Machine learning positioning aspects for AIML model at network nodes

Abstract

Various examples of embodiments described herein relate to apparatuses and methods for machine learning positioning. One such example of an embodiment relates to an apparatus comprising at least one processor, and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: determine a positioning configuration request based on machine learning related capabilities of a plurality of network nodes capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment; send the positioning configuration request to at least one network node of the plurality of network nodes; receive at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration; and configure at least one positioning configuration to the at least one network node of the plurality of network nodes based on the at least one response received from the at least one network node of the plurality of network nodes.

Description

[0001] MACHINE LEARNING POSITIONING ASPECTS FOR AIML MODEL AT NETWORK NODES

[0002] Field of the Invention

[0003] The present invention relates to an apparatus, a method and a computer program product for ML (machine learning) Positioning.

[0004] Related background Art

[0005] The following meanings for the abbreviations used in this specification apply:

[0006] Al Artificial Intelligence

[0007] AIML Artificial Intelligence and Machine Learning

[0008] AMF Access and Mobility Function

[0009] BW Bandwidth

[0010] CHO Conditional handover

[0011] CIR Channel Impulse Response

[0012] CNN Convolutional Neural Network

[0013] CSI Channel State Information

[0014] DL Downlink

[0015] DP Delay Profile

[0016] HO Handover

[0017] LCM - Lifecycle Management

[0018] LCS Location Services

[0019] LMF Location Management Function

[0020] LPP LTE Positioning Protocol

[0021] ML Machine Learning

[0022] NF Network Function

[0023] NRPPa NR Positioning Protocol A

[0024] NW Network

[0025] PDP Power Delay Profile

[0026] PRS Positioning Reference Signal

[0027] ResNet Residual Neural Network

[0028] RNN Recurrent Neural Network SRS Sounding Reference Signals

[0029] ToA Time of Arrival

[0030] TRP Transmission and Reception Point

[0031] TX Transmission

[0032] UE - User Equipment

[0033] UL Uplink

[0034] UPF User Plane Function

[0035] Example embodiments, although not limited to this, relate to UE positioning procedures. UE positioning can be carried out using AI / ML techniques. When applying AI / ML techniques, some problems may occur.

[0036] Summary of the Invention

[0037] Example embodiments address this situation aim to provide an improved handling of UE positioning.

[0038] Several aspects of the various example embodiments will be described with respect to certain aspects. These aspects are not intended to indicate key or essential features of the various example embodiments, nor are they intended to be used to otherwise limit the scope of the subject disclosure. Other features, aspects and elements of the various example embodiments will be readily apparent to a person skilled in the art in view of the subject disclosure.

[0039] According to a first aspect, an apparatus is provided which comprises at least one processor, and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: determine a positioning configuration request based on machine learning related capabilities of a plurality of network nodes capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, send the positioning configuration request to at least one network node of the plurality of network nodes, receive at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration, and configure at least one positioning configuration to the at least one network node of the plurality of network nodes based on the at least one response received from the at least one network node of the plurality of network nodes.

[0040] The first aspect may be modified as follows:

[0041] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: determine machine learning related capabilities of a network node of the plurality of network nodes by requesting, from the network node, information concerning its machine learning related capabilities and receiving a response from the network node including the information concerning the machine learning related capabilities.

[0042] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by each network node of the plurality of network nodes.

[0043] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model and / or machine learning functionality, and / or information concerning computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

[0044] The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature. The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format.

[0045] The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0046] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: determine the positioning configuration request to the plurality of network nodes based on the machine learning related capabilities of the plurality of network nodes by aligning a machine learning model and / or machine learning functionality to be applied by the plurality of network nodes, aligning a format of intermediate features to be applied for the aligned machine learning model, and aligning uplink sounding reference signals configuration.

[0047] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: configure the positioning configuration to the at least one network node of the plurality of network nodes based on responses received from all network nodes of the plurality of network nodes by determining which positioning configuration is or which positioning configurations are supported by all network nodes of the plurality of network nodes.

[0048] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: configure all network nodes of the plurality of network nodes with the same positioning configuration or configure the network nodes of the plurality of network nodes individually with positioning configuration.

[0049] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: receive positioning related measurement results from the plurality of network nodes, and determine the position of the user equipment based on a combination of the received positioning related measurement results.

[0050] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: combine the intermediate features by applying weights to the positioning related measurement results.

[0051] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: adjust the weights of intermediate features received from the plurality of network nodes based on a movement of the user equipment.

[0052] The positioning related measurement results may comprise intermediate features.

[0053] According to a second aspect, an apparatus is provided which comprises: at least one processor, and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: receive a positioning configuration request from a network control element, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, select at least one positioning configuration of the positioning configuration request based on machine learning related capabilities of the apparatus, send a response to the network control element including the at least one selected positioning configuration, receive information concerning a positioning configuration to be applied, and perform positioning related measurements by using a machine learning model based on the positioning configuration to be applied.

[0054] The second aspect may be modified as follows:

[0055] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: receive a request from the network control element to provide information concerning machine learning related capabilities of the apparatus, and send a response to the network control element including the information concerning machine learning related capabilities.

[0056] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by the apparatus.

[0057] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model, and / or information concerning the machine learning model computational complexity and / or interference latency of the machine learning model and / or machine learning functionality. The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature.

[0058] The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format.

[0059] The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0060] The instructions stored by the at least one memory may further cause the apparatus, when executed by the at least one processor, at least to: send a result of the positioning related measurements to the network control element.

[0061] The result may be an intermediate feature.

