Utilizing metadata to enable a model host node hosting an artificial intelligence / machine learning (ai / ML) model to position a target

Abstract

A method, system and apparatus are disclosed. In some embodiments, a method implemented in a model host node hosting an artificial intelligence / machine learning (AI / ML) model for estimating a location of a target user equipment (UE) is provided, the model host node configured to communicate with a requesting node. The method includes collecting data to form a training dataset, compiling context information of the collected data and attach the context information to the training dataset as model metadata, sending the model metadata to the requesting node, and receiving an indication as to whether the model host node can activate the AI / ML model for estimating the location of the target UE, the indication being based on the model metadata.

Description

[0001] UTILIZING METADATA TO ENABLE A MODEL HOST NODE HOSTING AN ARTIFICIAL INTELLIGENCE / MACHINE LEARNING (AI / ML) MODEL TO POSITION A TARGET DEVICE

[0002] FIELD

[0003] The present disclosure relates to wireless communications, and in particular, to the use of metadata for artificial intelligence (Al) and / or machine learning (ML) location estimate models.

[0004] BACKGROUND

[0005] The Third Generation Partnership Project (3GPP) has developed and is developing standards for Fourth Generation (4G) (also referred to as Long Term Evolution (LTE)) and Fifth Generation (5G) (also referred to as New Radio (NR)) wireless communication systems. Such systems provide, among other features, broadband communication between network nodes, such as base stations, and mobile user equipments (UE), as well as communication between network nodes and between UEs. The 3GPP is also developing standards for Sixth Generation (6G) wireless communication networks.

[0006] Artificial Intelligence (Al), Machine Learning (ML) have been investigated as promising tools to optimize the design of air-interface in wireless communication networks in both academia and industry. Example use cases include: using autoencoders for Channel State Information (CSI) compression to reduce the feedback overhead and improve channel prediction accuracy; using deep neural networks for classifying Line-Of-Sight (LOS) and Non-Line-Of-Sight (NLOS) conditions to enhance the positioning accuracy; using reinforcement learning for beam selection at the network side and / or the User Equipment (UE) side to reduce the signaling overhead and beam alignment latency; and using deep reinforcement learning to learn an optimal precoding policy for complex Multiple Input Multiple Output (MIMO) precoding problems.

[0007] In Third Generation Partnership Project (3GPP) New Radio (NR) standardization work, the release 18 study item on AI / ML for NR air interface has been completed. This study item explores the benefits of augmenting the air-interface with features enabling improved support of AI / ML based algorithms for enhanced performance and / or reduced complexity / overhead. Through studying a few selected use cases (e.g., CSI feedback, beam management and positioning), this study item aims at laying the foundation for future air-interface use cases leveraging AI / ML techniques.One AI / ML Physical Layer (PHY) use case is the positioning of a target UE. The 3GPP Release 18 study demonstrated that AI / ML based positioning methods are able to achieve improved positioning accuracy compared to legacy methods, especially in a medium-to-heavy non-line-of-sight environment. Consequently, the Release 19 work item proceeds to specify the AI / ML based positioning methods identified in the Release 18 study item.

[0008] Overview of AI / ML for wireless communications

[0009] Building an AI / ML model includes several development steps where the actual training of the Al model is just one step in a training pipeline. One part in AI / ML developing is the AI / ML model lifecycle management as illustrated in FIG. 1 where FIG.

[0010] 1 is an illustration of training and inference pipelines, and their interactions within a model lifecycle management procedure. The Al model lifecycle management typically consists of:

[0011] • A training (re-training) pipeline,

[0012] - With data ingestion referring to gathering raw (training) data from a data storage. After data ingestion, there may also be a step that controls the validity of the gathered data.

[0013] With data pre-processing referring to some feature engineering applied to the gathered data, e.g., it may include data normalization and possibly a data transformation required for the input data to the AI / ML model.

[0014] - With the actual model training steps where a model is obtained using the training dataset.

[0015] With model evaluation referring to benchmarking the performance to some baseline. The iterative steps of model training and model evaluation continue until the acceptable level of performance is achieved.

[0016] With model registration referring to register the AI / ML model, including any corresponding AI / ML-metadata that provides information on how the AI / ML model was developed, and possibly AI / ML model evaluations performance outcomes.

[0017] • A deployment stage to make the trained (or re-trained) AI / ML model part of the inference pipeline.

[0018] • An inference pipeline,

[0019] With data ingestion referring to gathering raw (inference) data from a data storage.With data pre-processing stage that is typically identical to corresponding processing that occurs in the training pipeline.

[0020] With model operational referring to using the trained and deployed model in an operational mode.

[0021] With data and model monitoring referring to validate that the inference data is from a distribution that aligns well with the training data, as well as monitoring model outputs for detecting any performance, or operational, drifts.

[0022] • A drift detection stage that informs about any drifts in the model operations.

[0023] AI / ML for positioning

[0024] One AI / ML PHY use case is the positioning of a target UE. Both positioning approaches below have been shown to be effective in obtaining the target UE's location.

[0025] Direct AI / ML positioning, where the AI / ML model output is the UE location. Direct AI / ML positioning typically refers to radio fingerprinting, where channel observation is used as the input of AI / ML model.

[0026] AI / ML assisted positioning, where the AI / ML model output is a new measurement and / or enhancement of an existing measurement. The model output can be one or more of the following: LOS / NLOS identification, timing and / or angle measurement, likelihood or reliability of the measurement. The model input is also channel observations.

[0027] When applying the direct and assisted AI / ML positioning to an NR wireless communication network, the following cases are further identified for investigation.

[0028] • Case 1 : UE-based positioning with UE-side model, direct AI / ML or AI / ML assisted positioning;

[0029] • Case 2a: UE-assisted / Location Management Function (LMF)-based positioning with UE-side model, AI / ML assisted positioning;

[0030] • Case 2b: UE-assisted / LMF-based positioning with LMF-side model, direct AI / ML positioning;

[0031] • Case 3a: New Generation-Random Access Network (NG-RAN) node assisted positioning with gNB (“base station” or “network node”)-side model, AI / ML assisted positioning; and

[0032] • Case 3b: NG-RAN node assisted positioning with LMF-side model, direct AI / ML positioning.

[0033] Case 2b above is further illustrated in FIG. 2. In FIG. 2, the dashed line represents the downlink reference signal (e.g., Positioning Reference Signal (PRS)) and the solid linerepresents the measurement report from the UE to the LMF. In this positioning use case, the UE performs measurements on reference signals (for example, the downlink positioning reference signals) from multiple transmission and reception points (TRPs). The UE then prepares and signals positioning related reports over the radio network to the location management function (LMF). AI / ML models are employed at the LMF to process the positioning related reports to generate an estimate of the UE position.

[0034] Case 3b is further illustrated in FIG. 3. In FIG. 3, the dashed line represents the reference signal and the solid line represents the measurement report. In this positioning use case, multiple TRPs perform measurements on reference signals (for example, the uplink sounding reference signals) from a UE. The TRPs or the network node controlling the TRPs then prepare and send signal positioning related reports to the location management function (LMF). AI / ML models are employed at the LMF to process the positioning related reports to generate an estimate of the UE position.

[0035] In addition to the above operation modes for using AI / ML models for accurate UE positioning (i.e., for model inference), the positioning related measurement reports may also be collected by an ML measurement data collection node to compile suitable training datasets for training high performance ML positioning models (i.e., for model training). Therefore, another example is a general system setup where the ML measurement report contains channel measurements corresponding to model input.

[0036] As discussed above, the ML data measurement node can be a UE measuring downlink reference signals transmitted from the radio network nodes (such as the TRPs) or a radio network node (such as a TRP or a gNB) measuring uplink reference signals transmitted from a UE.

[0037] Ensure consistency between training and inference

[0038] To keep consistency between the training and inference phases, the node operating the model can secure that the assistance data and signals received during inference are the same as what was used for training the model, or within the acceptable tolerance. For this purpose, in addition to the list of training data samples, the training dataset also needs to record the context information of the collected data. Such context information is attached to the training dataset as its metadata. The context information may include basic information such as:

[0039] Where the training data is collected.

[0040] When the training data is collected, for example, year / month / date.

[0041] Which nodes provided the measurement data, for example, TRP ID, UE ID.• Which nodes provided the label data, for example, Positioning Reference Unit (PRU) ID, UE ID, LMF ID.

[0042] • What kind of reference signals are used in generating the measurement data, for example, PRS configuration for DL measurement, Sounding Reference Signals (SRS) configuration for UL measurement.

[0043] • Type and format of the measurement data, for example,

[0044] - Type (e.g., Data Protocol / Packet Data Protocol (DP / PDP) and size, (e.g., Nt, N't) of the measurement data.

[0045] Quantization granularity used in recording the measurement values. • Type and format of the label data, for example,

[0046] - Type of information for label data, for example, UE location or information corresponding to model output of AI / ML assisted positioning models.

[0047] - Format used to record the location information (for labels).

[0048] 2D or 3D location coordinates.

[0049] • Criteria (if any) applied before a data sample can be accepted as part of the training dataset, for example, label quality requirement, Signal-to-Interference-Noise-Ratio (SINR) requirement, etc.

[0050] • Version ID of the training dataset, if the training dataset may be updated over time.

[0051] However, there are no existing mechanisms for communicating data of an AI / ML model where the data provides information on the measurement environment condition and quality.

[0052] SUMMARY

[0053] Some embodiments advantageously provide methods, systems, and apparatuses for using metadata for AI / ML positioning models such as, for example, to determine if a model is appropriate for a development environment.

[0054] In some embodiments, methods are provided on reporting information related to: (A) Reference signal quality; and

[0055] (B) Network timing error.

[0056] The reference signal quality includes downlink signal quality for methods using downlink reference signal (e.g., PRS) measurement, and uplink signal quality for methods using uplink reference signal (e.g., SRS) measurement.The information on (A) and / or (B) during training data collection needs to be collected and recorded with the training dataset.

