Method and apparatus for ai / ml data collection and reporting in a wireless communication system

Abstract

The disclosure relates to a 5G or 6G communication system for supporting a higher data transmission rate. There is disclosed a method, for a UE, for AI / ML data collection in a network. The method comprises: receiving, from the network, a data collection request message, wherein the data collection request message comprises a data collection assistance information; in response to receiving the data collection request message, performing data collection based on the data collection assistance information; and transmitting, to the network, a data collection report comprising collected data and / or information related to the data collection. There is also disclosed a method, for a base station, for AI / ML data collection in a network. The method comprises: transmitting, to a UE, a data collection request message, wherein the data collection request message comprises a data collection assistance information; and receiving, from the UE, a data collection report comprising data collected by the UE based on the data collection assistance information and / or information related to the data collection.

Description

METHOD AND APPARATUS FOR AI / ML DATA COLLECTION AND REPORTING IN A WIRELESS COMMUNICATION SYSTEM

[0001] Certain examples of the present disclosure provide one or more techniques for Artificial Intelligence (AI) / Machine Learning (ML) data collection and / or reporting in a network, for example a 3rdGeneration Partnership Project (3GPP) 5thGeneration (5G) New Radio (NR) network.

[0002] 5G mobile communication technologies define broad frequency bands such that high transmission rates and new services are possible, and can be implemented not only in "Sub 6GHz" bands such as 3.5GHz, but also in "Above 6GHz" bands referred to as mmWave including 28GHz and 39GHz. In addition, it has been considered to implement 6G mobile communication technologies (referred to as Beyond 5G systems) in terahertz (THz) bands (for example, 95GHz to 3THz bands) in order to accomplish transmission rates fifty times faster than 5G mobile communication technologies and ultra-low latencies one-tenth of 5G mobile communication technologies.

[0003] At the beginning of the development of 5G mobile communication technologies, in order to support services and to satisfy performance requirements in connection with enhanced Mobile BroadBand (eMBB), Ultra Reliable Low Latency Communications (URLLC), and massive Machine-Type Communications (mMTC), there has been ongoing standardization regarding beamforming and massive MIMO for mitigating radio-wave path loss and increasing radio-wave transmission distances in mmWave, supporting numerologies (for example, operating multiple subcarrier spacings) for efficiently utilizing mmWave resources and dynamic operation of slot formats, initial access technologies for supporting multi-beam transmission and broadbands, definition and operation of BWP (BandWidth Part), new channel coding methods such as a LDPC (Low Density Parity Check) code for large amount of data transmission and a polar code for highly reliable transmission of control information, L2 pre-processing, and network slicing for providing a dedicated network specialized to a specific service.

[0004] Currently, there are ongoing discussions regarding improvement and performance enhancement of initial 5G mobile communication technologies in view of services to be supported by 5G mobile communication technologies, and there has been physical layer standardization regarding technologies such as V2X (Vehicle-to-everything) for aiding driving determination by autonomous vehicles based on information regarding positions and states of vehicles transmitted by the vehicles and for enhancing user convenience, NR-U (New Radio Unlicensed) aimed at system operations conforming to various regulation-related requirements in unlicensed bands, NR UE Power Saving, Non-Terrestrial Network (NTN) which is UE-satellite direct communication for providing coverage in an area in which communication with terrestrial networks is unavailable, and positioning.

[0005] Moreover, there has been ongoing standardization in air interface architecture / protocol regarding technologies such as Industrial Internet of Things (IIoT) for supporting new services through interworking and convergence with other industries, IAB (Integrated Access and Backhaul) for providing a node for network service area expansion by supporting a wireless backhaul link and an access link in an integrated manner, mobility enhancement including conditional handover and DAPS (Dual Active Protocol Stack) handover, and two-step random access for simplifying random access procedures (2-step RACH for NR). There also has been ongoing standardization in system architecture / service regarding a 5G baseline architecture (for example, service based architecture or service based interface) for combining Network Functions Virtualization (NFV) and Software-Defined Networking (SDN) technologies, and Mobile Edge Computing (MEC) for receiving services based on UE positions.

[0006] As 5G mobile communication systems are commercialized, connected devices that have been exponentially increasing will be connected to communication networks, and it is accordingly expected that enhanced functions and performances of 5G mobile communication systems and integrated operations of connected devices will be necessary. To this end, new research is scheduled in connection with eXtended Reality (XR) for efficiently supporting AR (Augmented Reality), VR (Virtual Reality), MR (Mixed Reality) and the like, 5G performance improvement and complexity reduction by utilizing Artificial Intelligence (AI) and Machine Learning (ML), AI service support, metaverse service support, and drone communication.

[0007] Furthermore, such development of 5G mobile communication systems will serve as a basis for developing not only new waveforms for providing coverage in terahertz bands of 6G mobile communication technologies, multi-antenna transmission technologies such as Full Dimensional MIMO (FD-MIMO), array antennas and large-scale antennas, metamaterial-based lenses and antennas for improving coverage of terahertz band signals, high-dimensional space multiplexing technology using OAM (Orbital Angular Momentum), and RIS (Reconfigurable Intelligent Surface), but also full-duplex technology for increasing frequency efficiency of 6G mobile communication technologies and improving system networks, AI-based communication technology for implementing system optimization by utilizing satellites and AI (Artificial Intelligence) from the design stage and internalizing end-to-end AI support functions, and next-generation distributed computing technology for implementing services at levels of complexity exceeding the limit of UE operation capability by utilizing ultra-high-performance communication and computing resources

[0008] The present disclosure relates to controlling a plurality of reference signal ports.

[0009] It is an aim of certain examples of the present disclosure to address, solve and / or mitigate, at least partly, at least one of the problems and / or disadvantages associated with the related art, for example at least one of the problems and / or disadvantages described herein. It is an aim of certain examples of the present disclosure to provide at least one advantage over the related art, for example at least one of the advantages described herein.

[0010] The present invention is defined in the independent claims. Advantageous features are defined in the dependent claims. Embodiments or examples disclosed in the description and / or figures falling outside the scope of the claims are to be understood as examples useful for understanding the present invention.

[0011] Other aspects, advantages and salient features of the invention will become apparent to those skilled in the art from the following detailed description taken in conjunction with the accompanying drawings.

[0012] Aspects of the present disclosure provide efficient communication methods in a wireless communication system.

[0013] FIGURE 1 illustrates an example of different stages of data preparation for AI / ML model according to various embodiments of the present disclosure;

[0014] FIGURE 2 illustrates an example of a data collection and data reporting procedure taking into account assistance information in the data collection template according to various embodiments of the present disclosure;

[0015] FIGURE 3 illustrates a flow chart, for a UE, for AI / ML data collection in a network according to various embodiments of the present disclosure;

[0016] FIGURE 4 illustrates a flow chart, for a base station, for AI / ML data collection in a network according to various embodiments of the present disclosure;

[0017] FIGURE 5 illustrates a block diagram of an exemplary network entity according to various embodiments of the present disclosure

[0018] FIGURE 6 illustrates a block diagram illustrating a structure of a UE according to various embodiments of the present disclosure; and

[0019] FIGURE 7 illustrates a block diagram illustrating a structure of a base station according to various embodiments of the present disclosure, as disclosed herein.

[0020] Herein, the following document may be referenced:

[0021] [1] 3GPP RP-213599, Study on Artificial Intelligence (AI) / Machine Learning (ML) for NR Air Interface.

[0022] [2] 3GPP TS 38.413, Technical Specification Group Radio Access Network; NG-RAN; NG Application Protocol (NGAP).

[0023] [3] 3GPP TS 38.423, NG-RAN; Xn Application Protocol (XnAP).

[0024] [4] https: / cloud.google.com / architecture / mlops-continuous-delivery-and-automation-pipelines-in-machine-learning

[0025] Various acronyms, abbreviations and definitions used in the present disclosure are defined at the end of this description.

[0026] FIGURE 1 illustrates an example of different stages of data preparation for AI / ML model according to various embodiments of the present disclosure.

[0027] Overview of AI / ML Operations

[0028] Machine Learning Operations (MLOps) is a system of processes for the end-to-end AI / ML lifecycle management (LCM) at scale. These processes ensure that the AI / ML model can be scaled for a large user base and achieve accurate performance. The MLOps processes can be split into three categories [4].

[0029] -Data Preparation:

[0030] This category involvesdata collection, data reporting, data analysis, data validation, feature engineering and data splitting. Using data collection to collect data from a variety of sources, data analysis to analysing data to identify patterns and relationships, data validation is the process of verifying that the data is accurate and consistent. Through feature engineering, new features can be developed from existing data and this can enhance the model's accuracy. Using data splitting to separate the data into train, test, and validation sets to ensure the model is correctly trained.

[0031] -Model Development:

[0032] This category involvesmodel configuration, model training and model validation. The data scientist implements different algorithms with the prepared data to train various ML models. In addition, you subject the implemented algorithms to hyper-parameter tuning to obtain the best performing ML model. The output of this step is a trained model. Next step to validate model to be deployed in the production environment.

[0033] -Rollout:

[0034] This category involvesmodel deploying, model serving, model monitoring　and model re-training. The validated model is deployed to a production environment to serve predictions. The model deployment can be REST API micro service to serve online predictions or an embedded model to an edge or mobile device. The model predictive performance is monitored to potentially invoke a new iteration in the ML process.

[0035] Data is the backbone of AI / ML algorithms. Therefore, high-quality data is needed for any AI / ML project, a low-quality dataset never produces a high-quality AI model. Quality Data is described as the filtered, cleansed and contextualized form of massive dataset. Modern AI / ML models have become data-hungry and computationally expensive. Such models keep performing better as the size of the training data increases. However, training on large datasets has significantly increased end-to-end training times, computational power and energy consumption.

[0036] Many AI projects struggle with the data collection because of several common issues. For example:

[0037] - Incomplete data

[0038] - Unstructured data

[0039] - Duplicated data

[0040] - Biased data

[0041] - Imbalanced data

[0042] - Noisy data

[0043] - High-dimensional data

[0044] 3GPP Background Information

[0045] The following provides a selection of 3GPP agreements on AI / ML work in working groups RAN1, RAN2 and RAN3, that are related to the present disclosure.

[0046] - RAN1 agreements on Study Item on AI / ML for NR Air Interface [RP-213599] - refer to Annex I.

[0047] - RAN2 agreements on Study Item on AI / ML for NR Air Interface [RP-213599] - refer to Annex II.

[0048] - RAN3 agreement on AI / ML for NG-RAN work item - refer to Annex III

[0049] The above information is presented as background information only to assist with an understanding of the present disclosure. No determination has been made, and no assertion is made, as to whether any of the above might be applicable as prior art with regard to the present invention.

[0050] The following description of examples of the present disclosure, with reference to the accompanying drawings, is provided to assist in a comprehensive understanding of the present invention, as defined by the claims. The description includes various specific details to assist in that understanding but these are to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the examples described herein can be made without departing from the scope of the invention.

[0051] The same or similar components may be designated by the same or similar reference numerals, although they may be illustrated in different drawings.

[0052] Detailed descriptions of techniques, structures, functions, operations or processes known in the art may be omitted for clarity and conciseness, and to avoid obscuring the subject matter of the present invention.

[0053] The terms and words used herein are not limited to the bibliographical or standard meanings, but, are merely used to enable a clear and consistent understanding of the invention.

[0054] Throughout the description and claims of this specification, the words "comprise", "include" and "contain" and variations of the words, for example "comprising" and "comprises", means "including but not limited to", and is not intended to (and does not) exclude other features, elements, components, integers, steps, processes, operations, functions, characteristics, properties and / or groups thereof.

[0055] Throughout the description and claims of this specification, the singular form, for example "a", "an" and "the", encompasses the plural unless the context otherwise requires. For example, reference to "an object" includes reference to one or more of such objects.

[0056] Throughout the description and claims of this specification, language in the general form of "X for Y" (where Y is some action, process, operation, function, activity or step and X is some means for carrying out that action, process, operation, function, activity or step) encompasses means X adapted, configured or arranged specifically, but not necessarily exclusively, to do Y.