[0062] According to a third aspect, a method is provided which comprises determining a positioning configuration request based on machine learning related capabilities of a plurality of network nodes capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, sending the positioning configuration request to at least one network node of the plurality of network nodes, receiving at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration, and configuring at least one positioning configuration to the at least one network node of the plurality of network nodes based on the at least one response received from the at least one network node of the plurality of network nodes.

[0063] The third aspect may be modified as follows:

[0064] The method may further comprise: determining machine learning related capabilities of a network node of the plurality of network nodes by requesting, from the network node, information concerning its machine learning related capabilities and receiving a response from the network node including the information concerning the machine learning related capabilities.

[0065] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by each network node of the plurality of network nodes.

[0066] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model and / or machine learning functionality, and / or information concerning computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

[0067] The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature.

[0068] The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format.

[0069] The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0070] The method may further comprise: determining the positioning configuration request to the plurality of network nodes based on the machine learning related capabilities of the plurality of network nodes by aligning a machine learning model and / or machine learning functionality to be applied by the plurality of network nodes, aligning a format of intermediate features to be applied for the aligned machine learning model, and aligning uplink sounding reference signals configuration.

[0071] The method may further comprise: configuring the positioning configuration to the at least one network node of the plurality of network nodes based on responses received from all network nodes of the plurality of network nodes by determining which positioning configuration is or which positioning configurations are supported by all network nodes of the plurality of network nodes.

[0072] The method may further comprise: configuring all network nodes of the plurality of network nodes with the same positioning configuration or configuring the network nodes of the plurality of network nodes individually with positioning configuration.

[0073] The method may further comprise: receiving positioning related measurement results from the plurality of network nodes, and determining the position of the user equipment based on a combination of the received positioning related measurement results.

[0074] The method may further comprise: combining the intermediate features by applying weights to the positioning related measurement results.

[0075] The method may further comprise: adjusting the weights of intermediate features received from the plurality of network nodes based on a movement of the user equipment.

[0076] The positioning related measurement results may comprise intermediate features.

[0077] According to a fourth aspect, a method, in a network node, is provided which comprises: receiving a positioning configuration request from a network control element, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, selecting at least one positioning configuration of the positioning configuration request based on machine learning related capabilities of the network node, sending a response to the network control element including the at least one selected positioning configuration, receiving information concerning a positioning configuration to be applied, and performing positioning related measurements by using a machine learning model based on the positioning configuration to be applied. The fourth aspect may be modified as follows:

[0078] The method may further comprise: receiving a request from the network control element to provide information concerning machine learning related capabilities of the network node, and sending a response to the network control element including the information concerning machine learning related capabilities.

[0079] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by the network node.

[0080] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model, and / or information concerning the machine learning model computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

[0081] The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature.

[0082] The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format.

[0083] The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0084] The method may further comprise: sending a result of the positioning related measurements to the network control element.

[0085] The result may be an intermediate feature.

[0086] According to a fifth aspect of the present invention a computer program product is provided which comprises code means (or a set of instructions) for performing a method according to the third or fourth aspects and / or their modifications when run on a processing means or module. The computer program product may be embodied on a computer-readable medium, and / or the computer program product may be directly loadable into the internal memory of the computer and / or transmittable via a network by means of at least one of upload, download and push procedures.

[0087] According to a sixth aspect, an apparatus is provided which comprises means for determining a positioning configuration request based on machine learning related capabilities of a plurality of network nodes capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, means for sending the positioning configuration request to at least one network node of the plurality of network nodes, means for receiving at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration, and means for configuring at least one positioning configuration to the at least one network node of the plurality of network nodes based on the at least one response received from the at least one network node of the plurality of network nodes.

[0088] The sixth aspect may be modified as follows:

[0089] The apparatus may further comprise: means for determining machine learning related capabilities of a network node of the plurality of network nodes by requesting, from the network node, information concerning its machine learning related capabilities and receiving a response from the network node including the information concerning the machine learning related capabilities.

[0090] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by each network node of the plurality of network nodes.

[0091] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model and / or machine learning functionality, and / or information concerning computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

[0092] The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature.

[0093] The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format. The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0094] The apparatus may further comprise: means for determining the positioning configuration request to the plurality of network nodes based on the machine learning related capabilities of the plurality of network nodes by aligning a machine learning model and / or machine learning functionality to be applied by the plurality of network nodes, aligning a format of intermediate features to be applied for the aligned machine learning model, and aligning uplink sounding reference signals configuration.

[0095] The apparatus may further comprise: means for configuring the positioning configuration to the at least one network node of the plurality of network nodes based on responses received from all network nodes of the plurality of network nodes by determining which positioning configuration is or which positioning configurations are supported by all network nodes of the plurality of network nodes.

[0096] The apparatus may further comprise: means for configuring all network nodes of the plurality of network nodes with the same positioning configuration or means for configuring the network nodes of the plurality of network nodes individually with positioning configuration.

[0097] The apparatus may further comprise: means for receiving positioning related measurement results from the plurality of network nodes, and means for determining the position of the user equipment based on a combination of the received positioning related measurement results.

[0098] The apparatus may further comprise: means for combining the intermediate features by applying weights to the positioning related measurement results.

[0099] The apparatus may further comprise: means for adjusting the weights of intermediate features received from the plurality of network nodes based on a movement of the user equipment.

[0100] The positioning related measurement results may comprise intermediate features.