[0057] In some embodiments, after performing model training with the training dataset, the value range of reference signal quality (A) and / or network timing error (B) supported by the trained model may need to be attached to the model as its metadata. This includes the aggregate value ranges when using mixed training dataset with one or more different value ranges of reference signal quality (A), and one or more different value ranges of network timing error (B).

[0058] In some embodiments, when performing model inference, the metadata of the model under consideration is compared with reference signal quality (A) and / or network timing error (B) encountered at deployment. If they are consistent (e.g., reference signal quality (A) and / or network timing error (B) at deployment, are within the tolerance range of the trained model), then the model may be considered appropriate for the deployment environment, and model inference may proceed. Otherwise, the model may be considered inappropriate for the deployment environment. An error cause may be declared if the model inference cannot proceed due to the inconsistency or incompatibility.

[0059] In the discussion below, AI / ML based positioning is used as a representative use case to illustrate the metadata reporting of reference signal quality (A) and network timing error (B), and the associated procedures. In general, one or more embodiments described herein may be generally applicable to other uses cases where AI / ML model is used to support functions in a wireless communication system.

[0060] In accordance with one aspect of the present disclosure, a method implemented in a model host node hosting an artificial intelligence / machine learning (AI / ML) model for estimating a location of a target device, the model host node configured to communicate with a requesting node, is provided. The method includes collecting data to form a training dataset, compiling context information of the collected data and attach the context information to the training dataset as model metadata, sending the model metadata to the requesting node, and receiving an indication as to whether the model host node can activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata.

[0061] In accordance with another aspect of the present disclosure, a method implemented in a requesting node that is configured to communicate with a model host node hosting an AI / ML model for estimating a location of a target device, is provided. The method includes receiving model metadata from the model host node, the model metadataincluding a training dataset of collected data and context information of the collected data, and indicating to the model host node whether the model host node can activate the AI / ML model for estimating the location of the target UE, the indication being based on the model metadata.

[0062] In accordance with another aspect of the present disclosure, a model host node hosting an AI / ML model for estimating a location of a target device, the model host node configured to communicate with a requesting node, is provided. The model host node is configured to collect data to form a training dataset, compile context information and attach the context information to the training dataset as model metadata, send the model metadata to the requesting node, and receive an indication to activate the AI / ML model for estimating the location of the target UE, the indication being based on the model metadata.

[0063] In accordance with another aspect of the present disclosure, a requesting node that is configured to communicate with a model host node hosting an AI / ML model for estimating a location of a target device, is provided. The requesting node is configured to receive model metadata from the model host node, the model metadata including a training dataset of collected data and context information of the collected data, and indicate to the model host node whether the model host node can activate the AI / ML model for estimating the location of the target UE, the indication being based on the model metadata.

[0064] BRIEF DESCRIPTION OF THE DRAWINGS

[0065] A more complete understanding of the present embodiments, and the attendant advantages and features thereof, will be more readily understood by reference to the following detailed description when considered in conjunction with the accompanying drawings wherein:

[0066] FIG. 1 illustrates training and inference pipelines, and their interactions within a model lifecycle management procedure;

[0067] FIG. 2 illustrates a use case of direct AI / ML positioning;

[0068] FIG. 3 illustrates another use case of direct AI / ML positioning;

[0069] FIG. 4 is a schematic diagram of an example network architecture illustrating a communication system according to principles disclosed herein;

[0070] FIG. 5 is a schematic diagram of another example network architecture illustrating a communication system according to principles disclosed herein;FIG. 6 is a block diagram of a requesting node in communication with a model host node over a wireless connection according to some embodiments of the present disclosure;

[0071] FIG. 7 is a schematic diagram of another example network architecture illustrating a communication system according to principles disclosed herein;

[0072] FIG. 8 is a flowchart of an example process in a requesting node according to some embodiments of the present disclosure;

[0073] FIG. 9 is a flowchart of another example process in a requesting node according to some embodiments of the present disclosure;

[0074] FIG. 10 is a flowchart of an example process in a model host node according to some embodiments of the present disclosure;

[0075] FIG. 11 is a flowchart of another example process in a model host node according to some embodiments of the present disclosure; and

[0076] FIG. 12 is a block diagram illustrating a virtualization environment according to some embodiments of the present disclosure.

[0077] DETAILED DESCRIPTION

[0078] As discussed above, currently there is no mechanisms to record and report metadata of an AI / ML model, where the metadata provides information on the measurement environment condition and quality. Without this information, at the model inference stage, it is unclear whether a trained model is appropriate for the deployment environment or not.

[0079] Therefore, there is a need to introduce mechanisms to ensure metadata consistency (e.g., metadata Group 3 consistency) for training data collection, model training, and model inference.

[0080] It should be noted that throughout this disclosure, the term “target device” may refer to any device or node, such as, for example, a UE, or a network node, etc.

[0081] One or more embodiments described herein solve one or more problems discussed above. Some solutions (e.g., embodiments), as described herein, use a set of existing measurements to capture metadata associated with AI / ML positioning. The solutions are based on the measurement made on the communication signals, such as Channel State Information-Reference Signal (CSI-RS), Synchronization Signal Block (SSB), or SRS associated with positioning signals in assistance data (via the Quasi-Colocation (QCL) property). Such measurements are typically available to a serving network node and canbe requested by the LMF using legacy procedures, such as Enhanced Cell Identification (E-CID). Additionally, the solutions proposed herein also propose to track timing changes by recording timing information provided as assistance data, such as Round Trip Delay (RTD) info.

[0082] To compare the environment conditions during training and assess whether the model is fit for inference (e.g., should be used for inference), some embodiments disclosed herein:

[0083] - Have the model host collect and report the range of values for timing and power for which the model was trained during the model training stage.

[0084] Exchange the range of values with the LMF via capability signaling, and potentially update the values if further data collection is performed during inference.

[0085] Assess the suitability of the model in the LMF based on the reported range for measurements made during training from the model host, and comparing them to the timing and power measurements for the same signals, available to the LMF, prior to inference.

[0086] The solutions described herein allow for the savings of a potentially large amount of overhead, as it may enable the LMF to determine whether a positioning session can occur in advance, instead of first configuring a positioning session and then receiving unreliable results.

[0087] The solutions described herein may use existing signals and / or measurements, and may not require additional measurements to be defined in, for example, 3GPP standards and / or specifications.

[0088] The solutions described herein may use existing interfaces and introduce moderate additional overhead in capability signaling.

[0089] Before describing in detail example embodiments, it is noted that the embodiments reside primarily in combinations of apparatus components and processing steps related to using metadata for AI / ML positioning models such as to, for example, determine if a model is appropriate for a development environment. Accordingly, components have been represented where appropriate by conventional symbols in the drawings, showing only those specific details that are pertinent to understanding the embodiments so as not to obscure the disclosure with details that will be readily apparent to those of ordinary skill in the art having the benefit of the description herein.As used herein, relational terms, such as “first” and “second,” “top” and “bottom,” and the like, may be used solely to distinguish one entity or element from another entity or element without necessarily requiring or implying any physical or logical relationship or order between such entities or elements. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the concepts described herein. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and / or “including” when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0090] In embodiments described herein, the joining term, “in communication with” and the like, may be used to indicate electrical or data communication, which may be accomplished by physical contact, induction, electromagnetic radiation, radio signaling, infrared signaling or optical signaling, for example. One having ordinary skill in the art will appreciate that multiple components may interoperate and modifications and variations are possible of achieving the electrical and data communication.

[0091] In some embodiments described herein, the term “coupled,” “connected,” and the like, may be used herein to indicate a connection, although not necessarily directly, and may include wired and / or wireless connections.

[0092] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the concepts described herein. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and / or “including” when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0093] The term “network node” used herein can be any kind of network node comprised in a radio network which may further comprise any of base station (BS), radio base station, base transceiver station (BTS), base station controller (BSC), radio network controller (RNC), g Node B (gNB), evolved Node B (eNB or eNodeB), Node B, multistandard radio (MSR) radio node such as MSR BS, multi-cell / multicast coordination entity(MCE), relay node, donor node controlling relay, radio access point (AP), transmission points, transmission nodes, Remote Radio Unit (RRU) Remote Radio Head (RRH), a core network node (e.g., mobile management entity (MME), self-organizing network (SON) node, a coordinating node, positioning node, MDT node, etc.), an external node (e.g., 3rd party node, a node external to the current network), nodes in distributed antenna system (DAS), a spectrum access system (SAS) node, an element management system (EMS), etc. The network node may also comprise test equipment. The term “radio node” used herein may be used to also denote a user equipment (UE) such as a wireless device (WD) or a radio network node.

[0094] In some embodiments, the non-limiting terms wireless device (WD) or a user equipment (UE) are used interchangeably. The UE herein can be any type of user equipment capable of communicating with a network node or another UE over radio signals, such as a wireless device (WD). The UE may also be a radio communication device, target device, device to device (D2D) UE, machine type UE or UE capable of machine to machine communication (M2M), low-cost and / or low-complexity UE, a sensor equipped with UE, Tablet, mobile terminals, smart phone, laptop embedded equipped (LEE), laptop mounted equipment (LME), USB dongles, Customer Premises Equipment (CPE), an Internet of Things (loT) device, or a Narrowband loT (NB-IOT) device etc.

[0095] Also, in some embodiments the generic term “radio network node” is used. It can be any kind of a radio network node which may comprise any of base station, radio base station, base transceiver station, base station controller, network controller, RNC, evolved Node B (eNB), Node B, gNB, Multi-cell / multicast Coordination Entity (MCE), relay node, access point, radio access point, Remote Radio Unit (RRU) Remote Radio Head (RRH).

