Improvements in and relating to user equipment privacy in a telecommunication system

Abstract

Operating a User Equipment (UE) communicatively coupled to a telecommunication network, wherein the UE is arranged: to make measurements in relation to the network, and to store the measurement result

Description

5 The present invention relates to method for ensuring privacy of certain data within a telecommunication system. It further relates to improved performance of a User Equipment, UE, via use of on-device Artificial Intelligence, Al, techniques. Artificial Intelligence (Al) based data analytics is a significantly important function in current 10 Fifth Generation (5G) networks and key enabler for Sixth Generation (6G). 3GPP has proposed a network data analytics function (NWDAF) which collects data from various network functions based on subscriptions to events and provides analytics service to any network functions and the Operations and Management (OAM) function. The list of Network Function (NF) services provided by NWDAF is provide in Table 1, below: Service Name Service Operations Operation Semantics Example Consumer(s) Nnwdaf_AnalyticsSubscri ption Subscribe Subscribe / Notify PCF, NSSF, AMF, SMF, NEF, AF, OAM, CEF, NWDAF, DCCF Unsubscribe PCF, NSSF, AMF, SMF, NEF, AF, OAM, CEF, NWDAF, DCCF Notify PCF, NSSF, AMF, SMF, NEF, AF, OAM, CEF, NWDAF, DCCF, MFAF Transfer Request / Response NWDAF Nnwdaf_Analyticslnfo Request Request / Response PCF, NSSF, AMF, SMF, NEF, AF, OAM, CEF, NWDAF, DCCF ContextTransfer Request / Response NWDAF Nnwdaf_DataManagement Subscribe Subscribe / Notify NWDAF, DCCF Notify NWDAF, DCCF, MFAF, ADRF Fetch Request / Response NWDAF, DCCF, MFAF, ADRF Nnwdaf_MLModelProvision Subscribe Subscribe / Notify NWDAF Unsubscribe NWDAF Notify NWDAF N nwdaf_M LModell nfo Request Request / Response NWDAF Additionally, 3GPP defined a management data analytics (MDA) function which provides analytics service based on input data from different network functions e.g. data reported from a base station (gNB) and / or specific core network functions. MDA can use the analytics from NWDAF depending on the use cases. MDA can act as cross-domain (Radio Access Network, RAN, and Core) MDA MnS producer e.g. coordinating data analytics between NWDAF, gNB and MDA MnS producer(s) in different sub-domains as illustrated in Figure 1 which shows an example of coordination between NWDAF, gNB and MDAS (MDA MnS) producer. There is currently work ongoing to optimize Minimization of Drive Tests (MDT) for data collection in AI / ML systems. User-centric network is a key enabler for 6G, and it requires extensive User Equipment (UE) level data analytics to be able to provision network optimisation at the UE level. UE data analytics predicting UE behaviour is important to support a number of use cases like energy efficiency, load balancing, handover prediction and anomaly detection in fully autonomous 6G network. A number of use cases based on UE related analytics are identified for NWDAF in 3GPP, such as UE mobility prediction where UE location is predicted at tracking area (TA) or cell level or abnormal behaviours related network data analytics. Further UE behaviour / service analytics use cases such as NWDAF-assisted Radio Access Technology (RAT)Zfrequency selection and detection of anomaly events are introduced. An important enabler for detailed user level analytics is the availability of user specific radio measurements. 3GPP proposes collection of UE specific radio measurements using minimisation of drive test (MDT) procedures. Logged MDT procedure is used to collect UE measurement in RRCJNACTIVE and RRCJDLE mode and immediate MDT is used to report measurements in RRC-CONNECTED mode. However, MDT measurement collection provides user specific information to the network. Due to privacy and legal requirements, user consent is required before activating MDT for a specific International Mobile Subscriber Identifier (IMSI) or International Mobile Equipment Identifier (IMEI) number. Embodiments of the invention aim to address the above problem on preserving user privacy, private information safety and security concerns and need for user consent. Other problems, whether identified herein or not are also addressed. According to the present invention there is provided an apparatus and method as set forth in the appended claims. Other features of the invention will be apparent from the dependent claims, and the description which follows. According to a first aspect of the invention, there is provided a method of operating a User Equipment, UE, communicatively coupled to a telecommunication network, wherein the UE is arranged to make certain measurement in relation to the network, and to store the measurement results locally to the UE, wherein the UE is further arranged to not send the measurement results onto the network, thereby preserving privacy associated with the measurement results and to locally perform data analytics, inference or prediction using the measurement results. In an embodiment, the analytics are obtained using an Artificial Intelligence, Al, model that is trained locally to the UE or globally by the network, but using local measurement data. In an embodiment, the measurements include Minimization of Drive Tests, MDT, data. In an embodiment, the UE is configured by the network to generate particular analytics reports by providing one or more of specific configurations, parameters, conditions and