Enhancing UE capability
Patent Information
- Application Number
- PCT/EP2024/061333
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-25
- Filing Date
- 2024-04-25
- Publication Date
- 2024-11-28
Smart Images

Figure EP2024061333_28112024_PF_FP_ABST
Abstract
Description
[0001]Enhancing UE capability Technical Field The present disclosure relates to UE capability enhancement. Abbreviations 3GPP 3rdGeneration Partnership Project 5G / 6G / 7G 5th / 6th / 7thGeneration AI Artificial Intelligence AIML Artificial Intelligence – Machine Learning AoD Angle of Departure BW Bandwidth CA Carrier Aggregation CIR Channel Impulse Response CSI Channel State Information gNB Next Generation (5G) Node B ID Identifier KPI Key Performance Indicator LMC Location Management Component LMF Location Management Function LOS Line of Sight LPHAP Low-Power High-Accuracy Positioning LPP LTE Positioning Protocol MAC Medium Access Control ML Machine Learning NLOS Non Line of Sight NR New Radio NRPPa New Radio Positioning Protocol A NW Network NWDAF Network Data Analytics Function PRS Positioning Reference Signal PRU Positioning Reference Unit RF Radio Frequency RRC Radio Resource Control RSTD Reference Signal Time Difference SIB System Information Block TDOA Time Difference of Arrival ToA Time of Arrival UE User Equipment Glossary (partly taken from 3GPP R1-2205695 and 3GPP R1-2300603) AI / ML model testing A subprocess of training, to evaluate the performance of a tuning of the model. UE-side (AI / ML) model An AI / ML Model whose inference is performed entirely at the UE Network-side (AI / ML) An AI / ML Model whose inference is performed entirely at model the network One-sided (AI / ML) A UE-side (AI / ML) model or a Network-side (AI / ML) model model Model monitoring A procedure that monitors the inference performance of the AI / ML model Functionality A process / method of identifying an AI / ML functionality for identification the common understanding between the NW and the UE Note: Information regarding the AI / ML functionality may be shared during functionality identification. Direct AIML positioning Computing the location (e.g. of a UE) AIML assisted Extract and potentially pre-evaluate positioning positioning measurements and provide them to another entity to compute the location (e.g. of a UE) Background In several AI / ML techniques for positioning purposes the training is carried out at a central ML unit, referred to henceforth as central unit (or central ML unit). The central unit may be, for example, the location management function (LMF) residing at the core network, which is conventionally the responsible entity for coordinating positioning. As another example, the central ML unit may be the 5G network data analytics function (NWDAF), which runs data analytics to generate insights and take action to enhance user experience, including positioning use cases. The training in the central ML unit is typically done according to the following process: - A set of data collection devices is deployed in carefully chosen locations. An example of such data collection devices are the so-called positioning reference units (PRUs): PRUs are reference devices at known locations, which can provide ground truth measurements. PRUs may be suitable to accommodate real-world measurements and label measurement data for AIML based learning. These PRU data can be used to refine the location of other target UEs in the same area, thereby increasing their positioning accuracy. Henceforth data collection devices are referred to as PRUs for simplicity, yet this can be generalized to any type of data collection device. A PRU may be or may include, for example, a UE or other node device that may have one or more known characteristics (e.g., a known position and / or known LOS / NLOS classification, and / or a known positioning measurement based on certain reference signals), which may perform positioning measurement(s) based on reference signals. - PRUs conduct field positioning measurements and report the measurements to the central unit. - In addition to measurements collected from UE / PRU or instead thereof, the central unit may use emulation tools to generate (emulated) positioning measurements. For example, the central unit may use simulated / synthetic data for training the initial model. - The central unit combines the above positioning measurements to train a generic localization ML model. - The AIML model is deployed at the network entities running ML processes and / or algorithms. Henceforth, such entities are referred to as host types. Host types carrying out ML processes can be the target UE, the PRUs, LMF, and potentially the radio access network (e.g., the gNB and or a location management component - LMC) to enhance the positioning accuracy. Summary It is an object to improve the prior art. According to a first aspect, there is provided an apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: preparing a functionality; and at least one of providing the prepared functionality to a terminal; or broadcasting the prepared functionality or an availability indication indicating that the functionality is prepared; wherein the functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following direct positioning of the terminal, or assisted positioning of the terminal; and wherein the preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; - validating the trained ML model; or - updating the ML model. The training of the ML model may be based on data received from the terminal. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving respective data from each of plural terminals; training the ML model based on the data received from the plural terminals. The instructions, when executed by the one or more processors, may further cause the apparatus to perform determining the set of conditions; sending the set of conditions along with the at least one of the broadcasting and the providing the functionality. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving a request to provide the prepared functionality to the terminal; providing the prepared functionality to the terminal in response to the request; inhibiting providing the prepared functionality to the terminal in absence of the request. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving, from the terminal, the set of the one or more conditions. The instructions, when executed by the one or more processors, may further cause the apparatus to perform requesting the terminal to make the prepared functionality executable. The instructions, when executed by the one or more processors, may further cause the apparatus to perform continuing the preparing the functionality after the functionality has obtained an updated ML model; and at least one of the following providing the updated functionality or an incremental update of the functionality to the terminal; or broadcasting the updated functionality or the incremental update of the functionality or an update availability indication indicating that the updated functionality or the incremental update is available. According to a second aspect of the invention, there is provided an apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: receiving a request for a functionality from a terminal; providing assistance to the terminal in response to the request such that the terminal has a capability to prepare the functionality; wherein the functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning of the terminal, or assisted positioning of the terminal; and wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; and - validating the trained ML model; . The providing the assistance may comprise configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving a set of conditions of the terminal; determining whether the functionality is executable by the terminal under the set of conditions; inhibiting providing the assistance based on determining that the functionality is not suitable executable by the terminal