Data collection and reasoning

Abstract

Various examples of the present disclosure relate to data collection and reasoning and provide an apparatus (10, 110) comprising: a model (212u) wherein the model is associated with a first set of association identifiers (IDs) (205u) wherein the first set of association IDs is associated with a first set of network side conditions; means (11) for sending first information (502) to a node (130, 140) of the core network, the first information indicating a first set of association IDs; means (11) for receiving second information (505) from a node of the core network, the second information indicating a degree of match (506) between: a first set of association IDs, and a second set of association IDs (205bs), where the second set of association IDs is associated with a second set of one or more network-side conditions; and means (11) for determining whether to select the model, wherein the determining is based at least in part on the second information.

Description

Technical Field

[0001] Some examples of this disclosure relate to apparatus, methods, and computer programs for data collection. Some examples of this disclosure relate to apparatus, methods, and computer programs for inference. Some examples relate to apparatus, methods, and computer programs for collecting training data for training artificial intelligence / machine learning (AI / ML) models. Some examples relate to apparatus, methods, and computer programs for performing inference using trained AI / ML models. Background Technology

[0002] The conventional apparatus and processes used to collect training data for training AI / ML models (e.g., for AI / ML localization) are not always optimal. In some cases, it may be desirable to provide improved apparatus, methods, and computer programs for collecting training data for training AI / ML models.

[0003] Conventional apparatuses and processes for performing inference using trained AI / ML models (e.g., for AI / ML localization) are not always optimal. In some cases, it may be desirable to provide improved apparatuses, methods, and computer programs for performing inference using trained AI / ML models.

[0004] In some cases, it may be desirable to provide consistency between training an AI / ML model and performing inference using the trained AI / ML model. In some cases, it may be desirable to provide consistency between the network-side additional conditions used by radio access network (RAN) node entities involved in the collection of training data for training(multiple) AI / ML models and the network-side additional conditions of RAN node entities involved in inference using the trained(multiple) AI / ML models (such as a localization process utilizing a set of RAN node entities). In some cases, it may be desirable to provide consistency between the network-side additional conditions used by RAN node entities during the collection of training data for training(multiple) AI / ML models. Summary of the Invention

[0005] This invention is defined in the independent claims.

[0006] Based on various, but not necessarily all, examples provided in this disclosure are examples claimed under the appended claims. Any examples and features described in this specification that do not fall within the scope of the independent claims should be interpreted as examples useful for understanding the various embodiments of the invention.

[0007] According to at least some examples of this disclosure, an apparatus is provided, comprising: The model is associated with the first set of association identifiers ID, and the first set of association IDs is associated with the first set of network-side conditions. A component used to send first information to nodes in the core network, the first information indicating a first set of associated IDs; A component for receiving second information from nodes in the core network, the second information indicating the degree of matching between the following two: The first set of associated IDs, and The second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in the second set; The component used to determine whether to select a model is based at least in part on second information.

[0008] Based on various, but not necessarily all, examples of this disclosure, a method is provided, including: The device sends first information to the nodes of the core network indicating a first set of associated ID models, wherein the device includes models associated with a first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions. At the device, a second piece of information indicating the degree of matching between the following two is received from the nodes of the core network: The first set of associated IDs, and The second set of associated IDs is associated with one or more network-side conditions in the second set. At the device, it is determined whether to select a model, wherein the determination is based at least in part on the second information.

[0009] According to various, but not necessarily all, examples of this disclosure, a computer program is provided that includes instructions that, when executed by a device, cause the device to perform the methods described above.

[0010] According to various, but not all, embodiments, an apparatus is provided, comprising: At least one processor; and At least one memory includes computer program code, and at least one memory stores instructions that cause the device to execute when executed by at least one processor. Send first information to the nodes of the core network indicating a first set of associated ID models, wherein the device includes models associated with a first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; Receive second information from the nodes in the core network indicating the degree of matching between the following two: The first set of associated IDs, and The second set of associated IDs is associated with one or more network-side conditions in the second set. Determine whether to select a model, wherein the determination is based at least in part on the second information.

[0011] According to various, but not all, examples of this disclosure, a non-transitory computer-readable medium is provided, which is encoded with instructions that, when executed by at least one processor of the device, cause the device to perform: Send first information to the nodes of the core network indicating a first set of associated ID models, wherein the device includes models associated with a first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; Receive second information from the nodes in the core network indicating the degree of matching between the following two: The first set of associated IDs, and The second set of associated IDs is associated with one or more network-side conditions in the second set. Determine whether to select a model, wherein the determination is based at least in part on the second information.

[0012] The following sections of this “Summary of the Invention” section describe various features that may be features of any of the examples described in the preceding sections of this “Summary of the Invention” section. The description of the function should also be considered as disclosing any components suitable for performing the function, or any instructions stored in at least one memory that, when executed by at least one processor, cause the device to perform the function.

[0013] In some, but not all, examples, the determination of whether to select a model is at least partially based on the matching degree satisfying the condition.

[0014] In some, but not all, examples: The device includes multiple models, each model being associated with a corresponding set of association IDs, and each set of association IDs being associated with a corresponding set of network-side conditions; The first information includes an indication of each of the multiple corresponding sets of associated IDs; The second information includes: for each of the multiple corresponding sets of associated IDs, an indication of the degree of matching between the following two: A corresponding set of associated IDs, and The second set of associated IDs; and The determination involves selecting a model from multiple models, at least in part, based on second information.

[0015] In some, but not all, examples, each association ID represents at least one of the following: One or more network-side additional conditions; One or more network-side hypotheses; One or more explicit assumptions about the nodes of the core network; or When collecting data for training the model, one or more network sides are appropriately configured by one or more entities.

[0016] In some, but not all, examples, the model includes at least one of the following: Artificial intelligence (AI) models; Machine learning (ML) models; Model for the device; or Models for AI / ML positioning on the user equipment (UE) side.

[0017] In some, but not necessarily all, examples, the instructions cause the device to also execute while being executed by at least one processor: The component receives configuration information from the nodes of the core network. The configuration information is used to configure the device to perform the positioning process (206), wherein the configuration information includes information indicating the following: One or more position reference signal (PRS) configurations (203), wherein the one or more PRS configurations include information indicating a second group of one or more entities; and A component that performs the positioning process based at least in part on configuration information, wherein the positioning process is performed using the selected model.

[0018] According to at least some examples of this disclosure, an apparatus is provided, comprising: A component for receiving first information from a user equipment (UE), the first information indicating a first set of association IDs associated with a model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; The component used to obtain the second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in the second set; A component for evaluating second information based at least in part on a first set of associated IDs and a second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and A component used to send a second message to the UE so that the UE can determine whether to select a model.

[0019] Based on various, but not necessarily all, examples of this disclosure, a method is provided, including: The device receives first information from the user equipment (UE), the first information indicating a first set of association IDs associated with the model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Obtain a second set of associated IDs at the device, wherein the second set of associated IDs is associated with one or more network-side conditions in a second set; At the device, the second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and The device sends a second message to the UE so that the UE can determine whether to select a model.

[0020] According to various, but not necessarily all, examples of this disclosure, a computer program is provided that includes instructions that, when executed by a device, cause the device to perform the methods described above.

[0021] According to various, but not all, embodiments, an apparatus is provided, comprising: At least one processor; and At least one memory including computer program code, the at least one memory storing instructions that, when executed by at least one processor, cause the device to perform: Receive first information from the user equipment (UE), the first information indicating a first set of association IDs associated with the model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Obtain the second set of associated IDs, where the second set of associated IDs is associated with one or more network-side conditions in the second set; The second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and A second message is sent to the UE so that the UE can determine whether to select a model.

[0022] According to various, but not all, examples of this disclosure, a non-transitory computer-readable medium is provided, which is encoded with instructions that, when executed by at least one processor of the device, cause the device to perform: Receive first information from the user equipment (UE), the first information indicating a first set of association IDs associated with the model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Obtain the second set of associated IDs, where the second set of associated IDs is associated with one or more network-side conditions in the second set; The second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and A second message is sent to the UE so that the UE can determine whether to select a model.

[0023] The following sections of this “Summary of the Invention” section describe various features that may be features of any of the examples described in the preceding sections of this “Summary of the Invention” section. The description of the function should also be considered as disclosing any components suitable for performing the function, or any instructions stored in at least one memory that, when executed by at least one processor, cause the device to perform the function.

[0024] In some, but not all, examples, retrieving the second set of associated IDs includes: Send a request for one or more associated IDs to one or more entities in the second group of entities; and One or more entities from the second group of entities respond to the request to receive one or more associated IDs.

[0025] In some, but not all, examples, the request includes information indicating at least one of the following: One or more PRS configurations; or One or more network-side additional conditions.

[0026] In some, but not necessarily all, examples, the instructions cause the device to execute when executed by at least one processor: Associating at least one of one or more association IDs with at least one of one or more entities, wherein the association is based at least in part on determination of which particular association ID was received from which particular one or more entities.

[0027] In some, but not all, examples: UE includes multiple models, where each model is associated with a corresponding set of association IDs, and each set of association IDs is associated with a corresponding set of entities; The first information includes an indication of each of the multiple corresponding sets of associated IDs; The second information includes: for each of the multiple corresponding sets of associated IDs, an indication of the degree of matching between the following two: A corresponding set of associated IDs, and The second set of associated IDs.

[0028] According to various, but not necessarily all, examples of this disclosure, an apparatus, module, circuit, chipset including processing circuitry, device and / or system configured to perform at least a portion (or include components for thereto) one or more methods described herein.

[0029] The description of functions and / or actions in this specification should be additionally construed as disclosing any components suitable for performing these functions and / or actions. The functions and / or actions described herein can be performed using any suitable method and in any suitable manner.

[0030] Examples claimed in the appended claims are provided according to various, but not necessarily all, embodiments.

[0031] Although the examples and optional features of this disclosure have been described separately, it should be understood that their provisions in all possible combinations and permutations are included within this disclosure. It should be understood that the various examples of this disclosure may include any or all of the features described with respect to other examples of this disclosure, and vice versa. Furthermore, it should be understood that any one or more features in any combination may be implemented / included / executable by means, methods, and / or computer program instructions as needed and suitably. The description of the function should also be considered as disclosing any means suitable for performing that function. Attached Figure Description

[0032] Some examples will now be described with reference to the accompanying drawings, in which: Figure 1 Examples of radio access networks suitable for use with embodiments of this disclosure are shown; Figure 2 An example of a method for collecting data using associated IDs is shown; Figure 3 An example of a method for obtaining the associated ID is shown; Figure 4 An example of a signaling diagram used to obtain the associated ID for data collection is shown; Figure 5 An example of a method for selecting a model for inference based on the association ID is shown; Figure 6 An example of a signaling graph used for inference with associated IDs is shown; Figure 7 Another example of a method for selecting a model for inference based on the association ID is shown; Figure 8 Another example of a signaling graph used for inference with associated IDs is shown; Figure 9 Examples of apparatuses used in conjunction with embodiments of this disclosure are shown; and Figure 10 An example of a computer program used in conjunction with embodiments of this disclosure is shown.

[0033] The accompanying drawings are not necessarily drawn to scale. For clarity and brevity, some features and views in the drawings may be shown schematically or enlarged to scale. For example, the dimensions of some elements in the drawings may be enlarged relative to other elements to aid illustration. Similar reference numerals are used in the drawings to indicate similar features. For clarity, not all reference numerals may appear in all drawings.

[0034] In the specification and figures, reference numerals without subscripts (e.g., 123) can be used as general references to features or categories / sets of features. Reference numerals with subscripts (e.g., 123_1) can be used as specific markers, for example, to distinguish different instances of a feature or category / set of features. A subscript may include two numbers, including a first number marking a group of instances and a second number marking different instances within that group. Numerical type subscript indices (e.g., 123_1) can be used to indicate specific instances of a category / member of a set; and reference numerals with variable type subscript indices (e.g., 123_i) can be used to mark non-specific instances of that category (members of a set).

[0035] Abbreviations / Definitions 3GPP Third Generation Partnership Project 5G fifth generation BS base station CN core network or core network functions gNB next-generation NodeB, 5G / NR base station LMF location management function LPP LTE positioning protocol NR New Radio NRPP New Radio Positioning Protocol NRPPa New Radio Positioning Protocol A NW Network RAN Radio Access Network TRP Transmitter / Receiver Point UE User Equipment Detailed Implementation Figure 1 An example of a network 100 applicable to the present disclosure is illustrated schematically. The network (which may be referred to as NW) includes multiple network entities (which may be referred to as NE), including: Terminal device 110 (which may be referred to as terminal node or user equipment UE). Access device 120 (which may be referred to as access node, gNodeB, gNB, or base station BS); and One or more core network devices 130 (which may be referred to as core nodes, core functions, core entities, core network entities - one of the core nodes / functions / entities is the location management function LMF 140).