[0101] According to a seventh aspect, an apparatus is provided which comprises: means for receiving a positioning configuration request from a network control element, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, means for selecting at least one positioning configuration of the positioning configuration request based on machine learning related capabilities of the apparatus, means for sending a response to the network control element including the at least one selected positioning configuration, means for receiving information concerning a positioning configuration to be applied, and means for performing positioning related measurements by using a machine learning model based on the positioning configuration to be applied.

[0102] The seventh aspect may be modified as follows:

[0103] The apparatus may further comprise: means for receiving a request from the network control element to provide information concerning machine learning related capabilities of the apparatus, and means for sending a response to the network control element including the information concerning machine learning related capabilities.

[0104] The machine learning related capabilities may comprise information concerning a machine learning model and / or machine learning functionality applied by the network node.

[0105] The information concerning the machine learning model and / or machine learning functionality may comprise a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model, and / or information concerning the machine learning model computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

[0106] The outputs of the machine learning model and / or machine learning functionality may comprise at least one supported intermediate feature.

[0107] The inputs of the machine learning model and / or machine learning functionality may comprise at least one of a supported uplink sounding reference signals configurations, a supported channel impulse format, a supported power delay profile format, and / or a supported delay profile format.

[0108] The at least one positioning configuration indicated by the positioning configuration request may further comprise at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

[0109] The method may further comprise: sending a result of the positioning related measurements to the network control element.

[0110] The result may be an intermediate feature.

[0111] In all aspects and their modifications described above, the intermediate feature may comprise information, which is usable by the network control element for determining a position of the user equipment.

[0112] In all aspects and their modifications described above, the positioning method may be a machine learning based positioning method or a non-machine learning based positioning method.

[0113] Brief Description of the Drawings

[0114] These and other objects, features, details and advantages will become more fully apparent from the following detailed description of example embodiments, which is to be taken in conjunction with the appended drawings, in which:

[0115] Fig. 1A shows an LMF 1 according to an example embodiment,

[0116] Fig. IB shows a procedure carried out by the LMF 1 according to the example embodiment,

[0117] Fig. 2A shows a node 2 according to an example embodiment,

[0118] Fig. 2B shows a procedure carried out by the node 2 according to the example embodiment,

[0119] Fig. 3 illustrates a signal flow for a conditional handover, Figs. 4A to 4B illustrate different AI / ML positioning use cases,

[0120] Fig. 5 illustrates ML-enabled Features, in particular usage of functionality and model ID with applicable conditions information, and

[0121] Fig. 6 consisting of Figs. 6A and 6B illustrates exemplary steps for case 3a according to an example embodiment.

[0122] Detailed Description of example embodiments

[0123] In the following, description will be made to example embodiments. It is to be understood, however, that the description is given by way of example only, and that the described example embodiments are by no means to be understood as limiting the present invention thereto.

[0124] Before describing example embodiments, in the following, a technical context of example embodiments and problems of the prior art are discussed in some more detail.

[0125] Some example embodiments relate to a ML positioning use case. For example, positioning of a UE is required for a conditional handover (CHO), which is illustrated in Fig. 3.

[0126] The first processes B1-B9 are similar to the baseline handover of NR. Rel. 15 [TS 38.300]. A configured event triggers the UE to send a measurement report. Based on this report, the source node can prepare one or more target cells for the handover (CHO Request + CHO Request Acknowledge) and then sends an RRC Reconfiguration (CHO command) to the UE.

[0127] For baseline handover of NR Rel. 15, the UE will immediately access the target cell to complete the handover. Instead, for CHO, the UE will only access the target cell once an additional CHO execution condition expires (i.e., the HO preparation and execution phases are decoupled). The condition is configured by the source node in HO Command.

[0128] Once the UE completes the handover execution to the target cell (e.g., UE has sent RRC Reconfiguration Complete), the target cell sends to the source cell "Handover Success" indication. When receiving this indication from target cell, source cell stops its TX / RX to / from UE and starts data forwarding to target cell in step 18. Moreover, the source may release the CHO preparations in other target nodes / cells (which are no longer needed) when it receives "HO Success" indication.

[0129] The advantage of the CHO is that the HO command can be sent very early, when the UE is still safe in the source cell, without risking the access in the target cell and the stability of its radio link. That is conditional handover provides mobility robustness.

[0130] Moreover, some example embodiments relate to Rel-18 Study Item (SI) on Artificial Intelligence (AI) / Machine Learning (ML) for NR Air Interface [3GPP RP- 213599].

[0131] The SI aims at exploring the benefits of augmenting the air interface with features enabling support of AI / ML-based algorithms for enhanced performance and / or reduced complexity / overhead. This Si's target is to lay the foundation for future air-interface use cases leveraging AI / ML techniques. The initial set of use cases to be covered include CSI feedback enhancement (e.g., overhead reduction, improved accuracy, prediction), Beam management (e.g., beam prediction in time, and / or spatial domain for overhead and latency reduction, beam selection accuracy improvement), positioning accuracy enhancements. For those use cases the benefits shall be evaluated (utilizing developed methodology and defined KPIs) and potential impact on the specifications shall be assessed including PHY layer aspects, protocol aspects. This SI started in May RAN1 / RAN2 meeting in 3GPP.