[0057] Features, elements, components, integers, steps, processes, operations, functions, characteristics, properties and / or groups thereof described or disclosed in conjunction with a particular aspect, embodiment, example or claim are to be understood to be applicable to any other aspect, embodiment, example or claim described herein unless incompatible therewith.

[0058] The skilled person will appreciate that the techniques described herein may be used in any suitable combination.

[0059] Certain examples of the present disclosure provide one or more techniques for Artificial Intelligence (AI) / Machine Learning (ML) data collection and / or reporting in a network, for example a 3rdGeneration Partnership Project (3GPP) 5thGeneration (5G) New Radio (NR) network. However, the skilled person will appreciate that the present invention is not limited to these examples, and may be applied in any suitable system or standard, for example one or more existing and / or future generation wireless communication systems or standards, including any existing or future releases of the same standards specification, for example 3GPP 5G.

[0060] The functionality of the various network entities and other features disclosed herein may be applied to corresponding or equivalent entities or features in the same or any other suitable communication systems or standards. Corresponding or equivalent entities or features may be regarded as entities or features that perform the same or similar role, function or purpose within the network. For example, the functionality of a NG-RAN node (e.g. a base station, eNB or gNB) in the examples below may be applied to any other suitable type of entity performing RAN functions.

[0061] A particular network entity may be implemented as a network element on a dedicated hardware, as a software instance running on a dedicated hardware, and / or as a virtualised function instantiated on an appropriate platform, e.g. on a cloud infrastructure.

[0062] The skilled person will appreciate that the present invention is not limited to the specific examples disclosed herein. For example:

[0063] - The techniques disclosed herein are not limited to 3GPP 5G.

[0064] - One or more entities in the examples disclosed herein may be replaced with one or more alternative entities performing equivalent or corresponding functions, processes or operations.

[0065] - One or more of the messages in the examples disclosed herein may be replaced with one or more alternative messages, signals or other type of information carriers that communicate equivalent or corresponding information.

[0066] - One or more further elements or entities may be added to the examples disclosed herein.

[0067] - One or more non-essential elements or entities may be omitted in certain examples.

[0068] The functions, processes or operations of a particular entity in one example may be divided between two or more separate entities in an alternative example.

[0069] - The functions, processes or operations of two or more separate entities in one example may be performed by a single entity in an alternative example.

[0070] - Information carried by a particular message in one example may be carried by two or more separate messages in an alternative example.

[0071] - Information carried by two or more separate messages in one example may be carried by a single message in an alternative example.

[0072] - The order in which operations are performed and / or the order in which messages are transmitted may be modified, if possible, in alternative examples.

[0073] Certain examples of the present disclosure may be provided in the form of an apparatus / device / network entity configured to perform one or more defined network functions and / or a method therefor. Certain examples of the present disclosure may be provided in the form of a system (e.g. network or wireless communication system) comprising one or more such apparatuses / devices / network entities, and / or a method therefor.

[0074] 3GPP has started studying the benefits of integration of AI / ML solutions to communications networks, in order to improve network operations.

[0075] 3GPP working groups RAN1 and RAN2 are studying data collection requirements, for different AI / ML model Life Cycle Management (LCM) purposes per use case, for aspects such as data content, data size, and data latency. For example, data for model training may have less latency constraint on data collection compared with model inference or model monitoring.

[0076] However, in addition to the above data collection requirements identified in 3GPP, there are other data collection issues, inherent to AI / ML operations, that need to be addressed by 3GPP working groups to enable integration of AI / ML operations into wireless networks. For example, the following are some of those issues that require solutions that are suitable (or tailored / customised) for wireless network:

[0077] 1) Data monitoring: for example,

[0078] a. How to detect and report issues related to data drift, pollution, aging, other.

[0079] 2) Data preparation: for example,

[0080] b. How to categorize (or differentiate) data collection according to the AI / ML LCM purpose and / or model use case. I.e., data collection for model training, data collection for model inference and / or monitoring.

[0081] c. How to ensure data collection according to data collection requirements of different LCM purposes and use cases.

[0082] 3) Existing 3GPP data collection mechanisms and procedures may lack framework, procedures and / or network entities that would be able to handle AI / ML data collection in wireless communications networks.

[0083] 4) The behaviour of the network and the UE (or group of UEs) that need to handle data collection according to a set of data collection requirements. For example,

[0084] d. the behaviour of the UE if it decides that a given data is not valid for the specific / related model LCM purpose.

[0085] e. How the network informs the UE (or a group of UEs) of how to handle data collection process in wireless communications networks.

[0086] Additionally, the following are some of the issues that have not been addressed by RAN1 / RAN2 study item on AI / ML for NR air interface:

[0087] 1) Dataset(s) for training, validation, testing, and inference

[0088] 2) Potential specification impact related to the AI Model lifecycle management, and dataset construction for training, validation and test for the selected use cases

[0089] The present disclosure proposes various techniques, including for controlling data collection in communications networks. The skilled person will appreciate that one or more of the following techniques may be applied individually or in any suitable combination.

[0090] Certain examples of the present disclosure provide a method, for a UE, for AI / ML data collection in a network, the method comprising: receiving, from the network, a data collection request message, wherein the data collection request message comprises a data collection assistance information; in response to receiving the data collection request message, performing data collection based on the data collection assistance information; and transmitting, to the network, a data collection report comprising collected data and / or information related to the data collection.

[0091] In certain examples, the data collection report may be generated based on data reporting assistance information received from the network.

[0092] In certain examples, the data reporting assistance information and the data collection assistance information may be one or more of: part of the same assistance information; separate assistance information; transmitted together; and transmitted separately.

[0093] In certain examples, the data collection report may comprise an offset timestamp (e.g. a time stamp of one or more measurements and / or collected data) indicating the time elapsed between triggering of the data collection (e.g. the UE receiving the data collection request message) and the data collection reporting (e.g. the UE transmitting the data collection report).

[0094] In certain examples, the method may further comprise transmitting, to the network, information (e.g. parameters, requirements and / or values) used for data collection and / or data reporting.

[0095] In certain examples, the data collection assistance information may comprise configuration for configuring data collection requirements (e.g. UE measurements).

[0096] In certain examples, the data collection assistance information may comprise one or more of: one or more parameters; one or more conditions; one or more thresholds; one or more rules; one or more policies; data handling information; one or more requirements (e.g. data collection requirements per LCM and / or use case for a given model and / or functionality); information related to one or more requirements (e.g. data content, data size and / or data latency for a model and / or functionality); and metadata (e.g. individual metadata for an individual model, specific data collection requirement and / or use case; and / or group metadata for a group-based model, all data collection requirements and / or use cases).

[0097] In certain examples, the data collection may comprise a data collection procedure from the UE to the network (e.g. gNB, LMF and / or OAM).

[0098] In certain examples, the data collection may be performed for AIML model LCM (e.g. training, inference and / or management).

[0099] In certain examples, the AIML model may be one or more of: a network-sided model; a UE-sided model; and a two-sided model.

[0100] In certain examples, the data collection and / or reporting assistance information may comprise: one or more data collection durations for determining an allowable time for data collection; and / or one or more data reporting validity durations for determining an allowable time for data reporting.

[0101] In certain examples, the method may further comprise receiving (e.g. in the data collection assistance information and / or RRC and / or NAS signalling and / or messages) one or more timer values comprising one or more of: a value indicating a duration that the UE waits after receiving the data collection request message to start data collection; a value indicating a duration that the UE collects data in a data collection session; a value indicating a duration until the UE reports the collected data; and a value indicating a total duration that the UE used to collect and report the data.

[0102] In certain examples, the data collection may be performed in a data collection session associated with a unique ID; and / or the data reporting may be performed in a data reporting session associated with a unique ID (e.g. the same as, or different from, the ID associated with the corresponding data collection session).

[0103] In certain examples, the method may comprise: receiving, from the network, one or more data collection related configurations and one or more associated IDs; performing collection of data corresponding to the associated IDs; and reporting, to the network, information of one or more AI / ML models corresponding to the associated IDs (e.g. AI / ML models developed (e.g. trained and / or updated) at the UE side based on the collected data corresponding to the associated IDs).

[0104] In certain examples, the data collection assistance information may comprise configuration for configuring the UE with data collection requirements: per AI / ML model-based LCM and / or functionality-based LCM purposes; and / or for a given AI / ML model and / or functionality per specific use case.

[0105] In certain examples, the data collection may comprise tagging the collected data according to one or more of: time and / or timestamp; location; and / or type of location (e.g. cell, TA, country).

[0106] In certain examples, the tagging may be performed in response to receiving, from the network, a tagging indication.

[0107] In certain examples, the data collection report may be generated in a format specified by the network.

[0108] In certain examples, the method may further comprise reporting, to the network, a failure to perform the data collection and / or data reporting according to the data collection assistance information.

[0109] In certain examples, if the data collection request message includes a predetermined IE (e.g. AIML Data Collection Profile IE), the method may comprise: storing the information included in the predetermined IE; performing the data collection based on the stored information; and performing the data reporting based on the stored information.

[0110] In certain examples, the method may further comprise performing data collection verification based on the data collection assistance information.

[0111] In certain examples, the network may comprise a base station, and wherein the UE transmits and / or receives information to and / or from the network via the base station.

[0112] Certain examples of the present disclosure provide a method, for a base station, for AI / ML data collection in a network, the method comprising: transmitting, to a UE, a data collection request message, wherein the data collection request message comprises a data collection assistance information; and receiving, from the UE, a data collection report comprising data collected by the UE based on the data collection assistance information and / or information related to the data collection.

[0113] In certain examples, the data collection assistance information may be based on information received by the base station from a core network entity (e.g. AMF).

[0114] In certain examples, the method may further comprise transmitting, to another base station, information based on the data collection assistance information (e.g. via Xn messages and / or signalling).

[0115] Certain examples of the present disclosure provide a UE configured to perform a method according to any example, aspect, embodiment and / or claim disclosed herein.

[0116] Certain examples of the present disclosure provide a base station configured to perform a method according to any example, aspect, embodiment and / or claim disclosed herein.

[0117] Certain examples of the present disclosure provide a network (or wireless communication system) comprising a UE and a base station according to any examples, aspects, embodiments and / or claims disclosed herein.

[0118] Certain examples of the present disclosure provide a computer program comprising instructions which, when the program is executed by a computer or processor, cause the computer or processor to carry out a method according to any example, aspect, embodiment and / or claim disclosed herein.

[0119] Certain examples of the present disclosure provide a computer or processor-readable data carrier having stored thereon a computer program according to any example, aspect, embodiment and / or claim disclosed herein.

[0120] Various examples will now be described in more detail.

[0121] Control of data collection and data reporting

[0122] -A set of existing (or newly defined) network entities (and / or network functions (NFs)), a UE (or groups of UEs), are involved in the control (monitoring and / or verification) of the data collection and data reporting processes of AI / ML operations in communications networks. Additionally, other external network entities (and / or NFs), OAM, server, cloud, and / or application function(s) may be involved in the control (monitoring and / or verification) of the data collection and data reporting.

[0123] -The control (monitoring and / or verification) of the data collection and data reporting maybe performed based on assistance information (parameters, rules, conditions, thresholds, and / or policies, other) that are available and / or provided to the entity performing the control (monitoring and / or verification) of the data collection and data reporting.

[0124] -The control (monitoring and / or verification) of the data collection and data reporting maybe performed at one entity (i.e. centralised control), for example, namely data collection control entity (DCE) (or any other suitable naming).

[0125] -The control (monitoring and / or verification) of the data collection and data reporting maybe performed at different entities (e.g. distributed and / or joint control), and / or at the UE (or among a group of UEs), other mix of the above.

[0126] -In one example, the control (monitoring and / or verification) of the data collection control and / or data reporting is performed by the entity (or NF(s) or UE(s)) receiving (or requesting) the data collection from at least one other entity (or NF(s) or UE(s)).

[0127] -In another example, the control (monitoring and / or verification) of the data collection control and / or data reporting is performed by the entity (or NF(s) or UE(s)) performing the data collection and data reporting.