[0036] Terminal node 110 and access node 120 communicate with each other. Access node 120 communicates with core node 130. Access node 120 and LMF 140 can communicate directly with each other. In some, but not all, examples, one or more access nodes 120 can communicate with each other. In some, but not all, examples, one or more core network nodes 130 can communicate with each other.

[0037] exist Figure 1 In the example shown, network 100 includes a radio telecommunications network in which at least some of the terminal nodes 110 and access nodes 120 communicate with each other using the transmission / reception of radio waves. In this respect, network 100 includes a radio access network (RAN), such as a cellular network comprising multiple cells 122, each cell served by access node 120. Access node 120 includes a cellular radio transceiver. Terminal node 110 includes a cellular radio transceiver.

[0038] In the examples shown and discussed below, Network 100 is the 3GPP New Radio NR network and its fifth-generation 5G New Radio NR technology. However, it should be understood that in other examples, Network 100 could be a network beyond 5G, such as the next-generation (i.e., sixth-generation 6G) radio network currently under development (i.e., NR network and its evolution into 5G technology).

[0039] The interface between terminal node 110 and access node 120 is radio interface 124 (e.g., Uu interface). The interface between access node 120 and one or more core nodes 130 is backhaul interface 128 (e.g., S1 and / or next-generation NG interface). The interface between one or more location servers 140 and one or more core nodes 130 is backhaul interface 132 (e.g., NL interface).

[0040] Depending on the specific deployment scenario, access node 120 can be a RAN node, such as an NG-RAN node. NG-RAN nodes can be gNodeBs or gNBs that provide NG user plane and control plane protocol termination to the UE. The gNB connects to the 5G core 5GC via the NG interface, specifically, for example, to the Access and Mobility Management Function (AMF) via the NG control plane NG-C, and to the User Plane Function (UPF) via the NG user plane NG-U interface. Access nodes 120 can interconnect with each other via the Xn interface 126.

[0041] Cellular network 100 can be configured to operate in licensed or unlicensed frequency bands, particularly unlicensed bands such as those that rely on the transmitting device to sense radio resources / mediums before transmission begins, such as via the Listen-Before-Speak (LBT) process; and the 60 GHz unlicensed band, where beamforming may be required to achieve the desired coverage.

[0042] Access node 120 can be deployed in NG standalone operation / scenarios. Access node 120 can be deployed in NG non-standalone operation / scenarios. Access node 120 can be deployed in carrier aggregation (CA) operation / scenarios. Access node 120 can be deployed in dual-connectivity DC operation / scenarios (i.e., multiple radio access technologies - dual connectivity, MR-DC, or NR-DC). Access node 120 can be deployed in multi-connectivity MC operation / scenarios.

[0043] In such a non-standalone / dual-connectivity deployment, access nodes 120 can interconnect with each other via X2 or Xn interfaces and connect to the Evolved Packet Core (EPC) via the S1 interface or to the 5GC via the NG interface.

[0044] In addition to communicating via access node 120 of network 100 (i.e., communicating with other terminal nodes), terminal node 110 can also be enabled and configured to communicate directly with one or more other terminal nodes. In this regard, the terminal node can be enabled and configured to perform device-to-device (D2D) communication – which can be referred to as sidelink (SL) communication. This D2D / SL communication can use the PC5 interface. PC5 refers to the reference point where a terminal node communicates directly with another terminal node via a direct channel (i.e., without needing to communicate via an access node). D2D communication can be short-range, network-free, and direct. SL in New Radio (NR) is defined in 3GPP Release 16 for 5G NR.

[0045] exist Figure 1 In the example, core node 130 is shown as a single entity. In some examples, core node 130 may be distributed across multiple entities. For example, core node 130 may be cloud-based or distributed in any other suitable manner. Core nodes / core entities may provide one or more functionalities, specifically such as: User Plane Function (UPF), Session Management Function (SMF), Policy Control Function (PCF), Application Function (AF), and Location Management Function 140.

[0046] Access node 120 is a network element responsible for radio transmission and reception to or from one or more cells 122 of terminal node 110. Access node 120 is a network terminal of a radio link. Each access node can be a Transmitter-Receiver Point (TRP) or can host one or more TRPs.

[0047] Access node 120 can be implemented as a single network device, or as a split architecture with different functional split architectures and different interfaces, which are decomposed / distributed across two or more access nodes (such as central unit CU, distributed unit DU, remote radio head end RRH).

[0048] Terminal node 110 is a user-side network element in the network that terminates the radio link. They are devices that allow access to network services. The functions of terminal node 110 can also be performed by a mobile terminal MT (part of an integrated access and backhaul (IAB) node). Terminal node 110 can be referred to as user equipment (UE), mobile device, mobile terminal, or mobile station.

[0049] The term "user equipment" can be used to specify a mobile device that includes components for authentication / encryption, such as a smart card (e.g., a Subscriber Identity Module (SIM)). The SIM / SIM card can be a memory chip, module, or Universal Subscriber Identity Module (USIM). In some examples, the term "user equipment" can be used to specify a location / location tag, a super / smart sensor, or a mobile device that includes circuitry embedded as part of the user equipment for authentication / encryption, such as a software SIM.

[0050] Location server 140 is a device that manages support for various location services for the UE, including UE location and delivery of auxiliary data to the UE. The location server can connect to the core node and the Internet. The location server can be implemented as one or more servers. The location server is configured to support one or more location services for UE 110 that can connect to location server 140 via core network 130 and / or via the Internet. The location server may be referred to as a Location Management Function (LMF), a network function defined within 5GC.

[0051] In the following description, the terminal node may be referred to simply as UE 110.

[0052] In the following description, access devices / access nodes that access the RAN (e.g., cellular networks, especially next-generation RANs such as 5G or 6G) may be interchangeably referred to as BS 120 or gNB 120.

[0053] In the following description, the core device / node / function used to manage and process location information may be referred to simply as LMF140.

[0054] Now, let's briefly discuss positioning in radio telecommunications networks (such as RANs).

[0055] The location of a target UE within the RAN can be determined by a positioning server / location server (such as an LMF) using various conventional network-based positioning techniques, such as the LTE Positioning Protocol LPP (defined in TS 37.355), the New Radio Positioning Protocol NRPP, or NRPPa. Conventional techniques may involve exchanging RS via the Uu interface in the NR spectrum (e.g., transmitting an Orthogonal Frequency Division Multiplexing (OFDM) Positioning Reference Signal (PRS) from the RAN node to the target UE for DL ​​positioning; and transmitting an OFDM Probe Reference Signal (SRS) from the target UE to the RAN node for UL positioning). Such reference signals RS are received, detected, and measured by the gNB (for UL positioning) or the UE (for DL ​​positioning). The positioning server receives measurements from the gNB or the UE. This measurement information is received by the positioning server via the Access and Mobility Management Function (AMF) through a backhaul interface (e.g., the NL interface). The positioning server then uses this received measurement information to calculate the UE's location. The NR Positioning Protocol NRPPa carries positioning information between the NG-RAN node and the positioning server via the NG control plane interface (e.g., the NG-C interface).

[0056] Such positioning techniques based on Radio Access Technology (RAT) can utilize the following methods: uplink angle of arrival (UL-AoA), downlink angle of departure (DL-AoD), ranging based on time variance of arrival (TOA), uplink time difference of arrival (UL-TDOA), downlink time difference of arrival (DL-TDOA), and multi-cell round-trip time (multi-RTT).

[0057] The 5G NR positioning / addressing process is standardized in the 5G NR LPP specification. Traditional positioning sessions rely on receiver measurements of the positioning RS (PRS in DL and SRS in UL), which are scheduled by the network on specific time-frequency-space-code resources. The allocation of resources for such transmissions is coordinated across multiple UEs and gNBs via the LPP and NRPPa interfaces to ensure that the RS is unique and interference-free. This is done so that the receiver (UE in DL and gNB in ​​UL) can determine / calculate / extract positioning measurements, which are reported back to the network (in the case of UE-assisted positioning) or used locally (for UE-based positioning) to calculate the UE's location.

[0058] Target UEs can be located using the aforementioned positioning techniques, which can be referred to as NRUu-based positioning. Target UEs can also be located using AI / ML (or AI / ML-assisted) positioning.

[0059] Now, let's briefly discuss AI / ML-based or AI / ML-assisted localization.

[0060] AI / ML-based or AI / ML-assisted localization can include: Direct AI / ML localization, where the output of the AI / ML model is the UE's location. This can involve, for example, fingerprinting based on channel observation as input to the AI / ML model. The model can be a UE-side model for UE-based localization. Alternatively, it can be an LMF-side model for LMF-based localization.

[0061] AI / ML-assisted localization, where the output of the AI / ML model is new measurements and / or enhancements to existing measurements, such as line-of-sight Loss (LoS) or non-line-of-sight NLOS, identification, timing, and / or the angle of measurement and the probability of measurement. This model can be a UE-side model or a gNB-side model for AI / ML-assisted localization.

[0062] Some use cases for AI / ML-based or AI / ML-assisted localization include: Scenario 1: UE-based localization using a UE-side model for direct AI / ML localization.

[0063] Case 2a: UE-assisted / LMF-based localization using UE-side models for AI / ML-assisted localization.

[0064] Case 2b: UE-assisted / LMF-based localization using an LMF-side model for direct AI / ML localization.

[0065] Case 3a: Localization assisted by NG-RAN nodes using gNB-side models for AI / ML-assisted localization.

[0066] Case 3b: Localization assisted by NG-RAN nodes using the LMF side model for direct AI / ML localization.

[0067] AI / ML models can be trained based on collected training data. Training data can be collected based on measurements of PRS transmitted from one or more gNBs. In some cases, it may be desirable to associate (i.e., link / map / assign) the trained AI / ML model with the gNBs used to train the AI / ML model in the collection of training data. Such associations allow the model (trained in a collection of training data involving a specific set of gNBs) to be selected when performing an inference process involving that specific set of gNBs.

[0068] Now, let's briefly discuss network-side additional conditions / NW-side additional conditions.

[0069] For AI / ML-enabled features / feature groups, additional conditions can refer to any aspect that is assumed to be used for training the AI / ML model but is not part of the UE's capability to enable the AI / ML feature / feature group. Additional conditions are not necessarily specified. In this respect, additional conditions can be effectively considered as a set of aspects / conditions / parameters that exist / are satisfied / are applied during model training (and furthermore, they will be implicitly assumed to exist / are satisfied / are applied during model inference / use) (which are not explicitly signaled, for example, in Radio Resource Control (RRC) signaling).

[0070] Additional conditions can be divided into two categories: network-side (NW-side) additional conditions and UE-side additional conditions.

[0071] NW-side additional conditions (which may also be referred to as NW-side conditions or NW assumptions) may, but need not, be used to refer to network-side information (e.g., gNB or TRP implementation assumptions and / or proprietary settings) that are never disclosed to the UE or core node entity / core function (especially such as LMF). NW-side additional conditions can be considered to represent a set of settings or parameters that are not defined in the 3GPP specifications but are implemented / used / defined by the NW entity itself, which may have a direct impact on the measurements implemented in the UE entity.

[0072] Examples of additional conditions on the NW side include: physical beam azimuth, any NW supplier-specific settings known only to the NW supplier, any implementation details of the power amplifier or a set of antennas known only to the NW supplier.

[0073] Ensuring consistency of NW-side conditions between training and inference would be advantageous. In this regard, to ensure consistency between training and inference, information or indications related to additional NW conditions may be required at the UE. However, disclosing explicit information related to additional NW conditions could lead to proprietary issues.

[0074] Therefore, it is advantageous to implicitly represent a set of NW-side conditions via identifiers. Such identifiers for one or more NW-side conditions can be called "association IDs," meaning that an association ID can represent an identifier / label for one or more NW-side conditions. The association ID itself can be associated with one or more RAN nodes (especially, for example, the RAN nodes of a specific NW provider) so that mappings can be made between the association ID (and therefore the set or more NW-side conditions it represents) and the one or more RAN nodes associated with the NW-side conditions. It should be noted that the association ID can be exposed to the UE or core node entity / core function without revealing what the underlying NW-side conditions are.

[0075] Typically, (multiple) associated IDs can represent gNB or TRP implementation assumptions. The exact mapping between gNB / TRP implementation assumptions and (multiple) associated IDs may only be known to the gNB / TRP (or NW supplier).

