Updating machine learning models for positioning

Abstract

A method is disclosed, the method including: obtaining a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmitting information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the device; receiving at least one of a message indicating the updated machine learning model or the second training data; and transmitting the updated machine learning model to at least one or more second network nodes.

Description

[Technical Field]

[0001] The following exemplary embodiments relate to updating machine learning models for wireless communications and positioning. [Background technology]

[0002] Positioning techniques may be used to estimate the physical location of a device. It is desirable to improve positioning accuracy to more accurately estimate the location of a device. Summary of the Invention

[0003] The scope of protection sought for various exemplary embodiments is set forth in the independent claims. The exemplary embodiments and features described herein that do not fall within the scope of the independent claims, if any, are to be construed as examples useful for understanding the various embodiments.

[0004] According to one aspect, an apparatus is provided that includes at least one processor and at least one memory that stores instructions that, when executed by the at least one processor, cause the apparatus to at least: obtain a machine learning model for positioning, where the machine learning model is trained based on first training data associated with one or more first network nodes; transmit information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; receive at least one of a message indicating the updated machine learning model or the second training data; and transmit the updated machine learning model to at least one or more second network nodes.

[0005] According to another aspect, an apparatus is provided, the apparatus comprising: means for acquiring a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; means for transmitting information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; means for receiving at least one of a message indicating the updated machine learning model or the second training data; and means for transmitting the updated machine learning model to at least one or more second network nodes.

[0006] According to another aspect, a method is provided, the method including: obtaining, by an apparatus, a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmitting, by the apparatus, information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; receiving, by the apparatus, at least one of a message indicating the updated machine learning model or the second training data; and transmitting, by the apparatus, the updated machine learning model to at least one or more second network nodes.

[0007] According to another aspect, a computer program product is provided that includes instructions that, when executed by an apparatus, cause the apparatus to at least: obtain a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmit information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; receive at least one of a message indicating the updated machine learning model or the second training data; and transmit the updated machine learning model to at least one or more second network nodes.

[0008] According to another aspect, a computer-readable medium is provided that includes program instructions that, when executed by an apparatus, cause the apparatus to at least: obtain a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmit information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; receive at least one of a message indicating the updated machine learning model or the second training data; and transmit the updated machine learning model to at least one or more second network nodes.

[0009] According to another aspect, a non-transitory computer-readable medium is provided that includes program instructions that, when executed by an apparatus, cause the apparatus to at least: obtain a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmit information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide second training data for updating the machine learning model at the apparatus; receive at least one of a message indicating the updated machine learning model or the second training data; and transmit the updated machine learning model to at least one or more second network nodes.

[0010] According to another aspect, an apparatus is provided that includes at least one processor and at least one memory that stores instructions that, when executed by the at least one processor, cause the apparatus to: receive information including at least one of a machine learning model for positioning, a request to update the machine learning model at the apparatus based on second training data, or a request to provide second training data for updating the machine learning model, where the machine learning model has been trained based on the first training data associated with one or more first network nodes; obtain the second training data; and transmit at least one of a message indicating the updated machine learning model or the second training data.

[0011] According to another aspect, an apparatus is provided, comprising: means for receiving information including at least one of a machine learning model for positioning, a request to update the machine learning model at the apparatus based on second training data, or a request to provide second training data for updating the machine learning model, wherein the machine learning model has been trained based on first training data associated with one or more first network nodes; means for obtaining the second training data; and means for transmitting at least one of a message indicating the updated machine learning model or the second training data.

[0012] According to another aspect, a method is provided, the method including: receiving, by a device, information including at least one of a machine learning model for positioning, a request to update the machine learning model at the device based on second training data, or a request to provide second training data for updating the machine learning model, where the machine learning model has been trained based on first training data associated with one or more first network nodes; obtaining, by the device, the second training data; and transmitting, by the device, at least one of a message indicating the updated machine learning model or the second training data.

[0013] According to another aspect, a computer program product is provided that includes instructions that, when executed by an apparatus, cause the apparatus to receive information including at least one of a machine learning model for positioning, a request to update the machine learning model at the apparatus based on second training data, or a request to provide second training data for updating the machine learning model, where the machine learning model has been trained based on the first training data associated with one or more first network nodes; obtain the second training data; and transmit at least one of a message indicating the updated machine learning model or the second training data.

[0014] According to another aspect, a computer-readable medium is provided that includes program instructions that, when executed by an apparatus, cause the apparatus to receive information including at least one of a machine learning model for positioning, a request to update the machine learning model at the apparatus based on second training data, or a request to provide second training data for updating the machine learning model, where the machine learning model has been trained based on the first training data associated with one or more first network nodes; obtain the second training data; and transmit at least one of a message indicating the updated machine learning model or the second training data.

[0015] According to another aspect, a non-transitory computer-readable medium is provided that includes program instructions that, when executed by an apparatus, cause the apparatus to receive information including at least one of a machine learning model for positioning, a request to update the machine learning model at the apparatus based on second training data, or a request to provide second training data for updating the machine learning model, where the machine learning model has been trained based on the first training data associated with one or more first network nodes; obtain the second training data; and transmit at least one of a message indicating the updated machine learning model or the second training data.

[0016] In the following, various exemplary embodiments will be described in more detail with reference to the accompanying drawings. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 illustrates an example of a cellular communication network. [Figure 2] FIG. 1 illustrates an example of a positioning scenario. [Figure 3] 1 is a signal transmission diagram in accordance with an exemplary embodiment; [Figure 4]1 is a signal transmission diagram in accordance with an exemplary embodiment; [Figure 5] 1 is a flowchart according to an example embodiment. [Figure 6] 1 is a flowchart according to an example embodiment. [Figure 7] FIG. 1 illustrates an example of an apparatus. [Figure 8] FIG. 1 illustrates an example of an apparatus. [Figure 9] FIG. 1 illustrates an example of an apparatus. [Figure 10] FIG. 1 illustrates an example of an artificial neural network. [Figure 11] FIG. 2 illustrates an example of a computing node. DETAILED DESCRIPTION OF THE INVENTION

[0018] The following embodiments are illustrative. Although this specification may refer to "an," "one," or "some" embodiments in several places in the text, this does not necessarily mean that each reference is to the same embodiment or that a particular feature only applies to a single embodiment. Single features of different embodiments may be combined to provide other embodiments.