[0096] Note that although terminology from one particular wireless system, such as, for example, 3GPP LTE and / or New Radio (NR) and / or 6G, may be used in this disclosure, this should not be seen as limiting the scope of the disclosure to only the aforementioned system. It is contemplated that other 3GPP systems may utilize the concepts and arrangements disclosed herein. For example, a disclosure relating to NR may also be implementable in a 6G system and / or an LTE system, a disclosure relating to 6G may also be implementable in a NR and / or LTE system, and a disclosure relating to LTE may also be implementable in a NR and / or 6G system. Other wireless systems, including without limitation Wide Band Code Division Multiple Access (WCDMA), Worldwide Interoperability for Microwave Access (WiMax), Ultra Mobile Broadband (UMB) andGlobal System for Mobile Communications (GSM), may also benefit from exploiting the ideas covered within this disclosure.

[0097] Note further, that functions described herein as being performed by a user equipment or a network node may be distributed over a plurality of user equipments and / or network nodes. In other words, it is contemplated that the functions of the network node and user equipment described herein are not limited to performance by a single physical device and, in fact, can be distributed among several physical devices.

[0098] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. It will be further understood that terms used herein should be interpreted as having a meaning that is consistent with their meaning in the context of this specification and the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0099] Some embodiments are directed to using metadata for AI / ML positioning models to determine if a model is appropriate for a development environment.

[0100] Referring again to the drawing figures, in which like elements are referred to by like reference numerals, there is shown in FIG. 4 is a schematic diagram of a communication system 2, according to one or more embodiments. Communication system 2 includes at least one requesting node 4 and a plurality of model host nodes 6a-6n (referred to collectively as model host 6). In one or more embodiments, requesting node 4 may be a network node, an LMF, a UE, or a TRP. In one or more embodiments, model host node 6 may be a UE, a network node, or a TRP.

[0101] In one or more embodiments, requesting node 4 may include comparing unit 8 that is configured to perform one or requesting node 4 functions described herein. In one or more embodiments, model host 6 may include range unit 9 that is configured to perform one or more model host 6 functions described herein.

[0102] FIG. 5 is a schematic diagram of a communication system 10, according to an embodiment, such as a 3 GPP -type cellular network that may support standards such as LTE and / or NR (5G) and / or 6G, which comprises an access network 12, such as a radio access network, and a core network 14. The core network 14 may include one or more core network nodes 15 such as a location management function (LMF) 15. The access network 12 comprises a plurality of network nodes 16a, 16b, 16c (referred to collectively as network nodes 16), such as NBs, eNBs, gNBs or other types of wireless access points, each defining a corresponding coverage area 18a, 18b, 18c (referred to collectively ascoverage areas 18). Each network node 16a, 16b, 16c is connectable to the core network 14 over a wired or wireless connection 20. A first user equipment (UE) 22a located in coverage area 18a is configured to wirelessly connect to, or be paged by, the corresponding network node 16a. A second UE 22b in coverage area 18b is wirelessly connectable to the corresponding network node 16b. While a plurality of UEs 22a, 22b (collectively referred to as user equipments 22) are illustrated in this example, the disclosed embodiments are equally applicable to a situation where a sole UE is in the coverage area or where a sole UE is connecting to the corresponding network node 16. Note that although only two UEs 22, three network nodes 16 and one core network node (e.g., LMF 15) are shown for convenience, the communication system may include many more UEs 22, network nodes 16 and / or core network nodes 15.

[0103] As one example, in certain embodiments, access network 12 may contain some access network nodes 16 that support 3 GPP radio access technologies (RAT), such as LTE or NR, while other access network nodes 16 support (or the same access network nodes 16 additionally support) non-3GPP RATs, such as Wi-Fi or a proprietary RAT. As another example, communication system 10 may support multiple generations of related communication standards (e.g., 4G, 5G and 6G 3GPP communication standards) and, as a result, may include an access network 12 and / or a core network 14 that supports multiple different standard generations or may include multiple access networks 12 and / or multiple core networks 14 with individual networks supporting different standards generations.

[0104] Also, it is contemplated that a UE 22 can be in simultaneous communication and / or configured to separately communicate with more than one network node 16 and more than one type of network node 16. For example, a UE 22 can have dual connectivity with a network node 16 that supports LTE and the same or a different network node 16 that supports NR. As an example, UE 22 can be in communication with an eNB for LTE / E-UTRAN, a gNB for NR / NG-RAN (i.e. being configured for multiradio dual connectivity (MR-DC), such as E-UTRAN (Evolved-UMTS Terrestrial Radio Access Network) New Radio - Dual Connectivity (EN-DC) and / or Wi-Fi.

[0105] In one or more examples, requesting node 4 may be implemented by network node 16 (e.g., network node 16a) or a UE 22. In one or more embodiments, model host node 6 may be implemented by a UE 22 (e.g., UE 22a).

[0106] Example implementations, in accordance with an embodiment, of model host node 6 and requesting node 4 discussed in the preceding paragraphs will now be described with reference to FIG. 6.The communication system 2 includes a requesting node 4 provided in a communication system 2 and including hardware 28 enabling it to communicate with the model host node 6. The hardware 28 may include a communication interface 29 comprising a radio interface 30 communicating with the model host node 6. The radio interface 30 may be formed as or may include, for example, one or more RF transmitters, one or more RF receivers, and / or one or more RF transceivers. The radio interface 30 includes an array of antennas 34 to radiate and receive signal(s) carrying electromagnetic waves.

[0107] In the embodiment shown, the hardware 28 of the requesting node 4 further includes processing circuitry 36. The processing circuitry 36 may include a processor 38 and a memory 40. In particular, in addition to or instead of a processor, such as a central processing unit, and memory, the processing circuitry 36 may comprise integrated circuitry for processing and / or control, e.g., one or more processors and / or processor cores and / or FPGAs (Field Programmable Gate Array) and / or ASICs (Application Specific Integrated Circuitry) adapted to execute instructions. The processor 38 may be configured to access (e.g., write to and / or read from) the memory 40, which may comprise any kind of volatile and / or nonvolatile memory, e.g., cache and / or buffer memory and / or RAM (Random Access Memory) and / or ROM (Read-Only Memory) and / or optical memory and / or EPROM (Erasable Programmable Read-Only Memory).

[0108] Thus, the requesting node 4 further has software 42 stored internally in, for example, memory 40, or stored in external memory (e.g., database, storage array, network storage device, etc.) accessible by the requesting node 4 via an external connection. The software 42 may be executable by the processing circuitry 36. The processing circuitry 36 may be configured to control any of the methods and / or processes described herein and / or to cause such methods, and / or processes to be performed, e.g., by requesting node 4. Processor 38 corresponds to one or more processors 38 for performing requesting node 4 functions described herein. The memory 40 is configured to store data, programmatic software code and / or other information described herein. In some embodiments, the software 42 may include instructions that, when executed by the processor 38 and / or processing circuitry 36, causes the processor 38 and / or processing circuitry 36 to perform the processes described herein with respect requesting node. For example, processing circuitry 36 of the requesting node 4 may include comparing unit 8 which is configured to perform one or more requesting node 4 functions as described herein.In certain alternative embodiments, requesting node 4 may be capable of wireless communication but does not include separate radio front-end circuitry, instead, the processing circuitry 36 includes radio front-end circuitry and is connected to the antenna 34. Similarly, in some embodiments, all or some of the RF receivers, transmitters and / or transceivers are part of the radio interface 30. In still other embodiments, the communication interface 29 includes one or more ports or terminals, the radio interface 30, and the RF receiver, transmitter and / or transceiver, and the communication interface 31 communicates with baseband processing circuitry, which is part of a digital unit (not shown).

[0109] The antenna 34 may include one or more antennas, or antenna arrays, configured to send and / or receive wireless signals. The antenna 34 may be coupled to the radio front-end circuitry in radio interface 30 and may be any type of antenna capable of transmitting and receiving data and / or signals wirelessly. In certain embodiments, the antenna 34 is separate from the requesting node 4 and connectable to the requesting node 4 through one or more interfaces or ports.

[0110] Core network node 15 (e.g., LMF) can include one or more components described above with respect to requesting node 4, e.g., communication interface 29, radio interface 30, antenna 34, ports, processing circuitry 36, processor 38, memory 40 and software 42. These elements of core network node 15 can be arranged such that core network node 15 can perform various core network functions. Core network node 15 can communicate wirelessly or via a wired connection with requesting nodes 4 via communication link 59.

[0111] The communication system 2 further includes the model host node 6 already referred to. The model host node 6 may have hardware 44 that may include a radio interface 46 for communicating with requesting node 4. The radio interface 46 may be formed as or may include, for example, one or more RF transmitters, one or more RF receivers, and / or one or more RF transceivers. The radio interface 46 includes an array of antennas 48 to radiate and receive signal(s) carrying electromagnetic waves.

[0112] Communication functions of the radio interface 46 may include cellular communication, Wi-Fi communication (e.g., according to an IEEE 802.11 family standard), LPWAN communication, data communication, voice communication, multimedia communication, short-range communications such as Bluetooth, near-field communication, location-based communication such as the use of the global positioning system (GPS) to determine a location, another like communication function, or any combination thereof. Communications may be implemented according to one or morecommunication protocols and / or standards, such as IEEE 802.11, Code Division Multiplexing Access (CDMA), Wideband Code Division Multiple Access (WCDMA), GSM, LTE, New Radio (NR), UMTS, WiMax, Ethernet, transmission control protocol / intemet protocol (TCP / IP), synchronous optical networking (SONET), Asynchronous Transfer Mode (ATM), QUIC, Hypertext Transfer Protocol (HTTP), and so forth.

[0113] The hardware 44 of model host node 6 further includes processing circuitry 50. The processing circuitry 50 may include a processor 52 and memory 54. In particular, in addition to or instead of a processor, such as a central processing unit, and memory, the processing circuitry 50 may comprise integrated circuitry for processing and / or control, e.g., one or more processors and / or processor cores and / or FPGAs (Field Programmable Gate Array) and / or ASICs (Application Specific Integrated Circuitry) adapted to execute instructions. The processor 52 may be configured to access (e.g., write to and / or read from) memory 54, which may comprise any kind of volatile and / or nonvolatile memory, e.g., cache and / or buffer memory and / or RAM (Random Access Memory) and / or ROM (Read-Only Memory) and / or optical memory and / or EPROM (Erasable Programmable Read-Only Memory).