[0076] When performing AIML use cases under the RRC protocol (the protocol for establishing a radio connection between the UE and the NG-RAN), the direct application / use of (multiple) association IDs can be easily implemented. This can thereby provide indirect details / information / indications of NW-side additional conditions related to the setting / configuration of the gNB / TRP, i.e., by using the association IDs conveyed to the UE by the gNB / TRP via RRC.

[0077] However, with the increasing demand for AIML applications at the physical layer, some other higher-level protocols between the core network and the UE may require information related to NW-side additional conditions to ensure consistency of NW-side conditions between training and inference. For example, in regular AIML localization, the LMF currently cannot access the NW-side additional condition information available in gNB / TRP (NG-RAN).

[0078] Therefore, in order to ensure consistency between training and inference, new procedures and signaling are required to enable the use of a higher-level protocol, different from the RRC protocol, to leverage (multiple) associated IDs (which are used to represent additional conditions (and any other related configurations) on the NW side).

[0079] In some cases, it may be desirable to provide procedures and signaling for associating training data with (multiple) association IDs of RAN entities (e.g., TRPs) involved in the collection of the training data, i.e., such that a model trained on the collected training data can also be associated with (multiple) association IDs involved in the collection of the training data.

[0080] In some cases, it may be desirable to provide procedures and signaling for selecting a training model for an inference process involving a set of RAN entities, each having its own corresponding association ID, such that the selected training model is a model trained on training data, the collection of which involves a set of RAN entities having the same / matching association IDs.

[0081] Figure 2The diagram schematically illustrates process 200 between a first device (UE 110 in this example) and a second device (core node 130 in this example, such as LMF 140), and an example of the signaling framework used to support the process. As will be discussed below, process 200 is used to signal a new type of identifier used during the data collection process, where the identifier represents one or more NW-side additional conditions of one or more RAN entities (e.g., TRPs) used during the data collection process. In the following text, such identifiers may be interchangeably referred to simply as IDs or “associated IDs”.

[0082] In block 201, UE 110 receives first information 202 from LMF 140, which is used to support UE in performing data collection procedure 206.

[0083] As will be discussed in further detail below, a data collection process can be a process for collecting training data (i.e., datasets / training datasets) for training AI / ML models. The data collection process can involve not only the collection of one or more datasets, but also (implicitly or explicitly) associating each dataset with one or more association IDs 205, where the association ID represents a set of NW-side additional conditions for a set of RAN entities involved in the data collection process itself (e.g., a set of RAN entities transmitting signals (such as PRS) to the UE as part of the data collection process), where measurements of received signals by the UE form the dataset to be collected.

[0084] First Information 202 can provide: Auxiliary information used to assist the UE in performing the data collection process, used to collect training data, namely dataset 208 used to train model 212; and This is auxiliary information used to associate each dataset collected during the data collection process with one or more associated IDs.

[0085] In this regard, the first information includes information indicating one or more Location Reference Signal (PRS) configurations 203, wherein each PRS configuration includes information indicating multiple entities of the RAN. RAN entities can be one or more of the following: TRPs, gNBs, and / or cells. Information indicating such RAN entities can be a list of individual identifiers, i.e., a set of TRP IDs, cell IDs, or NR cell global identity (NCGI). In this regard, each PRS configuration may include a list of individual identifiers 204 for each RAN entity to be involved in the data collection process (the data collection process involves the RAN entity sending a PRS to the UE for measurement by the UE, wherein the collected dataset is based at least in part on measurements of the PRS received from each RAN entity).

[0086] The first information also includes information indicating one or more association IDs 205, wherein each association ID represents one or more NW-side additional conditions of one or more RAN entities indicated in the PRS configuration.

[0087] One or more associated IDs can represent at least one of the following: One or more network-side additional conditions for one or more RAN entities; One or more assumptions adopted by one or more RAN entities; One or more explicit assumptions made by the LMF for any auxiliary data after receiving any association ID from the RAN entity (in this respect, the association ID received from the RAN entity / gNB can be used to select any auxiliary data to support the UE, where such auxiliary data may include explicitly exposed assumptions (e.g., using a container referred to as the association ID); or One or more appropriate settings of one or more RAN entities will be used by one or more RAN entities during the data collection process.

[0088] Each PRS configuration received by the UE can be associated with a set of associated IDs (also received by the UE). The first information may include multiple PRS configurations and corresponding sets of associated IDs. The UE can associate each collected dataset with a set of associated IDs from the multiple sets of associated IDs (e.g., via links, tags, or mappings), where the association is based on which particular set of associated IDs is associated with a particular PRS configuration used by the UE to receive one or more PRSs to collect a particular dataset.

[0089] The association between an association ID and a PRS configuration can be implicit or explicit. For example, there can be an explicit indication of one or more association IDs to a specific PRS configuration's association / link / mapping / assignment.

[0090] The first information may also include configuration information for configuring the UE to perform the data collection process.

[0091] As will be discussed in further detail below, the LMF can select the RAN entities to be used during the UE's data collection process (e.g., based on the UE's approximate location). The LMF can then indicate the ID of the selected RAN entity (e.g., cell ID, NCGI, TRP ID) in PRS configuration 203. The LMF can have a mapping of associated IDs to RAN entities. The LMF can select associated IDs that correspond to / belong to the RAN entities selected for use during data collection, and the LMF can then include the selected associated IDs for the selected RAN entities in the first information.

[0092] Information transmission between LMF and UE can be achieved via at least one of the following: The interface between LMF and UE, such as the N1 interface; Non-access stratum (NAS) signaling; Non-RRC / higher-layer protocols between LMF and UE, such as LTE Positioning Protocol (LPP); New Radio Positioning Protocol (NRPPa); or High-level protocols between nodes and devices in the core network.

[0093] The UE then (e.g., when triggered) performs the data collection process 206 based at least in part on the first information 202.

[0094] As will be discussed further below, the data collection process includes: Collect one or more datasets 208, and The collected datasets, along with the associated IDs for each dataset, are sent to the training entity used to train the model.

[0095] In box 207, the UE collects one or more datasets 208, wherein each dataset is at least partially based on one or more PRS received from a set of RAN entities 204 indicated in one of the multiple PRS configurations 203 received in box 201. In this respect, the UE uses the received PRS configuration to receive PRS transmitted from the set of RAN entities indicated in the PRS configuration, and the UE performs measurements on the received PRS, such measurements forming datasets. Therefore, each dataset is based on measurements of PRS received based on the PRS configuration.

[0096] Each dataset may include at least one of the following: At least one channel measurement; At least one reference signal received power RSRP measurement; At least one reference signal receiving path power RSRPP measurement; At least one channel impulse response CIR measurement; At least one power delay distribution (PDP) measurement; or At least one delay distribution DP measurement.

[0097] In box 209, the UE sends second information 210 to training entity 211 for training model 212, wherein the second information includes information indicating the following: One or more datasets 208, and One or more associated IDs 205 received in box 201.

[0098] In this regard, the second information sent to training entity 211 for training model 212 includes information indicating one or more datasets 208 and one or more association IDs associated with each dataset, wherein the one or more association IDs are associated with one or more network-side additional conditions (i.e., network-side additional conditions) of the RAN entity that transmits the PRS for generating training data.

[0099] The model can be at least one of the following: Artificial intelligence (AI) models; Machine learning (ML) models; Models for UE; and A model for downlink-based positioning on the user equipment (UE) side.

[0100] The training entity can be any suitable entity that includes components for developing (e.g., training and / or updating) the model. The training entity can be the UE itself (e.g., for a UE-side AI / ML model hosted on the UE itself), or it can be a separate device located away from the UE.

[0101] In some examples, the training entity is the UE itself (e.g., its modules), and the UE develops (i.e., trains or updates) a model based at least in part on (multiple) datasets and (multiple) association IDs associated with each dataset. The UE can also associate the developed model with (multiple) association IDs. In this respect, the UE can explicitly link, label, or map one or more association IDs to its corresponding trained model.

[0102] In some examples, the UE may additionally associate the developed model with identifiers of the corresponding RAN entities involved in the collection of training data. For example, the UE may associate (multiple) cell IDs and / or TRP IDs (derived from the PRS configuration) with the corresponding model developed using such training data. In this regard, the UE may explicitly link, label, or map one or more RAN entity identifiers to the trained model.

[0103] The association between the associated ID and the dataset in box 209 can be implicit or explicit. The UE can provide explicit indication of the association / link / mapping / assignment of one or more associated IDs to a specific dataset in the second information.

[0104] The UE can associate each collected dataset with one or more associated IDs. In this regard, the UE can associate a specific associated ID with a specific dataset in the following ways: Determine which specific PRS configuration in the PRS configuration of the first information received in box 201 is used to receive the PRS measured to form a specific dataset; Based on the first information and its indication of which associated IDs are associated with which PRS configurations, determine the associated IDs that are associated with the specific PRS configuration; and Associate the identified association ID with a specific dataset.

[0105] Using the above process 200, the training entity used to train the AI / ML localization model can receive not only the dataset(s) used for training the model, but also the association(s) associated with each dataset(s). Advantageously, this allows the AI / ML localization model, properly trained using the dataset(s), to also be associated / mapped / linked / assigned with the corresponding association(s). The association(s), each representing a model training context, particularly one or more network-side appendages of the RAN entity used in the collection of training data for training the model, can be used as an indication of whether the trained model is effective / suitable for a specific context, such as whether the model is effective / suitable for the following: Perform the inference / localization process involving a set of RAN entities with a set of network-side additional conditions.

[0106] As will be discussed below, providing a trained AI / ML localization model associated with (multiple) association IDs enables the association ID-based selection of the AI / ML localization model to be used for inference processes, such as localization. For example, a UE may have multiple trained models, each associated with one or more association IDs. If a DL localization process is to be performed on a UE in a specific area (i.e., such that the localization process will involve a specific set of RAN entities in the specific area, each RAN entity having its own corresponding network-side additional conditions), the LMF can determine an association ID corresponding to the specific set of RAN entities. The LMF can then provide the UE with auxiliary information to support the UE in performing the DL localization process, where the auxiliary information includes an indication of which association ID the UE should use when selecting which of its multiple AI / ML localization models to use (each of the UE's AI / ML localization models is associated with one or more association IDs).

[0107] Advantageously, by linking datasets (collected for training a specific model) to one or more association IDs (which represent NW-side additional conditions of RAN entities involved in the data collection process), it becomes possible to select an AI / ML localization model for inference based on the association ID. This allows for the selection of an appropriate / optimal model based on the association ID, which represents the NW-side additional conditions of the RAN entities to be involved in the inference process. Therefore, the examples of this disclosure can support / enable consistency of NW-side conditions between training and inference without explicitly specifying or communicating the actual NW-side additional conditions to the UE or LMF, thus maintaining the confidentiality of the NW-side additional conditions.

[0108] Figure 3 The diagram schematically illustrates a process 300 and an example of a signaling framework supporting a first device (core node / function 130, such as LMF 140 in this example), a second device (BS 120 or one or more TRPs of a BS in this example), and a third device (UE 110 in this example). As will be discussed below, process 300 involves determining association IDs that represent one or more network-side appendages of one or more RAN entities (e.g., BS or TRPs) to be used during the data collection process, and also involves signaling to support the use of these association IDs during the data collection process.

[0109] In box 301, LMF 140 obtains one or more association IDs 205 from one or more BS 120, wherein the one or more association IDs are associated with one or more network-side additional conditions of the corresponding BS.

[0110] In some examples, each BS (each BS knowing its own network-side appendages) can generate its own association ID and map / associate that association ID to its network-side appendages. The BS can then store its set of network-side appendages mapped / associated to its association IDs.

[0111] LMF can send a request for their respective associated IDs to one or more BSs. In response, LMF can receive one or more associated IDs from one or more entities in response to the request.

[0112] It is worth noting that an association ID is a single identifier representing a set of NW-side additional conditions for at least one RAN entity. Therefore, a set of NW-side additional conditions can be identified and referenced via a single association ID (compared to explicitly listing and specifying each actual NW-side additional condition). Thus, using association IDs is a way to reference specific sets of NW-side additional conditions while avoiding the need to expose the underlying NW-side additional conditions represented by their respective association IDs, thereby maintaining the privacy of the NW-side additional conditions, which can be confidential / proprietary information of the NW supplier / NW-side entity (e.g., BS or TRP).

[0113] It is important to note that each BS only shares its corresponding association ID(s) with the LMF. The underlying network-side additional conditions represented by the association ID(s) are not shared with the LMF, thus maintaining their confidentiality.