[0019] In the following, different exemplary embodiments are described using radio access architectures based on Long Term Evolution Advanced (LTE-A), New Radio (NR, 5G), beyond 5G, or sixth generation (6G) as examples of access architectures to which the exemplary embodiments may be applied, without, however, limiting the exemplary embodiments to such architectures. It will be clear to those skilled in the art that the exemplary embodiments may also be applied to other types of communication networks with appropriate means by appropriately adjusting parameters and procedures. Some examples of other options for suitable systems may be a universal mobile telecommunications system (UMTS) radio access network (UTRAN or E-UTRAN), Long Term Evolution (LTE, essentially the same as E-UTRAN), a wireless local area network (WLAN or Wi-Fi), Worldwide Interoperability for Microwave Access (WiMAX), Bluetooth, personal communications services (PCS), ZigBee, wideband code division multiple access (WCDMA), systems using ultra-wideband (UWB) technology, sensor networks, mobile ad-hoc networks (MANET), and Internet Protocol multimedia subsystem (IMS), or any combination thereof.

[0020] 1 shows an example of a simplified system architecture showing several elements and functional entities, all of which are logical units whose implementation may differ from that shown. The connections shown in FIG. 1 are logical connections; the actual physical connections may differ. It will be apparent to one skilled in the art that the system may have functions and structures other than those shown in FIG. 1.

[0021] The exemplary embodiment is however not limited to the system given as an example, and a person skilled in the art may apply the solution to other communication systems that have the required characteristics.

[0022] The example of FIG. 1 shows a portion of an exemplary radio access network.

[0023] FIG. 1 illustrates an access node (AN) 104, such as an evolved Node B (eNB, abbreviated as eNodeB) or next generation Node B (gNB, abbreviated as gNodeB), providing a radio cell, and user devices 100 and 102 configured to be wirelessly connected with one or more communication channels within the radio cell. The physical link from the user device to the access node may be referred to as an uplink (UL) or reverse link, and the physical link from the access node to the user device may be referred to as a downlink (DL) or forward link. A user device may communicate directly with another user device via sidelink (SL) communication. It should be appreciated that the access node or its functionality may be performed by using any node, host, server, or access point, or other entity suitable for such use.

[0024] A communication system may include two or more access nodes, in which case the access nodes may be configured to communicate with each other through links (wired or wireless) designed for that purpose. These links may be used for signal transmission and also for routing data from one access node to another. An access node may be a computing device configured to control radio resources of a communication system to which the access node is coupled. An access node may also be referred to as a base station, base transceiver station (BTS), access point, or any other type of interfacing device, including a relay station, capable of operating in a wireless environment. An access node may include or be coupled to a transceiver. From the access node's transceiver, a connection may be provided to an antenna unit that establishes a bidirectional radio link to a user device. The antenna unit may include multiple antennas or antenna elements. The access node may be further connected to a core network 110 (core network (CN) or next generation core (NGC)). Depending on the deployed technology, the access node may be connected on the CN side to a serving gateway (S-GW, routes and forwards user data packets), packet data network gateway (P-GW) for providing user device connectivity to external packet data networks, user plane function (UPF), mobility management entity (MME), or access and mobility management function (AMF), etc.

[0025] With regard to positioning, the service-based architecture (core network) may comprise an AMF 111 and a location management function (LMF) 112. The AMF may provide location information for call processing, policy, and billing to other network functions in the core network and other entities that request positioning of terminal devices. The AMF may receive and manage location requests from several sources: mobile-originated location requests (MO-LR) from user devices and mobile-terminated location requests (MT-LR) from other functions in the core network or from other network elements. The AMF may select an LMF for a given request and use its positioning service to trigger a positioning session. The LMF may then perform positioning upon receiving such a request from the AMF. The LMF may manage resources and timing of positioning activities. The LMF may use the Namf_Communication service over the NL1 interface to request positioning of the user device from one or more access nodes, or the LMF may communicate with the user device through N1 for UE-based or UE-assisted positioning. The positioning may include an estimation of the location, and the LMF may also estimate the movement or accuracy of the location information when requested. In relation to the connection, the AMF is between the access node and the LMF and may therefore be closer to the access node than the LMF.

[0026] A user device represents one type of device to which resources over the air interface may be allocated and assigned, and therefore any features described herein with respect to a user device may also be implemented using a corresponding device such as a relay node.

[0027] An example of such a relay node may be a Layer 3 relay (self-backhauling relay) for an access node. A self-backhauling relay node may also be called an integrated access and backhaul (IAB) node. An IAB node may comprise two logical parts: a mobile termination (MT) part that handles the backhaul link (i.e., the link between the IAB node and a donor node, also known as a parent node), and a distributed unit (DU) part that handles the access link (i.e., the child link between the IAB node and a user device and / or between the IAB node and other IAB nodes (multi-hop scenarios).

[0028] Another example of such a relay node may be a Layer 1 relay, called a repeater, which may amplify signals received from an access node and forward the signals to a user device and / or may amplify signals received from a user device and forward the signals to the access node.

[0029] A user device may be referred to as a subscriber unit, mobile station, remote terminal, access terminal, user terminal, terminal device, or user equipment (UE), to name a few. A user device may refer to a portable computing device, including a wireless mobile communication device operating with or without a subscriber identification module (SIM), including, but not limited to, the following types of devices: mobile station (mobile phone), smartphone, personal digital assistant (PDA), handset, device using a wireless modem (such as an alarm or measurement device), laptop and / or touchscreen computer, tablet, game console, notebook, multimedia device, reduced capability (RedCap) device, wireless sensor device, or any device integrated into a vehicle.

[0030] It should be appreciated that a user device may be almost exclusively an uplink-only device, an example of which may be a camera or video camera that loads images or video clips onto the network. A user device may also be a device capable of operating in an Internet of Things (IoT) network, a scenario in which objects may have the ability to transfer data over a network without requiring human-to-human or human-to-computer interaction. A user device may utilize the cloud. In some applications, a user device may comprise a small, portable, or wearable device with wireless components (such as a watch, earphones, or glasses), and computation may be performed in the cloud or in another user device. A user device (or, in some exemplary embodiments, a Layer 3 relay node) may be configured to perform one or more of the user equipment functions.