under the set of conditions. According to a third aspect of the invention, there is provided an apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the desired functionality is available or not; requesting to provide the desired functionality based on determining that the capability is not available; receiving the desired functionality; making the received functionality executable ; wherein the functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; the functionality further comprises at least one of the following: direct positioning or assisted positioning; The required capability may comprise at least one of the following: - a capability to collect data for at least one of the training, the testing, and the validating of the ML model; - a processing power for at least one of the training, the testing, and the validating of the ML model; - an availability of the processing power for at least one of the training, the testing, and the validating of the ML model; - a memory for at least one of the training, the testing, and the validating the ML model; - an availability of the memory for at least one of the training, the testing, and the validating the ML model; or - a capability to obtain a ground truth for validating the ML model. The instructions, when executed by the one or more processors, may further cause the apparatus to perform determining a set of one or more conditions for making the functionality executable; sending the set of the one or more conditions along with the requesting the functionality. The instructions, when executed by the one or more processors, may further cause the apparatus to perform requesting the functionality in response to receiving a broadcasted availability indication indicating that the functionality is prepared The instructions, when executed by the one or more processors, may further cause the apparatus to perform providing data for preparing the functionality before the functionality is implemented, wherein the data are based on a performed measurement ; using the data for executing the functionality when the functionality is made executable. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving, along with the receiving the functionality, a set of one or more conditions for enabling the method executable; determining by verifying whether or not the one or more conditions are fulfilled. The instructions, when executed by the one or more processors, may further cause the apparatus to perform receiving an incremental update of the functionality; updating the functionality based on the incremental update. The prepared functionality may be received by receiving a broadcasted message comprising the functionality. The instructions, when executed by the one or more processors, may further cause the apparatus to perform the checking whether the required capability to prepare the functionality is available or not within a predetermined time. The instructions, when executed by the one or more processors, may further cause the apparatus to perform inhibiting the providing the data when the functionality is made executable. The instructions, when executed by the one or more processors, may further cause the apparatus to perform determining if the functionality is executable upon receiving the functionality; inhibiting the making the functionality executable based on determining that the functionality is not executable. According to a fourth aspect of the invention, there is provided an apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the functionality is available or not; requesting to provide assistance to make the capability available based on determining that the capability is not available; receiving the assistance; preparing the functionality based on the assistance; wherein the functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning or assisted positioning; wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; and validating the trained ML model. The providing the assistance may comprise configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal. The preparing the functionality may comprise additionally updating the trained ML model. The instructions, when executed by the one or more processors, may further cause the apparatus to perform informing on a set of one or more conditions for making the functionality executable; wherein the functionality is suitable for being executable under the condition that the set of conditions is fulfilled. According to a fifth aspect, there is provided a method comprising: preparing a functionality; and at least one of providing the prepared functionality to a terminal; or broadcasting the prepared functionality or an availability indication indicating that the functionality is prepared; wherein the functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following direct positioning of the terminal, or assisted positioning of the terminal; and wherein the preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; - validating the trained ML model; or - updating the ML model. The training of the ML model may be based on data received from the terminal. The method may further comprise receiving respective data from each of plural terminals; training the ML model based on the data received from the plural terminals. The method may further comprise determining the set of conditions; sending the set of conditions along with the at least one of the broadcasting and the providing the functionality. The method may further comprise receiving a request to provide the prepared functionality to the terminal; providing the prepared functionality to the terminal in response to the request; inhibiting providing the prepared functionality to the terminal in absence of the request. The method may further comprise receiving, from the terminal, the set of the one or more conditions. The method may further comprise requesting the terminal to make the prepared functionality executable. The method may further comprise continuing the preparing the functionality after the functionality has obtained an updated ML model; and at least one of the following providing the updated functionality or an incremental update of the functionality to the terminal; or broadcasting the updated functionality or the incremental update of the functionality or an update availability indication indicating that the updated functionality or the incremental update is available. According to a sixth aspect of the invention, there is provided a method comprising: receiving a request for a functionality from a terminal; providing assistance to the terminal in response to the request such that the terminal has a capability to prepare the functionality; wherein the functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning of the terminal, or assisted positioning of the terminal; and wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; and - validating the trained ML model; . The providing the assistance may comprise configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal. The method may further comprise receiving a set of conditions of the terminal; determining whether the functionality is executable by the terminal under the set of conditions; inhibiting providing the assistance based on determining that the functionality is not suitable executable by the terminal under the set of conditions. According to a seventh aspect of the invention, there is provided a method comprising: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the desired functionality is available or not; requesting to provide the desired functionality based on determining that the capability is not available; receiving the desired functionality; making the received functionality executable ; wherein the functionality comprises a configuration and a method executable by an apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; the functionality further comprises at least one of the following: direct positioning or assisted positioning; The required capability