[0114] In box 302, the LMF associates each acquired association ID with the specific RAN entity (e.g., BS or TRP from which the association ID was obtained.) In this respect, the LMF can generate and store a mapping / association between each acquired association ID and the specific BS (or specific TRP) from which the association ID was obtained. As will be discussed further below, the association ID can then be effectively used as an indication of valid network-side appending conditions to the model. For example, a model trained on one or more datasets associated with one or more association IDs can be effectively / suited for inference processes such as: The network-side additional conditions involve a group or more BS or TRPs represented by one or more corresponding associated IDs.

[0115] In box 201 (which effectively corresponds to) Figure 1 In box 201), the LMF sends first information 202 to the UE 110, which is used to support the UE in performing the data collection process.

[0116] As mentioned above Figure 1 The data collection process discussed can be a process for collecting training data (i.e., datasets / training datasets) to train AI / ML models, wherein the data collection process involves not only the collection of one or more datasets, but also associating each dataset with one or more associated IDs included in the first information (implicitly or explicitly).

[0117] First Information 202 provides: Assistance information for the UE to perform a data collection process, which involves collecting training data for training a model (this assistance information includes PRS configuration information, which itself indicates one or more IDs (e.g., cell ID, NCGI, TRP ID) of one or more RAN entities, which will be used during the data collection process to transmit (multiple) PRSs to the UE for measurement and to generate a dataset therefrom); and This auxiliary information is used to assist the UE in associating each collected dataset with one or more associated IDs (this auxiliary information includes one or more associated IDs of one or more RAN entities to be used in the data collection process).

[0118] When preparing / generating the first information, the LMF can select a set of RAN entities to be used during the data collection process (i.e., the PRS for transmitting measurements to the UE). The selection of RAN entities can be based at least in part on the estimated / coarse location of the UE. For example, the LMF can select a set of RAN entities near the UE to be used during the data collection process (e.g., TRPs) (i.e., one or more TRPs of the serving BS and neighboring BSs). The LMF can then determine a set of associated IDs for the set of RAN entities to be used during the data collection process and include them in the first information. The LMF can also determine a set of identifiers (e.g., cell ID, NCGI, TRP ID) for the set of RAN entities to be used during the data collection process and include them in the first information (i.e., the PRS configuration information generated by the LMF).

[0119] In some examples, the LMF can select a set of RAN entities to use during data collection and then generate PRS configuration information that includes a set of identifiers for the selected set of RAN entities. Requests for association IDs that the LMF can send to the RAN entities may include the PRS configuration information, which the RAN entities can use to determine a set of RAN entities (e.g., a set of TRPs) to use during data collection. Each RAN entity in the set can send its corresponding association ID to the LMF.

[0120] In some examples, the request for an association ID sent to the RAN entity includes information indicating one or more network-side additional conditions. In this regard, such information may include general high-level guidance on network-side additional conditions, rather than exact / specified network-side additional conditions. The RAN entity is free to use its own unique set of network-side additional conditions within the constraints configured by the LMF in the information indicating one or more network-side additional conditions. The RAN entity may select specific / specified network-side additional conditions based on the received general high-level guidance on network-side additional conditions and send an association ID for the selected specific / specified network-side additional conditions to the LMF.

[0121] Figure 4 An example of a signaling diagram is shown, which illustrates signaling and process 400 (between CN entity / function 130 (e.g., LMF140), multiple gNBs 12_1 to 120_3, and UE 110), which is used for data collection using association IDs for consistency purposes (i.e., consistency of network-side conditions between training and inference) (e.g., association IDs used to collect training data for the model, especially such as those used for "Case 1" type AI / ML positioning, i.e., direct AI / ML positioning - UE-based positioning with a UE-side model).

[0122] Despite Figure 4 The example shows three gNBs and a single UE, but it should be understood that different numbers of gNBs and UEs may be used in other examples.

[0123] Process 400 has certain aspects, features, and functions similar to Figure 2 and Figure 3 Certain aspects, features, and functions of processes 200 and 300, as well as the various other features and functions described above. Therefore, certain aspects, features, and functions of processes 200 and 300, as well as the various other features and functions described above, may be related to necessary changes to process 500, and should not be repeated / reiterated in detail.

[0124] In step 1, the UE provides a capability report to the CN indicating whether the UE supports associated IDs. In this regard, the UE indicates its ability to perform data collection procedures, including associating / labeling / mapping each collected dataset with an associated ID, where the associated ID represents one or more network-side conditions. The CN can determine whether to request the UE to use the associated ID to perform the data collection procedure based on whether the UE supports such operations.

[0125] In step 2, the CN performs initial steps for data collection based on a consistency process using additional conditions on the NW side. In this regard, a handshake / signaling exchange may exist between the UE and the CN, where data collection functionality can be set up between the CN and the UE to enable data collection. This may include reference signal configuration. The UE may trigger a request for auxiliary information from the CN, and the CN may respond regarding whether providing that information is feasible. The UE may send a request to the CN for an association ID that will be involved in the data collection (whereby the UE may subsequently use the association ID as reference information for verification between training and inference, and for selecting a training model for inference).

[0126] In step 3, the CN requests AIML consistency information (e.g., association ID) from some gNBs that are involved in the data collection for the UE. The association ID is only generated in / by the gNB, so the CN needs to request this information from the gNB.

[0127] In step 4 (broadly equivalent to) Figure 3 In box 301), each gNB indicates its associated ID(s) to the CN. Since the gNB / cell ID can include multiple TRPs belonging to a specific network area, where each TRP can have its own set of network-side conditions and each set of network-side conditions has its own associated ID, a gNB can have / be associated with multiple associated IDs.

[0128] In step 5 (broadly equivalent to) Figure 3 In box 302), the CN maps all received association IDs to their corresponding gNBs (i.e., each association ID is mapped to the corresponding gNB from which the association ID is received).

[0129] In step 6 (broadly equivalent to) Figure 2 and Figure 3 In box 201), auxiliary information is passed from the CN to the UE. The auxiliary information includes a set of (multiple) associated IDs (obtained in step 4) and PRS configuration, which includes gNB settings and the cell ID. Regarding the steps, the CN can be effectively considered as a routing entity of (multiple) associated IDs.

[0130] In step 7 (generally equivalent to box 206), data collection is performed, including a process to ensure consistency of additional conditions on the NW side, i.e., using the mapping / labeling of training data [such as channel measurements collected by the UE, e.g., Reference Signal Received Power (RSRP); Reference Signal Received Path Power (RSRPP); and Channel Impulse Response (CIR)] with the associated IDs. For consistency, the associated IDs are included in the channel measurements. In this regard, the training data can be categorized for each PRS configuration and associated ID, such that the final dataset is associated with a set of associated IDs.

[0131] In step 8, after the data collection process is completed, a UE-side model can be developed (e.g., trained / updated), and the developed model is associated with / mapped to / tagged with (multiple) association IDs of the training data used in the development of the model.

[0132] Figure 5 The diagram schematically illustrates process 500 between a first device (UE 110 in this example) and a second device (core node / entity / function, such as LMF 140 in this example) and an example of the signaling framework used to support the process. As will be discussed below, process 500 is used to signal a new type of identifier, namely an association ID, for inference in AI / ML positioning, where the identifier represents network-side conditions (i.e., network-side additional conditions of RAN entities (e.g., TRPs) to be used in the inference process).

[0133] Figure 5 The process is suitable for use with existing technology. Figure 2 , Figure 3 and Figure 4 The data collected during the data collection process is used in conjunction with the trained models, enabling the UE to have multiple trained models available to it. Each trained model is associated with one or more associated IDs representing a set of network-side conditions used to train the corresponding model during the collection of training data. Furthermore, the LMF can store or access a mapping of associated IDs to RAN entities.

[0134] As described above, UE 110 has multiple trained AI / ML models 212_u available to it (e.g., stored at the UE or accessible to the UE). Each of the multiple models of the UE is associated with one or more corresponding association IDs 205_u.

[0135] Each associated ID can represent at least one of the following: One or more network-side additional conditions; One or more network-side hypotheses; One or more explicit assumptions made by the LMF for any auxiliary data after receiving any associated ID from the RAN entity (in this respect, the associated ID received from the RAN entity / gNB can be used to select any auxiliary data to support the UE, wherein such auxiliary data may include explicitly exposed assumptions [e.g., using a container called the associated ID]); or Appropriate network-side settings are used by one or more RAN entities in the process of collecting data for training the model.

[0136] The UE model can be at least one of the following: Artificial intelligence (AI) models; Machine learning (ML) models; Model for the device; Models for downlink-based positioning on the user equipment (UE) side; and A trained model associated with / mapped to / tagged with one or more associated IDs, wherein the model is trained at least in part on at least one dataset associated with the associated IDs.

[0137] In block 501, the UE sends information 502 to the LMF indicating a first set of multiple associated IDs 205_u. In this respect, for each of the UE's multiple models 212_u, the UE may send information indicating each of one or more associated IDs associated with the UE model.

[0138] Communication between the UE and the LMF can be achieved through at least one of the following: The interface between LMF and UE; N1 interface; Non-access stratum (NAS) signaling; Non-Radio Resource Control Protocol; High-level protocol between LMF and UE; LTE Location Protocol (LPP).

[0139] In box 503, the LMF obtains a second set of association IDs 205_bs from the BS, which has been selected by the LMF for use in the inference process. In this respect, the LMF can send a request for association IDs(s) for each gNB, and the LMF can receive one or more association IDs in response to the request. The LMF can then associate each received association ID with the gNB from which it received the corresponding association ID.

[0140] In some instances, an LMF request may include information indicating at least one of the following: One or more PRS configurations; or One or more network-side additional conditions.

[0141] In some examples, the request for the association ID sent to the gNB includes information indicating one or more network-side additional conditions. In this regard, such information may include general high-level guidance on network-side additional conditions, rather than exact / specified network-side additional conditions. Each gNB is free to use its own unique set of network-side additional conditions within the constraints configured by the LMF in the information indicating one or more network-side additional conditions. The gNB may select specific / specified network-side additional conditions based on the received general high-level guidance on network-side additional conditions and send an association ID for the selected specific / specified network-side additional conditions to the LMF.

[0142] In box 504, LMF determines information 505 indicating the degree of matching between the following two: One or more associated IDs in the first group of associated IDs, and One or more associated IDs from the second set of associated IDs.

[0143] Such information may be referred to as a "consistency indicator" in this paper.506

[0144] In box 507, the LMF sends information 505 (including an indication of a matching degree / consistency indicator 506) to the UE to assist the UE in selecting one of its multiple models for the inference process (e.g., AI / ML-based positioning), wherein information 505 indicates the selected association ID 205_x.

[0145] In box 508, the UE uses the received information 505 (and its indication of the matching degree / consistency indicator 506) to determine whether to select one of its multiple models for the inference process (e.g., AI / ML-based localization).

[0146] The determination of whether to select one of the multiple models can be based on the degree of matching that satisfies the conditions, in particular, for example, where the consistency indicator has a value indicating a match in the first set of associated IDs and the second set of associated IDs (e.g., consistency indicator value = 1), or where the value of the consistency indicator exceeds a threshold, such as a threshold indicating the sufficient / necessary degree of matching of a set of associated IDs (and thus indicating the sufficient / necessary degree of consistency under the network-side conditions involved in the data collection used to train the model and to use the model for inference).

[0147] The UE can receive configuration information from the LMF for configuring the UE to perform inference procedures (e.g., localization procedures). The configuration information may include information indicating one or more Location Reference Signal (PRS) configurations, where the one or more PRS configurations include information indicating a second set of one or more entities (e.g., a list of cell IDs, NCGIs, or TRP IDs). The UE can perform the inference / localization procedures at least in part based on the configuration information using the model selected in step 508. For example, the inference procedure could be a localization procedure, and the UE could use the selected model to perform AI / ML-based UE localization.

[0148] Using the above process 500, the selection of a model for inference based on the association ID can be performed, where the association ID evaluation is completed at the LMF. In this respect, the LMF enables the inference process to be performed using a specific set of RAN entities (each RAN entity being associated with one or more association IDs representing one or more network-side conditions), and enables the UE to select one of its trained models that is effective / suitable for the inference process (each model being associated with one or more association IDs) in order to provide a sufficient / necessary level of consistency in the network-side conditions involved in data collection for training the model and using the model for inference.

[0149] Advantageously, the use of (multiple) association IDs in the collection of training data and the training of the model, combined with the use of (multiple) association IDs in the inference process, can achieve consistency between training and inference by enabling association ID-based selection of the model used for inference. Association IDs can be used to select a trained model that is effective / suitable for a specific context, for example, selecting a model that is effective / suitable for performing an inference / localization process involving a certain set of RAN entities (e.g., having a set of network side conditions that are the same as / sufficiently similar to the network side conditions used in a set of training data for the selected model), thereby advantageously providing consistency between training and inference.

