition. When the UE 3 is in the RRC idle mode or the RRC inactive mode and changes cells, the UE 3 may also continue to receive the AI / ML dataset and / or the AI / ML parameters without resuming the RRC connection if the RB configuration of the new cell can be acquired by the UE 3 after cell reselection. If the RB configuration is not provided in the new cell via system information or another common channel based transmission (e.g., multicast control channel (MCCH)), then the UE 3 may resume the RRC connection, to enter the RRC connected state, in order to obtain the RB configuration. The RAN node 5 may also be configured to provide, via system information broadcast in the cell or via any other suitable common channel, a list of neighbour cells that can provide the same AI / ML dataset and / or AI / ML model parameters to the UE 3. This means that the established RAN unicast RB, RAN broadcast RB or RAN multicast RB may be kept by the UE during / after RRC state transition, to continue to receive the AI / ML dataset and / or the AI / ML parameters.

[0114] < Devices of the Communication System> <User Equipment> 　　Fig. 15 is a schematic block diagram illustrating the main components of a UE 3 as shown in Fig. 1.

[0115] As shown, the UE 3 has a transceiver circuit 31 that is operable to transmit signals to and to receive signals from a RAN node 5 via one or more antennas 33 (e.g., comprising one or more antenna elements). The UE 3 has a controller 37 to control the operation of the UE 3. The controller 37 is associated with a memory 39 and is coupled to the transceiver circuit 31. Although not necessarily required for its operation, the UE 3 might, of course, have all the usual functionality of a conventional UE 3 (e.g., a user interface 35, such as a touch screen / keypad / microphone / speaker and / or the like for, allowing direct control by and interaction with a user) and this may be provided by any one or any combination of hardware, software, and firmware, as appropriate. Software may be pre-installed in the memory 39 and / or may be downloaded via the communication system 1 or from a removable data storage device (RMD), for example.

[0116] The controller 37 is configured to control overall operation of the UE 3 by, in this example, program instructions or software instructions stored within memory 39. As shown, these software instructions include, among other things, an operating system 41, and a communications control module 43.

[0117] The communication control module 43 is operable to control the communication between the UE 3 and its serving RAN node 5 (and other communication devices connected to the RAN node 5, such as further UEs 3 and / or core network nodes). The communication control module 43 is configured for the overall handling of uplink communications via associated uplink channels (e.g., via a physical uplink control channel (PUCCH), random access channel (RACH), and / or a physical uplink shared channel (PUSCH)) including both dynamic and semi-static signalling (e.g., SRS). The communication control module 43 is also configured for the overall handling of receipt of downlink communications via associated downlink channels (e.g., of DCI via a physical downlink control channel (PDCCH) and / or a physical downlink shared channel (PDSCH)) including both dynamic and semi-persistent scheduling (e.g., SPS). The communication control module 43 is responsible, for example: for determining where to monitor for downlink control information; for determining the resources to be used by the UE 3 for transmission / reception of UL / DL communications (including interleaved resources and resources subject to frequency hopping); for managing frequency hopping at the UE side; for determining how slots / symbols are configured (e.g., for UL, DL or full duplex communication, or the like); for determining which bandwidth parts are configured for the UE 3; for determining how uplink transmissions should be encoded and the like.

[0118] It will be appreciated that the communication control module 43 may include a number of sub-modules ('layers' or 'entities') to support specific functionalities. For example, the communication control module 43 may include a PHY sub-module, a MAC sub-module, an RLC sub-module, a PDCP sub-module, an RRC sub-module, etc.

[0119] The communication control module 43 is configured, in particular, to control the UE's communications, where applicable, in accordance with any of the methods described herein.