policies, to assist the UE in the generation of data analytics. In an embodiment, the particular analytics reports include one of particular measurements, particular analytics and particular inference and predictions. In an embodiment, the network provides the UE with assistance information on a validity time. In an embodiment, the network provides information on requirements of the data analytics for the UE, said requirements including one or more of: size of data analytics, latency of data analytics, security requirements, accuracy and Key Performance Indicators, KPIs. In an embodiment, the network may activate / deactivate or start, pause or stop the data analytics process at the UE. In an embodiment, the UE utilises a communication module arranged to enable the UE, or a group of UEs, to share securely the generated data analytics reports to the network without sharing raw user data, thus preserving user privacy. In an embodiment, the UE provides user consent to the network in connection with collecting analytics from the UE or sharing analytics beyond the UE. In an embodiment, the UE indicates to the network its capability to support analytics to the network, such that the network is able to configure the data analytics collection and reporting, taking into consideration the indicated capability. According to a second aspect of the invention, there is provided a method of operating a telecommunication network including the step of configuring a UE to locally perform data analytics using certain measurement results, and to share a resulting data analytics report with the network without compromising user privacy. In an embodiment, there is further provided the use of a module to receive data analytics reports from the UE. According to a third aspect of the invention, there is provided a UE arranged to perform the method of the first aspect. According to a fourth aspect of the invention, there is provided a telecommunication network arranged to perform the method of the second aspect. Embodiments of the present invention address problems in the prior art by introducing a new network function responsible for handling (e.g. producing, controlling, reporting, etc.) the user analytics at the UE. These analytics are obtained using a UE-side Al model(s) that is trained locally (using local data set(s)) or globally by the network (and / or externally (server, other)), based on MDT measurements and any other user specific information, for example, precise location etc. The present invention provides a novel method and system for implementing privacypreserving data analytics in 6G user-centric networks. By deploying AI / ML models at the UE level and generating analytics reports without sharing raw user data (or sensitive user data or information), user privacy is maintained and the need for user consent in measurement collection is potentially eliminated. In other words, the user data is kept at the UE and only analytics of this data (or some form of data representation, reshaping, reformulation, randomization, anonymization, and / or other format of the actual data or information) is shared with the network. For example, the analytics (i.e. all or part of the analytics maybe shared with at least one network entity (or network function). This entity (or function) can be a newly defined entity (or function) or existing entity (or function). Embodiments of the invention provide a method to provide user level analytics reporting from the UE to the network to enable more intelligent user-centric optimization decisions for a fully autonomous user-centric network. A key feature of embodiments of the invention is the ability to use key information from the UE, such as precise location, speed as well as collected measurements and produce accurate analytics reports while preserving the UE privacy, i.e. not sending UE data itself to the network but sending analytics reports only. Embodiments of the invention provide Al models residing at the UE and utilize sensitive UE information such as precise location information alongside measured network data to provide UE level analytics to the network when required while preserving user privacy. Methods also relate to how such Al models can be trained and analytics reports can be exchanged with the network using trained models inference. Multiple analytics report cases are also introduced alongside with their sample reporting criteria. Although a few preferred embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes and modifications might be made without departing from the scope of the invention, as defined in the appended claims. For a better understanding of the invention, and to show how embodiments of the same may be carried into effect, reference will now be made, by way of example only, to the accompanying diagrammatic drawings in which: Figure 1 shows an example of coordination between NWDAF, gNB and MDAS (MDA MnS) producer, according to an embodiment of the invention; Figure 2 shows example Signaling for Local Model Training at UE according to an embodiment of the invention; and Figure 3 shows example Signaling for Model Delivery and Analytics Configuration at UE according to an embodiment of the invention. Embodiments of the present invention comprise one or more of the following features: a) A trained AI / ML model that is capable of performing data analytics based on the specific user data available at the UE level. This model is deployed at the UE and