[0031] The various techniques described herein may be applied to cyber-physical systems (CPSs)—systems of cooperating computational elements that control physical entities. CPSs may enable the execution and utilization of vast numbers of interconnected ICT devices (sensors, actuators, processors, microcontrollers, etc.) embedded in physical objects at different locations. Mobile cyber-physical systems, in which the physical systems in question may have inherent mobility, are a subcategory of cyber-physical systems. Examples of mobile physical systems include mobile robotics and electronics transported by humans or animals.

[0032] Furthermore, although the device is shown as a single entity, it may be implemented as different units, processors and / or memory units (not all shown in FIG. 1).

[0033] 5G allows for many more base stations or nodes than LTE (the so-called small cell concept), including macro sites operating in cooperation with smaller stations using multiple input-multiple output (MIMO) antennas and employing various radio technologies depending on service needs, use cases, and / or available spectrum. 5G mobile communications may support a wide range of use cases and related applications, including video streaming, augmented reality, different methods of data sharing, and various forms of machine-type applications (such as massive machine-type communication (mMTC) including vehicle safety, different sensors, and real-time control). 5G may have multiple air interfaces, namely, sub-6 GHz, cm-wave, and mm-wave, and may also be integrated with existing legacy radio access technologies such as LTE. Integration with LTE may be performed, at least in early phases, as a system in which macro coverage may be provided by LTE and 5G air interface access may come from small cells through aggregation to LTE. In other words, 5G may support both inter-RAT interoperability (e.g., LTE-5G) and inter-RI interoperability (interoperability between air interfaces such as sub-6 GHz-cm wave-mm wave). One concept that may be used in 5G networks may be network slicing, in which multiple independent and dedicated virtual sub-networks (network instances) may be created within substantially the same infrastructure to run services with different requirements in terms of latency, reliability, throughput, and mobility.

[0034] The current architecture in LTE networks may be fully distributed in the radio and fully centralized in the core network. Low-latency applications and services in 5G may require content to be closer to the radio, which leads to local breakout and multi-access edge computing (MEC). 5G may enable analytics and knowledge generation to occur at the source of the data. This approach may require effective utilization of resources that may not be continuously connected to the network, such as laptops, smartphones, tablets, and sensors. MEC may provide a distributed computing environment for application and service hosting. MEC may have the ability to store and process content close to cellular subscribers for faster response times. Edge computing may cover a wide range of technologies, such as wireless sensor networks, mobile data acquisition, mobile signature analysis, collaborative distributed peer-to-peer ad hoc networking, and processing that can also be categorized as cloud / fog computing and grid / mesh computing, dew computing, mobile edge computing, cloudlets, distributed data storage and retrieval, autonomous self-healing networks, remote cloud services, augmented and virtual reality, data caching, Internet of Things (massive connectivity and / or latency critical), critical communications (autonomous vehicles, traffic safety, real-time analytics, time-critical control, healthcare applications).

[0035] The communications system may be able to communicate with or use services provided by one or more other networks 113, such as the public switched telephone network or the Internet. The communications network may be able to support the use of cloud services, e.g., at least part of the core network operations may be implemented as cloud services (this is illustrated in FIG. 1 by "cloud" 114). The communications system may comprise a central control entity or the like that provides facilities for networks of different operators to cooperate, e.g., in spectrum sharing.

[0036] The access node may be divided into a radio unit (RU) comprising a radio transceiver (TRX), i.e., a transmitter (Tx) and a receiver (Rx), one or more distributed units (DUs) 105 that may be used for so-called Layer 1 (L1) processing and real-time Layer 2 (L2) processing, and a central unit (CU) 108 (also known as a centralized unit) that may be used for non-real-time L2 and Layer 3 (L3) processing. The CU 108 may be connected to one or more DUs 105, for example, via an F1 interface. Such division may enable centralization of the CU with respect to the cell site and the DU, while the DU may be more decentralized and remain at the cell site. The CU and DU together may be referred to as baseband or baseband unit (BBU). The CU and DU may be included in a radio access point (RAP).

[0037] The CU 108 may be defined as a logical node that hosts higher layer protocols, such as radio resource control (RRC), service data adaptation protocol (SDAP), and / or packet data convergence protocol (PDCP), of an access node. The DU 105 may be defined as a logical node that hosts radio link control (RLC), medium access control (MAC), and / or physical (PHY) layers of an access node. The operation of the DU may be controlled at least in part by the CU. The CU may comprise a control plane (CU-CP), which may be defined as a logical node that hosts the RRC and control plane portions of the CU's PDCP protocol for the access node. The CU may further comprise a user plane (CU-UP), which may be defined as a logical node that hosts the user plane portions of the CU's PDCP and SDAP protocols for the access node.

[0038] A cloud computing platform may be used to execute the CU 108 and / or the DU 105. The CU may execute on the cloud computing platform, which may be referred to as a virtualized CU (vCU). In addition to the vCU, there may be a virtualized DU (vDU) executing on the cloud computing platform. Furthermore, a combination may exist, where the DU may use a so-called bare-metal solution, e.g., an application-specific integrated circuit (ASIC) or a customer-specific standard product (CSSP) system-on-a-chip (SoC) solution. It should also be understood that the distribution of functionality between the above-mentioned access node units or between different core network operations and access node operations may differ.

[0039] An edge cloud may be joined to a radio access network (RAN) by utilizing network function virtualization (NFV) and software defined networking (SDN). Using an edge cloud may mean that access node operations are at least partially implemented in a server, host, or node operatively coupled to a remote radio head (RRH) or radio unit (RU), or an access node comprising a radio portion. It is also possible that node operations may be distributed among multiple servers, nodes, or hosts. Application of a Cloud RAN architecture allows RAN real-time functions to be implemented on the RAN side (e.g., in the DU 105) and non-real-time functions to be implemented in a centralized manner (e.g., in the CU 108).

[0040] It should also be understood that the distribution of functions between core network operations and access node operations may differ from or even not exist in LTE. Some other technological advances that may be used include big data and all-IP, which may change the way networks are built and managed. 5G (or New Radio, NR) networks may be designed to support multiple hierarchies in which MEC servers may be located between the core and access nodes. It should be recognized that MEC may also be applied in 4G networks.