may comprise at least one of the following: - a capability to collect data for at least one of the training, the testing, and the validating of the ML model; - a processing power for at least one of the training, the testing, and the validating of the ML model; - an availability of the processing power for at least one of the training, the testing, and the validating of the ML model; - a memory for at least one of the training, the testing, and the validating the ML model; - an availability of the memory for at least one of the training, the testing, and the validating the ML model; or - a capability to obtain a ground truth for validating the ML model. The method may further comprise determining a set of one or more conditions for making the functionality executable; sending the set of the one or more conditions along with the requesting the functionality. The method may further comprise requesting the functionality in response to receiving a broadcasted availability indication indicating that the functionality is prepared The method may further comprise providing data for preparing the functionality before the functionality is implemented, wherein the data are based on a performed measurement ; using the data for executing the functionality when the functionality is made executable. The method may further comprise receiving, along with the receiving the functionality, a set of one or more conditions for enabling the method executable; determining by verifying whether or not the one or more conditions are fulfilled. The method may further comprise receiving an incremental update of the functionality; updating the functionality based on the incremental update. The prepared functionality may be received by receiving a broadcasted message comprising the functionality. The method may further comprise the checking whether the required capability to prepare the functionality is available or not within a predetermined time. The method may further comprise inhibiting the providing the data when the functionality is made executable. The method may further comprise determining if the functionality is executable upon receiving the functionality; inhibiting the making the functionality executable based on determining that the functionality is not executable. According to an eighth aspect of the invention, there is provided a method comprising: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the functionality is available or not; requesting to provide assistance to make the capability available based on determining that the capability is not available; receiving the assistance; preparing the functionality based on the assistance; wherein the functionality comprises a configuration and a method executable by an apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning or assisted positioning; wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; and validating the trained ML model. The providing the assistance may comprise configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal. The preparing the functionality may comprise additionally updating the trained ML model. The method may further comprise informing on a set of one or more conditions for making the functionality executable; wherein the functionality is suitable for being executable under the condition that the set of conditions is fulfilled. According to a ninth aspect of the invention, there is provided a computer program product comprising a set of instructions which, when executed on an apparatus, is configured to cause the apparatus to carry out the method according to any of the fifth to eighth aspects. The computer program product may be embodied as a computer-readable medium or directly loadable into a computer. According to some example embodiments, at least one of the following advantages may be achieved: ^ UE may execute functionalities it was previously not prepared for; ^ Restrictions for specific UEs may be taken into account in providing the functionality; ^ Network may control whether or not a certain functionality is available at a UE because it may or may not provide the prepared functionality to the UE ; ^ Signalling overhead may be reduced: ^ Collect much larger amounts of training data than each NR UE could. ^ Train better models since the computational complexity is not as limited as the UE’s ^ Tailor the model on demand, whenever the UE requires it, using UE’s constraints and hardware limitations. It is to be understood that any of the above modifications can be applied singly or in combination to the respective aspects to which they refer, unless they are explicitly stated as excluding alternatives. Brief description of the drawings Further details, features, objects, and advantages are apparent from the following detailed description of the preferred example embodiments which is to be taken in conjunction with the appended drawings, wherein: Fig.1 shows a message flow according to some example embodiments; Fig.2 shows an apparatus according to an example embodiment; Fig.3 shows a method according to an example embodiment; Fig.4 shows an apparatus according to an example embodiment; Fig.5 shows a method according to an example embodiment; and Fig.6 shows an apparatus according to an example embodiment; Fig.7 shows a method according to an example embodiment; Fig.8 shows an apparatus according to an example embodiment; Fig.9 shows a method according to an example embodiment; and Fig.10 shows an apparatus according to an example embodiment. Detailed description of certain example embodiments Herein below, certain example embodiments are described in detail with reference to the accompanying drawings, wherein the features of the example embodiments can be freely combined with each other unless otherwise described. However, it is to be expressly understood that the description of certain example embodiments is given by way of example only, and that it is by no way intended to be understood as limiting the disclosure to the disclosed details. Moreover, it is to be understood that the apparatus is configured to perform the corresponding method, although in some cases only the apparatus or only the method are described. Positioning based on AIML (“AIML positioning”) is considered to be more performant than conventional positioning without using AIML. However, AIML positioning is expected to be computationally more challenging to the UE than the conventional methods. In the same vein, AIML direct positioning is inherently more complex than the AIML assisted positioning, since the task the former solves (i.e., computing the UE location) is harder to solve than the latter task (i.e. extract positioning measurements (and potentially pre-evaluate them). Thus, not all UEs may support both types of AIML positioning. Some examples of UEs not supporting both types of AIML positioning may be ^ RedCap devices (i.e., devices with reduced capability). Some of these devices may implement both direct positioning and assisted positioning, but they are likely not capable of performing model update on the device. ^ LPHAP devices are battery limited, yet they require high accuracy positioning. Therefore they can greatly benefit from AIML positioning methods, as long as these methods are being delivered to the device at no (or hardly any) additional cost (in particular battery consumption). I.e., it is recommendable that the LPHAP device does not need to do model tuning, model update, storing large amounts of training data locally, etc. In particular, some UEs may not be capable to prepare AIML positioning, where preparing comprises at least one of the following (or all of the following): - collecting data for training a ML model; - training the ML model; - testing the ML model; - validating the trained ML model; or - updating the ML model. Data collection for training is lengthy and a complex problem. Therefore, it may consume quite a portion of entity’s (UE, gNB, LMF) resources. Therefore, one could consider to perform data