can be fetched (or requested or downloaded) from the network. b) A data analytics function located at the UE that uses the deployed AI / ML model to analyze user data and generate analytics reports. c) The network configures the UE to generate analytics reports using the data analytics function. d) The network may provide other assistance information to assist the UE (and / or data analytics function) in generating data analytics. e) The network may provide specific configurations, parameters, conditions, policies, and / or other information to help in the generation of data analytics. In related example, the network provide information related to the use case (or deployment scenario, model functionality, or other information) on which the UE need to provide (or generate analytics). f) The network may provide assistance information on the validity time g) (and / or location / area) of the generated data analytics. For example, the validity time a given set of analytics in a given cell / tracking area. h) The network provides information on the requirements of data analytics for the UE (and / or data analytics function). For example, size of data analytics, latency of data analytics (or End-to-End delays), security requirements, accuracy, Key Performance Indicators (KPIs), other requirements. i) The network may also activate / deactivate or start, pause or stop data analytics process at the UE. For example, the network (e.g. gNB or NG-RAN) sends a message to start the data analytics procedure in the UE. Additionally, the network (e.g. gNB or NG-RAN) may include information related to the data analytics (as mentioned above) and / or data analytics session ID. For example, the start, pause or stop could be indicated using a flag (included in an existing or newly defined RRC message) or a new IE (DataAnalytics IE = Enumerated {start, pause, stop, ...}, or Enumerated {activate, deactivate, ...}, or Data Analytics Activation IE, Data Analytics Deactivation IE, etc.. In another example, the indication can be in a newly defined RRC messages, Data Analytics Start Indication or Data Analytics Stop Indication, or Data Analytics Activation or Data Analytics Deactivation, other. j) The network may indicate the UE (or group of UEs) the failure of a given Data Analytics Session (or procedure) to the UE and / or other network entities (or functions). This indication may include the session ID and / or other assistance information related to this session. Additionally, the network may include a new cause value of the failure, a has failed for a UE. k) A communication module that enables the UE (or a group of UEs) to share securely the generated data analytics reports to the network without sharing raw user data, thus preserving user privacy. I) A network-side module that receives and processes the analytics reports, allowing the network to benefit from UE-level data analytics without compromising user privacy. m) A enhanced MDT procedure (or any other user measurements) that takes advantage of the privacy-preserving data analytics function at the UE level, eliminating (fully or partially or in some cases) the need for the network to obtain (or request or have as preconfigured) user consent to request and / or handle AI / ML data and / or measurements and enabling a wider range of UEs to participate in AI / ML measurements collection. n) The network and UE (or group of UEs) can exchange the generated data analytic using o Existing RRC (and / or NAS) signaling / messages, lEs, and / or via system information broadcast (periodic and / or on-demand) or o Newly defined RRC (and / or NAS) signaling / message, lEs, and / or via system information broadcast (periodic and / or on-demand) and / or newly defined SIBs, other. o A mix of the above. Note that throughout this application, step order and name only show one example of the of possible a number of solutions, method or procedure. In another example, other steps maybe added, modified, removed, and / or their order is changed. Additionally, all steps and / or message names may be replaced with other suitable names without changing the inherent meaning. Certain embodiments relate to user consent related to analytics: • In one embodiment, the network may not require to obtain a user consent for a given set of analytics (or other form of data collected at the UE) or part of the analytics, before requesting the analytics from the UE (or group of UEs) at a given time and / or location. • In another related embodiment, the network may require to obtain user consent for a given set of analytics (or other form of data collected at the UE) or part of the analytics, before requesting the analytics from the UE (or group of UEs) at a given time and / or location. • In another example, the network entity (or network function) may require to obtain a user consent on sharing the analytics with another network entity (or network function). • In another example, the network entity (or network function) may require to obtain a user consent on handling or processing (including extraction of information or update or modification of information, etc.) user analytics. • In another example, the network entity (or network function), that acquired the user analytics from the desired UE (or group of UEs), may verify (or request) the presence (or availability or existence) of user consent, for handling theses analytics, from at least another entity (or function). For example, target entity (e.g. target gNB, NG-RAN, eNB, other Core network entities or functions), before forwarding these analytics to target entity (or function). Certain other embodiments relate to UE capabilities related to data analytics: • In one embodiment, the UE indicates its capability to support data analytics to the network. • In another embodiment, the UE indicates it capability to support data analytics for a given use case, data type, configuration, scenarios, other. • In one example, the UE capability to support data analytics can be exchanged with the network during initial access and / or provided to the network at a later stage, based on (or following) a request from the network. Embodiments of the invention assume an enhanced version of MDT measurements (or any other user measurements) enabled at the UE, where the UE records and stores historical MDT measurement data. Enhanced MDT measurements are initiated with a measurement configuration message for example “LoggedMeasurementConfguration_RRC_Connected” message (or any other suitable message or Information Element, IE) sent from gNB to the UE(s) as shown in Figure 2. Here, Al model training at the UE is not limited to enhanced MDT measurements in RRC_CONNECTED state, but it may also involve any other UE measurements e.g. logged measurements in RRCJDLE and RRC-INACTIVE states or precise location information and UE speed. In embodiments of this invention, a data analytics function located at the UE is provided, where Al models can be trained locally using only the UE's data. An example of signaling for Al model training at the UE using local data is illustrated in Figure 2. When needed, the gNB can instruct individual UEs or groups of UEs with specific common attributes (such as geographical area, estimated speed, service type, etc.) to train certain Al models in order to generate specific analytics information. Upon receiving the model training request (including specific Model ID) from the gNB, the UE initiates the training process using its locally stored measurement data. Once the training is completed or fails for any reason, the UE sends a "training complete / failure" report back to the gNB (for example via RRC signaling or new signalling). Training failure may include a cause value for example “Not supported” (or any other cause values). Alternatively, Al models providing user-level analytics can be trained at a dedicated Network Function (NF) or a new logical entity using multiple UE data to achieve globally trained models. To train such global models, user data from UEs where consent is obtained can be used at this NF or logical entity to train models centrally. Additionally, Federated Learning (FL) techniques can be used to train a global model while preserving user privacy. This NF or logical entity can coordinate FL with a selection of UEs, where each UE uses its own data to train a local model and exchange model parameters with the NF or logical entity to update a global model. This way, UEs do not have to share actual data and hence preserve privacy while training a global model. In embodiments of this invention, the UE(s) may utilize locally trained models or request / download globally trained models from an existing or new NF or logical unit using OAM for inference. A procedure for configuring UE analytics reporting using RRC messages, creating new IE within an existing RRC message or new messages is provided. This enables user privacy preservation, as the local UE data required for inference stays at the UE, and only analytics information is shared with the network. In this way, globally trained models can be used by UEs that do not give consent for sharing their data with the network. Any network function may subscribe to UE analytics data generated from the UE and request the gNB to configure UE analytics reporting for a group of selected UEs. For example, an Al solution residing at the gNB (e.g., load balancing or energy saving) may require UE analytics for Al-assisted improved automated decisions. Below is an example procedure / message exchange protocol for configuring UE analytics reporting using RRC messages: 1) NF Subscription: a) The network function (e.g., Al for energy saving) sends a subscription request to the gNB, specifying the desired UE analytics data and a group of UEs based on common attributes (geographical area, estimated speed, service type, etc.). b) The gNB acknowledges the subscription request. 2) Model Training or Downloading: a) If the UE requires a globally trained model, it requests the model from the network using OAM, referring to a specific model using a "Model ID". b) The network sends the model to the UE. c) If the UE requires a locally trained model, it trains the model using its locally stored measurement data. 3) RRC Connection Reconfiguration for Analytics Reporting: a) The network (e.g., gNB or NG-RAN) initiates the configuration process by sending an RRC message (e.g., RRCConnectionReconfiguration or another newly defined message) to the selected UE(s). This message includes a new information element (IE), such as "UEAnalyticsConfigRequest" or another newly defined IE name, which contains details about: i) Desired Data analytics IDs. ii) Reporting configurations, such as periodic or event-triggered reporting. iii) Model parameters (e.g., globally trained model identifier or local training requirements). iv) Use case information. v) Validity time and location / area constraints. vi) Size