[0041] 5G may utilize non-terrestrial communications, such as satellite communications, to augment or complement 5G service coverage by providing backhauling. Possible use cases may include providing service continuity for machine-to-machine (M2M) or Internet of Things (IoT) devices or for passengers aboard vehicles, or ensuring service availability for critical communications and future rail, maritime, and aviation communications. Satellite communications may utilize geostationary earth orbit (GEO) satellite systems, as well as low earth orbit (LEO) satellite systems, particularly mega-constellations (systems in which hundreds of (nano)satellites are deployed). A given satellite 106 in a mega-constellation may cover several satellite-enabled network entities, creating on-ground cells. On-ground cells may be created through on-ground relay nodes or by access nodes 104 located on-ground or within the satellite.

[0042] 6G networks are expected to employ flexible decentralized and / or distributed computing systems and architectures and ubiquitous computing, with local spectrum licensing, spectrum sharing, infrastructure sharing, and intelligent automated management underpinned by mobile edge computing, artificial intelligence, short packet communications, and blockchain technologies. Key features of 6G may include intelligent connected management and control capabilities, programmability, integrated sensing and communications, reduced energy footprint, reliable infrastructure, scalability, and affordability. In addition, 6G also opens up new use cases covering the integration of location and sensing capabilities to system definitions for unifying user experiences across the physical and digital worlds.

[0043] It is apparent to those skilled in the art that the illustrated system is only a partial example of a wireless access system, and in practice, the system may include multiple access nodes, a user device may be able to access multiple wireless cells, and the system may include other devices such as physical layer relay nodes or other network elements, etc. At least one of the access nodes may be a Home eNodeB or a Home gNodeB.

[0044] Furthermore, within the geographical area of ​​the wireless communication system, multiple radio cells of different types may be provided. The radio cells may be macrocells (or umbrella cells), which may be large cells with a diameter of up to tens of kilometers, or small cells such as microcells, femtocells, or picocells. The access nodes of FIG. 1 may provide any type of these nodes. The cellular wireless system may be implemented as a multi-layer network including several types of radio cells. In a multi-layer network, one access node may provide one or more radio cells of one type, and therefore multiple access nodes may be required to provide such a network structure.

[0045] To achieve the need for improved deployment and performance of communication systems, the concept of a "plug-and-play" access node may be introduced. A network that may be able to use the "plug-and-play" access node may include a Home Node B gateway or HNB-GW (not shown in FIG. 1) in addition to a Home eNodeB or Home gNodeB. The HNB-GW, which may be installed in an operator's network, may aggregate traffic from multiple Home eNodeBs or Home gNodeBs back to the core network.

[0046] Positioning techniques may be used to estimate the physical location of a user device. Herein, a positioned user device is referred to as a target UE or target user device. For example, positioning techniques used in NR may be based on at least one of time difference of arrival (TDoA), time of arrival (TOA), time of departure (TOD), round trip time (RTT), angle of departure (AoD), angle of arrival (AoA), and / or carrier phase.

[0047] In Uu positioning (UL / DL positioning), multiple transmission and reception points (TRPs) within known locations may transmit and / or receive one or more positioning reference signals (PRS) to / from a target UE. In the uplink, a sounding reference signal (SRS) may be used as a positioning reference signal. For example, a multilateration technique may then be used to locate (position) the target UE with respect to the TRPs. One of these TRPs may be used as a positioning anchor, and the difference in TDoA may be calculated with respect to this positioning anchor. A positioning anchor may be referred to herein as an anchor, anchor node, multilateration anchor, or reference point.

[0048] Sidelink (SL) positioning refers to a positioning approach in which a target UE utilizes the sidelink (i.e., a direct device-to-device link) to position itself in an absolute manner (in the case of absolute positioning) or in a relative manner (in the case of relative positioning). In the case of UE-assisted positioning (in SL positioning and Uu positioning), the target UE may utilize the sidelink to obtain positioning measurements and report the measurements to a network entity such as a Location Management Function (LMF). Sidelink positioning may also be used to obtain ranging information.

[0049] Ranging refers to determining the distance between two UEs and / or the direction of one UE from another UE by direct device connection.

[0050] Absolute positioning refers to estimating the location of a target UE in two-dimensional or three-dimensional geographic coordinates (eg, latitude, longitude, and / or altitude) within a coordinate system.

[0051] Relative positioning refers to estimating the location of a target UE relative to other network elements or relative to other UEs.

[0052] SL positioning may be based on the transmission of sidelink positioning reference signals (SL PRS) by multiple anchor user devices (UEs), which are received and measured by the target UE to enable location of the target UE within the precise latency and accuracy requirements of the corresponding SL positioning session. Alternatively or additionally, the target UE may transmit SL PRS that are received and measured by the anchor UE.

[0053] An anchor UE may be defined as a UE-supported positioning of a target UE, for example, by transmitting and / or receiving reference signals (e.g., SL PRS) for positioning over the SL interface. This may be similar to UL / DL-based positioning, where the gNB may serve as an anchor to transmit and / or receive reference signals to / from the target UE for positioning.

[0054] SL PRS refers to the reference signal transmitted over SL for positioning purposes.

[0055] Additionally, a positioning reference unit (PRU) may be used in a positioning session to improve the positioning accuracy for positioning the target UE. A PRU is a reference device at a known location that makes measurements used to refine a location estimate of the target UE within its area and thereby generate correction data that may be used to improve the positioning accuracy for positioning the target UE. For example, a UE with a known location may be used as a PRU.

[0056] A PRU at a known location may perform positioning measurements, such as reference signal time difference (RSTD), reference signal received power (RSRP), and UE received transmission time difference measurements, and report these measurements to a location server such as an LMF. Additionally, the PRU may transmit an UL SRS for positioning to enable a TRP to measure and report UL positioning measurements (e.g., relative time of arrival, UL-AoA, gNB received transmission time difference, etc.) from a PRU at a known location. The PRU measurements may be compared by the location server with expected measurements at the known PRU location to determine correction data for other nearby target UEs. DL and / or UL location measurements for other target UEs may then be corrected based on the predetermined correction data.

[0057] The PRU may serve as a positioning anchor for the target UE, or the PRU may simply provide correction data (eg, to the LMF) to assist in positioning the target UE.