[0114] Thus, the model host node 6 may further comprise software 56, which is stored in, for example, memory 54 at the model host node 6 or stored in external memory (e.g., database, storage array, network storage device, etc.) accessible by the model host node 6. The software 56 may be executable by the processing circuitry 50. The software 56 may include a client application 58. The client application 58 may be operable to provide a service to a human or non-human user via the model host node 6.

[0115] The processing circuitry 50 may be configured to control any of the methods and / or processes described herein and / or to cause such methods, and / or processes to be performed, e.g., by model host node 6. The processor 52 corresponds to one or more processors 52 for performing model host node 6 functions described herein. The model host node 6 includes memory 54 that is configured to store data, programmatic software code and / or other information described herein. In some embodiments, the software 56 and / or the client application 58 may include instructions that, when executed by the processor 52 and / or processing circuitry 50, causes the processor 52 and / or processing circuitry 50 to perform the processes described herein with respect to model host node 6. For example, the processing circuitry 50 of the model host node 6 may include range unit9 which is configured to perform one or more model host node 6 functions described herein.

[0116] In some embodiments, the inner workings of the requesting node 4 and the model host node 6 may be as shown in FIG. 6 and independently, the surrounding network topology may be that of FIG. 4.

[0117] The wireless connection 32 between the model host node 6 and the requesting node 4 is in accordance with the teachings of the embodiments described throughout this disclosure. More precisely, the teachings of some of these embodiments may improve the data rate, latency, and / or power consumption and thereby provide benefits such as reduced user waiting time, relaxed restriction on file size, better responsiveness, extended battery lifetime, etc. In some embodiments, a measurement procedure may be provided for the purpose of monitoring data rate, latency and other factors on which the one or more embodiments improve.

[0118] Although FIGS. 4 and 6 show various “units” such as comparing unit 8 and range unit 9 as being within a respective processor, it is contemplated that these units may be implemented such that a portion of the unit is stored in a corresponding memory within the processing circuitry. In other words, the units may be implemented in hardware or in a combination of hardware and software within the processing circuitry.

[0119] FIG. 7 is another example of a communication system 10 according to some embodiments. For example, requesting node 4 may be implemented in AP 60 or STA 62 while model host node 6 may be implemented in STA 62. As used herein, the communication system 10 of FIG. 3 includes multiple access points (APs) 60 (with four example APs 60a, 60b, 60c, and 60d being depicted) and multiple wireless devices, referred to in the context of communication system 10 of FIG. 3 as stations (STAs) 62 (referred to individually as STA 62a, STA 62b, STA 62c, STA 62d, and STA 62e). STA 62a is served by AP 60a in a first basic service set (BSS) 64a. STA 62b and STA 62c are served by AP 60b in a second BSS, BSS 64b. STA 62d is served by AP 60c in a third BSS, BSS 64c. STA 62e is served by AP 60d in a fourth BSS, BSS 64d. Stations 62 may be non-AP STAs and correspond to various kinds of wireless devices, for example, user terminals, such as mobile or stationary computing devices like smartphones, laptop computers, desktop computers, tablet computers, gaming devices, head-mounted displays (HMDs) for Augmented Reality (AR) or Virtual Reality (VR), or the like, including UEs 22 that are shown and described with respect to FIGS. 1 and 2. In other words, in some embodiment, STA 62 is a UE 22. Further, stations 62 could, for example, correspond toother kinds of equipment like smart home devices, printers, multimedia devices, data storage devices, or the like.

[0120] Each of STAs 62 may connect through a radio link to one of APs 60. For example, depending on location or channel conditions experienced by a given STA 62, the STA may select an appropriate AP and BSS for establishing the radio link. The radio link may be based on one or more orthogonal frequency-division multiplexing (OFDM) carriers from a frequency spectrum that is shared on the basis of a contention-based mechanism, e.g., an unlicensed or license exempt band like 2.4 GHz Industrial, Scientific, and Medical (ISM) band, the 5 GHz band, the 6 GHz band, or the 60 GHz band.

[0121] Each AP 60 may provide data connectivity to STAs 62 connected to a particular AP 60. As illustrated, APs 60 may be connected to a data network 66. In this way, APs 60 may also provide data connectivity between STAs 62 and other entities, e.g., to one or more servers, service providers, data sources, data sinks, user terminals, or the like.

[0122] Accordingly, the radio link established between a given STA 62 and its serving AP 60 may be used for providing various kinds of services to STA 62, e.g., a voice service, a multimedia service, or other data service. Such services may be based on applications that are executed on STA 62 and / or on a device linked to STA 62. By way of example, FIG. 7 illustrates an application service platform 68 provided in data network 66. The application(s) executed on STA 62 and / or on one or more other devices linked to STA 62 may use the radio link for data communication with one or more other STA 62 and / or the application service platform 68, thereby enabling utilization of the corresponding service(s) at STA 62.

[0123] FIG. 8 is a flowchart of an example process in a requesting node 4 according to some embodiments of the present disclosure. One or more blocks described herein may be performed by one or more elements of requesting node 4 such as by one or more of processing circuitry 36 (including the comparing unit 8), processor 38, and / or radio interface 30. Requesting node 4 is configured to obtain (Block SI 00) measurements. Requesting node 4 is further configured to receive (Block SI 02) model metadata from a model host node 6. Requesting node 4 is further configured to determine (Block SI 04) whether the model host node 6 can activate the AI / ML model based on the obtained measurements and the model metadata.

[0124] In some embodiments, the obtained measurements comprise at least timing and power measurements.In some embodiments, the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0125] In some embodiments, the range of AI / ML values are updated while the AI / ML model is deployed.

[0126] In some embodiments, the model metadata is received via a capability exchange. In some embodiments, the capability exchange is performed upon request from the requesting node via LPP.

[0127] In some embodiments, the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0128] In some embodiments, the model metadata is received from the model host node 6 via an NRPPa protocol.

[0129] In some embodiments, requesting node 4 is one of a network node 16, UE 22, an LMF 15 and aTRP.

[0130] FIG. 9 is a flowchart of another example process in a requesting node 4 according to some embodiments of the present disclosure. Requesting node 4 is configured to communicate with model host node 6 hosting an AI / ML model for estimating a location of a target device, One or more blocks described herein may be performed by one or more elements of requesting node 4 such as by one or more of processing circuitry 36 (including the comparing unit 8), processor 38, and / or radio interface 30. Requesting node 4 is configured to receive (Block SI 06) model metadata from the model host node 6, the model metadata including a training dataset of collected data and context information of the collected data. Requesting node 4 is further configured to indicate (Block SI 08) to the model host node 6 whether the model host node 6 can activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata.

[0131] In some embodiments, the model metadata includes assistance data to enable reference signal measurements.

[0132] In some embodiments, the model metadata includes elements that provide environment condition and quality of wireless channel measurement.

[0133] In some embodiments, the elements that provide environment condition and quality of wireless channel measurement include one or more network synchronization error range, reception and transmission timing error range, interference range, and channel estimation error range.In some embodiments, the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0134] In some embodiments, the range of AI / ML values are updated while the model is deployed.

[0135] In some embodiments, the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

[0136] In some embodiments, the model metadata is received the model host node 6 via a capability exchange.

[0137] In some embodiments, the capability exchange is performed upon request from the requesting node via Long Term Evolution (LTE) Positioning Protocol (LPP).

[0138] In some embodiments, the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0139] In some embodiments, the requesting node is one of a network node, a UE 22, a Location Management Function (LMF), and a transmission and reception point (TRP).

[0140] In some embodiments, the target device is a UE (22).

[0141] In some embodiments, the estimated location is used for one or more of positioning the target device and sensing the target device.

[0142] FIG. 10 is a flowchart of an example process in a model host node 6 according to some embodiments of the present disclosure. One or more blocks described herein may be performed by one or more elements of model host node 6 such as by one or more of processing circuitry 50 (including the range unit 9), processor 52, and / or radio interface 46. Model host node 6, such as via processing circuitry 50 and / or processor 52 and / or radio interface 46 is configured to collect (Block SI 10) model metadata of the AI / ML model. Model host node 6 is further configured to send (Block SI 12) the model metadata to the requesting node 4. Model host node 6 is further configured to receive (Block SI 14) an indication to activate the AI / ML model, where the indication is based on the model metadata.

[0143] In some embodiments, the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0144] In some embodiments, the range of AI / ML values are updated while the AI / ML model is deployed.In some embodiments, the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

[0145] In some embodiments, the model metadata is sent to the requesting node 4 via a capability exchange.

[0146] In some embodiments, the capability exchange is performed upon request from the requesting node 4 via LPP.

[0147] In some embodiments, the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0148] In some embodiments, the model host node 6 is UE 22.

[0149] In some embodiments, the UE 22 voluntarily contacts the requesting node 4 with an update of the model metadata for the model.

[0150] In some embodiments, the model host node 6 provides the model metadata as a type of UE assistance information.

[0151] In some embodiments, the model host node 6 is a network node 16.

[0152] In some embodiments, the model host node 6 sends the model metadata to the requesting node 4 via an NRPPa protocol.

[0153] In some embodiments, the model host node 6 is a TRP.

[0154] FIG. 11 is a flowchart of another example process in a model host node 6 according to some embodiments of the present disclosure. Model host node 6 hosts an AI / ML model for estimating a location of a target device. One or more blocks described herein may be performed by one or more elements of model host node 6 such as by one or more of processing circuitry 50 (including the range unit 9), processor 52, and / or radio interface 46. Model host node 6, such as via processing circuitry 50 and / or processor 52 and / or radio interface 46 is configured to collect (Block S 116) data to form a training dataset. Model host node 6 is further configured to compile (Block SI 18) context information and attach the context information to the training dataset as model metadata. Model host node 6 is further configured to send (Block S120) the model metadata to the requesting node 4. Model host node 6 is further configured to receive (Block SI 22) an indication to activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata.