[0120] < RAN node> 　　Fig. 16 is a schematic block diagram illustrating the main components of a RAN node 5-1 for the communication system 1 shown in Fig. 1. As shown, the RAN node 5-1 has a transceiver circuit 51 for transmitting signals to and for receiving signals from the communication devices (such as UEs 3) via one or more antennas 53 (e.g., a single or multi-panel antenna array / massive antenna), and a core network interface 55 (e.g., comprising the N2, N3 and other reference points / interfaces) for transmitting signals to and for receiving signals from network nodes in the core network 7. Although not shown, the RAN node 5-1 may also be coupled to other RAN nodes 5 via an appropriate interface (e.g., the so-called 'Xn' interface in NR). The RAN node 5-1 has a controller 57 to control the operation of the RAN node 5-1. The controller 57 is associated with a memory 59. Software may be pre-installed in the memory 59 and / or may be downloaded via the communication system 1 or from a removable data storage device (RMD), for example. The controller 57 is configured to control the overall operation of the RAN node 5-1 by, in this example, program instructions or software instructions stored within memory 59.

[0121] As shown, these software instructions include, among other things, an operating system 61, and a communications control module 63.

[0122] The communications control module 63 is operable to control the communication between the RAN node 5-1 and UEs 3 and other network entities that are connected to the RAN node 5-1. The communications control module 63 is configured for the overall control of the reception and decoding of uplink communications, via associated uplink channels (e.g., via a physical uplink control channel (PUCCH), a random-access channel (RACH), and / or a physical uplink shared channel (PUSCH)) including both dynamic and semi-static signalling (e.g., SRS). The communications control module 63 is also configured for the overall handling the transmission of downlink communications via associated downlink channels (e.g., via a physical downlink control channel (PDCCH) and / or a physical downlink shared channel (PDSCH)) including both dynamic and semi-static signalling (e.g., channel state information reference signal (CSI-RS), synchronisation signal blocks (SSBs) etc.). The communications control module 63 is also responsible, for example, for determining and scheduling the resources to be used by the UE 3 for receiving in DL / transmitting in UL, for configuring slots / symbols appropriately (e.g., for UL, DL, flexible, full duplex communication, or the like), for configuring one or more bandwidth parts for the UE 3, and for providing related configuration signalling to the UE 3.

[0123] It will be appreciated that the communications control module 63 may include a number of sub-modules (or 'layers') to support specific functionalities. For example, the communications control module 63 may include a PHY sub-module, a MAC sub-module, an RLC sub-module, a PDCP sub-module, an SDAP sub-module, an IP sub-module, an RRC sub-module, etc.

[0124] The communication control module 63 is configured, in particular, to control the RAN Node's communications, where applicable, in accordance with any of the methods described herein.

[0125] <Modifications and Alternatives> 　　Detailed examples been described above. As those skilled in the art will appreciate, a number of modifications and alternatives can be made to the above examples whilst still benefiting from the concepts embodied therein.

[0126] It will be appreciated that description of features of and actions performed by a RAN node (base station), apply equally to distributed type base stations as to non-distributed type base stations.

[0127] It will also be appreciated that whilst information elements having specific names have been described differently named information elements but having a similar purpose may be used.

[0128] In the above description the UE and the base station are described for ease of understanding as having a number of discrete functional components or modules. Whilst these modules may be provided in this way for certain applications, for example where an existing system has been modified to implement the disclosed enhancements, in other applications, for example in systems designed with the inventive features in mind from the outset, these modules may be built into the overall operating system or code and so these modules may not be discernible as discrete entities.

[0129] In the above examples, a number of software modules were described. As those skilled in the art will appreciate, the software modules may be provided in compiled or un-compiled form and may be supplied to the UE or base station as a signal over a computer network, or on a recording medium. Further, the functionality performed by part, or all, of this software may be performed using one or more dedicated hardware circuits. However, the use of software modules is preferred as it facilitates the updating of the UE or the base station in order to update their functionalities.

[0130] Each controller may comprise any suitable form of processing circuitry including (but not limited to), for example: one or more hardware implemented computer processors; microprocessors; central processing units (CPUs); arithmetic logic units (ALUs); input / output (IO) circuits; internal memories / caches (program and / or data); processing registers; communication buses (e.g. control, data and / or address buses); direct memory access (DMA) functions; hardware or software implemented counters, pointers and / or timers; and / or the like. Various other modifications will be apparent to those skilled in the art and will not be described in further detail here.