[0132] One of the key expected outcomes of the SI is "The AI / ML approaches for the selected sub use cases need to be diverse enough to support various requirements on the gNB-UE collaboration levels." It must be noted that in the WI phase of "AI / ML for air interface", additionally other use cases might also be addressed. Starting from Release 18, it is very likely companies will propose a large variety of use cases and applications on ML in the gNB and UE. The goal is to explore the benefits of augmenting the air-interface with features enabling improved support of AI / ML-based algorithms for enhanced performance and / or reduced complexity / overhead. The enhanced performance here depends on the considered use cases and could be, e.g., improved throughput, robustness, accuracy, or reliability, etc. The goal is that sufficient use cases will be considered to enable the identification of a common AI / ML framework, including functional requirements of AI / ML architecture, which could be used in subsequent projects. The study should also identify areas where AI / ML could improve the performance of air-interface functions. Specification impact will be assessed in order to improve the overall understanding of what would be required to enable AI / ML techniques for the air interface.

[0133] In the following, some AIML positioning use cases are described.

[0134] Fig. 4A illustrates Case 1 - UE-based positioning with UE-side model, (a) direct AI / ML (b) or AI / ML assisted positioning. In this case, the model deployment is done on the UE-side. Here, the direct AI / ML and AI / ML assisted positioning subuse cases can be deployed to get as output the horizontal position or intermediate feature, respectively. For both cases, the parameter measurement used as input is done in the UE (downlink positioning), and the final position estimation is also on the UE-side. An illustration indicating the reference signal used to generate measurements and the entities involved in each case are presented in Fig. 4A.

[0135] Fig. 4B illustrates Case 2a - UE-assisted / LMF-based positioning with UE-side model, AI / ML assisted positioning. In this case, the position estimation is done on the LMF-side. However, the model is deployed on the UE-side to perform the AI / ML assisted positioning method, which estimates the intermediate feature. Here, the measurement is done in the UE (downlink positioning). This intermediate feature is reported to the LMF to calculate the position. An illustration of this case is represented in Fig. 4B. Fig. 4C illustrates Case 2b - UE-assisted / LMF-based positioning with LMF-side model, direct AI / ML positioning. The difference of this case with Case 2a is that the model is deployed on the LMF-side to enable a direct AI / ML positioning method. As in the previous case, the measurement is done in the UE (downlink positioning). An illustration of this case is represented in Fig. 4C.

[0136] Fig. 4D illustrates Case 3a - NG-RAN node assisted positioning with gNB-side model, AI / ML assisted positioning. The difference between this case and Case 2a is that the model is deployed in the gNB-side to enable an AI / ML assisted positioning method. Another difference is that the measurement is done in the gNB (uplink positioning). Here, the output is an intermediate feature that should be reported to the LMF. Finally, the LMF performs the position estimation. An illustration of this case is represented in Fig. 4D.

[0137] Fig. 4E illustrates Case 3b - NG-RAN node assisted positioning with LMF-side model, direct AI / ML positioning. In this case, the model output is the horizontal position, which is estimated in the LMF using measurements already obtained in the gNB (uplink positioning) and model deployed in the LMF using a direct AI / ML positioning. An illustration of this case is represented in Fig. 4E.

[0138] In the following some terminologies as used herein are defined based on RANI agreements on the list of terminologies used for AI / ML (source Rl-2304148).

[0139] In the following, applicable conditions for positioning use case are listed, based on

[0140] Nokia contribution Rl-2304680:

[0141] RAN1#112 has discussed the following FL proposal on applicable conditions ([FL-4] Proposal 5-8i): At least for UE-side models and UE-part of two-sided models, it is proposed to study how to define and study a (set of) applicable conditions for functionalities / [models], wherein applicable conditions may be used to enable development of scenario / configuration / [site]-specific models [and, if needed, report the models' applicability to the Network]. - whether and how to define performance targets (possibly as a part of applicable conditions) for functionality / [models], and

[0142] - whether and how UE reports a (set of) applicable conditions for supported functionalities (and if needed, for supported models) and / or supported set of functionalities.

[0143] As the FL proposal mentions, RANI shall first identify the applicable conditions for supported functionality / functionalities of a given sub-use case (ML-enabled feature). In Functionality identification and functionality-based LCM, knowing the UE conditions (including parameters / configurations) is required at the network as the first step prior to any other step, as this shall reveal the background conditions when using ML models for supporting a given ML-enabled feature. These applicable conditions may be depended on different sub-use cases, but we expect that at least a common set of applicable conditions (i.e., the definitions of the parameters are common, but the parameter values might be different depending on the Functionality or Feature) can be derived across all sub-use cases that are under discussion in Rel-18.

[0144] Fig. 5 illustrates ML-enabled Features, in particular usage of functionality and model ID with applicable conditions information. That is, in Fig. 5, the potential use of the applicable conditions is shown. Each Functionality can be identified by a subset of applicable conditions, which does not prevent individual applicable conditions to be re-used in different sets and / or different Functionalities.

[0145] In the following, the technical problem to be solved by some example embodiments is described.

[0146] To set a background for the technical problems to be solved, it is noted that from a ML positioning use-case perspective, the ML model can be deployed on the UE- side, gNB-side or LMF-side. Depending on the underlying case, either the DL PRS or the UL SRS is used as a reference signal for positioning measurements. Furthermore, a "positioning measurement" or an "intermediate feature" is used (within UE or network) to calculate the final position of the UE.

[0147] When a UE undergoes mobility in a NR network it physically moves. This movement requires the UE to be served by different cells / TRPs. To support positioning, the LMF has to prepare the gNB(s)

[0148] - to configure radio resources towards the UE that allows the UE to measure a set of DL PRS for a given set of candidate cell(s) or TRP(s) and report the measurements, and

[0149] - to configure radio resources towards the UE that allows the UE to transmit UL SRS which are measured by a given set of candidate cell(s) or TRP(s) and report the measurements. The LMF can prepare N such TRP(s) which may be spanning across more than one gNB.