collection for training, model tuning, model update and also the positioning at the central unit only such that the UEs contribute to assisted positioning only. However, if all UEs require assistance signaling and procedure to facilitate generation / collection of training data and ground truths, the network overhead might be tremendous. Transfer of training data between network entities could lead to significant overhead and result in waste of network resources. Positioning AIML models are typically specific to a region of interest and may not work in all regions in the network. Therefore, switching from one positioning method (e.g., AIML assisted positioning) to another positioning method (e.g., AIML direct positioning) when the UE roams from one region to another region may require re-training of a new model (or models) associated with the other positioning method. For proprietary models (e.g., UE-vendor specific models), models are typically trained outside 3GPP network and may consider different datasets (RF environment, synthetic data), which may not be valid for the given scenario or region of interest. In such conditions re-training or re-tuning of models is mandatory to meet the positioning requirements. According to some example embodiments, the network (represented e.g. by the central unit) may enhance the capabilities of a UE. For example, the network may enable a UE supporting only assisted positioning but being capable of supporting direct positioning (at least under certain conditions) to support direct positioning, too (and vice versa). Thus, the system of network and UE may reap the full benefits of an AIML framework. Some example embodiments provide a framework for enhancing an NR UE’s capabilities for AIML positioning that allows the UE to enrich its original set of capabilities with a new capability (or plural new capabilities) developed by the NW and adapted to the constraints of the respective UE. Thus, the NW may switch the configuration of a UE between: a. AIML direct positioning and b. AIML assisted positioning for any UE, although the UE may not initially possess both capabilities a) and b). In some example embodiments, the UE has to agree to undergo a NW-led capability enhancement to acquire the desired capability. According to some example embodiments, the ML model is prepared (in particular: created and trained and typically also verified) at the central unit before it is provided to the UE for running inference from the ML model. Preparing the model at the central unit (i.e. by the network) has typically at least one of the following advantages: o The NW may collect much larger amounts of training data than each UE could do. Additionally, data availability is much better at NW side. o Better models with higher complexity may be trained since the computational capabilities are not as limited as the UE’s o On demand of the UE, the NW may tailor the model such that it fits to constraints of the UE, e.g. in terms of hardware and / or software limitations, and also based on available radio resources. Fig. 1 shows the actions performed according to an example embodiment. In this example embodiment, LMF is the central unit. The actions are as follows: 1: LMF asks UE for its capabilities for positioning, in particular AIML positioning. 2: UE determines its capabilities for AIML positioning. 3: UE determines a AIML positioning functionality not executable by the UE but desired to be executable. 4: If the UE determines a desired but not executable functionality, it checks whether the UE is capable of executing the functionality under the condition that the prepared functionality is provided to the UE. UE may also check whether it is capable to prepare the functionality (at least to collect data, to train, and to validate the ML model on which the functionality is based). Apart from checking whether the UE is capable to prepare the functionality at all, it may check whether it is capable to prepare the functionality within a predetermined time. In the example embodiment of Fig.1, it is assumed that the UE either does not check whether it is capable of preparing the functionality, or that the check for preparing the functionality has a negative result. 5: The UE provides feedback to the request of 1 to the LMF. In particular, the UE indicates the desired but not executable AIML functionality to the LMF if the UE is capable of executing the prepared functionality (according to the check in 4). In addition, the UE may indicate the AIML positioning functionalities executable at the UE. 6: LMF evaluates the report provided by the UE in 5. It determines that the functionality Y is desired but not executable at the UE. 7 and 8: LMF prepares (data collection, training, validation) the functionality for the UE. 9: LMF provides the prepared functionality Y to UE. 10: Optionally, UE checks whether it actually supports the prepared functionality. For example, some portion of the processing power and / or the memory may have been occupied by some other task in the meantime such that the prepared functionality Y is not executable by the UE any more. 11: If the test in 10 is positive (or the test in 10 is not performed), UE indicates to the LMF that the functionality Y is now executable at the UE. 12: Upon receipt of the message 11, LMF determines the configuration for making the functionality Y executable on the UE and provides it to the UE. 13: AIML positioning for the UE may now be based on the functionality Y. In some example embodiments, UE may indicate (request) the set of one or more desired AIML positioning modes and methods without having received the capability request from the central unit before. Accordingly, in such example embodiments, the actions 1, 5, and 6 in Fig. 1 are not performed. In action 6 of Fig.1, in such example embodiments, LMF determines only the one or more desired positioning capabilities. Actions 11 to 13 are optional in such example embodiments. For example, at least actions 12 and 13 may be performed only if LMF wants to receive positioning information for the UE based on the desired positioning functionality. The capability of the UE may include certain conditions which the UE may fulfill in view of its hardware configuration and in view of other tasks to be performed by the UE, such as a maximum amount of processing power or a maximum amount of memory being available for performing the functionality. In some example embodiments, the UE may determine these conditions and inform the LMF accordingly. The LMF may take these conditions into account when it prepares the functionality. The actions of the example embodiment of Fig.1 and a scenario are explained hereinafter again with some different wording. In the scenario, LMF may need to obtain the location of the ^^^UE. In traditional positioning (Rel-17, non-AIML based), the LMF may request the UE to report either direct position information or assistance data, which could aid the LMF to estimate the UE location. Unlike the traditional positioning mechanism (non AIML-based), much of the complexity involved with the AIML based positioning framework is attributed to the functionality training. The functionality here refers to the AIML direct or assisted AIML positioning. ^ As an example, after getting the positioning capability request from the LMF, the UE which may not have the direct AIML positioning functionality, will respond to the LMF request with the capability information. o The UE may use the AIML functionality or classical functionality to determine the list of intermediate features, such as, CIR, AoD, ToA, RSTD, LOS / NLOS indicator, soft information / high resolution of RSTD, etc., using the PRS signal transmitted by multiple gNBs in its neighborhood. o UE may determine supported functionalities and desired functionalities with relevant set of inputs / outputs. ^ UE may indicate the capability including set of available AIML positioning modes including supported and desired functionalities