[0155] In some embodiments, the model metadata includes assistance data to enable reference signal measurements.In some embodiments, the model metadata includes elements that provide environment condition and quality of wireless channel measurement.

[0156] In some embodiments, the elements that provide environment condition and quality of wireless channel measurement include one or more network synchronization error range, reception and transmission timing error range, interference range, and channel estimation error range.

[0157] In some embodiments, the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0158] In some embodiments, the range of AI / ML values are updated while the model is deployed.

[0159] In some embodiments, the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

[0160] In some embodiments, the model metadata is sent to the requesting node via a capability exchange.

[0161] In some embodiments, the capability exchange is performed upon request from the requesting node via Long Term Evolution (LTE) Positioning Protocol (LPP).

[0162] In some embodiments, the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0163] In some embodiments, the model host node is the target device.

[0164] In some embodiments, the model host node 6 is a network node 16, the model host node 6 is further configured to send the model metadata to the requesting node an NRPPa protocol.

[0165] In some embodiments, the model host node 6 is a TRP.

[0166] In some embodiments, the target device is a UE (22).

[0167] In some embodiments, the estimated location is used for one or more of positioning the target device and sensing the target device.

[0168] For example, in some embodiments, the telecommunication system 10 includes one or more Open-RAN (ORAN) network nodes 16. An ORAN network node 16 is a node in the telecommunication system 10 that supports an ORAN specification (e.g., a specification published by the O-RAN Alliance, or any similar organization) and may operate alone or together with other nodes to implement one or more functionalities of any node in the telecommunication system 10, including one or more network nodes 16 in the access network 12 and / or core network nodes 14.Examples of an ORAN network node 16 include an open radio unit (O-RU), an open distributed unit (O-DU), an open central unit (O-CU), including an O-CU control plane (O-CU-CP) or an O-CU user plane (O-CU-UP), a RAN intelligent controller (near-real time or non-real time) hosting software or software plug-ins, such as a near-real time control application (e.g., xApp) or a non-real time control application (e.g., rApp), or any combination thereof (the adjective “open” designating support of an ORAN specification). The network node may support a specification by, for example, supporting an interface defined by the ORAN specification, such as an Al, Fl, Wl, El, E2, X2, Xn interface, an open fronthaul user plane interface, or an open fronthaul management plane interface. Moreover, an ORAN access node may be a logical node in a physical node. Furthermore, an ORAN network node may be implemented in a virtualization environment (described further below) in which one or more network functions are virtualized. For example, the virtualization environment may include an O-Cloud computing platform orchestrated by a Service Management and Orchestration Framework via an 0-2 interface defined by the O-RAN Alliance or comparable technologies. The network nodes 16 facilitate direct or indirect connection of user equipment (UE), such as by connecting UEs 22a, 22b, 22c, and 22d (one or more of which may be generally referred to as UEs 22) to the core network 14 over one or more wireless connections.

[0169] FIG. 12 is a block diagram illustrating a virtualization environment 94 in which functions implemented by some embodiments may be virtualized. In the present context, virtualizing means creating virtual versions of apparatuses or devices which may include virtualizing hardware platforms, storage devices and networking resources. As used herein, virtualization can be applied to any device described herein, or components thereof, and relates to an implementation in which at least a portion of the functionality is implemented as one or more virtual components. Some or all of the functions described herein may be implemented as virtual components executed by one or more virtual machines (VMs) implemented in one or more virtual environments 94 hosted by one or more of hardware nodes, such as a hardware computing device that operates as a network node, UE, core network node, or host. Further, in embodiments in which the virtual node does not require radio connectivity (e.g., a core network node or host), then the node may be entirely virtualized. In some embodiments, the virtualization environment 94 includes components defined by the 0-RAN Alliance, such as an O-Cloud environment orchestrated by a Service Management and Orchestration Framework via an 0-2 interface.Applications 96 (which may alternatively be called software instances, virtual appliances, network functions, virtual nodes, virtual network functions, etc.) are run in the virtualization environment 94 to implement some of the features, functions, and / or benefits of some of the embodiments disclosed herein.

[0170] Hardware 98 includes processing circuitry, memory that stores software and / or instructions executable by hardware processing circuitry, and / or other hardware devices as described herein, such as a network interface, input / output interface, and so forth. Software may be executed by the processing circuitry to instantiate one or more virtualization layers 100 (also referred to as hypervisors or virtual machine monitors (VMMs)), provide VMs 102a and 102b (one or more of which may be generally referred to as VMs 102), and / or perform any of the functions, features and / or benefits described in relation with some embodiments described herein. The virtualization layer 100 may present a virtual operating platform that appears like networking hardware to the VMs 102.

[0171] The VMs 102 comprise virtual processing, virtual memory, virtual networking or interface and virtual storage, and may be run by a corresponding virtualization layer 100. Different embodiments of the instance of a virtual appliance 96 may be implemented on one or more of VMs 102, and the implementations may be made in different ways. Virtualization of the hardware is in some contexts referred to as network function virtualization (NFV). NFV may be used to consolidate many network equipment types onto industry standard high volume server hardware, physical switches, and physical storage, which can be located in data centers, and customer premise equipment.

[0172] In the context of NFV, a VM 102 may be a software implementation of a physical machine that runs programs as if they were executing on a physical, non-virtualized machine. Each of the VMs 102, and that part of hardware 98 that executes that VM, be it hardware dedicated to that VM and / or hardware shared by that VM with others of the VMs, forms separate virtual network elements. Still in the context of NFV, a virtual network function is responsible for handling specific network functions that run in one or more VMs 102 on top of the hardware 98 and corresponds to the application 96.

[0173] Hardware 98 may be implemented in a standalone network node with generic or specific components. Hardware 98 may implement some functions via virtualization. Alternatively, hardware 98 may be part of a larger cluster of hardware (e.g. such as in a data center or CPE) where many hardware nodes work together and are managed via management and orchestration 104, which, among others, oversees lifecycle management of applications 96. In some embodiments, hardware 98 is coupled to one or more radio unitsthat each include one or more transmitters and one or more receivers that may be coupled to one or more antennas. Radio units may communicate directly with other hardware nodes via one or more appropriate network interfaces and may be used in combination with the virtual components to provide a virtual node with radio capabilities, such as a radio access node or a base station. In some embodiments, some signaling can be provided with the use of a control system 106 which may alternatively be used for communication between hardware nodes and radio units.

[0174] Having described the general process flow of arrangements of the disclosure and having provided examples of hardware and software arrangements for implementing the processes and functions of the disclosure, the sections below provide details and examples of arrangements for using metadata for AI / ML positioning models to determine if a model is appropriate for a development environment.

[0175] Some embodiments provide for using metadata for AI / ML positioning models to determine if a model is appropriate for a development environment. One or more requesting node 4 functions described below may be performed by one or more of processing circuitry 36, processor 38, comparing unit 8, communication interface 29, etc. One or more model host node 6 functions described below may be performed by one or more of radio interface 46, antenna 48, processing circuitry 50, processor 52, range unit 9, etc.

[0176] Type of metadata information elements

[0177] Regarding training data collection of Part A (reference signal quality), the model host node 6 (e.g., TRP, network node 16, LMF 15 or UE 22) integrates measurements together with related metadata. To enable accurate inference results, the metadata should ideally not differ between the training data set and the measurement performed during inference, or the difference is within a limited, acceptable range.

[0178] The metadata that need to be recorded for Part A (reference signal quality) of the training dataset can be grouped as follows:

[0179] • Group 1, assistance data. In the positioning framework, assistance data is sent from LMF 15 to either network nodes 16 or UEs 22 to enable reference signal measurements. For training data collection, the LMF 15 provides suitable assistance data to the measuring nodes.

[0180] • Group 2, signal measurement configuration, for example, PDP or DP, measurement parameters {Nt, Nt', k}. For Case 2b, 3b, and potentially Case 3a, this is the information included the measurement request, sent from LMF 15 to UE 22or network node 16, for model inference. For cases where the channel measurement is not sent to LMF 15 for model inference (Case 1, 2a, and potentially 3a), the analogous signal measurement information should be recorded as well.

[0181] • Group 3, measurement environment condition and quality. This may include, for example: network synchronization error range, UE 22 / network node 16 RX and TX timing error range, interference range, channel estimation error range, etc. For Group 3, Rel-18 study item evaluations show that such context information also affects the validity of the trained model in the deployment scenario. Thus, such context information about the training dataset is also recorded as part of the metadata, including:

[0182] - Range of SNR / SINR.

[0183] Range of Channel estimation error.

[0184] Range of NW synchronization error.

[0185] Range of UE 22 / network node 16 receive (RX) and transmit (TX) timing error. Deployment type, for example, various indoor factory scenarios, the clutter parameters on density, height, width of clutters on the factory floor.

[0186] One or more parameters from Groups 1, 2 and 3 may need to be consistent between training and inference (or the deviation is within a limited, acceptable range), so that positioning methods can be used with an AI / ML model. Ideally, the consistency condition is established before the LMF 15 starts a positioning procedure, in order not to waste overhead requesting measurements that would be either impossible for the network node 16 / UE 22 to produce, or unreliable.

[0187] Group 1 : There is already a mechanism to assert whether the assistance data during inference is consistent with the model trained, re-using on- demand positioning procedure. However, this requires a large overhead of message exchange between LMF 15 and UE 22. It is not possible as of now for a non-serving network node 16 to request a specific SRS configuration to fit the model.

[0188] Group 2: consistency of group 2 parameters can be left to implementation. The LMF 15 can request the appropriate measurements in case 3b, and case l / 3a can also maintain Group 2 parameters within the model host implementation.Group 3: network conditions at the training stage as well as measurement range of the model are not known by the LMF 15. Hence for efficient positioning procedures with reduced overhead, reporting of these information is beneficial.