and latency requirements. vii) Security requirements. viii) Desired accuracy or other performance metrics. ix) KPIs. x) Any other relevant parameters or assistance information. b) Upon receiving the RRC message, the UE acknowledges the configuration by sending an RRC message (e.g., RRCConnectionReconfigurationComplete or another newly defined message) back to the gNB. This message includes a new IE, such as "UEAnalyticsConfigResponse" or another suitable IE, which contains the UE's capabilities for supporting the requested analytics reporting configuration, such as supported AI / ML models, reporting triggers, and compliance with the specified requirements. c) Based on the UE's capabilities reported in the previous step, the network may send a new or updated "RRCConnectionReconfiguration" message (or another newly defined message) to the UE to finalize the analytics reporting configuration. This message may include an updated "UEAnalyticsConfigRequest" IE (or another suitable IE) with adjusted reporting parameters, specific instructions for initiating AI / ML model training, fetching the required model(s), or initiating analytics reporting, and any updated assistance information. 4) Analytics Generation: a) The UE starts generating analytics reports based on the received configuration instructions, taking into account the specified use case, validity time, location / area, and other requirements. b) Depending on the configuration, the UE may initiate AI / ML model training, fetch the required model(s) from the network or an external source, or directly start analyzing local user data. c) The UE generates the requested analytics using either the locally trained model or the downloaded globally trained model. d) The UE stores the generated analytics locally and prepares them for reporting. 5) Analytics Reporting: a) Based on the specified reporting configuration, the UE sends the generated analytics to the gNB using an RRC message (or another newly defined message), such as a new IE:"UEAnalyticsReport" (or any other suitable IE name), including: i) Data Analytics ID. ii) Timestamp of the generated analytics. iii) The generated analytics data. b) The network (e.g., gNB, NG-RAN) processes the received analytics and shares the information with the subscribed network function. 6) If the network needs to modify the analytics reporting configuration, it can send an updated message, such as "RRCConnectionReconfiguration" or another suitable message, with a new or modified IE, such as "UEAnalyticsConfigRequest", to the UE. 7) The UE responds with a new message, such as "RRCConnectionReconfigurationComplete" or another suitable message, containing an updated IE, such as "UEAnalyticsConfigResponse", to acknowledge the updated configuration. 8) The network can also send control messages by including a new IE, such as "UEAnalyticsControl" or another suitable IE, in the "RRCConnectionReconfiguration" message (or another newly defined message), to activate, deactivate, start, pause, or stop the data analytics process at the UE. 9) The UE responds with a corresponding acknowledgment message, "RRCConnectionReconfigurationComplete" (or another newly defined message), containing a new IE ACK message, such as "UEAnalyticsControlAck" or another suitable IE name. 10) Decision-making by Network Function: The network function utilizes the received analytics for Al-assisted decision-making in SON use cases, such as load balancing or energy efficiency. This detailed procedure set out above ensures that only analytics information is shared with the network while preserving user privacy, as the local UE data required for inference remains at the UE. This example message exchange procedure is depicted in Figure 3. With the use of embodiments of this invention, UE analytics generated from the UE can be used in any NF or logical function in the network to make enhanced decisions. In the following, example use cases are provided where UE analytics generated from UE can significantly improve decisions of network functions. An Al model at the UE can predict required handover (HO) earlier, making use of learnt UE behaviour from a trained Al model, the local MDT measurements, UE speed and precise UE location. The Al model can predict HO performance (failure rate) and the expected Quality of Service (QoS) for the UE at each potential target cell based on historical performance of the UE used fortraining. In the case of globally trained model, trained model will learn from other UEs with the similar behaviour (location, speed, 5G QoS Identifier (5QI) etc). Predicting the handover earlier can increase the handover success rate as HO procedure can be initiated earlier. Moreover, predicted handover analytics produce target cells taking into account predicted UE QoS will help maintain and improve UE QoS at the target cell after the handover, avoiding any degradation as a result of the handover. Additionally, it can prevent ping-pong handovers for high speed UEs where the predicted time on the target cell is very short (i.e. a small cell target is not selected for handover). In an embodiment, only the list of potential target cell IDs are shared with the network but other information like local MDT measurement, UE speed and precise location and expected QoS etc are not shared with the network but used for Al model inference within the UE to produce target cells for handover. Consequently, UE privacy is respected while making use of