[0150] In a practical deployment, specifically for Case 3a (described above by referring to Fig. 4D), for ML based positioning, the different gNB(s) / TRP(s) participating in the location determination may be hosting one or more ML model(s) or ML function(s). Due to practical constraints the different gNB(s) may host different ML model(s) or ML function(s) which produce different sets of "intermediate features". While this is the general mode of operation and is highly desirable from practical deployments this presents some practical issues for ML based positioning:

[0151] Issue 1 : Due to the mismatch in ML functionality between the source and neighbour cell(s) - the notion of an "intermediate feature" may not be seamless across TRP(s) or gNB(s). This may impact ML positioning latency when the requirements demand very high accuracy and low latency (e.g., 20 msec, for XR / VR services with cm level accuracy).

[0152] Issue 2: Aspect of harmonization of ML functionality between the source and neighbour cell(s) for ML positioning use case - for example how to align the differences of functionality (of ML configurations) between the source and target cell ML models that need to be resolved.

[0153] Some example embodiments aim to provide solutions for the above issues and define the signalling procedures suitable to handle the above.

[0154] In the following, a general overview of some example embodiments is described by referring to Figs. 1A, IB, 2A and 2B.

[0155] Fig. 1A shows an LMF 1 according to the present example embodiment. The LMF 1 is an example for an apparatus, which could be or be a part of a network control element carrying out a location management function, for example. A procedure carried out by the LMF 1 is illustrated in Fig. IB. The LMF 1 shown in Fig. 1A comprises at least one processor 11 and at least one memory 12 storing instructions that, when executed by the at least one processor 11, cause the apparatus to: determine a positioning configuration request based on machine learning related capabilities of a plurality of network nodes (e.g., gNB, TR.P) capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which indicates at least a positioning method to be applied to estimate a position of a user equipment (Sil in Fig. IB), send the positioning configuration request to at least one network node of the plurality of network nodes (S12 in Fig. IB), receive at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration (S13 in Fig. IB), and configure a positioning configuration to the plurality of network nodes based on the responses received from the network nodes (S14 in Fig. IB).

[0156] Fig. 2A shows a node 2 according to the present example embodiment. The node 2 is an example for an apparatus, which could be or be a part of a network node such as serving node or a neighbour or assisting node. For example the node may be a gNB or a TR.P. A procedure carried out by the node 2 is illustrated in Fig. 2B. The LMF 2 shown in Fig. 2A comprises at least one processor 21 and at least one memory 22 storing instructions that, when executed by the at least one processor 21, cause the apparatus to: receive a positioning configuration request from a network control element (e.g., LMF 1), the at least one positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment (S21 in Fig. 2B), select at least one positioning configuration of the positioning configuration request based on machine learning related capabilities of the apparatus, (S22 in Fig. 2B), send a response to the network control element including the at least one selected positioning configuration (S23 in Fig. 2B), receive information concerning a positioning configuration to be applied (S24 in Fig. 2B), and perform positioning related measurements by using a machine learning model based on the positioning configuration to be applied (S25 in Fig. 2B). Thus, according to example embodiments, the network control element such as the LMF 1 can align ML functionality between the different network nodes, which are applied for positioning a UE. Hence, no mismatch in ML functionality can occur between the different nodes.

[0157] The apparatuses 1 and 2 shown in Figs. 1A and 2A may comprise more components than described above, and may further comprise I / O units 13, 23, which are capable of transmitting to and receiving from other network elements.

[0158] The LMF 1 may determine the ML-related capabilities of a network node of the plurality of network nodes by requesting ML-related capabilities from the network node and receiving a response from the network node including the information concerning ML-related capabilities. This may be carried out when a network node is added or updated, for example.

[0159] The network nodes may be AI / ML enabled.

[0160] The network nodes may also indicate their legacy capabilities (combining of intermediate features can happen across ML and non-ML enabled component), for example.

[0161] The LMF 1 may select network nodes to be involved in positioning the user equipment ML-related capabilities based on the ML-related capabilities of the network node as the plurality of network nodes to which the positioning configuration request is sent.

[0162] The LMF 1 may configure all network nodes of the plurality of network nodes with the same positioning configuration or may configure the network nodes individually with an individual positioning configuration. For example, the LMF 1 may apply different positioning configurations for the network nodes due to different ML capabilities or non-ML capabilities of the network nodes, while ensuring that results achieved by the network nodes based on the different positioning configuration can be combined in order to estimate the position of the UE. In the following, some example embodiments are described more detail.

[0163] For Case 3a - NG-RAN assisted AIML positioning with intermediate feature reported to LMF (via. LPP) the following novel aspects are defined by some example embodiments:

[0164] Alignment of ML functionality between serving and neighbour cell(s) Negotiating the format of intermediate features produced by each gNB Alignment of UL SRS based on the alignment of ML functionality between serving and neighbour cell(s)

[0165] Combining of "intermediate features" at the LMF

[0166] A more detailed signal flow is shown in Fig. 6, which consists of Figs. 6A and 6B. In particular, signal exchange between an LMF, an AIML enabled serving node, AIML enabled neighbour node(s) and an UE is shown. "AIML enabled" means that the corresponding node applies AIML (or ML) for its operation, and in more detail, applies an ML model or ML functionality for carrying out positioning related measurements or the like. The AIML enabled serving node is referred to in the following also as "serving node" and can be a gNB, for example. The AIML enabled neighbour node(s) are referred to in the following also as "neighbour node(s)" and can be a TPR, for example.