to LMF. ^ LMF determines supported and desired functionalities based on the capability report from the given UE. LMF may determine / configure the configuration parameters for the supported functionality ^ LMF may configure the UE with supported functionality for the ML-enabled positioning. ^ If the UE indicates that it is not capable to prepare a desired functionality (e.g. ID Y) but capable of executing the desired functionality (e.g. ID Y), LMF may locally instantiate the desired functionality (ID Y) and trigger data collection and training based on the existing feedback from multiple UEs in the network. ^ Once the desired functionality is trained / tested and validated at the NW side (e.g., LMF), it provides assistance to UE for enabling / activation of the desired functionality Y. It may include a transfer of the AIML model associated with the functionality Y from LMF to UE. ^ Upon reception of the desired functionality and UE capability, UE may verify that the desired functionality is supported. In some example embodiments, UE may simply assume that the desired functionality is supported because the UE provided its conditions to the LMF which took them into account. ^ UE may indicate the change of the set of available AIML positioning modes and methods to LMF (including the updated set of supported functionalities); the indication may comprise further desired functionalities. ^ Based on the indication from UE, LMF may determine / configure the configurations / parameters for the supported new functionality (e.g., ID Y). Functionality refers to an AI / ML-enabled Feature / FG enabled by configuration(s), where configuration(s) is(are) supported based on conditions indicated by UE capability. in other words, the functionality comprises a configuration of the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions. If the method is executable, then the corresponding functionality would be executable as well. In other words, an executable method implies that the functionality relating to the method is executable as well.” The execution of the method is being realized while the UE has a certain configuration. For the configuration, the UE may have a set of conditions which are UE specific. Therefore, the functionality is the subject of said conditions, too. For example, a configuration is: a neural network with X = 3 hidden layers, where X = 3 is a parameter setting. The UE might have told the NW that it can at most have Xmax = 10 hidden layers, and thus X<Xmax is a condition which the network (central unit) should obey. The UE capability may comprise one or more of the following: - a capability to collect sufficient data for at least one of the training, the testing, and the validating of the ML model. That is, the amount of data which the UE may collect for the preparing the functionality must be sufficient such that the trained ML model is expected to have a desired quality. - a capability to collect appropriate data for at least one of the training, the testing, and the validating of the ML model; That is, the data collected by the UE must be data to be inputted into the ML model. - a processing power for at least one of the training, the testing, and the validating of the ML model; - an availability of the processing power for at least one of the training, the testing, and the validating of the ML model; - a memory for at least one of the training, the testing, and the validating the ML model; - an availability of the memory for at least one of the training, the testing, and the validating the ML model; or - a capability to obtain a ground truth for validating the ML model. In the example embodiment, the desired functionality is prepared by the LMF on request from the UE. In some example embodiments, the desired functionality is prepared independent from any request from a UE and provided to one or more UEs (e.g. by broadcasting). In some of these example embodiments, LMF may fine-tune the prepared functionality according to conditions indicated by a specific UE before it is provided to the specific UE. In some example embodiments, LMF may just broadcast an indication of availability indicating that the prepared functionality is available. Upon receipt of the availability indication, a UE may fetch the prepared functionality from a LMF indicated in the availability indication. For example, LMF may instantiate a AIML model (also denoted “reflection model”) for AIML direct positioning functionality and train the same with the measurement reports provided by multiple UEs. It is up to the network how to implement the AIML model (e.g., number of inputs, number of layers, architecture of the neural network, etc.). Typically, network has the knowledge of all possible inputs, i.e., reports it may collect from the set of multiple UEs. This way, the functionality can be trained with the UE locations of all the multiple UEs and can be trained over a period of time to ensure the generalization of the functionality. Since the functionality is a reflection of UE sided functionality, we refer this as reflection functionality. This new functionality can be visible to the network as new UE capability. The reflection functionality may be a new functionality to the network, too: For example, NW may group the UEs based region of interest (ROI) and capability and create relevant reflection functionality. Direct AIML positioning-1 representing Group 1 – ROI-1, group UE based on existing assisted positioning within in ROI-1 and UE capabilities. Direct AIML positioning-2 representing Group 2 – ROI-2, group UE based on existing assisted positioning within ROI-2 and UE capabilities. During the process of functionality training, the actual ground truth information may be obtained by the traditional positioning estimate (NR RAT-based, and / or non-NR based) using the available subset of measurements reports provided by the UEs. Thus, the functionality training, which is carried out at the LMF, considers the UE locations and channel characteristics experienced by each of the multiple UEs. Since the functionality uses the assistance information as list of inputs, i.e., CIR, PDP, AoD, DL-TDOA, etc., the model can easily be used by the UEs with minimal effort once made available from outside. Furthermore, the effect of any UE impairments may be averaged out due to a large number of training samples. Since the LMF instantiates, in such example embodiments, the AIML model while a corresponding model may be implemented in one or more of the UEs, the AIML model may also be denoted “reflection model”. LMF may provide the AIML model trained at LMF to one or more dedicated UEs. As another option, LMF may broadcast the model such that any UE may use the AIML model. However, since the AIML model may have a large data volume, in some example embodiments, LMF broadcasts just an indication that the functionality is available, e.g. via SIB-XX or via Pos-SIB. Then, each UE may decide whether or not it wants to download the AIML model. If the UE decides to download the AIML model, the UE may send a corresponding request to LMF to receive the AIML model. As another option, the broadcast message may comprise a link to a location from where the UEs may download the AIML model. The location may be considered as belonging to the central unit. When LMF is aware that UE received the AIML model prepared (created, trained, and validated) by LMF (e.g. because UE requested the functionality from LMF), LMF may request to switch to the functionality, e.g. from assisted AIML positioning to direct AIML positioning. LMF may provide to UE a list of applicable inputs to the AIML model. This list may be based on the capability of the UE. Once the functionality is deployed at the UE, initial fine-tuning and inference may be needed to ensure the functionality performance. This is mainly needed to mitigate potential impairments both at the UE and the NW side. As another option, UE(s) may proactively request a new capability / functionality from