[0131] The User Equipment (or "UE," "mobile station," "mobile device" or "wireless device") in the present disclosure is an entity connected to a network via a wireless interface.

[0132] It should be noted that the present disclosure is not limited to a dedicated communication device and can be applied to any device having a communication function as explained in the following paragraphs. The terms "User Equipment" or "UE" (as the term is used by 3GPP), "mobile station", "mobile device", and "wireless device" are generally intended to be synonymous with one another, and include standalone mobile stations, such as terminals, cell phones, smart phones, tablets, cellular IoT devices, IoT devices, and machinery. It will be appreciated that the terms "mobile station" and "mobile device" also encompass devices that remain stationary for an extended period of time.

[0133] A UE may, for example, be an item of equipment for production or manufacture and / or an item of energy related machinery (for example equipment or machinery such as: boilers; engines; turbines; solar panels; wind turbines; hydroelectric generators; thermal power generators; nuclear electricity generators; batteries; nuclear systems and / or associated equipment; heavy electrical machinery; pumps including vacuum pumps; compressors; fans; blowers; oil hydraulic equipment; pneumatic equipment; metal working machinery; manipulators; robots and / or their application systems; tools; moulds or dies; rolls; conveying equipment; elevating equipment; materials handling equipment; textile machinery; sewing machines; printing and / or related machinery; paper converting machinery; chemical machinery; mining and / or construction machinery and / or related equipment; machinery and / or implements for agriculture, forestry and / or fisheries; safety and / or environment preservation equipment; tractors; precision bearings; chains; gears; power transmission equipment; lubricating equipment; valves; pipe fittings; and / or application systems for any of the previously mentioned equipment or machinery etc.).

[0134] A UE may, for example, be an item of transport equipment (for example transport equipment such as: rolling stocks; motor vehicles; motorcycles; bicycles; trains; buses; carts; rickshaws; ships and other watercraft; aircraft; rockets; satellites; drones; balloons etc.).

[0135] A UE may, for example, be an item of information and communication equipment (for example information and communication equipment such as: electronic computer and related equipment; communication and related equipment; electronic components etc.).

[0136] A UE may, for example, be a refrigerating machine, a refrigerating machine applied product, an item of trade and / or service industry equipment, a vending machine, an automatic service machine, an office machine or equipment, a consumer electronic and electronic appliance (for example a consumer electronic appliance such as: audio equipment; video equipment; a loud speaker; a radio; a television; a microwave oven; a rice cooker; a coffee machine; a dishwasher; a washing machine; a dryer; an electronic fan or related appliance; a cleaner etc.).

[0137] A UE may, for example, be an electrical application system or equipment (for example an electrical application system or equipment such as: an x-ray system; a particle accelerator; radio isotope equipment; sonic equipment; electromagnetic application equipment; electronic power application equipment etc.).

[0138] A UE may, for example, be an electronic lamp, a luminaire, a measuring instrument, an analyser, a tester, or a surveying or sensing instrument (for example a surveying or sensing instrument such as: a smoke alarm; a human alarm sensor; a motion sensor; a wireless tag etc.), a watch or clock, a laboratory instrument, optical apparatus, medical equipment and / or system, a weapon, an item of cutlery, a hand tool, or the like.

[0139] A UE may, for example, be a wireless-equipped personal digital assistant or related equipment (such as a wireless card or module designed for attachment to or for insertion into another electronic device (for example a personal computer, electrical measuring machine)).

[0140] A UE may be a device or a part of a system that provides applications, services, and solutions described below, as to "internet of things (IoT)," using a variety of wired and / or wireless communication technologies.

[0141] Internet of Things devices (or "things") may be equipped with appropriate electronics, software, sensors, network connectivity, and / or the like, which enable these devices to collect and exchange data with each other and with other communication devices. IoT devices may comprise automated equipment that follow software instructions stored in an internal memory. IoT devices may operate without requiring human supervision or interaction. IoT devices might also remain stationary and / or inactive for an extended period of time. IoT devices may be implemented as a part of a (generally) stationary apparatus. IoT devices may also be embedded in non-stationary apparatus (e.g., vehicles) or attached to animals or persons to be monitored / tracked.