the UE specific data for handover prediction. UEs can be configured to perform model inference periodically or based on a triggering criteria such as source cell Reference Signal Received Power (RSRP) <threshold. The UE can then report back analytics information to the network based on a separate triggering criteria e.g. whenever there is an expected handover. Analytics report can include a list of target cells for handover. When a cell is highly loaded, one goal is to offload some of the UEs to lightly loaded cells for load balancing. It is often a difficult task to decide for which UEs to initiate handover to relevant target cells, where the target cells are not highly loaded and can still satisfy the QoS requirements for the UEs. When the cell load is above a threshold, the cell can enable mobility analytics for all UEs connected to the cell using the procedure defined earlier. Mobility analytics can include potential target cells for the UE which can satisfy QoS requirements but not necessarily the best serving cell. Target cells are identified based on UE analytics at the UE, using UE current / predicted location, speed, 5QI and QoS demand and learnt predicted QoS and predicted cell load at each potential target cell for the current / predicted UE behaviour. Mobility analytics provide a list of target cells suitable for handover for load balancing or energy saving purposes. Once mobility analytics is received from all UEs in RRC_CONNECTED state at the gNB, the gNB can then choose the UEs for handover based on target cell current load values, i.e. in the load balancing case, target cells are chosen if the target cell load is less than a threshold. Similar to load balancing, mobility analytics reporting can be used for the energy efficiency use case where a cell is lightly loaded, and the goal can be to handover the remaining connected UEs to other cells so that the cell can be switched off for energy saving. Energy Saving function can process the received mobility analytics reports to understand if all UEs have an available target cell to handover while maintaining the QoS and hence the cell can be switched off after the handover. UEs can send mobility analytics report periodically until load balancing / energy saving load conditions disappear, and the cell can initiate the procedure defined earlier to stop mobility analytics reporting from the UEs. In another embodiment, UEs can be configured to make regular inference on an Al model to detect any anomalies and send anomaly detection analytics report to the network when an anomaly is detected. Al models trained with historic UE behaviour will be able to differentiate abnormal UE behaviour such as unexpected QoS experience and report back to the network. These analytics can help to find abnormal UEs or fault finding at any network function. At least some of the example embodiments described herein may be constructed, partially or wholly, using dedicated special-purpose hardware. Terms such as ‘component’, ‘module’ or ‘unit’ used herein may include, but are not limited to, a hardware device, such as circuitry in the form of discrete or integrated components, a Field Programmable Gate Array (FPGA) or Application Specific Integrated Circuit (ASIC), which performs certain tasks or provides the associated functionality. In some embodiments, the described elements may be configured to reside on a tangible, persistent, addressable storage medium and may be configured to execute on one or more processors. These functional elements may in some embodiments include, by way of example, components, such as software components, object-oriented software components, class components and task components, processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. Although the example embodiments have been described with reference to the components, modules and units discussed herein, such functional elements may be combined into fewer elements or separated into additional elements. Various combinations of optional features have been described herein, and it will be appreciated that described features may be combined in any suitable combination. In particular, the features of any one example embodiment may be combined with features of any other embodiment, as appropriate, except where such combinations are mutually exclusive. Throughout this specification, the term “comprising” or “comprises” means including the component(s) specified but not to the exclusion of the presence of others. Attention is directed to all papers and documents which are filed concurrently with or previous to this specification in connection with this application and which are open to public inspection with this specification, and the contents of all such papers and documents are incorporated herein by reference. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and / or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and / or steps are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise. Thus, unless expressly stated otherwise, each feature disclosed is one example only of a generic series of equivalent or similar features. The invention is not restricted to the details of the foregoing embodiment(s). The invention extends to any novel one, or any novel combination, of the features disclosed in this specification (including any accompanying claims, abstract and drawings), or to any novel one, or any novel combination, of the steps of any method or process so disclosed.