[0128] -In another example, the entity (or NF(s) or UE(s)) requesting the data collection (and / or data reporting) and the entity (of NF(s) or UE(s)) performing the data collection (and / or data reporting) is the same entity (or NF(s) or UE(s)). In an alternative example, at least two different entities (or NF(s) or UE(s)) are handling the actions for the request, collection and / or reporting of the data.

[0129] -In another example, an entity (or NF(s) or UE(s)) performs control (monitoring and / or verification) of the data collection and another entity (or NF(s) or UE(s)) performs the control (monitoring and / or verification) of the data reporting.

[0130] -In one example, the network entity(-ies) (or NF(s)) performing the control (monitoring and / or verification) of the data collection and / or data reporting, maybe part of (or included in or co-located with) the RAN and / or CN and / or UE(s).

[0131] Data collection and reporting profile (assistance information, metadata)

[0132] -The assistance information that is used to control (monitor and / or verify) the data collection and / or data reporting, for example, may be termed data collection profile or data collection template, or data collection and / or reporting profile or data collection and / or reporting template (and / or any other suitable naming).

[0133] -In one example, the network (e.g. RAN and / or CN) and / or the UE(s), needs to control (monitor and / or verify) that the reported data is collected and / or reported according to (the requirements of the) data collection and reporting profile (or template) (or metadata).

[0134] -In one example, the network (e.g. RAN and / or CN) and / or the UE, needs to collect data according to the reported data collection requirements, as part of the data collection template (or metadata).

[0135] -In another example, the entity receiving the data, for example, the network (e.g. RAN and / or CN) and / or the UE, needs to verify that this data is collected and / or reported according to data collection template (or metadata).

[0136] -In one example, the data collection and reporting profile (or metadata) may include one or more of (but not limited to) the following information (parameters, conditions, thresholds, rules, and / or policies, other data handling info) related to the data collection and data collection:

[0137] A. Timestamp (on data collection and / or reporting):

[0138] - Absolute timestamp:indicates the current time (e.g. UTC, other);

[0139] - Offset timestamp:

[0140] 1)The UE reports the collected data including information on time offset (i.e. time elapsed) from a given time (e.g. UTC, or a time that the NW indicated to the UE to report the collected data, other).

[0141] 2)In one example, the UE may decide (or calculate) the offset by using a timer that it starts when the network triggers (or requests) a data collection session and / or data reporting session from the UE (or groups of UEs).

[0142] 3)The UE may calculate the time offset in the following options:

[0143] 3-1)Option 1 (time due to data collection process): time elapsed from the start to end of data collection session.

[0144] 3-2)Option 2 (time data available): time elapsed since the data has been available at the UE. I.e. after completing the data collection session.

[0145] 3-3)Option 3: sum of offset in Option 1 and 2. I.e., time elapsed since the network triggers data collection until the UE starts reporting the collected data the network.

[0146] B. Data collection and data reporting validity duration:

[0147] -The network configures a set (one or more) of data collection and / or data reporting validity duration(s) for the data collection and / or data reporting at a given UE (or groups of UEs).

[0148] -The network may configure two separate sets of validity durations, the first set for the data collection validity duration and the second set for data reporting validity duration.

[0149] -The network configures values for the data collection and / or data reporting validity duration(s) based on the information related to the data collection session (e.g. data collection for model monitoring, model inference, and / or use case).

[0150] -The network may configure the value(s) of the data collection and / or data reporting validity duration(s) based on the different data collection requirements for different model (or functionality-based) LCM purpose (e.g. monitoring, inference, training, etc.) and / or data collection use case.

[0151] -The network may configure the value(s) of the data collection and / or data reporting validity duration(s) based on assistance information related to the UE (or group of UEs), such as the UE type (or category), UE capability, UE power, UE RRC connection mode / state (i.e. RRC connected, inactive and / or idle mode), UE location (e.g. cell, TA, location, country, other), UE mobility, UE trajectory (or movement direction), UE altitude, UE angle, and / or any other information related to the UE.

[0152] -In another example, the network configurations may consider the model side (i.e. UE-side, Network-side, and / or two-side model).

[0153] -In another example, the data validity duration is decided based on the data collection purpose and / or use case.

[0154] -The data collection and / or reporting validity duration(s), is used by the network and / or the UE (or groups of UEs) to determine (monitor and / or verify) the allowable (acceptable and / or permitted) time for the data collection and / or data reporting procedure for a given data collection and / or reporting session for different data collection purposes (e.g. session for monitoring or training, etc.), and / or use cases.

[0155] -The data collection and / or data reporting validity duration(s), is used by the network to determine (decide, monitor, and / or verify) whether to accept (acknowledge, admit and / or any other action related to data collection and / or data reporting) or reject (and / or fail) the data reported from the UE (or groups of UEs).

[0156] 1) Network accepts the reported data from the UE:

[0157] 1-1)The UE collects and reports the data, to the network, based on the information (or parameters or requirements) in the data collection and reporting template.

[0158] 1-2)The UE includes all (or part of) information (parameters, requirements and / or values) included in the data collection and reporting template, together with the collected data, when sending to the network.

[0159] 1-3)In one example, based on the information (parameters, requirements and / or values), e.g. if time of data collection is needed, then the UE includes (along with the data) a time parameter and the value for the time parameters which represents e.g. the time when the data was collected, etc.

[0160] 1-4)The network receives the data together with (or in separately, or in separate messages and / or procedures) the information (parameters, requirements and / or values) used for the data collection and data reporting. Optionally, the UE may also include (part of or all of) the data collection and data reporting template.

[0161] 1-5)The network may compare the information in the data collection and / or data reporting template (or profile) to aspects (or information or parameters) of the collected data and / or reported data.

[0162] 1-6)The network may accept the reported data if it fulfils (meets) the information (parameters, requirements and / or values) included in the data collection and reporting template.

[0163] 1-7)In one example, the network compares the data collection and / or data reporting validity duration to the total data collection duration, which is defined as the time elapsed from the time the network triggers (or sends a request to start) the data collection procedure, (process, session, etc.) at the UE, until the (successful) reception of collected data at the network (e.g. part of the data collection response message or data reporting message, and / or any other existing and / or newly defined messages). That is, the network accepts the data, or decides that the data is valid, if the data collection duration is less than (or equal to) the data collection validity duration(s) or if the data collection duration is within the validity duration which was provided to the UE (or which his associated with the particular use case, etc.).

[0164] 1-8)In another example, the network may compare other aspects also based on data collection / reporting template (or profile). For example, aspects of data collection and / or reporting location, data size, data latency, data content, data purpose (LCM), data use case, specified data quality (e.g. accuracy, granularity, etc.),and / or specific (or configured) data security and / or data privacy, etc.

[0165] 2) Network rejects the reported data from the UE:

[0166] 2-1)The UE collects and reports the data, to the network, based on the information (or parameters or requirements) in the data collection and reporting template.

[0167] 2-2)The UE includes all (or part of) information (parameters, requirements and / or values) included in the data collection and reporting template, together with the collected data, when sending to the network.

[0168] 2-3)In one example, based on the information (parameters, requirements and / or values), e.g. if time of data collection is needed, then the UE includes (along with the data) a time parameter and the value for the time parameters which represents e.g. the time when the data was collected, etc.

[0169] 2-4)The network receives the data together with (or in separately, or in separate messages and / or procedures) the information (parameters, requirements and / or values) used for the data collection and data reporting. Optionally, the UE may also include (part of or all of) the data collection and data reporting template.

[0170] 2-5)The network may compare the information in the data collection and / or data reporting template (or profile) to aspects (or information or parameters) of the collected data and / or reported data. The network may reject the reported data if it fulfils (meets) the information (parameters, requirements and / or values) included in the data collection and reporting template.

[0171] 2-6)In one example, the network compares the data collection and / or data reporting validity duration to the total data collection duration, which is defined as the time elapsed from the time the network triggers (or sends a request to start) the data collection procedure, (process, session, etc.) at the UE, until the (successful) reception of collected data at the network (e.g. part of the data collection response message or data reporting message, and / or any other existing and / or newly defined messages). That is, the network accepts the data, or decides that the data is valid, if the data collection duration is less than (or equal to) the data collection validity duration(s) or if the data collection duration is within the validity duration which was provided to the UE (or which his associated with the particular use case, etc.).

[0172] 2-7)In another example, the network may compare other aspects also based on data collection / reporting template (or profile). For example, aspects of data collection and / or reporting location, data size, data latency, data content, data purpose (LCM), data use case, specified data quality (e.g. accuracy, granularity, etc.),and / or specific (or configured) data security and / or data privacy, etc.

[0173] 2-8)In one example, the network compares the data collection and / or data reporting validity duration to the total data collection duration, and the network rejects the collected data, or decides that the data is not valid, if the data collection duration is larger than the data collection and / or data reporting validity duration(s).

[0174] 3) UE decides to accept or reject the collected data at the UE:

[0175] 3-1)The UE collects the data based on the information (or parameters or requirements) in the data collection and reporting template.

[0176] 3-2)The data may be available at the UE from a previous data collection and / or data reporting sessions. For example, data collected at the UE, and / or the data is collected and reported from the network to the UE.

[0177] 3-3)The UE and / or the network includes (part of or all of) the data collection and data reporting template, e.g. at least the information used for the data collection and / or data reporting.

[0178] 3-4)The UE verifies the collected data based on all (or part of a modified version of) information (parameters, requirements and / or values) included in the data collection and data reporting template and / or information stored (or included) with the collected and / or reported data.

[0179] 3-5)The UE may compare the information in the data collection and / or data reporting template (or profile) to aspects (or information or parameters) of the collected data and / or reported data. The UE may accept or reject the reported data if it fulfils or fails to fulfils the information (parameters, requirements and / or values) included in the data collection and reporting template.

[0180] 3-6)The UE should verify if the data collected meets the requirements for data collection and / or reporting based on the data collection and / or reporting template which the UE may optionally have stored locally.

[0181] 3-7)In one example, the UE may decide to delete the data available at the UE (e.g. collected at the UE (locally), or received from the network) if for any reason the related data collection and / or data reporting duration exceeds (larger than) the data collection and / or data reporting validity duration. Optionally, the UE may also report this failure to the network and / or include a suitable cause value (e.g. data collection failure, data reporting failure, data collection and reporting failure, data collection not supported, or any other suitable naming).

[0182] 3-8)In another example, the UE may decide to keep (or modify, or amend, etc.) the data, generated locally and / or received from the network, if the related data collection duration does not exceed (less than or equal to) the data collection validity duration.

[0183] 3-9)In another example, the UE may decide to delete, modify, amend, or add, or store, the locally collected data and / or data received from the network (i.e. data collected at the NW), regardless of whether the related data collection duration is less than (or equal to) or larger than the data collection validity duration.

[0184] 4) UE verifies the data reported from the network:

[0185] 4-1)The network collects and reports the data, to the UE, based on the information (or parameters or requirements) in the data collection and reporting template.

[0186] 4-2)The network includes all (or part of) information (parameters, requirements and / or values) included in the data collection and reporting template, together with the collected data, when sending to the UE.

[0187] 4-3)In one example, based on the information (parameters, requirements and / or values), e.g. if time of data collection is needed, then the network includes (along with the data) a time parameter and the value for the time parameters which represents e.g. the time when the data was collected, etc.

[0188] 4-4)The UE receives the data together with (or in separately, or in separate messages and / or procedures) the information (parameters, requirements and / or values) used for the data collection and data reporting. Optionally, the network may also include (part of or all of) the data collection and data reporting template.

[0189] 4-5)The UE may compare the information in the data collection and / or data reporting template (or profile) to aspects (or information or parameters) of the collected data and / or reported data.

[0190] 4-6)The UE may accept or reject the reported data if it fulfils or fails to fulfil the information (parameters, requirements and / or values) included in the data collection and reporting template.

[0191] 4-7)Similar examples (and / or texts) to the cases above can be considered also for this case.

[0192] C. Data collection and reporting timers:

[0193] -The network configures a set of new timers for data collection in AI / ML operations.