[0058] In other words, PRUs located at known locations may act as reference target UEs, whereby their calculated locations are compared to their known locations. Comparison of the known and estimated locations may yield correction data, which may be used for the location estimation processes of other target UEs in the vicinity, under the assumption that the same or similar accuracy-determining effects apply to both the PRU's location and the other target UE's locations. The correction data may then be used to fine-tune the target UE's location estimate, thereby increasing positioning accuracy.

[0059] The location of the target UE can be calculated, for example, in the LMF's network (in the case of LMF-based positioning) or in the target UE's own network (in the case of UE-based positioning). Measurements for positioning can be performed on the UE side (e.g., in the case of DL or SL positioning) or on the network side (e.g., in the case of UL positioning).

[0060] 2 illustrates an example in which one or more PRUs 202, 202A, 202B are used to position a target UE 200. The PRUs may be configured to transmit reference signals to be measured for positioning the target UE 200. The target UE 200 may further transmit reference signals for positioning the target UE 200. One or more access nodes 204, 204A, 204B may measure reference signals received from the PRUs 202, 202A, 202B and from the target UE 200. For sidelink positioning, the target UE 200 may measure reference signals received from the PRUs 202, 202A, 202B, and / or the PRUs 202, 202A, 202B may measure reference signals received from the target UE 200 and / or from other UEs or PRUs. The measured parameters (measurement data) derived from the received reference signals may include, for example, reference signal reception time, reference signal time difference (RSTD), reference signal angle of arrival, and / or RSRP. The measurement data may be reported to a network element acting as a location management function (LMF) 212 configured to perform positioning based on the measurement data. The LMF 212 may estimate the location of the target UE 200 based on the received measurement data and the known locations of the PRUs measured by the reporting access node. For example, a location estimation function used in real-time kinematic positioning (RTK) applications of global navigation satellite systems may be used. As an example, if measurements indicate that signals received from the target UE 200 and one of the PRUs 202 have a high correlation, the location of the target UE 200 may be estimated to be close to the PRU 202 and further away from the other PRUs 202A, 202B. The correction from the location of a given PRU 202A, 202B may be calculated based on measurement data, for example, by using the difference between the measurement data associated with the target UE 200 and the measurement data associated with the nearest PRU 202.For example, multilateration measurements (multiple measurements of RSRP, RSTD and / or other parameters) may indicate that the target UE 200 is in a particular direction from the nearest PRU 202 and a correction may be made for that direction.

[0061] The NR air interface may be augmented with features that enable support for artificial intelligence (AI) and / or machine learning (ML)-based algorithms for improved performance and / or reduced complexity and overhead. Some use cases for such AI / ML techniques may include (but are not limited to) channel state information (CSI) feedback enhancements (e.g., reduced overhead, improved accuracy, prediction), beam management (e.g., beam prediction in the time and / or spatial domain for reduced overhead and latency, improved beam selection accuracy), and positioning accuracy enhancements (e.g., in heavy non-line-of-sight scenarios, reduced positioning reference signal transmission and measurement reporting overhead, positioning accuracy scenarios with limited labeling data availability, scenarios where devices have significant RF impairments / impairments affecting positioning measurements, etc.).

[0062] Using AI / ML techniques for positioning accuracy improvement may involve training for positioning being performed in a central ML unit, such as a location management function (LMF) or a 5G network data analytics function (NWDAF). The NWDAF may perform data analytics to generate insights and take actions to enhance user experience, including positioning use cases.

[0063] Training of the ML model in the central ML unit may be performed, for example, according to the following process (described in steps 1 to 4 below). 1) A set of data collection devices may be deployed at selected locations. Alternatively, the data collection devices may be randomly selected from a given geographic area. For example, these data collection devices may be PRUs. In the following, the data collection devices will be referred to as PRUs for simplicity, but any other type of data collection device may be used instead of PRUs. 2) The PRU performs field positioning measurements and reports the measurements to the central ML unit. 3) The central ML unit generates (emulated) positioning measurements using an emulation tool. It should be noted that step 3 may be performed instead of or in addition to step 2. 4) The central ML unit uses the positioning measurements (reported and / or emulated measurements) to train a generic ML-based localization framework, which is denoted herein as GLoc.

[0064] The trained GLoc may then be deployed to a network node that executes the ML process and / or algorithm. Such a network node is referred to herein as a host. The host may be of different types, and the type may be defined in relation to the host's radio frequency (RF) and / or baseband capabilities, form factor, or target positioning key performance indicators (KPIs). The host that performs the ML process may be, for example, the target UE, PRU, and radio access network (e.g., gNB, TRP, and / or Location Management Component, LMC) to enhance positioning accuracy.

[0065] However, a problem with the generic ML-based localization framework (GLoc) is that it does not address the specific RF limitations (also called RF imperfections) of the deployed host type (e.g., handheld UE, roadside unit, or gNB). The RF limitations / imperfections may depend on hardware limitations such as different antenna configurations and form factors, analog-to-digital conversion (ADC) resolution, crystal oscillators, etc. Various RF imperfections may introduce combinations of carrier frequency offset, sampling time offset, transmit / receive beam offset, clock offset and drift, phase noise, etc.

[0066] These RF imperfections may translate into additional phase rotations and delays of the positioning signals through the RF chain, as observed at the baseband receiver. As a result, the positioning entity (e.g., UE, TRP, etc.) hosting the GLoc framework may experience certain RF-based signal delays / rotations that are not accounted for within the GLoc and are inaccurately absorbed into the positioning measurements. Such imperfections may be different for different host types. For example, the PRU, UE, or gNB hosting the GLoc will need to adapt the GLoc to their own RF-specific characteristics.

[0067] Therefore, currently, there may not be a one-size-fits-all GLoc framework that will meet the high positioning accuracy requirements of NR Release 18, for example, at the cm level. Therefore, a device of a given host type may need to adapt the GLoc to its own characteristics, such as antenna form factor and configuration.

[0068] Some example embodiments are described below using principles and terminology of 5G NR technology, however, without limiting the example embodiments to 5G NR communication systems.