the network. Based on such requests from (e.g. a subset of) UEs in the given region of interest, NW may instantiate a reflection functionality and train / test / validate it before providing the reflection functionality to the respective UE(s). In some example embodiments, when the UE receives the prepared functionality, it checks whether or not the functionality is actually executable at the UE before it tries to execute the functionality. For example, some of the processing power and / or memory considered to be available for the functionality might be occupied by some other task meanwhile. If the UE observes that its own conditions do not fulfill the requirements of the functionality, the UE does not execute the functionality upon receipt of the prepared functionality. In some example embodiments, before the functionality is executable at the UE, the UE collects data (in particular: performance data) and provides them to the LMF. The LMF uses these data to prepare the functionality. Once the functionality becomes executable at the UE, the UE may use these data on its own, for executing the functionality at the UE. In some example embodiments, the UE does not provide these data anymore to the LMF, thus saving signaling effort. Due to the time varying nature of the wireless environment, the reflection functionality trained at the LMF and used at the UEs in the NW may become outdated after some time. To ensure the desired performance (e.g. estimation accuracy) from the reflection functionality, LMF may continue to train the reflection functionality with the measurement reports from UEs that did not implement the reflection functionality. These UEs may have 1. entered the NW recently after the broadcast of the functionality, or 2. not be capable of using the functionality due to minimum capability requirements, or 3. may use the functionality but nevertheless may provide assistance measurements to monitor the functionality used. In some example embodiments, the UEs using the reflection functionality may not provide any assistance measurements to the LMF to save signaling effort. At certain periodicity, further trained functionality (or an indication of the availability of the further trained functionality) may be broadcasted to all the UEs to update their respective models. In order to reduce the overhead, LMF can provide a differential (incremental) update to reduce the signaling overhead. Thus, the updates can happen at certain periodicity compared to the full functionality transfer to reduce the overhead for the UEs already having deployed the trained functionality. As the functionality is created with the inputs that are already used for the assistance information by the UEs, the processing overhead involved in creating the inputs to the model may be rather small. However, the benefit of using the new direct AIML functionality by the UE with NW assistance can be attributed to the following. 1. Frequent reporting to NW can be reduced and the reporting can be used only when functionality monitoring is requested by the NW 2. Latest location information can be provided to the UE side application layer without NW signaling, thus reducing the location estimate error due to the latency arising from control plane signaling In the above described example embodiments, the UE indicates to the central unit that it is capable of executing the prepared AIML functionality. In some example embodiments, the UE may additionally indicated to the LMF whether or not the UE is capable of preparing the AIML functionality. In some example embodiments, the UE may be generally capable to prepare the AIML functionality (e.g. from a point of view of installed processing power and memory), too, but some information is missing to the UE. In such example embodiments, the central unit may provide assistance information to the UE in order to enable the UE to prepare (collect data, train the model, test the model, and validate the model) the AIML functionality. I.e., in these example embodiments, the preparing may be performed at the UE after the central unit provided some assistance information to the UE. Thus, signaling effort is even further reduced. After the UE prepared the functionality, the UE may execute the functionality. For example, the assistance information may comprise configuring the UE to measure and evaluate a certain reference signal which is used for the training but has not been measured by the UE yet. As another example, the network may reduce the amount of processing power and / or memory of the UE used by some other task executed at the UE such that the conditions of the UE meet the requirements for preparing the functionality. If the central unit learns from the capability information provided by the UE that the UE is not capable of preparing the functionality, it may not provide the assistance information to the UE. Fig.3 shows an apparatus according to an example embodiment. The apparatus may be a central unit (such as LMF or NWDAF) or an element thereof. Fig.4 shows a method according to an example embodiment. The apparatus according to Fig.3 may perform the method of Fig. 4 but is not limited to this method. The method of Fig.4 may be performed by the apparatus of Fig.3 but is not limited to being performed by this apparatus. The apparatus comprises means for preparing 110, means for providing 120, and means for sharing 130. The means for preparing 110, means for providing 120, and means for sharing 130 may be a preparing means, providing means, and sharing means, respectively. The means for preparing 110, means for providing 120, and means for sharing 130 may be a preparer, provider, and sharer, respectively. The means for preparing 110, means for providing 120, and means for sharing 130 may be a preparing processor, providing processor, and sharing processor, respectively. The means for preparing 110 prepares, by a central unit, a functionality (S110). The functionality is based on a ML model. The preparing may comprise creating, training, and verifying the ML model or a subset of these activities. If the functionality is prepared, at least one of the following may be performed: The means for providing 120 may provide the prepared functionality to a terminal for execution by the terminal (S120); or The means for sharing 130 may share (e.g. broadcast) the prepared functionality or an availability indication (S130). The availability indication indicates that the functionality is prepared. Fig.5 shows an apparatus according to an example embodiment. The apparatus may be a central unit (such as LMF or NWDAF) or an element thereof. Fig.6 shows a method according to an example embodiment. The apparatus according to Fig.5 may perform the method of Fig. 6 but is not limited to this method. The method of Fig.6 may be performed by the apparatus of Fig.5 but is not limited to being performed by this apparatus. The apparatus comprises means for receiving 210 and means for providing 220. The means for receiving 210 and means for providing 220 may be a receiving means and a providing means, respectively. The means for receiving 210 and means for providing 220 may be a receiver and a provider, respectively. The means for receiving 210 and means for providing 220 may be a receiving processor and a providing processor, respectively. The means for receiving 210 receives a request for a functionality from a terminal (S210). The means for providing 220 provides (S220) assistance to the terminal in response to the request such that the terminal has a capability to prepare the functionality. The functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning of the terminal, or assisted positioning of the terminal. Preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; and - validating the trained ML model; . Fig.7 shows an apparatus according to an example embodiment. The apparatus may be a terminal (such as a UE or a MTC device) or an element thereof. Fig. 8 shows a method according to an example