[0194] -The network configures a set of new timers for different data collection type (or session) and / or data collection LCM purpose and / or use case.

[0195] -Data collection timers are configured, e.g. per data collection session, data collection LCM purpose, and / or data collection use cases.

[0196] -The following are examples (but not limited to) data collection and reporting timers:

[0197] 1) time-to-collect duration: duration that the UE waits for, after receiving a data collection request (or trigger) from the network, to start data collection session(s). In one example, if the UE successfully starts to collect data, the UE starts the data-collection timer;

[0198] 2) data-collect-timer: duration that the UE requires to successfully complete the data collection process or session. In one example, the UE (re)-starts the data-collect-timer after completing (successfully) the data collection session. In one example, in case of failure to complete data collection session (or complete successfully) before expiry of data-collect-timer, the UE may delete the collected data and re-start the data-collect-timer.

[0199] 3) data-report-timer: duration until a UE starts to report the data to the network; the timer is restarted upon successful data reporting to the network. In one example, in case of failure to report (successfully) the collected data, the UE may delete the collected data and re-start both of the data-collect-timer and data-collect-timer. In another example, the UE may keep the collected data and re-starts the data-collect-timer, but add the duration elapsed in previous failed data collection session to the total data collection duration.

[0200] 4) data collection duration (or data collection cycle): specifies the total duration that the UE uses to collect and report the data successful to the network, i.e. the time elapsed during time-to-collect duration, data-collect-timer, and data-report-timer (and / or any other time accounting for any failure steps in data collection);

[0201] Other timers (e.g. data collection validity timer, data reporting validity timer):

[0202] -In a related example, the network and / or the UE (or groups of UEs) re-start the data collection validity timer after reporting of collected data.

[0203] -The network and / or the UE may decide to re-acquire or re-start data collection session upon expiry of data collection validity duration for a given data collection session.

[0204] -In one example the data collection session and / or data reporting session are given a separate ID (globally unique, or temporary, or locally unique, or unique within a well-known scope e.g. cell, serving AMF, set of NG-RAN nodes, TA(s), etc). Both the NW and the UE may exchange the data collection and / or data reporting session IDs, e.g. included in the data collection and reporting profile, and / or any other suitable existing and / or newly defined RRC (or NAS) signalling / messages and / or system information broadcast (periodic and / or on-demand) using existing and / or newly defined SIBs.

[0205] -In another example, the data collection session and data reporting session are assigned (or associated or allocated) the same ID (or part of the same ID).

[0206] -In another example, the information related to the data collection and data reporting session(s) are included in the metadata (or template) of the model (or models) and / or the model functionality (or functionalities).

[0207] -In one example, the network and the UE (or groups of UEs) exchange a list of data collection sessions IDs and / or data reporting sessions IDs.

[0208] -In another example, two or more NG-RANs (or gNBs or eNBs) exchange the list of data collection sessions IDs and / or data reporting sessions IDs via Xn (or X2) interface / signalling / messages.

[0209] -In another example, the UE provides the list of data collection session IDs and / or data reporting session IDs to the network as part of the UE capability support to AI / ML operations in general and / or specific support to AI / ML data collection and / or data reporting requirements.

[0210] -In another example, the AMF (or MME) and the NG-RAN (or gNB or eNB) exchange the list of data collection session IDs and / or data reporting session IDs via NG (or S1) interface / signalling / messages.

[0211] -The UE behaviour to re-acquire (or restart) data collection session(s) and / or data reporting session(s) may be configured by the NW.In a related example, the UE may discard the collected data upon expiry of data collection validity duration (or timer, or any other suitable naming) or if the timer for data collection elapses and the UE is not able to complete the collection process. In another example, the UE maintains and / or reports the data but indicates to the NW that this data may or not be valid (or meet the data collection requirements) as it the validity duration is expired or the UE was not able to complete the data collection process for any reason.

[0212] -In another example, the network indicates to the UE to only report the collected data within the data collection validity duration (or timer). Optionally, the network indicates to the UE to automatically delete the collected data after expiry of the validity duration.

[0213] -The UE behaviour upon any changes in the data collection and / or data reporting conditions (e.g. location, time, other) maybe part of the configurations included in the data collection and data reporting template.

[0214] -In one example, if a certain condition (or requirements or rule or policyor parameter, or configuration, etc.) in the data collection and data reporting template, is not met, the UE may behave according to configuration that can be part of the data collection and data reporting template. For example,if data is to be collected in one location but the UE moves to another location before the process is complete, then the UE may need to either:

[0215] 1)delete the collected data and indicates to the network the case of data deletion after (resuming) (re)-establishing RRC connection again (e.g. part of existing and / or newly defined messages and / or signalling)

[0216] 2)store the collected data and forward to the network after (resuming) (re)-establishing RRC connection again (e.g. part of existing and / or newly defined messages and / or signalling). For example, assuming that the collected data still meet the information in the data collection and data reporting template (e.g. data collection validity duration, location, etc.)

[0217] 3)store the collected data and resume data collection upon (resuming) (re)-establishing RRC connection again (e.g. part of existing and / or newly defined messages and / or signalling). Then the UE report the collected data as planned (as per network configurations).

[0218] 4)In all above case, the UE may verify that the collected and / or reported data meet the information (or conditions or configurations or parameters, or requirements, etc.) indicated in the data collection and data reporting template before it reports the data to the network.

[0219] -In one example, the UE may participate in one or more data collection sessions. For example, data collection for different LCM purposes and / or use case.

[0220] -In a related example, the network and / or the UE (or groups of UEs) may re-start the data collection validity timer after reporting of collected data.

[0221] - Data collection duration

[0222] 1)This is needed to enable the network to configure a duration of a measurement gap when a UE performs a data measurement in RRC connected mode.

[0223] 2)In one example, the UE may signal to the network that the data measurement was performed in a terrestrial network or a non-terrestrial network, or using non-3gpp access, or while using a certain RAT, or any combination of this. The UE may also include information for instance whether the data measurement was performed using assistance information.

[0224] - Capability of performing data measurement simultaneously during downlink and / or uplink transmissions

[0225] - Last time a data measurement was performed

[0226] - Data collection and data reporting failure

[0227] 1)Failure instances, failure frequency, and / or failure rate

[0228] 2)Information on of data measurement failure during a data measurement.

[0229] E. Location stamp

[0230] -Cell ID, location, TA, radio technology used, frequency, mobility, RRC state, NAS state, active sending data or during data measurement gaps, etc.

[0231] F. Dual connectivity case.

[0232] -The network indicates whether the UE is allowed to perform data collection and / or data reporting in the case of dual connectivity.

[0233] -The network configures the UE to perform data collection and / or data reporting on the Secondary Node (SN) and / or Master Node (MN).

[0234] G. New RNTI(s) for data collection and data reporting

[0235] -The network may configure a newly defined RNTI(s) associated with the data collection and / or data reporting.

[0236] H. Inter-RAT data collection and reporting

[0237] -The network configures the UE to perform data collection and / or data reporting on LTE and / (or on NR) or vice versa.

[0238] J. Information and parameters related to data collection requirements per LCM and / or use case:

[0239] -The network may configure the UE (or groups of UEs) with the data collection requirements per AI / ML model-based (and / or functionality-based) LCM purposes.

[0240] -The network may configure the UE (or groups of UEs) with the data collection requirements for a given model (or models) per specific use case.

[0241] -In one example, the data collection requirements per LCM and / or use case, may include (but not limited to) one or more of the following parameters (or information):

[0242] 1) Data content: (e.g. raw data, processed data and optionally how to process it, summary of data, features of data, part of data, or statistics or analytics derived from the data); or formatted according to a given data type, or requirement of the data, e.g. data formatted (or modified or processed) for model monitoring, model training, other.

[0243] 2) Data size: This indicates the size allowed for transfer (or reporting).

[0244] 2-1)In one example, the UE is only allowed to report the collected data when the data size is equal to (or less than) a minimum data size threshold, (pre)-configured by the NW. For example, configured considering the data collection purpose (or LCM or use case) or the UE categories, capability on collecting data, resources, power, etc. In other example, the data size threshold may be configured based on the congestion status and / or the importance of other on-going transmissions on the RRC (or NAS) bearers (or signalling or messages). In another example, the threshold may be decided based on the data collection (or transfer) solutions, i.e. UP and / or CP or transfer from a server (or OAM) and / or any other data transfer solutions.

[0245] 2-2)In another example, the network configures the UE such that the UE is only allowed to report the collected data if its size (i.e. the data size) is more than a max data size threshold, (pre)-configured by the NW. For example, configured considering the data collection purpose (or LCM or use case) or the UE categories, capability on collecting data, resources, power, etc. In other example, the data size threshold may be configured based on the congestion status and / or the importance of other on-going transmissions on the RRC (or NAS) bearers (or signalling or messages). In another example, the threshold may be decided based on the data collection (or transfer) solutions, i.e. UP and / or CP or transfer from a server (or OAM) and / or any other data transfer solutions.

[0246] 3) Data purpose. For example, if the data is collected for monitoring, inference, training (offline or online), or any other model-based and / or functionality-based LCM purposes.

[0247] 4) Data collection reporting periodicity(if reported periodically).

[0248] 5) Data collection triggering events(conditions, policies, rules, and / or other info related to the data collection event triggering).

[0249] 6)Other information related to data collection requirements for AI / ML operations.

[0250] K. Privacy and security requirements on collected and reported data:

[0251] -The network and the UE (or groups of UE) may exchange a set of data privacy (and / or security) requirements on part of or all of the data to be collected and / or the data to be reported. For example, the UE may only share information that does not included any private (or sensitive) information related to its detailed location at a given time.

[0252] -The network and / or UE (or groups of UEs) may define a given security-level (and / or privacy-level) that control the level of data sharing between the UE and the network and / or between different network entities (and / or network functions) and / or between different UEs (or groups of UEs).

[0253] -In another example, the UE (or groups of UEs) and / or the network may provide user consent on the collection and reporting of the UE data to the network.

[0254] -In another example, a network entity (and / or network function) may grant (or provide or forward) user consent on the use of (and / or collection and / or reporting) a UE data (e.g. partial use or full use) to another network entity (and / or network function).

[0255] -In another example, a network entity (and / or network function) may grant (or provide or forward) user consent on the use of (and / or collection and / or reporting) its data (e.g. partial use or full use) to another network entity (and / or network function) and / or UE (or groups of UEs).

[0256] -In one example, the user consent in the examples above, maybe provided for a given time period and / or location.

[0257] Other Examples

[0258] -Upon expiry of data collection (stored or available) at the UE and / or the network, the UE and / or the network may re-start a specific timer that indicates the allowed time to store or keep a given data. In one example, the data is deleted upon expiry of this timer.

[0259] -In one example, the UE may be moved (by the network) or the UE transitions (automatically, e.g. based on (pre)-configuration from the NW) to RRC_IDLE, upon completion of data collection and reporting sessions for AI / ML operations.

[0260] -In another example, the UE may report any data collected during RRC_IDLE with indication that this data is collected based on data collection and reporting profile, and / or data is collected for AI / ML, and / or data is collected for a specific LCM purpose and / or use case. Additionally, the UE may indicate any data collection and reporting validity timers (and / or validity durations) to the network. The UE may use a new RRC establishment cause when it is establishing an RRC connection for the purpose of reporting data e.g. 'AIML data reporting', or any other cause may be defined / used for this purpose.

[0261] -In one example, the UE stays in the RRC_CONNECTED mode, in order to complete data collection and / or data reporting session.

[0262] -In one example, the network may reject or accept the data reported from the UE in RRC_IDLE and / or RRC_CONNECTED modes.

[0263] -In one example, the UE may request transmission to RRC_CONNECTED mode, or the UE may establish an RRC connection (or access) with the network in order to transfer the data collected for AI / ML. Optionally, the network may configure the UE, with an time event (e.g. data collection validity duration or data collection validity timers) and / or a timer, upon its expiry the UE will attempt RRC resume or RRC establishment in order to report the data collected within the specified timer.