[0167] In process Al, the LMF sends a NRRPa TRP information request to the serving node, including an ML positioning information request (ML positioning info req). In process A2, the LMF sends a similar request to the neighbour node(s). In process A3, the serving node responds with a NRRPa TRP information response including an ML positioning information response (ML positioning info rsp). In process A4, the LMF receives a similar response from the neighbour node(s).

[0168] By the ML positioning information request sent in processes Al and A2, the LMF requests ML-related capabilities (i.e., applicable conditions) of each gNB / TRP (i.e., from the serving node and the neightbor node(s)), including at least:

[0169] • Supported intermediate feature(s) (for two-step positioning), i.e., inputs, output of their ML model: o LOS / NLOS indication o ToA, path phase o etc.,

[0170] • Supported UL SRS configurations (or UL SRS characteristics) for their measurement, i.e., input of their ML model : o UL SRS bandwidth, periodicity o Comb size, comb offset o etc., o CIR o Power delay profile (PDP) o Delay profile (DP)

[0171] • ML model complexity (so as to account for any delays in positioning) o Model size o Processing latency o etc.,

[0172] In processes A3 and A4, the nodes (gNB / TRPs) respond to LMF either by indicating:

[0173] • Requested information (process 1 and process 2)

[0174] • No ML-related capability o Indicate its legacy capabilities (combining of intermediate features can happen across ML and non-ML enabled component)

[0175] •

[0176] The processes A3 and A4 may comprise further some combinations of the capabilities that are supported by the nodes. An example is shown in the table below that lists capability combinations published by the nodes to the LMF.

[0177] In process A5, the LMF updates ML positioning configuration combinations. That is, in process A5, based on the received information, LMF determines a positioning method and UL SRS configuration accounting for different ML capabilities at different candidate gNBs / TRPs for positioning a target UE. For example, from the above table, LMF may choose capability combination N to configure the nodes for ML positioning feature.

[0178] In process A6, the LMF performs an LPP capability transfer, in which the LMF obtains UE capabilities from the UE.

[0179] In process A7, LMF refines its determination by taking the UE capabilities (e.g., supported UL SRS bandwidth) into account, which it obtains in process A6 via conventional LPP capability information exchange procedure. In case the UE is restricted with some capabilities further than a given capability combination, the LMF scales the capability combinations in its database accordingly considering a specific UE's capability.

[0180] In processes A8 and A9, LMF provides the configuration it determined to the nodes (serving node and neighbour node(s)) that it would like to involve in positioning the target UE. The LMF provides a list of configuration choices towards the nodes and collects the response in process A10 and process All. That is, in process A8, the LMF sends an NRPPa POS Information Request including an ML positioning configuration request (ML positioning config req) to the serving node (gNB), and in process A9, the LMF sends a similar request to the serving node(s). In process A10, the LMF receives an NRPPa POS Information Response including an ML positioning configuration response (ML positioning config rsp) from the serving node (gNB), and in process All, the LMF receives a similar response from the serving node(s).

[0181] In an example implementation, the LMF generates 3 configurations as follows:

[0182] Configuration 1 : LOS / NLOS indication=YES, ToA=YES, SRS BW = 100 MHz, SRS periodicity=20 ms, Comb size=l, Comb offset=l, CNN with 16 layers

[0183] Configuration 2: LOS / NLOS indication=YES, ToA=NO, SRS BW=200 MHz, SRS periodicity=20 ms, Comb size=2, Comb offset=2, RNN with 16 layers

[0184] Configuration 3: LOS / NLOS indication=YES, ToA=NO, SRS BW=400 MHz, SRS periodicity=20 ms, Comb size=4, Comb offset=4, ResNet with 32 layers

[0185] It is noted that a definition such as "ToA=YES" means that reporting of ToA is supported, whereas a definition "ToA=NO" means that this is not supported.

[0186] In processes A10 and All, the nodes may respond with the preferred configuration choices. E.g., the serving node may respond that it supports Configuration 1 and Configuration 2 while an assisting one (i.e., one of the neighbour nodes) only supports Configuration 2 and Configuration 3. In this case the LMF will configure only Configuration ID 2.

[0187] In process A12, the gNB (the serving node) determines UL SRS resources.

[0188] In process A13, the LMF updates ML positioning configuration combinations, and in process A14, the LMF activates a given ML positioning configuration. In process A15, the LMF sends a NRPPa POS Activation Request including the configuration ID to the serving node and the neighbour node(s). The configuration ID indicates the configuration selected by the LMF.

[0189] For example, in process 15 after combining the responses in process 13 and 14 the LMF will configure only Configuration ID 2 (i.e., configuration 2 as specified above).

[0190] In process A16, the serving node instructs the UE to activate UE SRS transmission.

[0191] In process A17, the serving node sends a NRPPa POS Activation Response to the LMF. In process A18, the LMF sends a NRPPA Measurement Request to the serving node, and in process A19, the LMF sends a NRPPA Measurement Request to the neighbour node(s). Thereafter, the serving node and the neighbour node(s) perform US SRS measurements in process 20, and in processes A21 and A22, they send the results in NRPPA Measurement Responses to the LMF.

[0192] Thus, in processes A16 - 2A2, the UL SRS transmission is generated by the serving node and transmitted towards the UE. The LMF requests NRPPa measurement of UL SRS towards both serving and assisting node.

[0193] In process A23, the LMF combines intermediate features and derives the final location.