[0189] For model metadata Group 3 (measurement environment condition and quality), there is no existing mechanism to provide it to the wireless nodes involved in model inference. However, evaluations have shown that Group 3 metadata provides important context information about the trained model. If substantial inconsistency exists between training and inference for such context condition, the AI / ML model is not expected to perform well.

[0190] Therefore, there is a need to introduce mechanisms to ensure metadata Group 3 consistency for training data collection, model training, and model inference.

[0191] One or more embodiments described herein focuses on metadata information elements that provide environment condition and quality of the wireless channel measurement. Typical type of information in this category include:

[0192] • Network synchronization error.

[0193] • UE 22 / network node 16 RX and TX timing error.

[0194] • Interference to the wireless signal under measurement.

[0195] • Channel estimation error.

[0196] Wireless signal measurement condition for metadata of the AI / ML model

[0197] The inference condition of the wireless signal under measurement can be quantified in a variety of formats, including:

[0198] • reference signal received power (RSRP).

[0199] • reference signal received quality (RSRQ).

[0200] • signal-to-noise and interference ratio (SINR).

[0201] Downlink wireless signal measurement condition

[0202] For DL measurement for positioning, positioning reference signal (PRS) is typically used. The PRS is typically quasi-collocated (QCL) with another DL reference signal, either a SSB or another DL-PRS resource. One example of signaling the QCL information of PRS is shown below.

[0203] DL-PRS -QCL-Info-r 16 ::= CHOICE {

[0204] ssb-r!6 SEQUENCE {

[0205] pci-rl6 NR-PhysCellID-rl6,

[0206] ssb-Index-rl6 INTEGER (0..63),rs-Type-rl6 ENUMERATED {typeC, typeD, typeC-plus-typeD}

[0207]

[0208] In order to capture the inference condition of PRS measurement, the interference condition of its QCL-ed RS can be used. Specifically, the quality of the secondary synchronization (SS) signal of the quasi-collocated SSB can be used. Such SS signal quality can be provided by:• SS reference signal received power (SS-RSRP). SS reference signal received power (SS-RSRP) is defined as the linear average over the power contributions (in [W], e.g., in Watts) of the resource elements that carry secondary synchronization signals.

[0209] • SS reference signal received quality (SS-RSRQ). Secondary synchronization signal reference signal received quality (SS-RSRQ) is defined as the ratio of N*SS-RSRP / NR carrier Received Signal Strength Indicator (RSSI), where N is the number of resource blocks in the NR carrier RSSI measurement bandwidth. In the above, RSSI comprises the linear average of the total received power (in [W]) observed only per configured Orthogonal Frequency-Division Multiple (OFDM) Access symbol and in the measurement bandwidth (e.g., indicated by higher layers or the defined channel bandwidth of the carrier). The total received power is observed by the UE 22 from all (or one or more) sources, including co-channel serving and non-serving cells, adjacent channel interference, thermal noise, etc.

[0210] • SS signal-to-noise and interference ratio (SS-SINR). SS signal-to-noise and interference ratio (SS-SINR), is defined as the linear average over the power contribution (in [W]) of the resource elements carrying secondary synchronisation signals divided by the linear average of the noise and interference power contribution (in [W]).

[0211] In a given serving cell, the SSB may also be QCL-ed with CSI-RS. Thus, it may be useful to include information on the CSI-RS signal quality, including:

[0212] • CSI Reference signal received power (CSI-RS RP). CSI reference signal received power (CSI-RSRP), is defined as the linear average over the power contributions (in [W]) of the resource elements of the antenna port(s) that carry CSI reference signals configured for RSRP measurements within the considered measurement frequency bandwidth in the configured CSI-RS occasions.

[0213] • CSI Reference Signal Received Quality (CSI-RSRQ). CSI reference signal received quality (CSI-RSRQ) is defined as the ratio of N*CSI-RSRP to CSI-RSSI, where N is the number of resource blocks in the CSI-RSSI measurement bandwidth. The measurements in the numerator and denominator is made over the same set of resource blocks.

[0214] • CSI signal-to-noise and interference ratio (CSI-SINR). CSI signal-to-noise and interference ratio (CSI-SINR), is defined as the linear average over the powercontribution (in [W]) of the resource elements carrying CSI reference signals divided by the linear average of the noise and interference power contribution (in [W]).

[0215] After obtaining the relevant downlink reference signal quality metric(s) described above, it is reported to provide the context or environment condition of the measurement for positioning. One example of signaling it as a part of measurement result is shown below, where RSRP and RSRQ are reported:

[0216] MeasuredResultsElement ::= SEQUENCE {

[0217] physCellld INTEGER (0 .503),

[0218] cellGloballd CellGloballdEUTRA-AndUTRA OPTIONAL, arfcnEUTRA ARFCN-ValueEUTRA,

[0219] systemFrameNumber BIT STRING (SIZE (10)) OPTIONAL, rsrp-Result INTEGER (0..97) OPTIONAL, rsrq-Result INTEGER (0..34) OPTIONAL,

[0220] In some embodiments, the RSRP and / or RSRQ is included with each measurement result in training data collection. Alternatively, the RSRP and / or RSRQ is recorded when it has fluctuated significantly, e.g., more than a predefined threshold.

[0221] In some embodiments, when performing model training, the RSRP range and / or RSRQ range of the training data samples used in model training are recorded and attached to the trained model as its metadata.

[0222] In some embodiments, when preparing for model inference, the RSRP range and / or RSRQ range of the trained model is compared with those observed at model inference stage. If the RSRP and / or RSRQ observed at model inference is compatible with those of the trained model, the model inference may proceed; otherwise, the model may be considered inappropriate and model inference is not expected to perform well. As used herein, in one or more embodiments, the compatibility refers to that the RSRP and / or RSRQ observed at model inference is within the acceptable tolerance of the RSRP and / or RSRQ range of the training data samples used to obtain the trained model. For example, if the metadata of the model indicates that it is trained to perform within RSRQ range of [RSRQ min, RSRQ max], then the acceptable tolerance may be: [RSRQ min - 8_rsrq, RSRQ max + S rsrq], Thus, in terms of RSRQ, compatibility can be declared if the observed RSRQ is within [RSRQ min - S rsrq, RSRQ max + S rsrq]; otherwise,incompatibility can be declared. Similar compatibility checking can be performed for RSRQ and SINR.

[0223] Uplink wireless signal measurement condition

[0224] For uplink reference signal for positioning, the uplink channel measurement is performed using the UL sounding reference signal (UL SRS). The quality of the UL SRS can be reflected by the UL SRS-RSRP measurement.

[0225] • UL SRS reference signal received power (UL SRS-RSRP). UL SRS reference signal received power (UL SRS-RSRP) is defined as linear average of the power contributions (in [W]) of the resource elements carrying sounding reference signals (SRS).

[0226] Considering downlink-uplink reciprocity, the downlink reference signal quality can be used to provide environment condition and quality of the wireless channel measurement for uplink SRS. Such downlink reference signal include (Synchronization Signal (SS) and CSI-RS. Hence, SS and CSI-RS signal quality metrics can be used, including: SS-Reference Signal Received Power (RSRP), SS-RSRQ, SS-SINR, CSI-RSRP, CSI-RSRQ, CSI-SINR.

[0227] After obtaining the reference signal quality metric(s), it is reported to provide the context or environment condition of the SRS measurement. One example of signaling it as a part of measurement result is shown below, where RSRP and RSRQ are reported:

[0228]

[0229]

[0230]

[0231]

[0232]

[0233] Signaling of network synchronization error for metadata of the AI / ML model Network synchronization error for downlink wireless signal measurement For an AI / ML model based on downlink reference signals (e.g., DL-PRS), IE NR-RTD-Info provides the time synchronization information between the reference TRP and neighbor TRPs. Specifically, IE subframeOffset is provided on the subframe boundary offset between the reference TRP and neighbor TRPs.

[0234] Thus NR-RTD-Info can serve as metadata on network synchronization error.• For training data collection, NR-RTD-Info is recorded and covers all (or one or more of) the related training data samples, since the network synchronization status is not expected to change fast nor frequently.

[0235] • For model training, the NR-RTD-Info of all training data samples used to train the model are recorded as metadata of the trained model. Note that if mixed training data sets are used in training, where each sub-training dataset is associated with a different NR-RTD-Info, then the range of aggregate RTD info is recorded as metada of the trained model.

[0236] • At model inference, the RTD info provided by network node 16 (or LMF 15) is checked against the RTD info of the trained model. If the RTD at model inference is within the acceptable tolerance of the RTD of the trained model, then model inference may proceed; otherwise, model inference may not proceed, since the trained model is not compatible with the network condition at inference, and the model inference is not expected to generate accurate output.

[0237] NR-RTD-Info

[0238] The information element (IE) NR-RTD-Info is used by the location server to provide time synchronization information between a reference TRP and a list of neighbour TRPs together with integrity information.