[0142] It will be appreciated that IoT technology can be implemented on any communication devices that can connect to a communication system for sending / receiving data, regardless of whether such communication devices are controlled by human input or software instructions stored in memory.

[0143] It will be appreciated that IoT devices are sometimes also referred to as Machine-Type Communication (MTC) devices or Machine-to-Machine (M2M) communication devices. It will be appreciated that a UE may support one or more IoT or MTC applications. Some examples of MTC applications are listed in the following table. This list is not exhaustive and is intended to be indicative of some examples of machine type communication applications.

[0144] Further, the above-described UE categories are merely examples of applications of the technical ideas and exemplary examples described in the present document. Needless to say, these technical ideas and examples are not limited to the above-described UE and various modifications can be made thereto.

[0145] Various other modifications will be apparent to those skilled in the art and will not be described in further detail here.

[0146] For example, the whole or part of the exemplary embodiments disclosed above can be described as, but not limited to, the following supplementary notes. 　　(Supplementary note 1) 　　A method performed by a User Equipment, UE, the method comprising: 　　receiving, from a Radio Access Network node, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and 　　performing at least one of a periodic logging or an event-triggered logging based on the configuration information. 　　(Supplementary note 2) 　　The method of supplementary note 1, wherein the configuration information is to log the measurements for at least either Layer 1 (L1) or Layer 3 (L3). 　　(Supplementary note 3) 　　The method of supplementary note 1 or 2, wherein the measurements are for beam management. 　　(Supplementary note 4) 　　The method of any one of supplementary notes 1-3, further comprising: 　　sending a first message to the RAN node when data related to the AI / ML stored in the UE reaches a threshold. 　　(Supplementary note 5) 　　The method of any one of supplementary notes 1-4, further comprising: 　　transmitting, to the RAN node, a second message to request data collection for UE-side AI / ML model training. 　　(Supplementary note 6) 　　The method of supplementary note 5, wherein the second message is further sent to an operations and maintenance, OAM. 　　(Supplementary note 7) 　　The method of any one of supplementary notes 1-6, wherein the configuration information is based on a third message in which an operations and maintenance, OAM, requests the RAN node configure data type for data collection. 　　(Supplementary note 8) 　　The method of supplementary note 6 or 7, wherein the second message comprises a first indication of a data type to be collected. 　　(Supplementary note 9) 　　The method of any one of supplementary notes 1-8, wherein the configuration information indicates that Radio Link Failure, RLF, events and Layer 3 (L3) measurements of one or more cells are to be collected. 　　(Supplementary note 10) 　　The method of any one of supplementary notes 1-9, wherein the configuration information comprises a start / stop indication, wherein the start / stop indication indicates when the UE is to start and / or stop to log the measurements. 　　(Supplementary note 11) 　　The method of any one of supplementary notes 1-10, wherein the configuration information indicates a time granularity for logging the measurements. 　　(Supplementary note 12) 　　The method of any one of supplementary notes 1-11, wherein the configuration information indicates an event for triggering the UE to log the measurements. 　　(Supplementary note 13) 　　The method of any one of supplementary notes 1-12, further comprising: receiving a paging information for reporting the measurements that the UE logged. 　　(Supplementary note 14) 　　The method of any one of supplementary notes 1-13, further comprising: establishing a specific RAN Data Radio Bearer, DRB, for AI / ML data reporting. 　　(Supplementary note 15) 　　The method of any one of supplementary notes 1-14, further comprising: collecting AI / ML data while the UE is in a Radio Resource Control, RRC, idle state. 　　(Supplementary note 16) 　　A method performed by a User Equipment, UE, the method comprising: 　　transmitting, to a Radio Access Network node, RAN, node, a first indication indicating that the UE supports a functionality of Channel State Information, CSI, compression based on an Artificial Intelligence / Machine Learning, AI / ML, dataset and / or an AI / ML model; 　　establishing a unicast Radio Bearer, RB, broadcast RB, or multicast RB; and receiving parameters for the AI / ML dataset and / or AI / ML model from the RAN node 5. 　　(Supplementary note 17) 　　The method of supplementary note 16, wherein the unicast RB, the broadcast RB, or the multicast RB is a RAN RB that is specific to AI / ML data transfer. 　　(Supplementary note 18) 　　The method of supplementary note 17, wherein parameters of a Quality of Service, QoS, for the RAN RB are: preconfigured at the RAN node, predefined at the RAN node for a specific type of transfer of data for the AI / ML dataset and / or AI / ML model, or determined by the RAN node. 　　(Supplementary note 19) 　　The method of any one of supplementary notes 16-18, wherein priority of the the unicast RB, the broadcast RB, or the multicast RB is configured by the RAN node using a Radio Resource Control, RRC, message transmitted from the RAN node to the UE. 　　(Supplementary note 20) 　　The method of any one of supplementary notes 17, wherein the RAN RB is established by unicast, multicast, or broadcast. 　　(Supplementary note 21) 　　The method of any one of supplementary notes 16-20, further comprising: switching to a Radio Resource Control, RRC idle mode or an RRC inactive mode; and receiving the parameters for the AI / ML dataset and / or the AI / ML model whilst in the RRC idle mode or the RRC inactive mode. 　　(Supplementary note 22) 　　The method of supplementary note 21, further comprising: changing cells; in a case where an RB configuration of a new cell can be acquired by the UE after cell reselection, receiving the parameters for the AI / ML dataset and / or the AI / ML without resuming an RRC connection; and in a case where RB configuration is not provided in the new cell, resuming the RRC connection. 　　(Supplementary note 23) 　　The method of any one of supplementary notes 16-22, further comprising: receiving via system information, from the RAN node, a list of neighbour cells that can provide parameters for the AI / ML dataset and / or AI / ML model. 　　(Supplementary note 24) 　　The method of any one of supplementary notes 16-23, further comprising: keeping the unicast RB, the broadcast RB, or the multicast RB during / after RRC state transition. 　　(Supplementary note 25) 　　A method performed by a Radio Access Network, RAN, node, the method comprising: 　　transmitting, to a User Equipment, UE, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and 　　receiving, from the UE, logged information of the measurement performed based on the configuration information, wherein the logged information is based on at least one of a periodic logging or an event-triggered logging.