[0150] Figure 6 An example of a signaling diagram is shown, illustrating signaling for consistency purposes (between UE110 (i.e., the target UE to be the subject of the inference / location process), gNB 120_1 to 120_3 and core node entity / function 130 (such as LMF 140)) and process 600, which is used for inference in AI / ML location (at least such as "Case 1" type AI / ML location, i.e., direct AI / ML location - UE-based location with UE-side model).

[0151] Figure 6 The process is suitable for use with existing technology. Figure 2 , Figure 3and Figure 4 The data collected during the data collection process is used in combination with the trained models, so that the UE has multiple training models that can be used by it, each of which is associated with one or more associated IDs representing a set of network-side conditions (i.e., TRPs) used to train the corresponding model in the collection of training data.

[0152] Process 600 has certain aspects, features, and functions similar to Figure 5 The process 500 includes certain aspects, features, and functions, as well as the various other features and functions described above. Therefore, certain aspects, features, and functions of process 500, as well as the various other features and functions described above, may be related to necessary changes to process 600, and should not be repeated / reiterated in detail.

[0153] Despite Figure 6 The example shows three gNBs and a single UE, but it should be understood that in other examples, different numbers of gNBs and UEs may be used.

[0154] In step 1, UE 110 provides a capability report to LMF 140 to indicate whether the UE supports association IDs. In this regard, the UE indicates its ability to perform inference processes (e.g., AI / ML positioning), including the selection of an AI / ML model used for performing inference / AI / ML positioning based on the association ID. LMF can determine whether to request the UE to perform inference in AI / ML positioning using the association ID based on whether the UE supports such operations.

[0155] In step 2, the AI / ML function setup process is performed. In this regard, the LMF may, at least in part, respond to the capability report received in step 1 by providing the UE with auxiliary data for inference in AI / ML positioning. This auxiliary data may be general auxiliary data used to support the UE in performing inference in AI / ML positioning (e.g., regular auxiliary data for inference in AI / ML positioning – compared to the specific auxiliary data including the consistency_indicator provided in step 10 below). The auxiliary data in step 2 may include PRS configuration information to enable the UE to receive and measure PRS transmitted from the gNB or TRP (and where PRS measurements can be input into the AI / ML model to output location). In this regard, the PRS configuration includes information indicating the RAN node entity (e.g., gNB or TRP) to which the PRS is to be transmitted (e.g., a list of cell IDs, NCGI, or TRP IDs).

[0156] In step 3, there is signal exchange between the UE and LMF to ensure consistency between training and inference.

[0157] In step 4 (which broadly corresponds to) Figure 5 In box 501, the LMF collects from the UE one or more association IDs linked / tagged to one or more AIML models of the UE. In this respect, the UE can report a set (or multiple) of association IDs linked / tagged to the corresponding AIML model for each AIML model. The UE can also report to the LMF information about the PRS configuration used to collect training data. In this respect, the LMF can collect information about the PRS configuration and association IDs reported by the UE / target UE. In this step, multiple pairs of PRS configuration + association IDs are collected to form Set_1.

[0158] In step 5, the LMF requests AIML consistency information from each gNB involved in the inference / localization process. In this regard, the LMF requests the association ID from each gNB.

[0159] In step 6, each gNB indicates one or more of its respective associated IDs to the LMF.

[0160] In step 7 (broadly equivalent to) Figure 5 In box 503), the LMF collects all (multiple) associated IDs from each gNB and maps each associated ID to its corresponding gNB to form a pair: gNB + associated ID. This pair is used to form Set_2.

[0161] In step 8 (which is broadly equivalent to) Figure 5 (See box 504). LMF can check for any matches between the associated IDs of Set_1 and Set_2. In this respect, Set_1 and Set_2 can be mapped, and such a mapping involves checking each gNB for a match with the expected associated ID.

[0162] In step 9, if a match exists in step 8, the LMF sends a flag to the UE, such as consistency_indicator = 1, to indicate a match. Otherwise, if no match exists in step 8, the LMF sends the flag consistency_indicator = 0 to the UE to indicate no match. In other examples, instead of a binary indicator with a binary value (e.g., binary flag bit = 0 or 1), consistency_indicator can have other values, such as any real number between 0 and 1, to indicate a specific level / matching degree / similarity between the associated IDs in Set 1 and Set 2.

[0163] In step 10 (broadly equivalent to) Figure 5In box 507, the consistency_indicator = {1, 0} value of (multiple) association IDs (i.e., association IDs of the UE's models) is sent from the LMF to the UE. In this respect, the LMF can provide the UE with auxiliary data, which includes information (i.e., consistency_indicator) to enable the UE to select which of its multiple AI / ML models to use during inference / AI / ML localization based on the auxiliary information.

[0164] It is anticipated that each AI / ML model deployed on the UE side will have at least one or more associated IDs. These associated IDs (forming Set 1) are IDs evaluated for consistency purposes (as opposed to the associated IDs in Set 2). Accordingly, consistency indication is accomplished by the complete set of associated IDs reported by the UE (in step 4 for forming Set 1). The consistency indication can be mapped to a unique AI / ML model of the UE, or it can be mapped to many AI / ML models of the UE (this mapping can be a decision made by the UE's implementation).

[0165] In step 11 (broadly equivalent to) Figure 5 In box 508, the UE performs applicable functional analysis based on the consistency_indicator values ​​of (multiple) associated IDs. In this respect, the UE performs a process in which the UE evaluates / selects the most suitable model to activate for the inference process. LMF uses information (i.e., the consistency_indicator value in step 10) to assist the UE in enabling the UE to evaluate available AI / ML models.

[0166] In step 12, feature activation and further steps are completed for inference. Specifically, if the UE's AIML model obtains consistency_indicator = 1, the feature established in step 2 is activated, and the UE uses the selected model to perform the feature, i.e., inference process / AIML localization. This feature will report the inference location, which is the primary target in AIML localization.

[0167] Figure 7 The diagram schematically illustrates process 700 between a first device (UE 110 in this example) and a second device (core node / entity / function, such as LMF 140 in this example) and an example of the signaling framework used to support the process. As will be discussed below, process 700 is used to signal a new type of identifier, namely an association ID, for inference in AI / ML positioning, where the identifier represents network-side conditions (i.e., network-side additional conditions of RAN entities (e.g., TRPs) to be used in the inference process).

[0168] Figure 7 The process is suitable for use with existing technology. Figure 2 , Figure 3 and Figure 4 The data collected during the data collection process is used in conjunction with the trained models, enabling the UE to have multiple trained models available to it. Each trained model is associated with one or more associated IDs representing a set of network-side conditions used to train the corresponding model during the collection of training data. Furthermore, the LMF can store or access a mapping of associated IDs to RAN entities.

[0169] As described above, UE 110 has multiple trained AI / ML models 212_u available to it (e.g., stored at the UE or accessible to the UE). Each of the UE's multiple models is associated with one or more corresponding association IDs 205_u. In this respect, the association IDs of the UE's models are targeted at a first set of association IDs.

[0170] Each associated ID can represent at least one of the following: One or more network-side additional conditions; One or more network-side hypotheses; One or more explicit assumptions made by the LMF for any auxiliary data after receiving any associated ID from the RAN entity (in this respect, the associated ID received from the RAN entity / gNB can be used to select any auxiliary data to support the UE, wherein such auxiliary data may include explicitly exposed assumptions [e.g., using a container called the associated ID]); or Appropriate network-side settings are used by one or more RAN entities in the process of collecting data for training the model.

[0171] The UE model can be at least one of the following: Artificial intelligence (AI) models; Machine learning (ML) models; Model for the device; Models for downlink-based positioning on the user equipment (UE) side; and A trained model associated with / mapped to / tagged with one or more associated IDs, wherein the model is trained at least in part on at least one dataset associated with the associated IDs.

[0172] In box 701, the LMF obtains a second set of association IDs 205_bs from the BS, which has been selected by the LMF for use in the inference process. In this respect, the LMF can send a request for association IDs(s) for each gNB, and the LMF can receive one or more association IDs in response to the request. The LMF can then associate each received association ID with the gNB from which it received the corresponding association ID.

[0173] In some instances, an LMF request may include information indicating at least one of the following: One or more PRS configurations; or One or more network-side additional conditions.

[0174] In some examples, the request for the association ID sent to the gNB includes information indicating one or more network-side additional conditions. In this regard, such information may include general high-level guidance on network-side additional conditions, rather than exact / specified network-side additional conditions. Each gNB is free to use its own unique set of network-side additional conditions within the constraints configured by the LMF in the information indicating one or more network-side additional conditions. The gNB may select specific / specified network-side additional conditions based on the received general high-level guidance on network-side additional conditions and send an association ID for the selected specific / specified network-side additional conditions to the LMF.

[0175] In box 702, the LMF sends information 703 to the UE indicating a set of associated IDs 205_bs obtained in box 701. This information is sent to assist the UE in selecting one of its multiple models for the inference process (e.g., AI / ML-based positioning).

[0176] Communication between the UE and the LMF can be achieved through at least one of the following: The interface between LMF and UE; N1 interface; Non-access stratum (NAS) signaling; Non-Radio Resource Control Protocol; High-level protocol between LMF and UE; LTE Location Protocol (LPP).

[0177] In box 704, the UE determines whether to select one of its multiple models 212_ux for the inference process (e.g., AI / ML-based localization) based at least in part on the following: The second set of associated IDs, and The first set of associated IDs.

[0178] In this regard, the UE can determine information indicating the degree of matching between the following two: One or more associated IDs in the first group of associated IDs, and One or more associated IDs from the second set of associated IDs.

[0179] Then, the UE can select one of its multiple models based on the determined information.

[0180] The determination of whether to select one of the multiple models can be based on the degree of matching that satisfies the conditions, in particular, for example, where the consistency indicator has a value indicating a match in the first set of associated IDs and the second set of associated IDs (e.g., consistency indicator value = 1), or where the value of the consistency indicator exceeds a threshold, such as a threshold indicating the sufficient / necessary degree of matching of a set of associated IDs (and thus indicating the sufficient / necessary degree of consistency under the network-side conditions involved in the data collection used to train the model and to use the model for inference).

[0181] The UE can receive configuration information from the LMF for configuring the UE to perform inference procedures (e.g., localization procedures). The configuration information may include information indicating one or more Location Reference Signal (PRS) configurations, where the one or more PRS configurations include information indicating a second set of one or more entities (e.g., a list of cell IDs, NCGIs, or TRP IDs). The UE can perform the inference / localization procedure at least in part based on the configuration information using the model selected in step 704. For example, the inference procedure could be a localization procedure, and the UE could use the selected model to perform AI / ML-based UE localization.

[0182] Using the above process 700, the selection of the model for inference based on the association ID can be performed, where the association ID evaluation is completed at the UE (compared to process 500 where the association ID evaluation is completed at the LMF).

[0183] Process 700 enables the inference process to be performed using a specific set of RAN entities (each RAN entity being associated with one or more association IDs representing one or more network-side conditions) and enables the UE to select one of its trained models that is effective / suitable for the inference process (each model being associated with one or more association IDs) in order to provide a sufficient / necessary level of consistency of the network-side conditions involved in the data collection for use in training the model and inference.

[0184] Advantageously, the use of (multiple) association IDs in the collection of training data and the training of the model, combined with the use of (multiple) association IDs in the inference process, can achieve consistency of network-side conditions between training and inference by enabling the selection of the model used for inference based on association IDs. Association IDs can be used to select a trained model that is effective / suitable for a specific context, for example, to select a model that is effective / suitable for performing an inference / localization process involving a certain set of RAN entities (e.g., having a set of network-side conditions that are the same as / sufficiently similar to the network-side conditions used in a set of training data for the selected model), thereby advantageously providing consistency of network-side conditions between training and inference.

[0185] Figure 8 An example of a signaling diagram is shown, illustrating signaling for consistency purposes (between UE 110 (i.e., the target UE that will be the subject of the inference / location process), gNB 120_1 to 120_3 and core node entity / function 130 (such as LMF 140)) and process 800, which is used for inference in AI / ML location (at least such as "Case 1" type AI / ML location, i.e., direct AI / ML location - UE-based location with UE-side model).

[0186] Figure 8 The process is suitable for use with existing technology. Figure 2 , Figure 3 and Figure 4 The data collected during the data collection process is used in combination with the trained models, so that the UE has multiple training models that can be used for it. Each training model is associated with one or more associated IDs representing a set of network side conditions (i.e., TRPs) used to train the corresponding model during the collection of training data.