[0239] - ASN1 START

[0240] NR-RTD-Info-rl6 ::= SEQUENCE {

[0241] referenceTRP-RTD-Info-r 16 ReferenceTRP-RTD-Info-rl 6,

[0242] rtd-InfoList-r!6 RTD-InfoList-rl6,

[0243] ReferenceTRP-RTD-Info-rl 6 ::= SEQUENCE {

[0244] dl-PRS-ID-Ref-rl6 INTEGER (0..255),

[0245] nr-PhysCellID-Ref-rl6 NR-PhysCellID-rl6 OPTIONAL, - Need ON

[0246] nr-CellGlobalID-Ref-rl6 NCGI-rl5 OPTIONAL, - Need ON

[0247] nr-ARFCN-Ref-rl6 ARFCN-ValueNR-rl5 OPTIONAL, - Need ON

[0248] refTime-rl6 CHOICE {systemFrameNumber-r!6 BIT STRING (SIZE (10)),

[0249] utc-r!6 UTCTime,

[0250] rtd-RefQuality-r!6 NR-TimingQuality-rl 6 OPTIONAL, —Need ON

[0251] RTD-InfoList-rl 6 ::= SEQUENCE (SIZE (L.nrMaxFreqLayers-rl6)) OF RTD- InfoListPerFreqLay er-rl 6

[0252] RTD-InfoListPerFreqLayer-rl6 ::= SEQUENCE (SIZE(l..nrMaxTRPsPerFreq-rl6)) OF RTD-InfoElement-rl 6

[0253] RTD-InfoElement-rl 6 ::= SEQUENCE {

[0254] dl-PRS-ID-r!6 INTEGER (0 .255),

[0255] nr-PhysCelllD-rl 6 NR-PhysCellID-rl6 OPTIONAL, - Need ON

[0256] nr-CellGloballD-rl 6 NCGI-rl5 OPTIONAL, - Need ON

[0257] nr-ARFCN-r!6 ARFCN-ValueNR-rl5 OPTIONAL, - Need ON

[0258] subframeOffset-r 16 INTEGER (0..1966079),

[0259] rtd-Quality-r!6 NR-TimingQuality-rl 6,

[0260] [[

[0261] nr-IntegrityRTD-InfoBounds-rl 8 NR-IntegrityRTD-InfoBounds-rl 8 OPTIONAL - Need OR

[0262] ]]

[0263]

[0264] meanRTD-rl8 INTEGER (0..255),stdDevRTD-r 18 INTEGER (0..31 ),

[0265] - ASN1STOP

[0266] subframeOffset

[0267] This field specifies the subframe boundary offset at the TRP antenna location between the reference TRP and this neighbour TRP in time units / ,. =

[0268]

[0269] ■ Nt) where A / max = 480 ■ 103

[0270]

[0271] The offset is counted from the beginning of a subframe #0 of the reference TRP to the beginning of the closest subsequent subframe of this neighbour TRP.

[0272] Scale factor 1 Tc.

[0273] Network synchronization error for uplink wireless signal measurement

[0274] For AI / ML model based on uplink reference signals (e.g., UL-SRS), IE “System Frame Number (SFN) Initialisation Time” of each TRP provides the time synchronization information between two TRPs.

[0275] Thus, “SFN Initialisation Time” of each TRP associated with UL-SRS measurement can serve as metadata on network synchronization error. Alternatively, the time synchronization error between a pair of TRPs can be obtained from their “SFN Initialisation Time” and used as metadata.

[0276] • For training data collection, ‘‘SFN Initialisation Time” of each TRP is recorded and covers all the related training data samples, since the network synchronization status is not expected to change fast nor frequently.

[0277] • For model training, “SFN Initialisation Time” of each TRP in all training data samples used to train the model are recorded as metadata of the trained model. Note that if mixed training data sets are used in training, where each sub- training dataset is associated with a different set of “ SFN Initialisation Time,” then the range of aggregate time synchronization error between the TRPs is obtained from the sets o “SFN Initialisation Time,” and recorded as metada of the trained model.

[0278] • At model inference, the time synchronization error between the TRPs can be obtained from each TRP’s “SFN Initialisation Time.” The time synchronization error at inference is checked against the network synchronization metadata of the trained model. If the time synchronization error at model inference is within the acceptable tolerance of the trained model, then model inference may proceed;otherwise, model inference may not proceed, since the trained model is not compatible with the network condition at inference, and the model inference is not expected to generate accurate output.

[0279]

[0280]

[0281] Signaling of metadata range

[0282] One or more embodiments described above relate to how a device hosting a model could record metadata when building the data set for training the model. Once data collection is completed, the range of metadata for which the model was trained constitutes capability of the model. Such information can be used by the LMF 15 to assess whether to use a AI / ML model based positioning method.

[0283] In some embodiments, a capability exchange is established between the wireless node hosting the model (e.g., model host node 6) and the requesting node 4 (e.g., LMF 15).

[0284] In one example, the wireless node hosting the model (e.g., model host node 6) is the UE 22. The UE 22 provides the metadata of its model to the requesting node 6 (e.g., LMF 15) via capability reporting, so that the LMF 15 may assess whether it is appropriate to activate the AI / ML model hosted at UE 22 for positioning. The LMF 15 may indicate this assessment to the UE 22 via, for example, the networknode 16. In a preferred embodiment, the capability exchange is performed upon request from the LMF 15 to the UE 22 via LTE Positioning Protocol (LPP). The UE 22 may also voluntarily provide the meta of its model as a type of UE 22 assistance information.

[0285] In another example, the wireless node hosting the model (e.g., model host node 6) is a network node 16 or a TRP. The network node 16 provides the metadata of its model to the requesting node 4 (e.g., LMF 15) over NR Positioning Protocol A (NRPPa) protocol, so that the LMF 15 may assess whether it is appropriate to activate the AI / ML model hosted at network node 16 for positioning.

[0286] In a related example, if the network node 16 uses a split architecture and the model is hosted at TRP (i.e., gNB-DU), the metadata of the model is also exchanged over the F1AP protocol between gNB-Distributed Unit (DU) and gNB-Centralized Unit (CU).

[0287] In another example, the capability exchange may be between a network node 16 or TRP hosting the model (i.e., model host node 6) and a UE 22.

[0288] In yet another example, the capability exchange may be between a network node 16 and a UE 22 hosting the model.

[0289] In one embodiment, the capability exchange may be for one or a plurality of capabilities. In one example, the requesting node 4 (e.g., network node 16, TRP, or LMF 15) requests the model host node 6 (e.g., network node 16, TRP or UE 22) to provide one or several of:

[0290] The operating range for one or several of RSRP, RSRQ, SINR for a given CSI-RS resource, or SS, or SRS resource as discussed above.

[0291] The range of timing error as provided by NR-RTD-Info or SFN initialization time.

[0292] The range for which the model was trained may also be updated while the model is deployed, if the model can be re-trained, or fine-tuned, or updated during deployment. For example, the model may leverage on additional collected data during inference in favorable conditions. In one embodiment, the capability exchange may also consist of an initial exchange of the metadata, followed by a periodic update of the metadata. The periodic update may be configured by the LMF 15 as part of the initial exchange, e.g., the periodicity of reporting the model’s metadata. Alternatively, the model host node 6 could on its own initiate the reporting for an metadata update, when the model metadatachanges. For example, if the model is hosted by the UE 22, the UE 22 may voluntarily contact the LMF 15 with an update of its model’s metadata.

[0293] Some additional examples include the following:

[0294] Example Al. A model host node 6 hosting an artificial intelligence / machine learning (AI / ML) model, the model host node 6 configured to communicate with a requesting node 4, the model host node 6 configured to, and / or comprising a radio interface 46 and / or processing circuitry 50 configured to:

[0295] collect model metadata of the AI / ML model;

[0296] send the model metadata to the requesting node 4; and

[0297] receive an indication to activate the AI / ML model, the indication being based on the model metadata.

[0298] Example A2. The model host node 6 of Example Al, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0299] Example A3. The model host node 6 of any of Examples A1-A2, wherein the range of AI / ML values are updated while the model is deployed.

[0300] Example A4. The model host node 6 of any of Examples Al -A3, wherein the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

[0301] Example A5. The model host node 6 of any of Examples A1-A4, wherein the model metadata is sent to the requesting node via a capability exchange.

[0302] Example A6. The model host node 6 of Example A5, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution, LTE, Positioning Protocol (LPP).

[0303] Example A7. The model host node 6 of any of Examples A5-A6, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0304] Example A8. The model host node 6 of any of Examples A1-A7, wherein the model host node 6 is a user equipment (UE) 22.

[0305] Example A9. The model host node 6 of Example A8, wherein the UE 22 voluntarily contacts the requesting node 4 with an update of the model metadata.

[0306] Example A10. The model host node 6 of any of Examples A8-A9, wherein the model host node 6 provides the model metadata as a type of UE22 assistance information.Example All. The model host node 6 of Examples A1-A7, wherein the model host node 6 is a network node 16.

[0307] Example A12. The model host node 6 of Example Al 1, wherein the model host node 6 sends the model metadata to the requesting node 4 via a New Radio Positioning Protocol a (NRPPa) protocol.

[0308] Example Al 3. The model host node6 of Examples A1-A7, wherein the model host node 6 is a transmission and reception point (TRP).

[0309] Example Bl . A method implemented in a model host node 6 that is configured to communicate with a requesting node 4, the method comprising:

[0310] collecting model metadata of an artificial intelligence / machine learning (AI / ML) model; and

[0311] sending the model metadata to the requesting node 4; and

[0312] receiving an indication to activate the AI / ML model, the indication being based on the model metadata.

[0313] Example B2. The method of Example Bl, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0314] Example B3. The method of any of Examples B1-B2, wherein the range of AI / ML values are updated while the model is deployed.

[0315] Example B4. The method of any of Examples B1-B3, wherein the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality

[0316] and network timing error.

[0317] Example B5. The method of any of Examples B1-B4, wherein the model metadata is sent to the requesting node via a capability exchange.

[0318] Example B6. The method of Example B5, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution (LTE) Positioning Protocol (LPP).

[0319] Example B7. The method of any of Examples B5-B6, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0320] Example B8. The model method of any of Examples B1-B7, wherein the model host node is a user equipment (UE) 22.Example B9. The method of Example B8, wherein the UE voluntarily contacts the requesting node with an update of the model metadata.

[0321] Example BIO. The method of Examples B8-B9, wherein the model host node 6 provides the model metadata as a type of UE assistance information.

[0322] Example Bl 1. The method of any of Examples B1-B7, wherein the model host node 6 is a network node 16.

[0323] Example Bl 2. The method of Example Bl 1, wherein the model host node 6 sends the model metadata to the requesting node via a New Radio Positioning Protocol a (NRPPa) protocol.

[0324] Example B13. The method of any of Examples B1-B7, wherein the model host node 6 is a transmission and reception point (TRP).