[0147] This application is based upon and claims the benefit of priority from Great Britain Patent Application No. 2501798.9, filed on February 6, 2025, the disclosure of which is incorporated herein in its entirety by reference.

[0148] 1 COMMUNICATION SYSTEM 3 USER EQUIPMENT 5 RAN NODE 7 CORE NETWORK 10 CONTROL PLANE FUNCTIONS 10-1 AMF 10-2 SMF 10-4 UDM 10-5 NEF 10-6 AF 10-7 NWDAF / DCF 11 USER PLANE FUNCTIONS 14 OPERATIONS AND MAINTENANCE 20 EXTERNAL DATA NETWORK 241 DATA COLLECTION 243 MODEL TRAINING 245 INFERENCE 247 ACTOR 249 MANAGEMENT 251 MODEL STRAGE 31 TRANSCEIVER CIRCUIT 33 ANTENNA 35 USER INTERFACE 37 CONTROLLER 39 MEMORY 41 OPERATING SYSTEM 43 COMMUNICATIONS CONTROL MODULE 51 TRANSCEIVER CIRCUIT 53 ANTENNA 55 CORE NETWORK INTERFACE 57 CONTROLLER 59 MEMORY 61 OPERATING SYSTEM 63 COMMUNICATIONS CONTROL MODULE

Claims

1. A method performed by a User Equipment, UE, the method comprising: 　　receiving, from a Radio Access Network node, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and 　　performing at least one of a periodic logging or an event-triggered logging based on the configuration information.

2. The method of claim 1, wherein the configuration information is to log the measurements for at least either Layer 1 (L1) or Layer 3 (L3).