[0069] Some exemplary embodiments may address the above problem by providing a method for adapting a generic GLoc framework for host-type-specific ML positioning. In other words, some exemplary embodiments may be used to adapt a generic machine learning model (e.g., trained using training data collected from different types of NR elements) into a machine learning model that is adapted to compensate for intrinsic errors that are specific to a given host type. As such, some exemplary embodiments may provide positioning accuracy improvement using AI / ML techniques.

[0070] Meta-learning in the context of AI / ML refers to adapting a general-purpose model (e.g., trained using features extracted from heterogeneous sources) to a specific type of entity and / or task. A sub-branch of meta-learning is transfer learning (TL), which aims to adjust an already trained model to perform the same task but on a different entity type.

[0071] Some example embodiments may provide a TL framework for positioning, through which a generic GLoc framework for positioning may be customized to a specific NR element host type. More specifically, before the GLoc framework is deployed on a large scale, the GLoc may be refined based on at least intrinsic characteristics (e.g., RF limitations) of the NR element type (e.g., target UE, PRU, or gNB).

[0072] Some exemplary embodiments allow a central ML unit (e.g., LMF) to select an NR head unit (NR-HU) as representative of a given NR element type and, therefore, a given expected intrinsic distortion range. The GLoc model may then be refined by or with the help of the NR-HU, so that the GLoc model is customized to compensate for the distortion specific to that NR element type.

[0073] To perform model adaptation, the central ML unit may provide the NR-HU with a generic GLoc framework along with parameters (such as the capabilities and RF imperfections of the device used to generate the GLoc) that are considered to obtain such framework and a corresponding training procedure.

[0074] The generic GLoc may then be refined based on the individual type of NR-HU (e.g., capabilities and RF deficiencies). Finally, the adapted model may be reported to the central ML unit along with the reasons behind refining the process. Based on such reasons, the central ML unit may iteratively further refine or validate the GLoc framework and provide the next refined version to the host entity.

[0075] As an example, a GLoc trained using the UL SRS collected by the TRP may be adapted to a static UE by using the DL PRS as input, as observed in the static UE baseband. The refined model, called static UE-GLoc, may be forwarded back to the LMF, which then distributes it to other static UEs.

[0076] As another example, a GLoc trained using UL SRS collected by a TRP may be adapted for high-speed UEs.

[0077] As another example, a GLoc trained in an outdoor environment may be adapted to an indoor environment to provide a refined model called an indoor GLoc.

[0078] As another example, a GLoc trained on samples collected from an urban scenario may be adapted to a suburban scenario to provide a refined model called a suburban GLoc.

[0079] Training in the NR-HU may be beneficial because the host (NR-HU) collects signals that are distorted similarly to other NR elements of the same type. TL by the NR-HU may be based on the fact that intrinsic signal distortions are specific to the positioning signals collected by the NR-HU, and that the NR-HU uses type-specific cost functions to refine the GLoc as well as type-specific model constraints (e.g., depth of an artificial or simulated neural network, available activation functions, etc.).

[0080] Alternatively or additionally, the central ML unit may request the NR-HU to collect timestamps, transfer its training data to the central ML unit, and specify its model constraints (if any) in order to refine the GLoc at the central ML unit.

[0081] 3 shows a signaling diagram according to an exemplary embodiment. Although two types (type x and type y) of network nodes are shown in FIG. 3, it should be noted that the number of types may be different from two. In other words, there may be one or more types of network nodes. Furthermore, the signaling procedure shown in FIG. 3 may be extended and applied according to the actual number of types. A central ML unit (e.g., LMF) may determine the actual number of types depending on its ability to group various network nodes based on their RF characteristics.

[0082] Referring to FIG. 3 , in block 301, a central ML unit, such as an LMF, obtains a machine learning model for positioning, where the machine learning model is trained based on first training data associated with one or more first network nodes. The machine learning model may be trained for positioning or for a similar task that may be used for positioning. The LMF may obtain the machine learning model by training the model at the LMF. Alternatively, the machine learning model may be trained at another entity from which the LMF may receive the machine learning model. The machine learning model may be referred to herein as a GLoc. For example, the machine learning model may include an artificial neural network (ANN). An example of an artificial neural network is shown in FIG. 10.

[0083] The first training data may include at least one of reference signal measurement information measured at one or more first network nodes from one or more received positioning reference signals (e.g., DL PRS, UL SRS, and / or SL PRS), emulated reference signal measurement information, or simulated reference signal measurement information related to the one or more first network nodes. For example, the reference signal measurement information may include at least channel impulse response (CIR) measurements that may be simulated or measured at the one or more first network nodes from the one or more received positioning reference signals. The emulated reference signal measurement information may be obtained, for example, by using an emulation tool such as ray tracing, digital twin, etc.

[0084] The one or more first network nodes may comprise one or more types of network nodes, hi one example, the one or more first network nodes may comprise multiple network nodes of different types.

[0085] At block 302, the LMF defines error ranges for type x and type y. The error ranges may indicate internal timing errors caused during measurement collection due to RF imperfections or impairments. For example, a transmit timing error may indicate the time delay from when a digital signal is generated at baseband to when the RF signal is transmitted from the transmit antenna. A receive timing error may indicate the time delay from when the RF signal arrives at the receive antenna to when the signal is digitized and time-stamped at baseband. Error ranges may be defined by the LMF to classify network nodes into type x and type y. For example, a type x host may be associated with an error range −x of + / −a nanoseconds (e.g., a=5 nanoseconds).

[0086] In block 303, the LMF selects a third network node of type x (denoted as type x NR-HU) that will perform transfer learning, i.e., update / refine the machine learning model (generic GLoc), to take into account RF impairments specific to type x. The type of the third network node may be different from the type of one or more of the first network nodes.

[0087] As such, a single Type-x NR-HU may be selected as a representative of all Type-x network nodes, and this single Type-x NR-HU may perform model adaptation, which limits the computational complexity and signaling overhead of transfer learning by avoiding a scheme in which each Type-x network node would perform model adaptation independently.

[0088] The LMF may select a fourth network node of type y (denoted as type y NR-HU) that will update / refine the generic GLoc to account for RF imperfections specific to type y. Herein, type y may refer to a type different from type x.

[0089] As used herein, the term "network node" may refer to, for example, a target user device, a positioning reference unit, an anchor user device, a TRP, or an access node (e.g., gNB) of a radio access network.