[0187] Process 800 has certain aspects, features, and functions similar to Figure 7 The process 700 includes certain aspects, features, and functions, as well as the various other features and functions described above. Therefore, certain aspects, features, and functions of process 700, as well as the various other features and functions described above, may be related to necessary changes to process 800, and should not be repeated / reiterated in detail.

[0188] Although Figure 8 The example shows three gNBs and a single UE, but it should be understood that in other examples, different numbers of gNBs and UEs may be used.

[0189] In step 1, UE 110 provides a capability report to LMF 140 to indicate whether the UE supports association IDs. In this regard, the UE indicates its ability to perform inference processes (e.g., AI / ML positioning), including the selection of an AI / ML model used for performing inference / AI / ML positioning based on the association ID. LMF can determine whether to request the UE to perform inference in AI / ML positioning using the association ID based on whether the UE supports such operations.

[0190] In step 2, the AI / ML function setup process is performed. In this regard, the LMF may, at least in part, respond to the capability report received in step 1 by providing the UE with auxiliary data for inference in AI / ML positioning. This auxiliary data may be general auxiliary data used to support the UE in performing inference in AI / ML positioning (e.g., regular auxiliary data for inference in AI / ML positioning – compared to the specific auxiliary data including a set of associated IDs provided in step 6 as described below). The auxiliary data in step 2 may include PRS configuration information to enable the UE to receive and measure PRS transmitted from the gNB or TRP (and where PRS measurements can be input into the AI / ML model to output location). In this regard, the PRS configuration includes information indicating the RAN node entity (e.g., gNB or TRP) to which the PRS is to be transmitted (e.g., a list of cell IDs, NCGI, or TRP IDs).

[0191] In step 3, there is signal exchange between the UE and LMF to ensure consistency of network-side conditions between training and inference.

[0192] In step 4, the LMF requests AIML consistency information from each gNB involved in the inference / localization process. In this regard, the LMF requests the association ID from each gNB.

[0193] In step 5, each gNB indicates one or more of its respective associated IDs to the LMF.

[0194] In step 6 (broadly equivalent to) Figure 7 In box 701, the LMF collects all (or multiple) association IDs from each gNB (or from each of one or more TRPs within each gNB). The LMF maps each association ID to the corresponding gNB (or TRP) that transmitted the association ID to form a pair: gNB (or TRP) + association ID. This pair is used to form Set_2. The gNB (or multiple TRPs) of Set_2 may correspond to the gNB (or multiple TRPs) indicated in the PRS configuration information.

[0195] In step 7 (which is broadly equivalent to) Figure 7In box 702, the LMF sends UE assistance data to help the UE select which AI / ML model in its AI / ML model to use for the inference / AI / ML positioning process. In this respect, the LMF sends assistance data to the UE including information indicating Set_2.

[0196] In step 8, the UE uses the Set_2 information to evaluate all models available to the UE under the specific function set in step 2. This evaluation includes checking whether all gNBs (or TRPs) and associated IDs (which may be referred to as Set_1 of gNBs (or TRPs) + associated IDs for all models available to the UE) associated with each of the UE's multiple available models match the information available in Set_2 (e.g., a pair: gNBs (or TRPs) + associated IDs of Set_2). The UE calculates the consistency_indicator (i.e., consistency_indicator represents the result of the evaluation) based on the evaluation and the matching degree.

[0197] In step 8, a matching degree / level can be evaluated to soften the mismatch assessment (i.e., a value indicating the matching degree / level can be evaluated, where the value is not necessarily a binary value of 0 or 1, but can be a real number between 0 and 1). Accordingly, the consistency_indicator can be any real number between 0 and 1, for example, consistency_indicator = {0, 0.1, 0.2, 0.8, 0.9, 1}, where a value of 0 indicates no / lowest matching degree and a value of 1 indicates full / highest matching degree. The value of the consistency indicator is calculated for all models deployed in the UE. In step 11, this matching degree / level can be indicated to provide the UE with additional flexibility to select a specific model and indicate the applicability of related functions.

[0198] In step 9 (broadly equivalent to) Figure 7 In box 704, the UE selects the model with the highest consistency_indicator value among its models.

[0199] In step 10, the function activation and further steps are completed for inference. In this regard, the function established in step 2 is activated, and the UE uses the selected model to perform the function, namely the inference process / AIML localization.

[0200] For the UE-side scenario (i.e., the model deployed on the UE side), the above example enables the use of (multiple) association IDs (e.g., AIML positioning) in the protocol between the UE and the CN. The example provides new signaling between RAN node entities (e.g., gNB) and CN entities / functions (e.g., LMF for positioning) to enable the use of (multiple) association IDs in a protocol different from RRC. These (multiple) association IDs can be used to improve / ensure consistency of network-side conditions for data collection (for collecting training data to train the model) and inference (using the trained model).

[0201] In this regard, for data collection, based on UE capabilities, CN functions or CN entities can assist UEs with processes involving (multiple) associated IDs to ensure consistency between training and inference. For this purpose, CN functions or CN entities can send process requests for AIML consistency information to all gNBs (NG-RAN) involved in the AIML application. The AIML consistency information can contain a list of (multiple) associated IDs that the gNB should use. In response to the request, the gNB can send an acknowledgment message with the applicable associated IDs.

[0202] In the first variant, the CN entity / function can provide the gNB with a list of NW-side additional conditions (e.g., the LMF can assign NW-side additional conditions applicable to the gNB) and information related to RS transmission (e.g., PRS configuration) for each NW-side additional condition. The gNB can examine its NW assumptions regarding RS transmission (in other words, information related to the NW-side additional conditions) for data collection and map the NW-side additional conditions from the list provided by the CN entity / function to associated IDs originating from the gNB. When mapping is completed at the gNB for one or more associated IDs, a response message can be sent to the CN entity / function to confirm that one or more NW-side additional conditions from the list of NW-side additional conditions can be supported using the given RS transmission information.

[0203] In the second variant, the CN entity / function can provide a list of NW-side additional conditions (e.g., beam information, etc.), and the gNB can check its NW-side additional conditions and determine whether it can support the NW-side additional conditions from the list provided by the CN entity / function. After determining possible support, the specific NW-side additional conditions are mapped to one or more associated IDs, and a response message indicating the associated ID can be sent to the CN entity / function.

[0204] For inference, during inference, depending on the process, the CN function or CN entity can check a list of (multiple) associated IDs from different gNBs (NG-RAN). For this purpose, a process request is sent from the CN function or CN entity to each gNB (NG-RAN) to request the associated IDs of each gNB, thus expecting the LMF to obtain a set of (multiple) associated IDs.

[0205] In some examples, a gNB can initiate signaling to report a list of supported association IDs (for data collection or inference). For instance, suppose the LMF assigns static association IDs to gNBs within a given area; since consistency in the use of association IDs within a gNB can be expected, the gNB can subsequently initiate signaling procedures based on the assigned list of association IDs.

[0206] The (multiple) association IDs are identifiers of NW assumptions or settings (NW provider implementation details) that have never been disclosed between the UE provider, NW provider, and CN provider. Therefore, the only information disclosed between the providers is the identifier (i.e., the association identifier), without details of what it represents.

[0207] In the second variant described above, the CN entity / function that provides information about the additional conditions of NW can be only a general high-level guideline on the NW assumptions, rather than the exact implementation details of gNB / TRP, and the NW supplier is still free to use only one set of NW assumptions within the constraints configured by the CN entity / function.

[0208] Figures 2 to 8 A signaling diagram can be considered to represent multiple methods; in a sense, each signaling diagram can be considered to represent one or more actions, procedures, or processes performed by / at multiple parties / entities (e.g., UE 110, LMF 140, and gNB 120). Therefore, a signaling diagram can be considered to represent multiple individual methods performed by each corresponding individual party / entity among multiple parties / entities.

[0209] The aforementioned component blocks and steps (e.g., steps in the signaling diagram) are functional, and the functionality, along with the further functions / functionalities described above, can be provided by a single physical entity (e.g., as referenced). Figure 9 The described function is performed by a device embodied in the UE, LMF, or gNB. The described function can also be implemented by a computer program (such as reference 1). Figure 10 The described process is for execution by the UE's processor and the gNB's LMF.

[0210] Figure 9 This schematically illustrates the means for performing the functions described in this disclosure and Figures 2 to 8The block diagram of the apparatus 10, which includes methods, procedures, processes, and signaling, is shown. In this respect, the apparatus can perform the role of an entity (such as a UE, gNB's LMF) in the methods shown and described.

[0211] Figure 9 The component boxes are functional, and the described functions can be performed by a single physical entity, such as a UE, LMF, or gNB.

[0212] The device includes a controller 11, which may be located within a device / entity, rather than at least such as a UE, LMF, or gNB.

[0213] The controller 11 may be embodied by a computing device, rather than at least those described above. In some, but not all, examples, the device may be embodied as a chip, chipset, circuit, or module, i.e., as used in any of the foregoing. The term "module" as used herein refers to a unit or device excluding certain parts / components that will be added by the end-user manufacturer or user.

[0214] The controller 11 can be implemented as a controller circuit. The controller 11 can be implemented solely in hardware, have certain aspects of software including firmware, or can be a combination of hardware and software (including firmware).

[0215] The controller 11 can be implemented using instructions that implement hardware functions, for example, by using executable instructions of a computer program 14 in a general-purpose or special-purpose processor 12, which can be stored on a computer-readable storage medium 13 (e.g., a memory or a disk) for execution by such a processor 12.

[0216] Processor 12 is used to read from and write to memory 13. Processor 12 may also include an output interface through which processor 12 outputs data and / or commands, and an input interface through which data and / or commands are input to processor 12. The device may be coupled to or include one or more other components 15 (not at least, for example, radio transceivers, sensors, input / output user interface elements, and / or other modules / devices / components for inputting and outputting data / commands).

[0217] Memory 13 stores instructions, such as computer program 14, which include such instructions (e.g., computer program instructions / code) that control the operation of device 10 when loaded into processor 12. The instructions of computer program 14 provide instructions that enable the device to perform the operations described in this disclosure and... Figures 2 to 8 The methods, procedures, and process logic and routines shown are illustrated. Processor 12 can load and execute computer program 14 by reading memory 13.

[0218] Instructions may be included in a computer program, a non-transitory computer-readable medium, a computer program product, or a machine-readable medium. As used herein, the term "non-transitory" refers to a limitation on the medium itself (i.e., tangible, not tactile) rather than on the persistence of data storage (e.g., RAM versus ROM). In some, but not necessarily all, examples, computer program instructions may be distributed across more than one computer program.

[0219] Although memory 13 is shown as a single component / circuit, it can be implemented as one or more separate components / circuits, some or all of which can be integrated / removable and / or provide permanent / semi-permanent / dynamic / cached storage.

[0220] Although processor 12 is shown as a single component / circuit, it can be implemented as one or more separate components / circuits, some or all of which can be integrated / removable. Processor 12 can be a single-core or multi-core processor.

[0221] The apparatus may include tools for implementing the features described herein. Figures 2 to 8 The methods, processes, and procedures shown herein may include one or more components. It is conceivable that the functionality of these components may be combined within one or more components or performed by other components having equivalent functionality. The description of the functionality should also be considered as disclosing any means suitable for performing that functionality.

[0222] Where a structural feature has been described, it may be replaced by a component that performs one or more functions of the structural feature, whether or not the function is explicitly or implicitly described.

[0223] Although examples of devices have been described above based on various components, it should be understood that components may be embodied as corresponding controllers or circuits (such as one or more processing elements or processors of the device) or otherwise controlled by corresponding controllers or circuits. In this regard, each of the components described above may be one or more of any device, component, or circuit embodied in hardware, software, or a combination of hardware and software, configured to perform the corresponding function of the corresponding component as described above.

[0224] The device can be, for example, a user equipment, base station, or network node of a mobile cellular telecommunications system. The device can be embodied in a computing device, not just those described above. However, in some examples, the device can be embodied as a chip, chipset, circuit, or module, i.e., used in any of the foregoing.

[0225] In one instance, the device is embodied in a client device, UE, mobile cellular phone, handheld portable electronic device, mobile communication device, wearable computing device, or personal digital assistant, which may additionally provide one or more audio / text / video communication functions (e.g., telecommunication, video communication, and / or text transmission (SMS / MMS / email) functions), interactive / non-interactive viewing functions (e.g., web browsing, navigation, TV / program viewing functions), music recording / playback functions (e.g., Motion Picture Experts Group-1 Audio Layer 3 (MP3) or other formats and / or (frequency modulation / amplitude modulation) radio broadcast recording / playback), data download / transmission functions, image capture functions (e.g., using (e.g., built-in) digital camera), and gaming functions, or any combination thereof.