3. The method of claim 1 or 2, wherein the measurements are for beam management.

4. The method of any one of claims 1-3, further comprising: 　　sending a first message to the RAN node when data related to the AI / ML stored in the UE reaches a threshold.

5. The method of any one of claims 1-4, further comprising: 　　transmitting, to the RAN node, a second message to request data collection for UE-side AI / ML model training.

6. The method of claim 5, wherein the second message is further sent to an operations and maintenance, OAM.

7. The method of any one of claims 1-6, wherein the configuration information is based on a third message in which an operations and maintenance, OAM, requests the RAN node configure data type for data collection.

8. The method of claim 6 or 7, wherein the second message comprises a first indication of a data type to be collected.

9. The method of any one of claims 1-8, wherein the configuration information indicates that Radio Link Failure, RLF, events and Layer 3 (L3) measurements of one or more cells are to be collected.

10. The method of any one of claims 1-9, wherein the configuration information comprises a start / stop indication, wherein the start / stop indication indicates when the UE is to start and / or stop to log the measurements.

11. The method of any one of claims 1-10, wherein the configuration information indicates a time granularity for logging the measurements.

12. The method of any one of claims 1-11, wherein the configuration information indicates an event for triggering the UE to log the measurements.

13. The method of any one of claims 1-12, further comprising: receiving a paging information for reporting the measurements that the UE logged.

14. The method of any one of claims 1-13, further comprising: establishing a specific RAN Data Radio Bearer, DRB, for AI / ML data reporting.

15. The method of any one of claims 1-14, further comprising: collecting AI / ML data while the UE is in a Radio Resource Control, RRC, idle state.

16. A method performed by a User Equipment, UE, the method comprising: 　　transmitting, to a Radio Access Network node, RAN, node, a first indication indicating that the UE supports a functionality of Channel State Information, CSI, compression based on an Artificial Intelligence / Machine Learning, AI / ML, dataset and / or an AI / ML model; 　　establishing a unicast Radio Bearer, RB, broadcast RB, or multicast RB; and receiving parameters for the AI / ML dataset and / or AI / ML model from the RAN node 5.

17. The method of claim 16, wherein the unicast RB, the broadcast RB, or the multicast RB is a RAN RB that is specific to AI / ML data transfer.

18. The method of claim 17, wherein parameters of a Quality of Service, QoS, for the RAN RB are: preconfigured at the RAN node, predefined at the RAN node for a specific type of transfer of data for the AI / ML dataset and / or AI / ML model, or determined by the RAN node.

19. The method of any one of claims 16-18, wherein priority of the the unicast RB, the broadcast RB, or the multicast RB is configured by the RAN node using a Radio Resource Control, RRC, message transmitted from the RAN node to the UE.

20. The method of any one of claims 17, wherein the RAN RB is established by unicast, multicast, or broadcast.

21. The method of any one of claims 16-20, further comprising: switching to a Radio Resource Control, RRC idle mode or an RRC inactive mode; and receiving the parameters for the AI / ML dataset and / or the AI / ML model whilst in the RRC idle mode or the RRC inactive mode.

22. The method of claim 21, further comprising: changing cells; in a case where an RB configuration of a new cell can be acquired by the UE after cell reselection, receiving the parameters for the AI / ML dataset and / or the AI / ML without resuming an RRC connection; and in a case where RB configuration is not provided in the new cell, resuming the RRC connection.

23. The method of any one of claims 16-22, further comprising: receiving via system information, from the RAN node, a list of neighbour cells that can provide parameters for the AI / ML dataset and / or AI / ML model.

24. The method of any one of claims 16-23, further comprising: keeping the unicast RB, the broadcast RB, or the multicast RB during / after RRC state transition.

25. A method performed by a Radio Access Network, RAN, node, the method comprising: 　　transmitting, to a User Equipment, UE, configuration information to log measurements related to Artificial Intelligence / Machine Learning, AI / ML; and 　　receiving, from the UE, logged information of the measurement performed based on the configuration information, wherein the logged information is based on at least one of a periodic logging or an event-triggered logging.

Images