[0090] As used herein, the term “type” may refer to, for example, a vendor-specific user device, a vendor-specific access node (e.g., gNB), a TRP with specific RF characteristics, a user device with a specific number of receive antennas, an industrial internet of things (IIoT) device, a low-power high-accuracy positioning (LPHAP) device, a reduced capability (RedCap) device, a handheld user device, or a road-side unit (RSU). For example, Type N may be a UE with N receive antennas, e.g., Type 1 may be a UE with one receive antenna, Type 2 may be a UE with two receive antennas, etc. A type may be defined with respect to both the target positioning accuracy and the inherent error range that a given network node of that type is expected to introduce.

[0091] At block 304, the LMF sends a request to a third network node (type x NR-HU) to update the machine learning model at the third network node based on the second training data (i.e., a request to support transfer learning). For example, the request may be sent in an information element of an LTE positioning protocol (LPP) request message. The third network node may accept or reject the request from the LMF based on the load condition and / or hardware limitations of the third network node.

[0092] The request may include information regarding the configuration of the machine learning model to be updated. For example, such information may define the machine learning model's output and cost function (i.e., model functionality), the type, size, and shape of the machine learning model's inputs, and the machine learning model's architecture.

[0093] As a non-limiting example, the output of the machine learning model may be time of arrival (TOA) information used for positioning, and the cost function may be mean squared error (MSE).

[0094] As a non-limiting example, the input of the machine learning model may be a certain number of strongest channel impulse response (CIR) complex gains.

[0095] As a non-limiting example, the architecture of the machine learning model may refer to a deep neural network (DNN) with N=10 hidden layers and an activation function of the rectified linear unit (ReLU).

[0096] In block 305, the LMF sends a request to a fourth network node (type-y NR-HU) to update the machine learning model at the fourth network node based on the third training data. For example, the request may be sent in an information element of an LPP request message. The fourth network node may accept or reject the request from the LMF based on the load condition and / or hardware limitations of the fourth network node.

[0097] The request may include information regarding the configuration of the machine learning model to be updated. For example, such information may define the machine learning model's output and cost function (i.e., model functionality), the type, size, and shape of the machine learning model's inputs, and the machine learning model's architecture.

[0098] In other words, in blocks 304 and 305, the LMF requests to configure the different types of NR-HUs involved in processing the particular positioning request.

[0099] At block 306, the third network node (Type x NR-HU) sends a reply message to the LMF to accept the request. For example, the acceptance may be indicated in an information element of the LPP reply message including a yes or no flag, or a conditional yes, where the model constraint is described therein. For example, the constraint may mean that the Type x NR-HU may support a maximum architecture different from the architecture configured for the GLoc. For example, the reply may include "model_constraint=Model architecture:DNN with N=4 hidden layers" to indicate that the Type x NR-HU supports a DNN with a maximum of four hidden layers, while the GLoc architecture may include a DNN with 10 hidden layers.

[0100] In block 307, the LMF transmits or forwards the machine learning model to the third network node (Type x NR-HU) in response to the third network node accepting the request. This transmission may indicate at least one of the structure of the machine learning model, one or more activation functions of the machine learning model, a set of weights for a layer of the machine learning model, a set of biases for a layer of the machine learning model, a cost function used to train the machine learning model, an input type and format of the machine learning model (i.e., how the inputs are arranged and what the inputs correspond to), and / or an output type and format of the machine learning model (e.g., probability vector or binary vector, vector length, etc.).

[0101] The inputs for the machine learning model may include at least one of received signal samples for one receive beam for the total number of beams (some of the entries may be zero-padded if unavailable), reference signal received power (RSRP) for one positioning source and / or for one beam, line-of-sight (LOS) indication or probability for one positioning source, etc. For example, the inputs may be a vector of CIRs having a particular length and entries arranged in order of decreasing magnitude.

[0102] For example, the output may be time of arrival (TOA) information expressed as a real scalar.

[0103] The LMF may transmit information indicating a set of constraints for updating the machine learning model at the third network node. In this manner, the LMF may at least partially parameterize the transfer learning procedure at the third network node. By way of example, the set of constraints may indicate at least one of updating the machine learning model using a reference signal from a selected physical resource block (PRB) pool for a given duration and / or whether a particular condition is met. For example, the condition may be met when the third network node considers itself to be static, undisturbed, etc. Alternatively or additionally, the set of constraints may indicate freezing a portion of the machine learning model and updating the remaining architecture, e.g., to update weights from layer L onward.

[0104] The LMF may send information about a reference training procedure to update the machine learning model at the third network node. For example, the reference training procedure may correspond to using training data collected from multiple network nodes. Therefore, the information about the reference training procedure may include at least a set of parameters (e.g., capabilities and RF deficiencies) used to construct the reference training procedure, and the set of parameters may be considered by the third network node when refining the reference training procedure based on its own capabilities and deficiencies.

[0105] The fourth network node (type-y NR-HU) sends a reply message to the LMF to accept the request, block 308. For example, the acceptance may be indicated in an information element of the LPP reply message that includes a yes or no flag, or a conditional yes, where the model constraints are described therein.

[0106] In block 309, the LMF transmits or forwards the machine learning model to the fourth network node (type-y NR-HU) in response to the fourth network node accepting the request. The LMF may transmit information indicating a set of constraints for updating the machine learning model at the fourth network node. The LMF may transmit information regarding a reference training procedure for updating the machine learning model at the fourth network node, the information regarding the reference training procedure including at least a set of parameters used to construct the reference training procedure.

[0107] The content of the transmission of block 309 may be similar to that described above for block 307 .

[0108] At block 310, the third network node (Type-x NR-HU) obtains a first updated machine learning model by updating the machine learning model (i.e., the original GLoc) based on the second training data. Updating may mean adjusting, refining, or retraining the machine learning model based on the second training data specific to the Type-x third network node. The third network node may validate the first updated machine learning model. The first updated machine learning model obtained by the third network node may be referred to herein as Type-x-GLoc.

[0109] The second training data may include reference signal measurement information measured at a third network node from one or more received positioning reference signals (e.g., DL PRS, UL SRS, and / or SL PRS), which may be different from the one or more first network nodes. For example, the reference signal measurement information may include at least channel impulse response (CIR) measurements measured at the third network node from the one or more received positioning reference signals.