embodiment. The apparatus according to Fig. 7 may perform the method of Fig.8 but is not limited to this method. The method of Fig.8 may be performed by the apparatus of Fig.7 but is not limited to being performed by this apparatus. The apparatus comprises first and second means for determining 310 and 320, means for requesting 330, means for receiving340, and means for making350. The first and second means for determining 310 and 320, means for requesting 330, means for receiving340, and means for making350 may be a first and second determining means, requesting means, receiving means, and making means, respectively. The first and second means for determining 310 and 320, means for requesting 330, means for receiving340, and means for making350 may be a first and second determiner, requestor, receiver, and maker, respectively. The first and second means for determining 310 and 320, means for requesting 330, means for receiving340, and means for making350 may be a first and second determining processor, requesting processor, receiving processor, and making processor, respectively. The first means for determining 310 determines a functionality not executable but desired to be executable (S310). The second means for determining 320 determines whether a capability to prepare the desired functionality is available or not (S320). The functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions. The functionality further comprises at least one of the following: direct positioning or assisted positioning; The means for requesting 330 requests to provide the desired functionality based on determining that the capability is not available (S330). The means for receiving 340 receives the desired functionality (S340). The means for making 350 makes the received functionality executable (S350). Fig.9 shows an apparatus according to an example embodiment. The apparatus may be a terminal (such as a UE or a MTC device) or an element thereof. Fig. 10 shows a method according to an example embodiment. The apparatus according to Fig. 9 may perform the method of Fig.10 but is not limited to this method. The method of Fig.10 may be performed by the apparatus of Fig.9 but is not limited to being performed by this apparatus. The apparatus comprises first and second means for determining 410 and 420, means for requesting 430, means for receiving 440, and means for preparing 450. The first and second means for determining 410 and 420, means for requesting 430, means for receiving 440, and means for preparing 450 may be a first and second determining means, requesting means, receiving means, and preparing means, respectively. The first and second means for determining 410 and 420, means for requesting 430, means for receiving 440, and means for preparing 450 may be a first and second determiner, requestor, receiver, and preparer, respectively. The first and second means for determining 410 and 420, means for requesting 430, means for receiving 440, and means for preparing 450 may be a first and second determining processor, requesting processor, receiving processor, and preparing processor, respectively. The first means for determining 410 determines a functionality not executable but desired to be executable (S410). The second means for determining 420 determines whether a capability to prepare the functionality is available or not (S420). The means for requesting 430 requests to provide assistance to make the capability available based on determining that the capability is not available (S430). The means for receiving 440 receives the assistance (S440). The means for preparing 450 prepares the functionality based on the assistance (S450). The functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions. The functionality further comprises at least one of the following: direct positioning or assisted positioning. Preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; and - validating the trained ML model. Fig.11 shows an apparatus according to an example embodiment. The apparatus comprises at least one processor 810, at least one memory 820 storing instructions that, when executed by the at least one processor 810, cause the apparatus at least to perform the method according to at least one of the following figures and related description: Fig.4, or Fig.6, or Fig.8, or Fig.10. Some example embodiments are explained with respect to 5G (NR). However, other example embodiments may be employed in other 3GPP generations, such as 4G, 6G, 7G, etc. A UE is an example of a terminal. Other examples are MTC devices. Each of the terminals may be implemented as a smartphone, a mobile phone, a laptop, a sensor device etc. LMF is an example of a central unit. Another example of a central unit is NWDAF. For the purpose of the present application, LMF, NWDAF and central unit may be considered as equivalent unless otherwise indicated or made clear from the context. Some example embodiments are explained with positioning as the relevant functionality. However, some example embodiments may provide another relevant functionality, such as CSI compression, in particular CI compression with two-sided model, BM with UE-sided model, CSI prediction with UE sided model. One piece of information may be transmitted in one or plural messages from one entity to another entity. Each of these messages may comprise further (different) pieces of information. Names of network elements, network functions, protocols, and methods are based on current standards. In other versions or other technologies, the names of these network elements and / or network functions and / or protocols and / or methods may be different, as long as they provide a corresponding functionality. The same applies correspondingly to the terminal. If not otherwise stated or otherwise made clear from the context, the statement that two entities are different means that they perform different functions. It does not necessarily mean that they are based on different hardware. That is, each of the entities described in the present description may be based on a different hardware, or some or all of the entities may be based on the same hardware. It does not necessarily mean that they are based on different software. That is, each of the entities described in the present description may be based on different software, or some or all of the entities may be based on the same software. Each of the entities described in the present description may be deployed in the cloud. According to the above description, it should thus be apparent that example embodiments provide, for example, a central unit (such as LMF or NWDAF) or a component thereof, an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s). According to the above description, it should thus be apparent that example embodiments provide, for example, a terminal (such as a UE, a MTC device etc.) or a component thereof, an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s). Implementations of any of the above described blocks, apparatuses, systems, techniques or methods include, as non-limiting examples, implementations as hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof. Each of the entities described in the present description may be embodied in the cloud. It is to be understood that what is described above is what is presently considered the preferred example embodiments. However, it should be noted that the description of the preferred example embodiments is given by way of example only and that various modifications may be made without departing from the scope of the disclosure as defined by the appended claims. The terms “first X” and “second X” include the options that “first X” is the same as “second X” and that “first X” is different from “second X”, unless otherwise specified. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, where the list of two or more elements are joined by “and” or “or”, mean at least any one of the elements, or at least any two or more of the elements, or at least all the elements.