[0226] In some examples (such as where the device is provided within UE 110), device 10 includes: At least one processor 12; and At least one memory 13 stores instructions that, when executed by at least one processor 12, cause the device to perform at least the following: Send first information to the nodes of the core network indicating a first set of associated ID models, wherein the device includes models associated with a first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; Receive second information from the nodes in the core network indicating the degree of matching between the following two: The first set of associated IDs, and The second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in the second set; and Determine whether to select a model, wherein the determination is based at least in part on the second information.

[0227] In some examples (such as where the device is provided within LMF 140), device 10 includes: At least one processor 12; and At least one memory 13 stores instructions that, when executed by at least one processor 12, cause the device to perform at least the following: Receive first information from the user equipment (UE), the first information indicating a first set of association IDs associated with the model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Obtain the second set of associated IDs, where the second set of associated IDs is associated with one or more network-side conditions in the second set; The second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and...

[0228] The examples above apply to enabled components for: telecommunications systems; tracking systems; automotive systems; electronic systems, including consumer electronics; distributed computing systems; media systems for generating or rendering media content, including audio, visual, and audiovisual content, as well as mixed, mediated, virtual, and / or augmented reality; personal systems, including personal health systems or personal fitness systems; navigation systems; user interfaces, also known as human-machine interfaces; networks, including cellular, non-cellular, and optical networks; self-organizing networks; the Internet of Things (IoT); vehicle-to-everything (V2X) networks; virtualized networks; and related software and services.

[0229] According to the examples of this disclosure, the device can be disposed in an electronic device, such as a mobile terminal. However, it should be understood that the mobile terminal is merely an illustration of an electronic device that will benefit from the implementations of this disclosure, and therefore should not be construed as limiting the scope of this disclosure. While the device may be provided in a mobile terminal in some implementation examples, other types of electronic devices (such as, but not limited to, mobile communication devices, handheld portable electronic devices, wearable computing devices, portable digital assistants (PDAs), pagers, mobile computers, desktop computers, televisions, gaming devices, laptop computers, cameras, video recorders, GPS devices, and other types of electronic systems) can readily adopt the examples of this disclosure. Furthermore, devices can readily adopt the examples of this disclosure regardless of their intention to provide mobility.

[0230] Figure 10 A computer program 14 that can be transmitted via delivery mechanism 20 is shown. Delivery mechanism 20 can be any suitable delivery mechanism, such as a machine-readable medium, a computer-readable medium, a non-transitory computer-readable storage medium, a computer program product, a memory device, a solid-state storage device, a recording medium such as an optical disc read-only memory (CD-ROM) or a digital versatile optical disc (DVD), or an article of manufacture that includes or tangibly embodies the computer program 14. The delivery mechanism can be a signal configured to reliably transmit the computer program. The apparatus can receive, propagate, or transmit the computer program as a computer data signal.

[0231] In some examples of this disclosure, a computer program is provided that includes instructions that, when executed by a device (e.g., UE 110), cause the device to perform at least the following operations or are used to cause the device to perform at least the following operations: Sending first information to nodes in the core network indicating a first set of associated ID models, wherein the apparatus includes models associated with a first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions. Receive second information from the nodes in the core network indicating the degree of matching between the following two: The first set of associated IDs, and The second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in the second set; and At the device, it is determined whether to select a model, wherein the determination is based at least in part on the second information.

[0232] In some embodiments of the present invention, a computer program is provided that includes instructions, which, when executed by a device (e.g., LMF 140), cause the device to perform at least the following operations or are used to cause the device to perform at least the following operations: Receive first information from the user equipment (UE), the first information indicating a first set of association IDs associated with the model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Obtain the second set of associated IDs, where the second set of associated IDs is associated with one or more network-side conditions in the second set; The second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs, and A second message is sent to the UE so that the UE can determine whether to select a model.

[0233] References to “computer program,” “computer-readable storage medium,” “computer program product,” “computer program tangibly embodied,” or “controller,” “computer,” “processor,” etc., should be understood to encompass not only computers with different architectures such as single / multiprocessor architectures and sequential (Von Neumann) / parallel architectures, but also special-purpose circuitry such as field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), signal processing devices, and other devices. References to computer programs, instructions, code, etc., should be understood to encompass software used in programmable processors or firmware, such as the programmable content of hardware devices, whether instructions for processors or configuration settings for fixed-function devices, gate arrays, or programmable logic devices.

[0234] As used in this application, the term "circuit" may refer to one or more or all of the following: (a) Hardware circuit implementation only (such as implementation in analog and / or digital circuits only); and (b) A combination of hardware circuitry and software, such as (if applicable): i) A combination of (multiple) analog and / or digital hardware circuits and software / firmware, and ii) Any part of a hardware processor having software (including (multiple) digital signal processors), software, and (multiple) memories works together to enable a device such as a mobile phone or server to perform various functions; and (c) (Multiple) hardware circuits and / or (multiple) processors, such as (multiple) microprocessors or a portion thereof, which require software (e.g., firmware) for operation, but may not exist when not required for operation.

[0235] This definition of "circuit" applies to all uses of the term in this application (including any claim). As another example, as used in this application, the term "circuit" also covers only the implementation of hardware circuitry or processors and (or their accompanying) software and / or firmware. The term "circuit" also covers, for example and if applicable to elements of a particular claim, baseband integrated circuits for mobile devices or similar integrated circuits in servers, cellular network devices, or other computing or networking devices.

[0236] Although various examples of this disclosure have been described in the preceding paragraphs, it should be understood that modifications may be made to the given examples without departing from the scope of the invention as set forth in the claims.

[0237] Figures 2 to 8 The boxes shown may represent actions in a method, functions performed by a device, and / or instruction / code segments in a computer program. The illustration of a specific order of boxes does not necessarily imply a required or preferred order, and the order and arrangement of boxes can vary. Furthermore, some boxes may be omitted.

[0238] It should be understood that Figures 2 to 8 Each box and combination of boxes shown, as well as the further functions described above, can be implemented in various ways, such as hardware, firmware, and / or software including one or more computer program instructions. For example, one or more of the functions described above can be implemented by appropriately configured means (e.g., means including units for performing the functions described above [such as...]). Figure 9 [As shown]) can be performed. One or more of the above functions / functions can be performed by a properly configured computer program (such as a computer program including computer program instructions [e.g. Figure 10 [As shown], the computer program instructions embody the above-mentioned functions / functionality and can be stored in a memory storage device and executed by a processor.

[0239] It should be understood that any such computer program instructions can be loaded onto a computer or other programmable device (i.e., hardware) to produce a machine, such that when the instructions are executed on the programmable device, they create means for implementing the function / functionality specified in the box. These computer program instructions can also be stored in a computer-readable medium that can instruct the programmable device to operate in a particular manner, such that the instructions stored in the computer-readable storage produce an article of writing including means of instruction implementing the function specified in the box. The computer program instructions can also be loaded onto the programmable device to cause a series of operations to be performed on the programmable device to produce a computer-implemented process, such that the instructions, which execute on the programmable device, provide actions for implementing the function / functionality specified in the box.

[0240] Various, but not all, examples of this disclosure may take the form of methods, apparatus, or computer programs. Accordingly, various, but not all, examples may be implemented in hardware, software, or a combination of hardware and software.

[0241] Flowcharts and schematic block diagrams are used to describe various, but not all, examples of this disclosure. It should be understood that each block (of the flowcharts and block diagrams), and combinations thereof, can be implemented by computer program instructions of a computer program. These program instructions can be provided to one or more processors, processing circuits, or controllers such that instructions executing on the same fabrication apparatus cause the function specified in the one or more blocks to be implemented, thus making the method computer-implemented. Computer program instructions can be executed by the processor(s) to cause a series of operation blocks / steps / actions to be performed by the processor(s) to produce a computer-implemented process, such that instructions executing on the processor(s) provide one or more blocks for implementing the function specified in the one or more blocks.

[0242] Therefore, a box supports: combinations of components for performing a specified function; combinations of actions for performing a specified function; and computer program instructions / algorithms for performing a specified function. It will also be understood that each box and combination of boxes can be implemented by a system based on dedicated hardware or a combination of dedicated hardware and computer program instructions that performs the specified function or action.

[0243] The various, but not necessarily all, examples of this disclosure provide a method and corresponding means including various modules, components, or circuits that provide functionality for performing / applying the method. The modules, components, or circuits may be implemented as hardware or as software or firmware executed by a computer processor. In the case of firmware or software, examples of this disclosure may be provided as a computer program product including a computer-readable storage structure embodying computer program instructions (i.e., software or firmware) for execution by a computer processor.

[0244] Although specific terms are used in this article, they are used only in a general and descriptive sense, and not for limiting purposes.

[0245] The features described above may be used in combinations other than those explicitly described.

[0246] Although some features have been described with reference to certain characteristics, these functions can be performed by other features, whether or not they are described.

[0247] Although features have been described with reference to certain examples, these features may also exist in other examples, whether or not they are described. Therefore, a feature described with respect to one example / aspect of this disclosure may include any or all of the features described with respect to another example / aspect of this disclosure, provided that they are not inconsistent with each other, and vice versa.

[0248] The term "includes" is used herein in a sense of inclusion rather than exclusivity. That is, any reference to X that includes Y indicates that X may include only one Y or may include more than one Y. If the intention is to use "includes" with an exclusive meaning, it will become clear in the context by referring to "includes only one..." or by using "comprises".

[0249] In this specification, the terms “connection” and “communication” and their derivatives mean operatively connecting / communicating. It should be understood that any number or combination of intermediate components (including no intermediate components) may be present to provide direct or indirect connection / coupling / communication. Any such intermediate component may include hardware and / or software components.

[0250] As used herein, the term "determine / determine" (and its grammatical variations) can include, in particular: evaluation, calculation, measurement, processing, derivation, measurement, investigation, identification, lookup (e.g., searching in a table, database, or other data structure), ascertainment, etc. Furthermore, "determine" can include receiving (e.g., receiving information), retrieving / accessing (e.g., retrieving / accessing data in memory), obtaining, etc. Additionally, "determine / determine" can include parsing, selecting, choosing, establishing, reasoning, etc.

[0251] As used herein, a description of an action should also be considered as disclosing enabling and / or causing and / or controlling the action. For example, a description of transmitting information should also be considered as disclosing enabling and / or causing and / or controlling the transmission of information. Similarly, for example, a description of a means of transmitting information should also be considered as disclosing at least one means or controller for enabling and / or causing and / or controlling the means of transmitting the information.

[0252] The term "component" as used in the specification and claims may refer to one or more individual elements configured to perform a corresponding recounted function, or it may refer to several elements performing such functions. Furthermore, the functions recounted in the claims may be performed by the same individual components or the same combination of components. For example, such functions may be performed in the device by a processor executing instructions stored in the device's memory.

[0253] References to parameters or parameter values ​​should be understood as referring to data that "indicates" the relevant parameter / value, "defines" the relevant parameter / value, or "represents" the relevant parameter / value (unless the context requires otherwise). Data may indicate the relevant parameter / value in any way, and may indicate the relevant parameter / value directly or indirectly.

[0254] Various examples have been referenced in this specification. Descriptions of features or functions of an example indicate which features or functions exist in that example. The use of the terms “example” or “for example,” “can,” or “may” in the text indicates, whether explicitly stated or not, that such features or functions exist in at least the described example, whether or not described as an example, and that they may, but not necessarily, exist in some or all other examples. Thus, “example,” “for example,” “can,” or “may” refers to a specific instance of a class of examples. An instance’s properties may be properties of only that instance, properties of the class, or properties of subclasses of the class that include some, but not all, instances of that class.

[0255] In this specification, unless otherwise expressly stated, references to “a / an / that” [feature, element, component, part] are used in an inclusive rather than exclusive sense and will be interpreted as “at least one” [feature, element, component, part]. That is, any reference to X that includes a / Y indicates that X may include only one Y or may include more than one Y, unless the context clearly indicates the opposite. If “a” or “that” is intended to have an exclusive meaning, it will become clear in the context. In some cases, the use of “at least one” or “one or more” may be used to emphasize the inclusive meaning, but the absence of these terms should not be taken as grounds for any exclusive meaning. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, wherein a list of two or more elements is combined by “and” or “or”, means at least any one of the elements, or at least any two or more of the elements, or at least all of the elements.

[0256] The presence of a feature (or combination of features) in a claim relates to the feature (or combination of features) itself, and also to features that achieve substantially the same technical effect (equivalent features). Equivalent features include, for example, features that are variations and achieve substantially the same result in substantially the same manner. Equivalent features include, for example, features that perform substantially the same function in substantially the same manner to achieve substantially the same result.