Claims

1. A method of operating a User Equipment, UE, communicatively coupled to a telecommunication network, wherein the UE is arranged to make certain measurement in relation to the network, and to store the measurement results locally to the UE, wherein the UE is further arranged to not send the measurement results onto the network, thereby preserving privacy associated with the measurement results and to locally perform data analytics, inference or prediction using the measurement results.

2. The method of claim 1 wherein the analytics are obtained using an Artificial Intelligence, Al, model that is trained locally to the UE or globally by the network, but using local measurement data.

3. The method of any preceding claim wherein the measurements include Minimization of Drive Tests, MDT, data.

4. The method of any preceding claim wherein the UE is configured by the network to generate particular analytics reports by providing one or more of specific configurations, parameters, conditions and policies, to assist the UE in the generation of data analytics.

5. The method of claim 4 wherein the particular analytics reports include one of particular measurements, particular analytics and particular inference and predictions.

6. The method of any preceding claim wherein the network provides the UE with assistance information on a validity time.

7. The method of any preceding claim wherein the network provides information on requirements of the data analytics for the UE, said requirements including one or more of: size of data analytics, latency of data analytics, security requirements, accuracy and Key Performance Indicators, KPIs.

8. The method of any preceding claim wherein the network may activate / deactivate or start, pause or stop the data analytics process at the UE.9 The method of any preceding claim wherein the UE utilises a communication module arranged to enable the UE, or a group of UEs, to share securely the generated data analytics reports to the network without sharing raw user data, thus preserving user privacy.

10. The method of any preceding claim wherein the UE provides user consent to the network in connection with collecting analytics from the UE or sharing analytics beyond the UE.5 11. The method of any preceding claim wherein the UE indicates to the network itscapability to support analytics to the network, such that the network is able to configure the data analytics collection and reporting, taking into consideration the indicated capability.

12. A method of operating a telecommunication network including the step of configuring a 10 UE to locally perform data analytics using certain measurement results, and to share a resulting data analytics report with the network without compromising user privacy.

13. The method of claim 12 comprising the use of a module to receive data analytics reports from the UE.1514. A UE arranged to perform the method of any of claims 1 to 11.

15. A telecommunication network arranged to perform the method of any of claims 12 to 13.

Images