Claims
Claims:
1. Apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: preparing a functionality; and at least one of the following providing the prepared functionality to a terminal; or broadcasting the prepared functionality or an availability indication indicating that the functionality is prepared; wherein the functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following direct positioning of the terminal, or assisted positioning of the terminal; and wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; - validating the trained ML model; or - updating the ML model.
2. The apparatus according to claim 1, the instructions, when executed by the one or more processors, further cause the apparatus to perform the training the ML model based on data received from the terminal.
3. The apparatus according to any of claims 1 and 2, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform receiving respective data from each of plural terminals; training the ML model based on the data received from the plural terminals.
4. The apparatus according to any of claims 1 to 3, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform determining the set of conditions;sending the set of conditions along with the at least one of the broadcasting and the providing the functionality.
5. The apparatus according to any of claims 1 to 4, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform receiving a request to provide the prepared functionality to the terminal; providing the prepared functionality to the terminal in response to the request; inhibiting providing the prepared functionality to the terminal in absence of the request.
6. The apparatus according to claim 5, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform receiving, from the terminal, the set of the one or more conditions.
7. The apparatus according to any of claims 1 to 6, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform requesting the terminal to make the prepared functionality executable.
8. The apparatus according to any of claims 1 to 7, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform continuing preparing the functionality after the functionality has obtained an updated ML model; and at least one of the following providing the updated functionality or an incremental update of the functionality to the terminal; or broadcasting the updated functionality or the incremental update of the functionality or an update availability indication indicating that the updated functionality or the incremental update is available.
9. Apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: receiving a request for a functionality from a terminal; providing assistance to the terminal in response to the request such that the terminal has a capability to prepare the functionality; whereinthe functionality comprises a configuration for the terminal and a method executable by the terminal if the terminal is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning of the terminal, or assisted positioning of the terminal; and wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; and - validating the trained ML model.
10. The apparatus according to claim 9, wherein the providing the assistance comprises configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal.
11. The apparatus according to any of claims 9 to 10, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform receiving a set of conditions of the terminal; determining whether the functionality is executable by the terminal under the set of conditions; inhibiting providing the assistance based on determining that the functionality is not suitable executable by the terminal under the set of conditions.
12. Apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the desired functionality is available or not; requesting to provide the desired functionality based on determining that the capability is not available; receiving the desired functionality; making the received functionality executable; wherein the functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; the functionality further comprises at least one of the following:direct positioning or assisted positioning.
13. The apparatus according to claim 12, wherein the required capability comprises at least one of the following: - a capability to collect data for at least one of the training, the testing, and the validating of the ML model; - a processing power for at least one of the training, the testing, and the validating of the ML model; - an availability of the processing power for at least one of the training, the testing, and the validating of the ML model; - a memory for at least one of the training, the testing, and the validating the ML model; - an availability of the memory for at least one of the training, the testing, and the validating the ML model; or - a capability to obtain a ground truth for validating the ML model.
14. The apparatus according to any of claims 12 to 13, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform determining a set of one or more conditions for making the functionality executable; sending the set of the one or more conditions along with the requesting the functionality.
15. The apparatus according to any of claims 12 to 14, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform requesting the functionality in response to receiving a broadcasted availability indication indicating that the functionality is prepared 16. The apparatus according to any of claims 12 to 15, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform providing data for preparing the functionality before the functionality is implemented, wherein the data are based on a performed measurement; using the data for executing the functionality when the functionality is made executable.
17. The apparatus according to any of claims 12 to 16, wherein the instructions, when executed by the one or more processors, further cause the apparatus to performreceiving, along with the receiving the functionality, a set of one or more conditions for enabling the method executable; determining whether or not the one or more conditions are fulfilled.
18. The apparatus according to any of claims 12 to 17, wherein the instructions, when executed by the one or more processors, further cause the apparatus to perform receiving an incremental update of the functionality; updating the functionality based on the incremental update.
19. Apparatus comprising: one or more processors, and memory storing instructions that, when executed by the one or more processors, cause the apparatus to perform: determining a functionality not executable but desired to be executable; determining whether a capability to prepare the functionality is available or not; requesting to provide assistance to make the capability available based on determining that the capability is not available; receiving the assistance; preparing the functionality based on the assistance; wherein the functionality comprises a configuration and a method executable by the apparatus if the apparatus is configured with the configuration and the configuration is under a set of conditions; and the functionality further comprises at least one of the following: direct positioning or assisted positioning; wherein preparing the functionality comprises at least one of the following: - collecting data for training a ML model; - training the ML model; - testing the ML model; and validating the trained ML model.
20. The apparatus according to claim 19, wherein the providing the assistance comprises configuring the terminal to receive a reference signal, and the collecting the data comprises evaluating the received reference signal.
Citation Information
Cited By
Time of arrival information for positioning
US12696227B2
Time of arrival information for positioning
US20250142512A1