[0257] In this specification, various examples have been referenced to describe the characteristics of the examples using adjectives or adjective phrases. Such descriptions of characteristics associated with an example indicate that the characteristic exists exactly as described in some examples, and substantially as described in others.

[0258] In the above description, the described apparatus may alternatively or additionally include apparatuses that, in some other examples, comprise a distributed system of apparatuses, such as a client / server apparatus system. In examples where the apparatus provides for forming a distributed system (or the method is implemented as a distributed system), each apparatus of a component and / or part forming the system provides (or implements) one or more features that collectively implement the examples of this disclosure. In some examples, the apparatus is reconfigured by an entity other than its initial manufacturer to implement the examples of this disclosure by, for example, by a user who downloads additional software, which, when executed, causes the apparatus to implement the examples of this disclosure (such implementation is entirely by the apparatus or as part of a system of apparatuses as described above).

[0259] The foregoing description illustrates some examples of this disclosure; however, those skilled in the art will recognize possible alternative structural and methodological features that provide equivalent functionality to specific examples of such structures and features described above, and have been omitted from the foregoing description for the sake of brevity and clarity. Nevertheless, unless such alternative structural or methodological features are expressly excluded in the foregoing description of the examples of this disclosure, the foregoing description should be understood to implicitly include such alternative structural and methodological features that provide equivalent functionality.

[0260] Although the foregoing specification avoids focusing on those features of the examples of this disclosure that are considered particularly important, it should be understood that the applicant claims protection for any patentable feature or combination of features shown in the foregoing references and / or the accompanying drawings, whether or not they have been specifically emphasized.

[0261] The examples and appended claims of this disclosure may be appropriately combined in any manner obvious to a person skilled in the art. Unless stated as such and / or unless obvious from the specification to a person skilled in the art, individual references to “example,” “in some examples,” etc., in the specification do not necessarily refer to the same example and are not mutually exclusive. For example, a feature, structure, process, block, step, action, etc., described in one example may also be included in other examples, but are not necessarily included.

[0262] Each and every claim is incorporated herein as a further disclosure, and the claims are embodiments of this disclosure. Furthermore, while the claims herein are provided to include specific dependencies, it is contemplated that any claim may depend on any other claim, and any such alternative embodiments and their equivalents are also within the scope of this disclosure to the extent that any alternative embodiment may arise from the combination, integration, and / or omission of features and / or alteration of the dependencies of various claims.

[0263] Furthermore, the various implementations of this disclosure can be described with reference to the following terms, and their features can be combined in any reasonable manner.

[0264] Clause 1. An apparatus (10, 110) comprising: a model (212_u) associated with a first set of association identifier IDs (205_u), wherein the first set of association IDs is associated with a first set of network-side conditions; a component (11) for sending first information (502) to a node (130, 140) of a core network, the first information (502) indicating the first set of association IDs; a component (11) for receiving second information (505) from the node of the core network, the second information (505) indicating a degree of matching (506) between the first set of association IDs and a second set of association IDs (205_bs), wherein the second set of association IDs is associated with a second set of one or more network-side conditions; and a component (11) for determining whether to select the model, wherein the determination is based at least in part on the second information.

[0265] Clause 2. The apparatus according to Clause 1, wherein the determination of whether to select the model is also based at least in part on the matching degree satisfying the condition.

[0266] Clause 3. An apparatus according to any one of the preceding clauses, wherein: the apparatus comprises a plurality of models, wherein each model is associated with a corresponding set of associated IDs, and each set of associated IDs is associated with a corresponding set of network-side conditions; the first information comprises an indication of each of the plurality of said corresponding sets of associated IDs; the second information comprises an indication of the degree of matching between the following two for each of the plurality of said corresponding sets of associated IDs: the corresponding set of associated IDs and the second set of associated IDs; and the determination comprises selecting a model from the plurality of models based at least in part on the second information.

[0267] Clause 4. An apparatus according to any one of the preceding clauses, wherein each associated ID represents at least one of the following: one or more network-side additional conditions; one or more network-side assumptions; one or more explicit assumptions of the nodes of the core network; or one or more network-side appropriate settings used by one or more entities when collecting data for training the model.

[0268] Clause 5. The device according to any one of the preceding clauses, wherein the model includes at least one of the following: an artificial intelligence (AI) model; a machine learning (ML) model; a model for the device; or a model for AI / ML positioning on the user equipment (UE) side.

[0269] Clause 6. The apparatus according to any one of the preceding clauses further includes: a component for receiving configuration information from the node of the core network, the configuration information for configuring the apparatus to perform a positioning process (206), wherein the configuration information includes information indicating one or more Location Reference Signal (PRS) configurations (203), wherein the one or more PRS configurations include information indicating a second group of one or more entities; and a component for performing the positioning process at least in part based on the configuration information, wherein the positioning process is performed using the selected model.

[0270] Clause 7. A method comprising: sending from a device to a node of a core network first information indicating a first set of associated ID models, wherein the device includes models associated with the first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; receiving from the node of the core network second information indicating a degree of matching between the first set of associated IDs and a second set of associated IDs, wherein the second set of associated IDs is associated with a second set of one or more network-side conditions; and determining at the device whether to select the model, wherein the determination is based at least in part on the second information.

[0271] Clause 8. A computer program comprising instructions that, when implemented by a device, cause the device to: send first information to a node in a core network indicating a first set of associated ID models, wherein the device includes models associated with the first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; receive from the node in the core network second information indicating a degree of matching between the first set of associated IDs and a second set of associated IDs, wherein the second set of associated IDs is associated with a second set of one or more network-side conditions; and determine at the device whether to select the model, wherein the determination is based at least in part on the second information.

[0272] Clause 9. An apparatus (10, 130, 140) comprising: a component (11) for receiving first information (502) from a user equipment (UE) (110), the first information (502) indicating a first set of association IDs (205_u) associated with a model (212_u) of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; a component (11) for acquiring a second set of association IDs (205_bs), wherein the second set of association IDs is associated with a second set of one or more network-side conditions; a component (11) for evaluating second information (505) at least in part based on the first set of association IDs and the second set of association IDs, wherein the second information indicates a degree of matching (506) between the first set of association IDs and the second set of association IDs; and a component (11) for sending the second information to the UE such that the UE can determine whether to select the model.

[0273] Clause 10. The apparatus according to Clause 9, wherein obtaining the second set of associated IDs comprises: sending a request for one or more associated IDs to one or more entities in the second set of entities; and receiving the one or more associated IDs from the one or more entities in the second set of entities in response to the request.

[0274] Clause 11. The apparatus according to Clause 10, wherein the request includes information indicating at least one of the following: one or more PRS configurations; or one or more network-side additional conditions.

[0275] Clause 12. The apparatus according to any one of Clauses 9 to 11 further includes: a component for associating at least one of the one or more association IDs with at least one of the one or more entities, wherein the association is based at least in part on determination of which particular association ID was received from which particular one or more entities.

[0276] Clause 13. An apparatus according to any one of Clauses 9 to 12, wherein: the UE includes a plurality of models, wherein each model is associated with a corresponding set of associated IDs, and each set of associated IDs is associated with a corresponding set of entities; the first information includes an indication of each of the plurality of said corresponding sets of associated IDs; and the second information includes: an indication of the degree of matching between the following two for each of the plurality of said corresponding sets of associated IDs: the corresponding set of associated IDs and the second set of associated IDs.

[0277] Clause 14. A method comprising: receiving, at a device, first information from a user equipment (UE), the first information indicating a first set of association IDs associated with a model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; obtaining, at the device, a second set of association IDs, wherein the second set of association IDs is associated with a second set of one or more network-side conditions; evaluating, at the device, second information based at least in part on the first set of association IDs and the second set of association IDs, wherein the second information indicates a degree of matching between the first set of association IDs and the second set of association IDs; and sending, from the device, the second information to the UE to enable the UE to determine whether to select the model.

[0278] Clause 15. A computer program comprising instructions, which, when implemented by a device, cause the device to: receive first information from a user equipment (UE), the first information indicating a first set of association IDs associated with a model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; obtain a second set of association IDs, wherein the second set of association IDs is associated with a second set of one or more network-side conditions; evaluate second information based at least in part on the first set of association IDs and the second set of association IDs, wherein the second information indicates a degree of matching between the first set of association IDs and the second set of association IDs; and send the second information to the UE to enable the UE to determine whether to select the model.

Claims

1. A communication apparatus (10, 110), comprising: Model (212_u), wherein the model is associated with a first set of association identifier IDs (205_u), wherein the first set of association IDs is associated with a first set of network-side conditions; A component (11) for sending first information (502) to nodes (130, 140) in the core network, the first information (502) indicating the first set of associated IDs; A component (11) for receiving second information (505) from the node in the core network, the second information (505) indicating the degree of matching (506) between the following two: The first set of associated IDs, and The second set of associated IDs (205_bs) is associated with one or more network-side conditions in the second set. as well as The component (11) is used to determine whether to select the model, wherein the determination is based at least in part on the second information.

2. The apparatus of claim 1, wherein the determination of whether to select the model is at least in part based on the matching degree satisfying the condition.

3. The apparatus according to any one of claims 1 to 2, wherein: The device includes multiple models, each model being associated with a corresponding set of association IDs, and each set of association IDs being associated with a corresponding set of network-side conditions. The first information includes an indication of each of the plurality of corresponding sets of associated IDs; The second information includes: for each of the plurality of corresponding sets of associated IDs, an indication of the degree of matching between the following two: The corresponding set of associated IDs, and The second set of associated IDs; and The determination includes selecting a model from the plurality of models, at least in part based on the second information.

4. The apparatus according to any one of claims 1 to 2, wherein each associated ID represents at least one of the following: One or more network-side additional conditions; One or more network-side hypotheses; One or more explicit assumptions about the nodes of the core network; or When collecting data for training the model, one or more network sides are appropriately configured by one or more entities.

5. The apparatus according to any one of claims 1 to 2, wherein the model comprises at least one of the following: Artificial intelligence (AI) models; Machine learning (ML) models; A model for the device; or Models for AI / ML positioning on the user equipment (UE) side.

6. The apparatus according to any one of claims 1 to 2, further comprising: A component for receiving configuration information from the node in the core network, the configuration information being used to configure the device to perform a positioning process (206), wherein the configuration information includes information indicating the following: One or more Position Reference Signal (PRS) configurations (203), wherein the one or more PRS configurations include information indicating a second group of one or more entities; and Components for performing the positioning process based at least in part on the configuration information, wherein the positioning process is performed using the selected model.

7. A method for communication, comprising: The device sends first information indicating a first set of associated IDs to nodes in the core network, wherein the device includes a model associated with the first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions. At the device, second information indicating the degree of matching between the following two factors is received from the node in the core network: The first set of associated IDs, and A second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in a second set; and At the device, it is determined whether to select the model, wherein the determination is based at least in part on the second information.

8. A computer program product comprising instructions that, when executed by a device, cause the device to perform: Sending first information indicating a first set of associated IDs to nodes in the core network, wherein the device includes a model associated with the first set of associated identifier IDs, and wherein the first set of associated IDs is associated with a first set of network-side conditions; Receive second information from the node in the core network indicating the degree of matching between the following two: The first set of associated IDs, and A second set of associated IDs, wherein the second set of associated IDs is associated with one or more network-side conditions in a second set; and At the device, it is determined whether to select the model, wherein the determination is based at least in part on the second information.

9. A communication apparatus (10, 130, 140), comprising: A component (11) for receiving first information (502) from a user equipment (UE) (110), the first information (502) indicating a first set of association IDs (205_u) associated with a model (212_u) of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; Component (11) for obtaining a second set of associated IDs (205_bs), wherein the second set of associated IDs is associated with a second set of one or more network-side conditions; A component (11) for evaluating second information (505) at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching (506) between the following two: The first set of associated IDs, and The second set of associated IDs; and A component (11) for sending the second information to the UE so that the UE can determine whether to select the model.

10. A method comprising: The device receives first information from the user equipment (UE), the first information indicating a first set of association IDs associated with a model of the UE, wherein the first set of association IDs is associated with a first set of one or more network-side conditions; A second set of associated IDs is obtained at the device, wherein the second set of associated IDs is associated with a second set of one or more network-side conditions; At the device, second information is evaluated at least in part based on the first set of associated IDs and the second set of associated IDs, wherein the second information indicates the degree of matching between the two: The first set of associated IDs, and The second set of associated IDs; and The device sends the second information to the UE so that the UE can determine whether to select the model.

Images