[0264] -The UE may be configured with a time duration to remain in idle mode (RRC_IDLE or 5GMM_IDLE) before the UE can go to connected mode (RRC_CONNECTED mode or 5GMM_CONNECTED mode) in order to report the data which has been collected. Note that the time may be set to the value zero (e.g. zero seconds) which means the UE can immediately go to connected mode (RRC_CONNECTED mode or 5GMM_CONNECTED mode) to report the data.

[0265] -In one example, the UE may prioritise the data collection and / or data reporting sessions (e.g. resources allocated, signalling and / or messages) on other transmission and reception communication sessions.

[0266] -In an alternative example, the UE may deprioritise the data collection and / or data reporting sessions (e.g. resources allocated, signalling and / or messages) on other transmission and reception communication session.

[0267] -the UE should report it capability to support data collection and data reporting validity timers, validity durations, capability to perform an RRC establishment (RRC re-establishment) or RRC resume based on a trigger (or timer expiry) for reporting available data.

[0268] -In one example, the UE and / or the network may delete the collected data, upon the UE performing TAU (or registration procedure) or RNA Update. In another example, the UE and / or the network may store the collected data. Alternatively, the UE or network may do so when the UE has moved into a new area which is not served by the same network entity (e.g. NG-RAN or AMF) that configured the UE for data reporting, etc.

[0269] -In the case of RLF during data collection, the network configures the UE to restart data collection (and all related data collection timers, according to metadata) and perform RRC connection re-establishment procedure for recovering from RLF and to continue data collection.

[0270] -The UE may declare Radio Link Failure (RLF) if it fails to collected and / or report the data to the network. In one example, the UE fails to collected and / or reported the data based on expiry of configured timers.

[0271] -The UE may delete the collected data in the case of RLF. In a related example, the UE will (automatically based on configuration from the network) resume or re-start the data collection and / or data reporting session, after recovery from RLC.

[0272] -The UE may delete the collected data in the following cases:

[0273] 1)Random access procedure failure; or

[0274] 2)RLC failure; or

[0275] 3)Handover and / or handover failure.

[0276] 4)Other

[0277] -The UE maintains a data-collection timer (or timer instance) per data collection session.

[0278] -In one example, the data-collection timer can be set to different values to support different data collection sessions (e.g. one or more data collection sessions per LCM purpose and / or use case).

[0279] -The network (pre)configures the values of data-collection timers. For example, the network may provide timer values to the UE using existing (and / or newly defined) dedicated RRC and / or NAS signalling / messages, and / or system information broadcast (e.g. periodic and / or on-demand) using existing (and / or newly defined SIBs).

[0280] -In one example, the data collection profile (and / or data reporting) may only include information on data collection requirements per LCM and / or use case.

[0281] -Metadata is configured as individual metadata or group metadata. That is, the metadata may be defined, for an individual model (or model functionality) or a group-based (models or models functionalities). In another example, individual metadata for a specific data collection requirement and / or use case, and a group metadata for all data collection requirements and / or use cases.

[0282] -Metadata may be assigned a unique global (or local, or temporal, or dynamic) ID.

[0283] -Metadata and / or any information related to the metadata may be exchanged between network entities (and / or functions) and / or UE(s) (and / or OAM, a sever, external entity, and / or application function) in the form of list of metadata IDs and related information.

[0284] -The network informs the UE in the case of failure to report according to the metadata (e.g. validity timers, validity durations, requirements, other info). Optionally, the network may indicate to the UE one of the following:

[0285] 1)The UE needs to delete existing (or available) data and report again following the data collection requirements (i.e. info in data collection and / or data reporting profile, other info).

[0286] 2)Alternatively, the UE needs to collect more data (or amend existing data) and / or report existing (or amended or updated or modified data), following the data collection requirements (i.e. info in data collection and / or data reporting profile, other info).

[0287] -The network may remove the UE (or UEs) from the set of data collection and data reporting UEs. Optionally, inform any other NW function (or entity) of the removal of the UE (or groups of UEs) and the reason, e.g. the UE (or groups of UEs) not able to (or fails to) collect and / or report data according to the data collection and reporting profile.

[0288] -The UE is configured to report the collected data as part of existing measurement report; and / or newly defined measurement objects and measurement reports. optionally using existing or newly defined IEs, e.g. AI / ML data IE (or any other suitable naming).

[0289] -The UE is configured to report the collected data separately from existing measurement report (e.g. measurements for other purposes than AI / ML operations); e.g. newly defined NAS / RRC signalling / messages and / or IEs.

[0290] -The UE is configured to report the collected data in a similar method of (or based on) MDT and / or using an enhanced MDT methods to handle the data collection requirements.

[0291] -In one example, the entity collecting the data, for example, the network (e.g. RAN and / or CN) and / or the UE, needs to collect data according to the reported data collection requirements, as part of the data collection template (or metadata).

[0292] -In a related example, the entity verifying the data informs the data collecting (and / or reporting) entity (or function) of the outcome of the verification.

[0293] 1)In one example, the entity performing the verification informs (or confirms to) the sending entity of the reception of the data that meet (or fulfil or according to) the data collection template (or metadata).

[0294] 2)In another example the entity performing the verification informs (or confirms to) the sending entity of the reception of the data that fail to meet (or fulfils or according to) the data collection template (or metadata). The failure message

[0295] 3)Subscription aspects: if subscription indicate not to use the UE (start, stop, ...) based on subscription.

[0296] 4)If stopping needs to inform the UE what to do with the collected data

[0297] 4-1)Delete or report what collected so far.

[0298] -Support collecting and handling data collection based on metadata provided by the NW.

[0299] -A new message (RRC and / or NAS) to report metadata, optionally a new IE / container to do so may be used in any RRC / NAS message

[0300] -A set of network entities (and / or network functions) that provide assistance information on data collection requirements, per model (or functionality) LCM purpose and / or per use case, to at least another existing (and / or newly defined) network entity (and / or network function) in wireless communication networks.

[0301] -In one related example, the assistance information includesdata collection requirements per LCM and / or use case for a given model (or multiple models) or functionality (or multiple functionalities).

[0302] -In one example, the assistance information is included in a model (or functionality) meta data. Optionally, the meta data may include information related to the configured data collection requirements, e.g. data content, data size, data latency, and / or other data related parameters for the model (or functionality) described by this meta data.

[0303] -In another example, the assistance information is included in a pre-defined data collection template. In a related example, the profile (or information in this profile) is created by at least one newly defined (and / or existing) network entity (and / or function) and / or UE (or a group of UEs).

[0304] -In one related example, the data collection profile (or template) and / or meta data related to data collection process, may include timing information related to data collection process. Optionally, the timing information (or time tag or timestamp) related to collected data may be defined as follows:

[0305] 1)Option 1 (Unified or general Time Tag): the time tag (timing information or timestamp) is used for data collected for all (or any of) data collection purposes and / or use cases (e.g. a unified time tag).

[0306] 2)Option 2 (LCM purpose or use case specific Time Tag): the time tag (timing information or timestamp) is specific to the data collection purpose (e.g. timing tag for data collection for training or inference or monitoring) and / or use case.

[0307] 3)Option 3 (Category specific Time Tag): the timing tag is specific for a given data collection category. For example, a group of LCM purposes (and / or use cases) that share similar data collection requirements, such as latency requirement on data collected for model monitoring or model inference.

[0308] 4)Option 4 (other mix of the options above).

[0309] 5)Option 5: data collection session ID and / or UE data report ID and / or UE group ID, and / or use case ID, etc

[0310] -In one example, the data collection profile (or template) and / or meta data related to data collection process, may include location information related to data collection process. Optionally, the location information (or location tag) maybe defined in similar options as that for the time information (or time tag).

[0311] -In another example, the meta data may contain the data collection template (or data collection profile, or any other suitable naming) or vice versa.

[0312] -Data collection template may refer to metadata or profile for data collection and / or reporting.

[0313] -The following is a possible example of behaviour of the UE(s) and gNB during the data collection and data reporting sessions:

[0314] 1)The gNB indicates (or configures or propose) to the UE (s) to perform one or more of the following actions during the data collection and reporting sessions:

[0315] 1-1)Data to be tagged according to the time and / or location (i.e. added a time tag and / or location tag).

[0316] 1-2)Data tagging order could be: Time tagging followed by location tagging, followed by tagging based on type of location (cell, TA, country, other).

[0317] 2)The gNB may also request the UE(s) to report according to given format of structure and / or other data collection requirements.

[0318] 3)If UE(s) does (do) not collected and / or reported the data according to the above the gNB (and / or network) configuration, and / or data collection and data reporting profile.

[0319] 3-1)The gNB may reject and / or discard the data and / or request the UE(s) to discard (and / or store all or part of the collected and / or reported data).

[0320] 4)The UE(s) may only report data collected to the NW following a request from the NW or if the requirements based on the data collection template are met, or if the UE(s) is(are) able to provide the values for each of the parameters of the template e.g. location, time, RAT, etc.

[0321] FIGURE 2 illustrates an example of a data collection and data reporting procedure taking into account assistance information in the data collection template according to various embodiments of the present disclosure. Figure 2 shows an example, of the data collection and reporting procedure between the UE and gNB. The following are example, steps in this Figure:

[0322] Step 1: the gNB sends the AI / ML data collection request to the UE (or groups of UEs). For example, in this request message, the gNB includes the data collection and data reporting profile or metadata related to the data collection requirements for a given model (or multiple models) and / or model functionality (or multiple model functionalities). Optionally, the gNB may also include in the request other information related to the model, model functionality, the use case, and / or the model-based (and / or functionality-based) LCM purpose.

[0323] Step 2(a): the UE start to collect the data according to the information included in the metadata, and / or data collection and reporting profile. In another example, the UE may report failure to perform the data collection and / or data reporting sessions (or procedures) according to the information (parameters and / or requirements) included in the data collection and data reporting profile. For example, the UE may not have the required resources (power, memory, other) to collect and report the requested data. The UE may include a failure cause in the AI / ML data collection failure message (e.g. datacollectionNotsupported, dataReportingNotsupported, and / or any other suitable naming).

[0324] Step2(b): The UE may indicate (or confirm) to the gNB, using an AI / ML data collection acknowledge / response message, that the UE will be able to provide (i.e. collect and / or report) the data according to the data collection requirements and / or other parameters related to the data collection and / or data reporting included in the data collection and data reporting profile.

[0325] Note that not that Step 2(a) and Step 2(b) may be provided in parallel or sequentially.

[0326] Step 3: the UE to report the collected data according to the information included in the data collection and data reporting profile.

[0327] In one example, the data collection and reporting duration includes all steps in Figure 2. For example, in Figure 2, in Step 1, the gNB sends the AI / ML DATA COLLECTION REQUEST message (or any other suitable message naming) to the UE. In turn the UE will perform Step2(a) (start data collection), Step2(b) (AI / ML DATA COLLECTION ACKNOWLEDGE / RESPONSE / FAILURE message, or any other suitable message naming), and Step 3 (AI / ML DATA COLLECTION REPORT message, or any other suitable message naming) to the gNB.

[0328] In an alternative example, in Step 1, the UE is sending the AI / ML DATA COLLECTION REQUEST message (or any other suitable message naming) to the gNB. Then the gNB will perform Step2(a) (start data collection), Step2(b) (AI / ML DATA COLLECTION ACKNOWLEDGE / RESPONSE / FAILURE message, or any other suitable message naming), and Step 3 (AI / ML DATA COLLECTION REPORT message, or any other suitable message naming) to the UE.

[0329] In another example, the gNB is replaced by the AMF, and / or any other suitable network entity (or network function).

[0330] In another example, other steps maybe included and / or steps in Figure 2 may be replaced with other steps, and / or steps may be split into different steps and / or steps order maybe changed.

[0331] Note that in another example, the data collection and data reporting may be performed at the gNB and reported to the UE. In this case, the data collection and data reporting profile maybe stored (or available) and / or provided to the gNB from another network entity (or function) (e.g. in RAN or CN) and / or the UE. In one example, this would be the case for the model located at the gNB-side and / or the UE side, and / or two-sided models.