[0194] As the positioning configuration is aligned between different anchor nodes (e.g., including serving and neighbouring nodes (e.g. gNBs or TRPS) as mentioned above), the LMF is able to combine the intermediate features across the nodes using an algorithm. The algorithm may simply combine the samples across serving and assisted nodes or may choose a weighting approach e.g., consider the serving gNB's measurements / features with 50% weight and 50% of the neighbouring nodes.

[0195] In mobility scenarios, while the UE moves, the serving gNB would change over time, and LMF may adjust weights measurements or intermediate features collected from different gNBs accordingly, e.g., by acquiring the serving cell information from AMF. Overall, due to the combining of the intermediate features the overall positioning feature can consistently work during the HO as the UE moves away from the serving cells.

[0196] It is noted that the processes Al to A23 do not have to be carried out in the order described above. In particular, it is noted that the processes Al to A5 related to obtaining ML-related capabilities from the serving node and the neighbour node(s) can be carried out separately. Once the LMF is aware of the ML-related capabilities of the nodes, it is not necessary to request the ML-related capabilities again. For example, the processes may only be carried out when a new node is established or when a node is updated.

[0197] Thus, summarizing, according to some example embodiments, the following is provided:

[0198] According to some embodiments, the LMF (Location Management Function) requests the machine learning capabilities from a serving and / or neighbouring node(s). The LMF receives the capability response, and determines a list of positioning configuration. The LMF then sends the list of positioning configuration to the serving and / or neighbouring node(s) to position the target UE, and receives the response from nodes with selected / preferred configuration from the list of configuration. Based on this, the LMF configures a positioning configuration to serving and / or neighbouring node(s).

[0199] According to some embodiments, the node (which can be a serving node or a neighbouring / assisting node) receives a machine learning capability request from the LMF, and sends a response of machine learning capability(ies) to the LMF. Thereafter, the node receives a list of positioning configurations, and sends the selected / preferred positioning configuration from the list of positioning configurations to the LMF. After this, the node receives a positioning configuration.

[0200] According to some embodiments, the LMF may request the machine learning capabilities of each gNB / TRP. The machine learning capabilities may include: supported intermediate feature(s) (for two-step positioning), i.e., inputs, output of their ML model such as LOS / NLOS indication, ToA, path phase etc.; supported UL SRS configurations for their measurement, i.e., input of their ML model such as UL SRS bandwidth, periodicity, Comb size, comb offset etc.; and ML model complexity (so as to account for any delays in positioning) such as Model size, Processing latency etc.

[0201] The capability response may consist of the requested information if present.

[0202] The LMF may determine the positioning configuration based on the machine learning capabilities of different candidate gNBs / TRPs for positioning a target UE.

[0203] The positioning configuration may also take into account UE capabilities e.g., supported UL SRS bandwidth, obtained via conventional LPP capability information exchange procedure.

[0204] The LMF may request NRPPa measurement of UL SRS towards both serving and neighbour (assisting) node.

[0205] The serving and neighbour (assisting) nodes may be AI / ML enabled.

[0206] Additionally, the capability response may also consist of some combination of the capabilities that are supported by the nodes.

[0207] The above-described example embodiments are only examples and may be modified.

[0208] For example, according to some example embodiments described above, the LMF 1 may configure all network nodes of the plurality of network nodes with the same positioning configuration. However, embodiments are not limited to this. The LMF 1 may also configure the network nodes individually with an individual positioning configuration. For example, while LMF 1 asks / configures the same output format for the ML-based estimation at the network nodes (gNBs, TRPs), it may configure (or ask) different types of input for the estimation at different network nodes, e.g., depending on an input capability (e.g., measurement bandwidth) of the network node.

[0209] Also in this case, it can be ensured to harmonize or at least combine multiple network nodes supporting ML, for a single estimation task (i.e., a single target UE) at least.

[0210] Names of network elements, protocols, and methods are based on current standards. In other versions or other technologies, the names of these network elements and / or protocols and / or methods may be different, as long as they provide a corresponding functionality.

[0211] In general, example embodiments may be implemented by computer software stored in the memory (memory resources, memory circuitry) 12, 22 and executable by the processor (processing resources, processing circuitry) 11, 21 or by hardware, or by a combination of software and / or firmware and hardware.

[0212] The terms "connected," "coupled," or any variant thereof, mean any connection or coupling, either direct or indirect, between two or more elements, and may encompass the presence of one or more intermediate elements between two elements that are "connected" or "coupled" together. The coupling or connection between the elements can be physical, logical, or a combination thereof. As employed herein two elements may be considered to be "connected" or "coupled" together by the use of one or more wires, cables and printed electrical connections, as well as by the use of electromagnetic energy, such as electromagnetic energy having wavelengths in the radio frequency region, the microwave region and the optical (both visible and invisible) region, as non-limiting examples.

[0213] The memory (memory resources, memory circuitry) 12, 22 may be of any type suitable to the local technical environment and may be implemented using any suitable data storage technology, such as semiconductor based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory, and non-transitory computer-readable media. The processor (processing resources, processing circuitry) 11, 21 may be of any type suitable to the local technical environment, and may include one or more of general purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) and processors based on a multi core processor architecture, as non-limiting examples.

[0214] Further, as used in this application, the term "circuitry" may refer to one or more or all of the following:

[0215] (a) hardware-only circuit implementations (such as implementations in only analog and / or digital circuitry) and

[0216] (b) combinations of hardware circuits and software, such as (as applicable):

[0217] (i) a combination of analog and / or digital hardware circuit(s) with software / firmware and

[0218] (ii) any portions of hardware processor(s) with software (including digital signal processor(s)), software, and memory(ies) that work together to cause an apparatus, such as a mobile phone or server, to perform various functions) and

[0219] (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessor(s), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation.