[0325] Example Cl . A requesting node 4 configured to communicate with a model host node 6, the model host node 6 hosting an artificial intelligence / machine learning (AI / ML) model, the requesting node 4 configured to, and / or comprising a radio interface 30 and / or comprising processing circuitry 36 configured to:

[0326] obtain measurements;

[0327] receive model metadata from the model host node 6; and

[0328] determine whether the model host node 6 can activate the AI / ML model based on the obtained measurements and the model metadata.

[0329] Example C2. The requesting node 4 of Example Cl, wherein the obtained measurements comprise at least timing and power measurements.

[0330] Example C3. The requesting node 4 of any of Examples C1-C2, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0331] Example C4. The requesting node 4 of any of Examples C1-C3, wherein the range of AI / ML values are updated while the model is deployed

[0332] Example C5. The requesting node 4 of any of Examples C1-C4, wherein the model metadata is received via a capability exchange.

[0333] Example C6. The requesting node 4 of Example C5, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution (LTE) Positioning Protocol (LPP).Example C7. The requesting node 4 of any of Examples C5-C6, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

[0334] Example C8. The requesting node 4 of any of Examples C1-C7, wherein the model metadata is received from the model host node 6 via a New Radio Positioning Protocol a (NRPPa) protocol.

[0335] Example C9. The requesting node 4 of any of Examples C1-C8, wherein the requesting node 4 is one of a network node 16, a user equipment (UE) 22 a Location Management Function (LMF) and a transmission and reception point (TRP).

[0336] Example DI. A method implemented in a requesting node 4 that is configured to communicate with a model host node 6 that host an artificial intelligence / machine learning (AI / ML) model, the method comprising:

[0337] obtaining measurements;

[0338] receiving model metadata from a model host node 6; and

[0339] determining whether the model host node 6 can activate the AI / ML model based on the obtained measurements and the model metadata.

[0340] Example D2. The method of Example DI, wherein the obtained measurements comprise at least timing and power measurements.

[0341] Example D3. The method of any of Examples D1-D2, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

[0342] Example D4. The method of any of Examples D1-D3, wherein the range of AI / ML values are updated while the model is deployed.

[0343] Example D5. The method of any of Examples D1-D4, wherein the metadata is received via a capability exchange.

[0344] Example D6. The method of Example D5, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution (LTE) Positioning Protocol (LPP).

[0345] Example D7. The method of any of Examples D5-D6, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.Example D8. The method of any of Examples D1-D7, wherein the model metadata is received from the model host node via a New Radio Positioning Protocol a (NRPPa) protocol.

[0346] Example D9. The method of any of Examples D1-D8, wherein the requesting node 4 is one of a network node 16, a user equipment (UE) 22, a Location Management Function (LMF), and a transmission and reception point (TRP).

[0347] As will be appreciated by one of skill in the art, the concepts described herein may be embodied as a method, data processing system, computer program product and / or computer storage media storing an executable computer program. Accordingly, the concepts described herein may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects all generally referred to herein as a “circuit” or “module.” Any process, step, action and / or functionality described herein may be performed by, and / or associated to, a corresponding module, which may be implemented in software and / or firmware and / or hardware. Furthermore, the disclosure may take the form of a computer program product on a tangible computer usable storage medium having computer program code embodied in the medium that can be executed by a computer. Any suitable tangible computer readable medium may be utilized including hard disks, CD-ROMs, electronic storage devices, optical storage devices, or magnetic storage devices.

[0348] Some embodiments are described herein with reference to flowchart illustrations and / or block diagrams of methods, systems and computer program products. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer (to thereby create a special purpose computer), special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0349] These computer program instructions may also be stored in a computer readable memory or storage medium that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructionmeans which implement the function / act specified in the flowchart and / or block diagram block or blocks.

[0350] The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0351] It is to be understood that the functions / acts noted in the blocks may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may in fact be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending upon the functionality / acts involved. Although some of the diagrams include arrows on communication paths to show a primary direction of communication, it is to be understood that communication may occur in the opposite direction to the depicted arrows.

[0352] Computer program code for carrying out operations of the concepts described herein may be written in an object oriented programming language such as Python, Java® or C++. However, the computer program code for carrying out operations of the disclosure may also be written in conventional procedural programming languages, such as the "C" programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer. In the latter scenario, the remote computer may be connected to the user's computer through a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0353] Many different embodiments have been disclosed herein, in connection with the above description and the drawings. It will be understood that it would be unduly repetitious and obfuscating to literally describe and illustrate every combination and subcombination of these embodiments. Accordingly, all embodiments can be combined in any way and / or combination, and the present specification, including the drawings, shall be construed to constitute a complete written description of all combinations and subcombinations of the embodiments described herein, and of the manner and process ofmaking and using them, and shall support claims to any such combination or subcombination.

[0354] It will be appreciated by persons skilled in the art that the embodiments described herein are not limited to what has been particularly shown and described herein above. In addition, unless mention was made above to the contrary, it should be noted that all of the accompanying drawings are not to scale. A variety of modifications and variations are possible in light of the above teachings without departing from the scope of the following claims.

Claims

1. Claims:

1. A method implemented in a model host node (6) hosting an artificial intelligence / machine learning, AI / ML, model for estimating a location of a target device, the model host node (6) configured to communicate with a requesting node (4), the method comprising:collecting data to form a training dataset (SI 16);compiling context information of the collected data and attach the context information to the training dataset as model metadata (SI 18);sending the model metadata to the requesting node (4) (S120); andreceiving an indication as to whether the model host node (6) can activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata (S 122).

2. A method implemented in a requesting node (4) that is configured to communicate with a model host node (6) hosting an artificial intelligence / machine learning, AI / ML, model for estimating a location of a target device, the method comprising:receiving model metadata from the model host node (6), the model metadata including a training dataset of collected data and context information of the collected data (SI 06); andindicating to the model host node (6) whether the model host node (6) can activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata (SI 08).

3. A model host node (6) hosting an artificial intelligence / machine learning (AI / ML) model for estimating a location of a target device, the model host node (6) configured to communicate with a requesting node (4), the model host node (6) configured to:collect data to form a training dataset;compile context information and attach the context information to the training dataset as model metadata;send the model metadata to the requesting node (4); andreceive an indication to activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata.

4. The model host node (6) of Claim 3, wherein the model metadata includes assistance data to enable reference signal measurements.

5. The model host node (6) of any one of Claims 3 and 4, wherein the model metadata includes elements that provide environment condition and quality of wireless channel measurement.

6. The model host node (6) of Claim 5, wherein the elements that provide environment condition and quality of wireless channel measurement include one or more network synchronization error range, reception and transmission timing error range, interference range, and channel estimation error range.

7. The model host node (6) of any one of Claims 3-6, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

8. The model host node (6) of Claim 7, wherein the range of AI / ML values are updated while the model is deployed.

9. The model host node (6) of any one of Claims 3-8, wherein the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

10. The model host node (6) of any one of Claims 3-9, wherein the model metadata is sent to the requesting node via a capability exchange.

11. The model host node (6) of Claim 10, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution, LTE, Positioning Protocol, LPP.

12. The model host node (6) of any one of Claims 10 and 11, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

13. The model host node (6) of any one of Claims 3-12 wherein the model host node (6) is the target device.

14. The model host node (6) of any one of Claims 3-12, wherein the model host node (6) is a network node (16).

15. The model host node (6) of Claim 14, wherein the model host node (6) is further configured to send the model metadata to the requesting node (4) via a New Radio Positioning Protocol a, NRPPa, protocol.

16. The model host node (6) of any one of Claims 3-12, wherein the model host node (6) is a transmission and reception point, TRP.

17. The model host node (6) of any one of Claims 3-16, wherein the target device is a user equipment, UE (22).

18. The model host node (6) of any one of Claims 3-17, wherein the estimated location is used for one or more of positioning the target device and sensing the target device.

19. A requesting node (4) that is configured to communicate with a model host node (6) hosting an artificial intelligence / machine learning, AI / ML, model for estimating a location of a target device, the requesting node (4) configured to:receive model metadata from the model host node (6), the model metadata including a training dataset of collected data and context information of the collected data; andindicate to the model host node (6) whether the model host node (6) can activate the AI / ML model for estimating the location of the target device, the indication being based on the model metadata.

20. The requesting node (4) of Claim 19, wherein the model metadata includes assistance data to enable reference signal measurements.

21. The requesting node (4) of any one of Claims 19 and 20, wherein the model metadata includes elements that provide environment condition and quality of wireless channel measurement.

22. The requesting node (4) of Claim 21, wherein the elements that provide environment condition and quality of wireless channel measurement include one or more network synchronization error range, reception and transmission timing error range, interference range, and channel estimation error range.

23. The requesting node (4) of any one of Claims 19-22, wherein the model metadata includes a range of AI / ML values for which the model was trained during a model training stage.

24. The requesting node (4) of Claim 23, wherein the range of AI / ML values are updated while the model is deployed.

25. The requesting node (4) of any one of Claims 19-24, wherein the model metadata comprises aggregate value ranges when using a mixed training dataset with one or more different value ranges of at least one of a reference signal quality and network timing error.

26. The requesting node (4) of any one of Claims 19-25, wherein the model metadata is received the model host node (6) via a capability exchange.

27. The requesting node (4) of Claim 26, wherein the capability exchange is performed upon request from the requesting node via Long Term Evolution, LTE, Positioning Protocol, LPP.

28. The requesting node (4) of any one of Claims 26 and 27, wherein the capability exchange comprises an initial exchange of the model metadata, followed by a periodic update of the model metadata.

29. The requesting node (4) of any of Claims 19-28, wherein the requesting node (4) is one of a network node (16), a user equipment, UE (22), a Location Management Function, LMF, and a transmission and reception point, TRP.

30. The requesting node (4) of any of Claims 19-29, wherein the target device is a user equipment, UE (22).

31. The requesting node (4) of any of Claims 19-30, wherein the estimated location is used for one or more of positioning the target device and sensing the target device.

Images