[0332] Mobility:

[0333] -how to handle data collection and reporting when the UE is moving across cells or TA.

[0334] -whether to discard data, e.g. local (cell-specific) or TA-specific data

[0335] -UE may report any collected data to the target eNB or NG-RAN node

[0336] -the source eNB or NG-RAN or gNB) forwards the data to target eNB (or NG-RAN or gNB).

[0337] -source eNB (or NG-RAN or gNB) only forwards to target if the target request.

[0338] -source eNB (or NG-RAN or gNB) may also forward to another entity.

[0339] -NG-based handover, source may forward to AMF and AMF forwards to target eNB (or NG-RAN or gNB).

[0340] -in another example, the source eNB (or NG-RAN or gNB) indicates to the UE to delete the data. This may be achieved usingsystem information and / or dedicated signalling (new or existing RRC messages / signalling):

[0341] 1)e.g. define a flag (1 / 0) delete / save data, this may be included in the RRC Release message (or RRC reconfiguration message), or any other RRC (existing and / or newly defined message / signalling and / or IEs and / or procedures).

[0342] -in another example, the source eNB (or NG-RAN or gNB) indicates to the UE to forward the collected data directly to target upon connection.

[0343] -UE is configured to resume / move to connected mode, if it has data to report.

[0344] -UE requests moving to connected mode to report data.

[0345] - Define new signalling procedures, messages, and / or IEs to support the data collection for different AI / ML model (or model functionality) LCM purposes and / or model (or model functionality) use case.

[0346] In the present disclosure, all examples, aspect, embodiments and / or claims that refer to a new network entity and / or network function may also be extended (or used) with existing network entities and / or network functions or a mix of new and existing network entities and network functions.

[0347] Other examples on Network configuration of data collection and / or data reporting requirements

[0348] Various examples of the present disclosure may include one or more of the following techniques, either individually or in any suitable combination, optionally in combination with one or more of the techniques described above.

[0349] Examples on network configurations of data collection requirements:

[0350] -The network (e.g. RAN, CN, and / or any other network entity (and / or function) provides configurations for data collection and / or data reporting requirements, per model (or functionality) LCM purpose and / or per use case, to a given UE (or group of UEs).

[0351] -The configurations for data collection and data reporting requirements may be provided by an external network entity (or function), OAM, server, cloud, and / or application function (AF).

[0352] -The configurations for data collection and data reporting requirements maybe provided directly to the UE (or a group of UEs) or indirectly via other network entity(-ies) (and / or function(s)).

[0353] -In one example, the UE (or group of UEs) receives from NG-RAN (or gNB) configurations for data collection and / or data reporting requirements via dedicated signalling, for example, using existing (or newly defined) RRC messages / signalling.

[0354] -In another example, the UE (or group of UEs) receives from NG-RAN (or gNB) configurations for data collection and / or data reporting requirements via broadcast (e.g. periodically and / or on-demand) of system information, using existing SIB and / or newly defined SIB.

[0355] -In another example, the UE (or group of UEs) receives from AMF (or another CN entity and / or function) configurations for data collection and / or data reporting requirements via dedicated signalling, for example, using existing (or newly defined) NAS messages / signalling.

[0356] -In another example, the network may provide to the UE (or a group of UEs) the data collection and / or data reporting requirements as part of UE registration procedure. In a related example, the UE may receive new data collection and / or data reporting requirements depending on the access time and / or access location.

[0357] -In another example, the network may provide to the UE (or a group of UEs) the data collection and / or data reporting requirements as part of TAU procedure. In a related example, the UE may receive new data collection and / or data reporting requirements depending on the access time and / or access location.

[0358] -In an example, the decision on (or setting of) data collection and / or data reporting requirements for a given model (or model functionality) may be decided by one or more network entity(-ies) (and / or functions).

[0359] -In another example, multiple models (or models functionality or functionalities) may share the same data collection requirements.

[0360] Examples on exchange of configurations of data collection and data reporting requirements via NG (or S1) or Xn (or X2):

[0361] -The data collection and / or data reporting requirements is provided to the NG-RAN (or gNB) from the CN. In one example, provided from AMF to the NG-RAN (or gNB) via existing (and / or newly defined) NG messages / signalling. In another example, via existing (and / or newly defined) S1 messages / signalling

[0362] -The data collection and / or data reporting requirements is exchanged between NG-RAN modes (or gNBs) via existing (and / or newly defined) Xn messages / signalling. In another example, via existing (and / or newly defined) X2 messages / signalling

[0363] Examples on network and UE behaviour on monitoring data collection and data reporting requirements:

[0364] -In one example, a new set of network entities (and / or network functions) is involved in monitoring (control and / or verification) of the data collection and / or data reporting process according to a given set of data collection requirements assigned to a given model (or multiple models) or model functionality (or functionalities).

[0365] -In a related example, at least one of the network entities (and / or functions) may send a request to another network entity to monitor (control and / or verification) the data collection and / or data reporting for a given model (or multiple models) or a given model functionality (or multiple functionalities).

[0366] -In a related example, one or more of the new network entities (or functions). Optionally, the request may be sent from / to existing (or newly defined network entity (and / or function)).

[0367] -In another example, the request for monitoring (control and / or verification) data collection and / or data reporting process may be sent to the network entity (or entities or NFs) performing the data collection and / or to another entity (or entities or NFs) that is (are) not performing the data collection.

[0368] -In a related example, the network (e.g. NG-RAN, AMF, other) may send the data collection monitoring request to the UE that collects data for a UE-side model, UE-part of model, network-part of model and / or network-side model.

[0369] -In an alternative example, the UE sends a request to the network side and / or network side that collects data for a UE-side model, UE-part of model, network-part of model and / or network-side model.

[0370] -The following are options for monitoring (control and / or verification) of data collection and data reporting requirements and / or sessions at the UE, the network, and / or other network entities (or NFs):

[0371] Option 1: UE-side model (monitoring at the UE and / or the network (e.g. gNB, AMF, SMF, other RAN and / or CN entities and / or functions):

[0372] -The model (or model functionality) is at the UE (or groups of UEs).

[0373] -In one example, the data collection is performed at the UE side.

[0374] -In another example, the data is already available at the UE, for example, previously collected at the UE and / or previously collected and reported, to the UE, by the network (and / or another UE (or groups of UEs)).

[0375] -The UE may report the collected data (and / or available data at the UE side), partially, in full, and / or modified to the network. For example, for data monitoring purposes (and / or other model and / or model functionality monitoring purposes, other).

[0376] -The Network may report the collected data (and / or available data at the network side), partially, in full, and / or modified to the network. For example, for data monitoring purposes (and / or other model and / or model functionality monitoring purposes, other).

[0377] -The monitoring (or verification or validation, etc.) of the data collection and reporting processes (or sessions) is performed at the UE (and / or the network).

[0378] -This monitoring of data collection and data reporting processes (sessions) maybe triggered by the UE and / or the network (e.g. the gNB or AMF).

[0379] -The UE monitors the data collection process according to the configured model data collection requirements (per LCM purpose and / or use case) (and / or other information in the data collection and data reporting profile).

[0380] -The UE decides based on data collection requirements (and / or other information in the data collection and data reporting profile) whether the collected data is valid (or useful, aligned, agrees with, fulfils, or meets) the requirements for the given LCM purpose (e.g. monitoring, inference, training) and / or use case.

[0381] Option 2: NW-side model (monitoring at the UE and / or the network (e.g. gNB, AMF, SMF, other RAN and / or CN entities and / or functions):

[0382] -The model (or model functionality) is at the network.

[0383] -In one example, the data collection is performed at the network side.

[0384] -In another example, the data is already available at the network, for example, previously collected at the network, and / or previously collected and reported, to the network, by the UE (or groups of UEs), and / or another network entity(-ies) (and / or network function(s)).

[0385] -The network may report the collected data (and / or available data) partially, in full, and / or modified to the another network entity (and / or function) and / or UE (or groups of UEs). For example, for data monitoring purposes (and / or other model and / or model functionality monitoring purposes, other).

[0386] -The monitoring (or verification or validation, etc.) of the data collection and reporting processes (or sessions) is performed at the network (and / or the UE).

[0387] -The monitoring at the network is triggered by the UE and / or the network.

[0388] -The network monitors the data collection process according to the configured model data collection requirements (per LCM purpose and / or use case), and / or based on data other information in the data collection and data reporting profile.

[0389] -The network decides whether the collected data is valid (or useful, aligned, agrees with, fulfils, or meets) the requirements (and / or other information in the data collection and data reporting profile) for the given LCM purpose (e.g. monitoring, inference, training) and / or use case.

[0390] Option 3: two-sided (monitoring at the UE and / or the network (e.g. gNB, AMF, SMF, other RAN and / or CN entities and / or functions):

[0391] - -The model (or model functionality) is at two sides (or split or divided) the UE and the network. That is, two-sided model (or two-sided functionality).

[0392] -In one example, the data collection is performed at the UE side and / or the network side.

[0393] -In another example, the data is already available at the UE and / or the network. For example, previously collected at the UE and / or the network and / or reported to the UE and / or the network.

[0394] -The monitoring (or verification or validation, etc.) of the data collection and data reporting processes (outcomes or sessions) is performed at the network and / or the UE (e.g. joint monitoring and / or single-side monitoring).

[0395] -The monitoring at the UE and / or network-part of model is triggered by the UE and / or the network.

[0396] -The network and / or the UE monitors the data collection process according to the configured model data collection requirements (per LCM purpose and / or use case), and / or based on other information in the data collection and data reporting profile.

[0397] -The network and / or the UE decides whether the collected data is valid (or useful, aligned, agrees with, fulfils, or meets) based on the requirements for the given LCM purpose (e.g. monitoring, inference, training) and / or use case, and / or based on other information in the data collection and data reporting profile.

[0398] -In one example, the triggering of the monitoring process of data collection and / or data reporting is achieved via existing (or newly defined) RRC signalling / messages, NAS signalling / messages, system information broadcast (periodic and / or on-demand) using existing or newly defined SIBs, and / or using MAC CE, and / or other methods to convey the trigger to the UE and / or the network.

[0399] FIGURE 3 illustrates a flow chart, for a UE, for AI / ML data collection in a network according to various embodiments of the present disclosure. Referring to Figure 3, in a first operation 301, the UE receives, from the network, a data collection request message. The data collection request message comprises data collection assistance information. In a second operation 302, in response to receiving the data collection request message, the UE performs data collection based on the data collection assistance information. In a third operation 303, the UE transmits, to the network, a data collection report comprising collected data and / or information related to the data collection.

[0400] FIGURE 4 illustrates a flow chart, for a base station, for AI / ML data collection in a network according to various embodiments of the present disclosure. Referring to Figure 4, in a first operation 401, the base station transmits, to a UE, a data collection request message. The data collection request message comprises data collection assistance information. In a second operation 402, the base station receives, from the UE, a data collection report comprising data collected by the UE based on the data collection assistance information and / or information related to the data collection.

[0401] In all examples, aspects, embodiments and / or claims disclosed herein, use cases may refer to one or more of the following non-limiting examples: CSI compression, beam management, positioning enhancement, load balancing, energy saving (or optimization), and / or mobility optimization, and / or other potential use cases and / or sub-use cases.

[0402] Example modification to 3GPP specifications

[0403] 3GPP TS 38.331:

[0404] New IEs data collection (ASN.1):

[0405] AIMLDataCollectionProfile

[0406] AIMLDataCollectionAndReportProfile

[0407] AIMLDataReportProfile

[0408] AIMLModelMetaData

[0409] AIMLDataCollectionMetaData

[0410] AIMLDataReportMetaData

[0411] New added text on data collection and / or data reporting:

[0412] In one example:

[0413] For each AI / ML model, if theAIML Data Collection ProfileIE (or AI / ML Data Collection and reporting IE, or Data Reporting IE) was included in theAIML Model Meta DataIE, contained in AI / ML DATA COLLECTION REQUEST message, the UE and / or the NG-RAN node stores this information, and, if supported, perform data collection, data reporting, and / or data collection verification according to the information included in theAIML Data Collection ProfileIE or in theAIML Model Meta DataIE.