[0220] This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in a server, a cellular network device, or other computing or network device.

[0221] The term "non-transitory", as used herein, is a limitation of the medium itself (i.e., tangible, not a signal) as opposed to a limitation on data storage persistency (e.g., RAM vs. ROM). It is noted that, as used herein, "at least one of the following: " and "at least one of " and similar wording, where the list of two or more elements are joined by "and" or "or", mean at least any one of the elements, or at least any two or more of the elements, or at least all the elements.

[0222] It is to be understood that the various example embodiments of the subject disclosure are illustrative and non-limiting and are not intended to be construed as limiting. Various modifications and applications may be apparent to those skilled in the art without departing from the spirit and scope of the various example embodiments of the subject disclosure.

Claims

CLAIMS1. An apparatus comprising: at least one processor, and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: determine a positioning configuration request based on machine learning related capabilities of a plurality of network nodes capable of performing positioning related measurements, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, send the positioning configuration request to at least one network node of the plurality of network nodes, receive at least one response from the at least one network node of the plurality of network nodes, wherein the at least one response includes at least one selected positioning configuration, and configure at least one positioning configuration to the at least one network node of the plurality of network nodes based on the at least one response received from the at least one network node of the plurality of network nodes.

2. The apparatus according to claim 1, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: determine machine learning related capabilities of a network node of the plurality of network nodes by requesting ,from the network node, information concerning its machine learning related capabilities and receiving a response from the network node including the information concerning the machine learning related capabilities.

3. The apparatus according to claim 1 or 2, wherein the machine learning related capabilities comprises information concerning a machine learning model and / or machine learning functionality applied by each network node of the plurality of network nodes.

4. The apparatus according to claim 3, wherein the information concerning the machine learning model and / or machine learning functionality comprises a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model and / or machine learning functionality, and / or information concerning computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

5. The apparatus according to claim 4, wherein the outputs of the machine learning model and / or machine learning functionality comprise at least one supported intermediate feature.

6. The apparatus according to any one of the claims 1 to 5, wherein the at least one positioning configuration indicated by the positioning configuration request further comprises at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

7. The apparatus according to claim 6, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: determine the positioning configuration request to the plurality of network nodes based on the machine learning related capabilities of the plurality of network nodes by aligning a machine learning model and / or machine learning functionality to be applied by the plurality of network nodes, aligning a format of intermediate features to be applied for the aligned machine learning model, and aligning uplink sounding reference signals configuration.

8. The apparatus according to any one of the claims 1 to 7, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: configure the positioning configuration to the at least one network node of the plurality of network nodes based on responses received from all network nodes of the plurality of network nodes by determining which positioning configuration is or which positioning configurations are supported by all network nodes of the plurality of network nodes.

9. The apparatus according to any one of the claims 1 to 8, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: configure all network nodes of the plurality of network nodes with the same positioning configuration or configure the network nodes of the plurality of network nodes individually with positioning configuration.

10. The apparatus according to any one of the claims 1 to 9, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: receive positioning related measurement results from the plurality of network nodes, and determine the position of the user equipment based on a combination of the received positioning related measurement results.

11. The apparatus according to claim 10, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: combine the intermediate features by applying weights to the positioning related measurement results.

12. The apparatus according to claim 11, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: adjust the weights of intermediate features received from the plurality of network nodes based on a movement of the user equipment.

13. The apparatus according to any one of the claims 10 to 12, wherein the positioning related measurement results comprise intermediate features.

14. An apparatus comprising: at least one processor, and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: receive a positioning configuration request from a network control element, the positioning configuration request indicating at least one positioning configuration, which includes at least a positioning method to be applied to estimate a position of a user equipment, select at least one positioning configuration of the positioning configuration request based on machine learning related capabilities of the apparatus, send a response to the network control element including the at least one selected positioning configuration, receive information concerning a positioning configuration to be applied, and perform positioning related measurements by using a machine learning model based on the positioning configuration to be applied.

15. The apparatus according to claim 14, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: receive a request from the network control element to provide information concerning machine learning related capabilities of the apparatus, and send a response to the network control element including the information concerning machine learning related capabilities.

16. The apparatus according to claim 14 or 15, wherein the machine learning related capabilities comprises information concerning a machine learning model and / or machine learning functionality applied by the apparatus.

17. The apparatus according to claim 16, wherein the information concerning the machine learning model and / or machine learning functionality comprises a type, characteristics, features and / or structure of the machine learning model and / or machine learning functionality, and / or inputs and / or outputs of the machine learning model, and / or information concerning the machine learning model computational complexity and / or interference latency of the machine learning model and / or machine learning functionality.

18. The apparatus according to claim 17, wherein the outputs of the machine learning model and / or machine learning functionality comprise at least one supported intermediate feature.

19. The apparatus according to any one of the claims 14 to 18, wherein the at least one positioning configuration indicated by the positioning configuration request further comprises at least one of a machine learning model and / or machine learning functionality to be applied, inputs and / or outputs of the machine learning model and / or machine learning functionality, and an uplink sounding reference signals configuration.

20. The apparatus according to any one of the claims 14 to 19, wherein the instructions stored by the at least one memory further cause the apparatus, when executed by the at least one processor, at least to: send a result of the positioning related measurements to the network control element.

Images