[0414] In another example:

[0415] For each AI / ML model, if theAIML Data Collection ProfileIE (or AI / ML Data Collection and reporting IE, or Data Reporting IE) was included in the AI / ML DATA COLLECTION REQUEST message, the UE and / or the NG-RAN node stores this information, and, if supported, perform data collection, data reporting, and / or data collection verification according to the information included in theAIML Data Collection ProfileIE.

[0416] AI / ML Model Data Collection and / or Data Reporting Parameters

[0417] IE / Group NamePresenceRangeIE type and referenceSemantics descriptionCriticalityAssigned CriticalityUE (or gNB) IDM-gNB (or UE) IDM-Data Collection Session ID (or data collection and reporting session ID) or Data collection session ID and data reporting session ID, etc.)M-Model ID (or list of model IDs)M-> Model informationOYESignoreFunctionality ID (or list of functionalities IDs)M> Functionality informationOYESignoreData collection profileIndicates information related to the data collection and / or data reporting requirements, validity timers, validity duration, other parameters and / or info mentioned in the present disclosure.> data collection session IDO> data reporting session IDO-> data collection and / or reporting timers and validityIndicates the data collection requirements per LCM and / or use casexxxx> data collection requirementsOModel informationOYESignoreFunctionality informationOYESignoreOther information related to data collection and / or data reportingOther information

[0418] FIGURE 5 illustrates a block diagram of an exemplary network entity that may be used in examples of the present disclosure according to various embodiments of the present disclosure. For example, a UE and / or base station in the examples of Figures 1 to 4 may comprise an entity of Figure 5. The skilled person will appreciate that a network entity may be implemented, for example, as a network element on a dedicated hardware, as a software instance running on a dedicated hardware, and / or as a virtualised function instantiated on an appropriate platform, e.g. on a cloud infrastructure.

[0419] The entity 500 comprises a processor (or controller) 501, a transmitter 503 and a receiver 505. The receiver 505 is configured for receiving one or more messages from one or more other network entities, for example as described above. The transmitter 503 is configured for transmitting one or more messages to one or more other network entities, for example as described above. The processor 501 is configured for performing one or more operations, for example according to the operations as described above.

[0420] FIGURE 6 illustrates a block diagram illustrating a structure of a UE according to various embodiments of the present disclosure.

[0421] As shown in FIG. 6, the UE according to an embodiment may include a transceiver 610, a memory 620, and a processor (e.g. controller) 630. The transceiver 610, the memory 620, and the processor 630 of the UE may operate according to a communication method of the UE described above. However, the components of the UE are not limited thereto. For example, the UE may include more or fewer components than those described above. In addition, the processor 630, the transceiver 610, and the memory 620 may be implemented as a single chip. Also, the processor 630 may include at least one processor.

[0422] The transceiver 610 collectively refers to a UE receiver and a UE transmitter, and may transmit / receive a signal to / from a base station. The signal transmitted or received to or from the base station may include control information and data. The transceiver 610 may include a RF transmitter for up-converting and amplifying a frequency of a transmitted signal, and a RF receiver for amplifying low-noise and down-converting a frequency of a received signal. However, this is only an example of the transceiver 610 and components of the transceiver 610 are not limited to the RF transmitter and the RF receiver.

[0423] Also, the transceiver 610 may receive and output, to the processor 630, a signal through a wireless channel, and transmit a signal output from the processor 630 through the wireless channel.

[0424] The memory 620 may store a program and data required for operations of the UE. Also, the memory 620 may store control information or data included in a signal obtained by the UE. The memory 620 may be a storage medium, such as read-only memory (ROM), random access memory (RAM), a hard disk, a CD-ROM, and a DVD, or a combination of storage media.

[0425] The processor 630 may control a series of processes such that the UE operates as described above. For example, the transceiver 610 may receive a data signal including a control signal transmitted by the base station, and the processor 630 may determine a result of receiving the control signal and the data signal transmitted by the base station.

[0426] FIGURE 7 illustrates a block diagram illustrating a structure of a base station according to various embodiments of the present disclosure.

[0427] As shown in FIG. 7, the base station according to an embodiment may include a transceiver 710, a memory 720, and a processor (e.g. controller) 730. The transceiver 710, the memory 720, and the processor 730 of the base station may operate according to a communication method of the base station described above. However, the components of the network entity are not limited thereto. For example, the base station may include more or fewer components than those described above. In addition, the processor 730, the transceiver 710, and the memory 720 may be implemented as a single chip. Also, the processor 730 may include at least one processor.

[0428] The transceiver 710 collectively refers to a base station receiver and a base station transmitter, and may transmit / receive a signal to / from a terminal. The signal transmitted or received to or from the terminal may include control information and data. The transceiver 710 may include a RF transmitter for up-converting and amplifying a frequency of a transmitted signal, and a RF receiver for amplifying low-noise and down-converting a frequency of a received signal. However, this is only an example of the transceiver 710 and components of the transceiver 710 are not limited to the RF transmitter and the RF receiver.

[0429] Also, the transceiver 710 may receive and output, to the processor 730, a signal through a wireless channel, and transmit a signal output from the processor 730 through the wireless channel.

[0430] The memory 720 may store a program and data required for operations of the base station. Also, the memory 720 may store control information or data included in a signal obtained by the base station. The memory 720 may be a storage medium, such as read-only memory (ROM), random access memory (RAM), a hard disk, a CD-ROM, and a DVD, or a combination of storage media.

[0431] The processor 730 may control a series of processes such that the network entity operates as described above. For example, the transceiver 710 may receive a data signal including a control signal transmitted by the terminal, and the processor 730 may determine a result of receiving the control signal and the data signal transmitted by the terminal.

[0432] The techniques described herein may be implemented using any suitably configured apparatus and / or system. Such an apparatus and / or system may be configured to perform a method according to any aspect, embodiment, example or claim disclosed herein. Such an apparatus may comprise one or more elements, for example one or more of receivers, transmitters, transceivers, processors, controllers, modules, units, and the like, each element configured to perform one or more corresponding processes, operations and / or method steps for implementing the techniques described herein. For example, an operation / function of X may be performed by a module configured to perform X (or an X-module). The one or more elements may be implemented in the form of hardware, software, or any combination of hardware and software.

[0433] It will be appreciated that examples of the present disclosure may be implemented in the form of hardware, software or any combination of hardware and software. Any such software may be stored in the form of volatile or non-volatile storage, for example a storage device like a ROM, whether erasable or rewritable or not, or in the form of memory such as, for example, RAM, memory chips, device or integrated circuits or on an optically or magnetically readable medium such as, for example, a CD, DVD, magnetic disk or magnetic tape or the like.

[0434] It will be appreciated that the storage devices and storage media are embodiments of machine-readable storage that are suitable for storing a program or programs comprising instructions that, when executed, implement certain examples of the present disclosure. Accordingly, certain examples provide a program comprising code for implementing a method, apparatus or system according to any example, embodiment, aspect and / or claim disclosed herein, and / or a machine-readable storage storing such a program. Still further, such programs may be conveyed electronically via any medium, for example a communication signal carried over a wired or wireless connection.

[0435] While the invention has been shown and described with reference to certain examples, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the invention, as defined by the appended claims.

[0436] RAN1 Study Item on AI / ML for NR Air Interface

[0437] RAN1#109-e Agreements and working assumptions:

[0438] Working list of terminologies

[0439] TerminologyDescriptionData collectionA process of collecting data by the network nodes, management entity, or UE for the purpose of AI / ML model training, data analytics and inferenceAI / ML ModelA data driven algorithm that applies AI / ML techniques to generate a set of outputs based on a set of inputs.AI / ML model trainingA process to train an AI / ML Model [by learning the input / output relationship] in a data driven manner and obtain the trained AI / ML Model for inferenceAI / ML model InferenceA process of using a trained AI / ML model to produce a set of outputs based on a set of inputsAI / ML model validationA subprocess of training, to evaluate the quality of an AI / ML model using a dataset different from one used for model training, that helps selecting model parameters that generalize beyond the dataset used for model training.AI / ML model testingA 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.UE-side (AI / ML) modelAn AI / ML Model whose inference is performed entirely at the UENetwork-side (AI / ML) modelAn AI / ML Model whose inference is performed entirely at the networkOne-sided (AI / ML) modelA UE-side (AI / ML) model or a Network-side (AI / ML) modelTwo-sided (AI / ML) modelA paired AI / ML Model(s) over which joint inference is performed, where joint inference comprises AI / ML Inference whose inference is performed jointly across the UE and the network, i.e. the first part of inference is firstly performed by UE and then the remaining part is performed by gNB, or vice versa.AI / ML model transferDelivery of an AI / ML model over the air interface, either parameters of a model structure known at the receiving end or a new model with parameters. Delivery may contain a full model or a partial model.Model monitoringA procedure that monitors the inference performance of the AI / ML model

[0440] RAN1#110bis-e Agreements

[0441] Study AI / ML model monitoring for at least the following purposes: model activation, deactivation, selection, switching, fallback, and update (including re-training).

[0442] Study at least the following metrics / methods for AI / ML model monitoring in lifecycle management per use case:

[0443] i. Monitoring based on inference accuracy, including metrics related to intermediate KPIs

[0444] ii. Monitoring based on system performance, including metrics related to system performance KPIs

[0445] iii. Other monitoring solutions, at least following 2 options.

[0446] * Monitoring based on data distribution

[0447] a) Input-based: e.g., Monitoring the validity of the AI / ML input, e.g., out-of-distribution detection, drift detection of input data, or SNR, delay spread, etc.

[0448] b) Output-based: e.g., drift detection of output data

[0449] * Monitoring based on applicable condition

[0450] Note: Model monitoring metric calculation may be done at NW or UE

[0451] RAN#113 Agreements

[0452] Consider at least the following aspects and if applicable, the corresponding potential specification impact related to data collection:

[0453] - Measurement configuration and reporting

[0454] - Contents, type and format of data including:

[0455] * Data related to model input

[0456] * Data related to ground truth

[0457] * Quality of the data

[0458] * Other information

[0459] - Signaling of assistance information for categorizing the data

[0460] * Note: The study should consider the feasibility of disclosure of proprietary information

[0461] - Signaling for data collection procedure

[0462] - Note 1: Use-case specific details can be studied in respective agenda items

[0463] - Note 2: Signaling mechanism details can be studied by appropriate working groups.

[0464] Remaining Open issues

[0465] - Complete AI / ML model, terminology and description to identify common and specific characteristics for framework investigations:

[0466] * ...

[0467] * Characterize lifecycle management of AI / ML model: e.g., model training, model deployment, model inference, model monitoring, model updating

[0468] * Dataset(s) for training, validation, testing, and inference

[0469] * Identify common notation and terminology for AI / ML related functions, procedures and interfaces

[0470] - Evaluate performance benefits of AI / ML based algorithms for the agreed use cases in the final representative set

[0471] - Assess potential specification impact, specifically for the agreed use cases in the final representative set and for a common framework:

[0472] * PHY layer aspects, e.g., (RAN1)

[0473] ** Consider aspects related to, e.g., the potential specification of the AI Model lifecycle management, and dataset construction for training, validation and test for the selected use cases

[0474] ** Use case and collaboration level specific specification impact, such as new signalling, means for training and validation data assistance, assistance information, measurement, and feedback

[0475] RAN2 Study Item on AI / ML for NR Air Interface

[0476] RAN2#122 agreements:

[0477] Related to data collection, the following was agreed:

[0478] - RAN 2 assumes that for the data collection in some scenarios (e.g., internal data up to implementation or the existing data are enough), possibly no RAN2 specification effort is needed in some scenarios, e.g. (not exhaustive):

[0479] * For model inference of UE-sided model, input data for model inference is available inside the UE.

[0480] * For UE-side (real time) monitoring of UE-sided model, performance metrics are available inside the UE. UE can independently monitor a model's performance without any data input from NW.

[0481] - For the latency requirement of data collection, RAN2 assumes:

[0482] * for all types of offline model training (i.e., UE- / NW- / two-sided model training), there is no latency requirement for data collection

[0483] * for model inference, when required data comes from other entities, there is a latency requirement for data collection

[0484] * for model monitoring, when required monitoring data (e.g., performance metric) comes from the other entities, there is a latency requirement for data collection.

[0485] - RAN2 assumes that the analysis / selection of the data collection frameworks should focus on the RRC_CONNECTED state (for both data generation and reporting). Analysis and potential enhancement on the non-connected state can be revisited when needed.

[0486] - For the data generation entity and termination entity deployed at different entities, RAN2 assumes:

[0487] * For CSI enhancement and beam management use cases:

[0488] ** For model training, training data can be generated by UE / gNB and terminated at gNB / OAM / OTT server

[0489] ** For NW-sided model inference, input data can be generated by UE and terminated at gNB.

[0490] ** For UE-side model inference, input data / assistance information can be generated by gNB and terminated at UE.

[0491] ** For model monitoring at NW side, performance metrics can be generated by UE and terminated at gNB.

[0492] * For positioning enhancement use case:

[0493] ** For model training, training data can be generated by UE / gNB and terminated at LMF / OTT server

[0494] ** For NW-sided model inference, input data can be generated by UE / gNB and terminated at LMF and / or gNB

[0495] ** For UE-side model inference, input data / assistance information can be generated by LMF / gNB and terminated at the UE

[0496] ** For model monitoring at NW side, performance metrics can be generated by UE / gNB and terminated at LMF.

[0497] Related to data collection, RAN2 has also agreed to send an LS to RAN1 in R2-2306906, asking RAN1 to express concerns (if any) on the above assumption, and to provide additional information (if any) on the above discussed topics. Additionally, RAN2 asks RAN1 to provide inputs on the following:

[0498] - The required data content per use case and per LCM purpose, when available, and to what extent said data would / should be specified (in detail).

[0499] - The reporting type (e.g., periodic, event triggered, other) of the identified data content

[0500] - The typical size (value or value range) of the identified data content.

[0501] - The typical latency requirement (value or value range) to transfer the identified data content.

[0502] Related to architectural discussion, the following was agreed:

[0503] - RAN2 will cover functional architecture in general, e.g. covering both be model based and / or functionality based LCM

[0504] - Figure 2 is R2-2305327 capturing architectural aspects is agreed with the following assumptions:

[0505] * "Model Storage" in the figure is only intended as a reference point (if any) for protocol terminations etc for model transfer / delivery etc. It is not intended to limit where models are actually stored. Add a note for this.

[0506] * Remove "Model" in Model Managemt and Model Inference and for the actions / the arrow form Management to Inference (to reduce the risk for misunderstanding).

[0507] * Management may be model based management, or functionality based management.

[0508] RAN#121bis agreements:

[0509] - Study the applicability (and limitations) of each identified data collection framework for each of the identified LCM purposes, i.e., inference, monitoring and (offline) training. FFS how we do the formatting / presentation of the results.

[0510] - With more progress on architectural discussion, consider the suitability of each identified data collection framework for the termination points and mapping with the location of LCM purposes / functions (inference, monitoring, (offline) training) considering:

[0511] - Model sidedness (UE side, NW side, two sided) FFS

[0512] - Use case mapping FFS

[0513] Remaining Open issues

[0514] Based on what was discussed during RAN2#122 the following are the open issues:

[0515] - Architecture; functionality-to-entity mapping,

[0516] - Life Cycle Management implications from a RAN2 point of view,

[0517] - Progress with data collection, including suitability analysis of identified collection frameworks, taking into account model sidedness (i.e. UE-sided, NW-sided), the LCM function, and the consumer of the data collection (e.g. UE, gNB, OAM, OTT server). Reflect such analysis in the previously endorsed table.

[0518] - Continue discussion on model ID handling, and model transfer / delivery

[0519] RAN3 Work Item on AI / ML for NG-RAN

[0520] RAN3#117bis-e agreements:

[0521] - Procedures used for AI / ML support in the NG-RAN shall be use case agnostic.

[0522] - Legacy information that are used to support AI / ML are transferred via existing legacy procedures (no need to signal them via other procedures)

[0523] - RAN3 will focus on non-split architecture use cases and procedures first and discuss split architecture use cases and procedures when completion for the non-split architecture use cases and procedures is achieved.

[0524] - Signalling describing the capability to support specific information predictions used for AI / ML is not pursued in this release

[0525] - Signalling describing the capability to supports specific AI / ML use cases is not pursued in this release

[0526] - AI / ML capability exchange in NG-RAN can be achieved by means of procedures for AI / ML information request, AI / ML information response and AI / ML Information Request Failure

[0527] - WA: Solutions for AI / ML information exchange over the NG interface are not considered as part of Rel18.

[0528] Xn interface:

[0529] - Introduce a new Class 1 procedure for initiating the reporting of AI / ML Related Information and a Class 2 procedure for Data Reporting of AI / ML Related Information.

[0530] - Reporting options for the new procedure used for AI / ML Related Information to be evaluated on a case-by-case basis. Possible reporting options are one-time and periodic reporting.

[0531] - The new procedure is non-UE associated procedure. If needed, the procedure can be used to capture UE-associated information.

[0532] - The response message of the new procedure for AI / ML Related Information indicates if the requested information can be provided.

[0533] - Support the following UE performance information to be sent for feedback purposes: Average Packet Delay, Average UE Throughput DL, Average UE Throughput UL, Average Packet Error Rate.

[0534] - How to indicate validity time (e.g., implicitly with a new prediction when the previous prediction becomes invalid, explicitly with every prediction in the AI / ML output or by the request to the prediction) shall be discussed on a case by case basis.

[0535] RAN3#118 agreements

[0536] WA: Procedures used for AI / ML support in the NG-RAN shall be "data type agnostic"

[0537] Xn interface:

[0538] The request in the new Class 1 procedure for initiating the reporting of AI / ML Related Information can include an ID assigned by the requesting NG-RAN node to request for reporting, which includes

[0539] -the reporting parameters

[0540] -list of cells to report

[0541] -reporting periodicity

[0542] The response in the new Class 1 procedure for initiating the reporting of AI / ML Related Information can include an ID assigned by the responding NG-RAN node which includes the confirmation on the reporting parameters requested.

[0543] The message in the Class 2 procedure for Data Reporting of AI / ML Related Information can include the corresponding IDs assigned by the NG-RAN nodes, reports result.

[0544] The "Energy Efficiency" metric should be measurable, produced and interpretable by the RAN.

[0545] Start with per node granularity EE and Per cell granularity EE could be considered if it is feasible.

[0546] WA: Take the EE defined in SA5 as the baseline for the energy efficiency of a gNB.What to be transfered between NG-RAN nodes is FFS.

[0547] UE Trajectory Prediction is transferred to the target gNB via the Handover Request.

[0548] Abbreviations / Definitions

[0549] In the present disclosure, the following acronyms / definitions are used.

[0550] 3GPP 3rdGeneration Partnership Project

[0551] 5G 5thGeneration

[0552] 5GMM 5G Mobility Management

[0553] AF Application Function

[0554] AI ( / ) ML Artificial Intelligence ( / ) Machine Learning

[0555] AMF Access and Mobility Management Function

[0556] API Application Programming Interface

[0557] CE Control Element

[0558] CN Core Network

[0559] CP Control Plane

[0560] CSI Chanel State Information

[0561] DCE Data Collection Control Entity

[0562] eNB Base Station

[0563] gNB 5G Base Station

[0564] ID Identity / Identification

[0565] IE Information Element

[0566] LCM Life Cycle Management

[0567] LMF Location Management Function

[0568] LTE Long Term Evolution

[0569] MAC Medium Access Control

[0570] MDT Minimisation of Drive Test

[0571] MLOps Machine Learning Operations

[0572] MME Mobility Management Entity

[0573] MN Master Node

[0574] NAS Non Access Stratum

[0575] NF Network Function

[0576] NG Next Generation

[0577] NGAP NG Application Protocol

[0578] NR New Radio

[0579] NW Network

[0580] OAM Operations, Administration and Maintenance

[0581] RAN Radio Access Network

[0582] RAT Radio Access Technology

[0583] REST Representational State Transfer

[0584] RLF Radio Link Failure

[0585] RNA RAN-based Notification Area

[0586] RNTI Radio Network Temporary Identifier

[0587] RRC Radio Resource Control

[0588] S1 Interface between RAN and CN

[0589] SIB System Information Block

[0590] SMF Session Management Function

[0591] SN Secondary Node

[0592] TA Tracking Area

[0593] TAU Tracking Area Update

[0594] TS Technical Specification

[0595] UE User Equipment

[0596] UP User Plane

[0597] UTC Coordinated Universal Time

[0598] Xn / X2 Interface between RAN nodes

[0599] XnAP Xn Application Protocol

Claims

1.A method performed by a user equipment (UE) for artificial intelligence / machine learning (AI / ML) data collection in a wireless communication system, the method comprising:receiving, from a base station, a data collection request message including data collection assistance information;in response to receiving the data collection request message, performing a data collection based on the data collection assistance information; andtransmitting, to the base station, a data collection report including collected data or information related to the data collection,wherein the data collection is performed for an AI / ML model life cycle management (LCM).2.A method of claim 1, wherein the data collection report is generated based on data reporting assistance information received from the base station.3.A method of claim 2, wherein the data reporting assistance information and the data collection assistance information are at least one of:part of the same assistance information;separate assistance information;transmitted together; ortransmitted separately.4.A method of claim 1, wherein the data collection report includes an offset timestamp indicating the time elapsed between triggering of the data collection and the data collection reporting.5.A method performed by a base station for artificial intelligence / machine learning (AI / ML) data collection in a wireless communication system, the method comprising:transmitting, to a user equipment (UE), a data collection request message including data collection assistance information, wherein the data collection assistance information is for a performing of a data collection by the UE; andreceiving, from the UE, a data collection report including collected data or information related to the data collection,wherein the data collection is performed for an AI / ML model life cycle management (LCM).6.A method of claim 5, wherein the data collection report is generated based on data reporting assistance information transmitted to the UE.7.A method of claim 6, wherein the data reporting assistance information and the data collection assistance information are at least one of:part of the same assistance information;separate assistance information;transmitted together; ortransmitted separately.8.A method of claim 5, wherein the data collection report includes an offset timestamp indicating the time elapsed between triggering of the data collection and the data collection reporting.9.A user equipment (UE) for artificial intelligence / machine learning (AI / ML) data collection in a wireless communication system, the UE comprising:a transceiver; anda controller couple with the transceiver, and configured to:receive, from a base station, a data collection request message including data collection assistance information,in response to receiving the data collection request message, perform a data collection based on the data collection assistance information, andtransmit, to the base station, a data collection report including collected data or information related to the data collection,wherein the data collection is performed for an AI / ML model life cycle management (LCM).10.A UE of claim 9, wherein the data collection report is generated based on data reporting assistance information received from the base station.11.A UE of claim 10, wherein the data reporting assistance information and the data collection assistance information are at least one of:part of the same assistance information;separate assistance information;transmitted together; ortransmitted separately.12.A UE of claim 9, wherein the data collection report includes an offset timestamp indicating the time elapsed between triggering of the data collection and the data collection reporting.13.A base station for artificial intelligence / machine learning (AI / ML) data collection in a wireless communication system, the base station comprising:a transceiver; anda controller couple with the transceiver, and configured to:transmit, to a user equipment (UE), a data collection request message including data collection assistance information, wherein the data collection assistance information is for a performing of a data collection by the UE, andreceive, from the UE, a data collection report including collected data or information related to the data collection,wherein the data collection is performed for an AI / ML model life cycle management (LCM).14.A base station of claim 13, wherein the data collection report is generated based on data reporting assistance information transmitted to the UE.15.A base station of claim 14, wherein the data reporting assistance information and the data collection assistance information are at least one of:part of the same assistance information;separate assistance information;transmitted together; ortransmitted separately.

Images