[0110] Updating the machine learning model may include the following steps 1 to 6. 1) Initializing a machine learning model using a model received from the LMF, where initialization may mean that the machine learning model is pruned or otherwise simplified according to the capabilities of the third network node. 2) Obtaining, e.g., collecting or accessing, second training data. 3) Cleaning the second training data and preparing the input to match the format defined by the LMF. 4) Selecting a cost function specific to the third network node, for example by using a variant of the generic cost function indicated by the LMF, or a cost function from a list of cost functions selected by the LMF. 5) Selecting an activation function provided by the LMF to generate an output format defined by the LMF. 6) Performing update / training under the set of constraints dictated by the LMF.

[0111] At block 311, the third network node (Type-x NR-HU) sends a message to the LMF indicating the first updated machine learning model (Type-x-GLoc) obtained by the third network node. The message may include the first updated machine learning model. Alternatively, the message may include an updated set of weights and biases associated with the first updated machine learning model. In other words, the refinement process may be reported as a “delta” to a provided reference process such that only the updated weights and biases may be reported (i.e., without reporting the entire model).

[0112] The message may be sent based on the estimated performance improvement of the updated machine learning model exceeding a threshold. In other words, the refined model may be reported if, when tested, it produces a performance improvement (compared to the original GLoc) greater than a given threshold. The threshold may be defined based on positioning accuracy and measurement granularity. This may reduce unnecessary reporting and therefore reduce network signaling.

[0113] At block 312, a fourth network node (Type-y NR-HU) obtains a second updated machine learning model by updating the machine learning model (i.e., the original GLoc) based on the third training data. The fourth network node may validate the second updated machine learning model. The second updated machine learning model obtained by the fourth network node may be referred to as Type-y-GLoc. The fourth network node may perform the update in a manner similar to that described above for block 310.

[0114] The third training data may include reference signal measurement information measured at a fourth network node from one or more received positioning reference signals (e.g., DL PRS, UL SRS, and / or SL PRS), where the fourth network node may be different from the one or more first network nodes and the third network node. For example, the reference signal measurement information may include at least a channel impulse response (CIR) measurement measured at the fourth network node from the one or more received positioning reference signals.

[0115] At block 313, the fourth network node (type-y NR-HU) sends a message to the LMF indicating the second updated machine learning model (type-y-GLoc) obtained by the fourth network node. The message may include the second updated machine learning model. Alternatively, the message may include an updated set of weights and biases associated with the second updated machine learning model.

[0116] At block 314, the LMF may validate or modify the first updated machine learning model and the second updated machine learning model. For example, prior to large-scale deployment, the LMF may cross-validate the updated model to ensure it remains robust and performs within target key performance indicators (KPIs). For example, the LMF may use stored test data to check that the updated model meets given KPI targets.

[0117] At block 315, the LMF transmits or distributes the first updated machine learning model (Type-x-GLoc) to one or more second network nodes of Type-x. In other words, the LMF distributes the first updated machine learning model to other entities of the same type, such as third network nodes (Type-x NR-HUs). The one or more second network nodes may be referred to as Type-x units herein. The difference between Type-x NR-HUs and Type-x units is that Type-x NR-HUs have the ability to train the updated machine learning model, i.e., the ability and computational resources to collect and label second training data, and are therefore selected to produce an updated machine learning model that works well for all Type-x units.

[0118] A given second network node (Type-x unit) may be configured to use the first updated machine learning model (Type-x-GLoc) to position a target UE. For example, if the Type-x unit is a gNB or anchor UE, the Type-x unit may measure reference signals received from the target UE, e.g., to obtain CIR measurements. Alternatively, if the Type-x unit is a target UE, the Type-x unit may measure reference signals received from the gNB or anchor UE, e.g., to obtain CIR measurements. The Type-x unit may then provide these measurements as input to the Type-x-GLoc. The output of the Type-x-GLoc may be a location estimate for the target UE. Alternatively, the output of the Type-x-GLoc may be some other useful positioning-related information or intermediate feature, such as time of arrival (TOA) and / or angle of arrival (AOA) of (possible) line-of-sight (LOS) and / or strong non-line-of-sight (NLOS) paths.

[0119] A type-x-GLoc may be fed the same input information and produce the same output type as the original GLoc. The difference involves the accuracy of these models; a type-x-GLoc model may provide higher accuracy for a specific type of device (i.e., for a type x unit) compared to the original GLoc because the type-x-GLoc model is refined based on the specific RF imperfections of this device type.

[0120] It should be noted that the LMF may transmit the first updated machine learning model to other types of network nodes, for example, to one or more first network nodes.

[0121] In block 316, the LMF transmits or distributes the type-y-GLoc to one or more fifth network nodes of type y. In other words, the LMF distributes the second updated machine learning model to other entities of the same type as the fourth network node (type-y NR-HU).

[0122] It should be noted that the LMF may transmit the second updated machine learning model to other types of network nodes, for example, to one or more of the first network nodes.

[0123] In this specification, the terms "first network node," "second network node," etc. are used to distinguish network nodes and do not necessarily imply a specific identifier of the network node.

[0124] 4 shows a signaling diagram according to another exemplary embodiment. In this exemplary embodiment, selected NR-HUs collect their training data and forward it back to the LMF. The LMF may then aggregate data from multiple network nodes of the same type, store it in memory, and generate an updated machine learning model. It should be noted that the LMF may use the collected training data and combine it across different network node types to be used when updating the generic GLoc model.

[0125] Although two types (type x and type y) of network nodes are shown in Figure 4, it should be noted that the number of types may be different from two. In other words, there may be one or more types of network nodes. Furthermore, the signaling procedure shown in Figure 4 may be extended and applied according to the actual number of types. A central ML unit (e.g., LMF) may determine the actual number of types depending on its ability to group various network nodes based on their RF characteristics.

[0126] Referring to FIG. 4 , in block 401, a central ML unit, such as an LMF, obtains a machine learning model for positioning, where the machine learning model is trained based on first training data associated with one or more first network nodes. The machine learning model may be trained for positioning or for a similar task that may be used for positioning. The LMF may obtain the machine learning model by training the model at the LMF. Alternatively, the machine learning model may be trained at another entity from which the LMF may receive the machine learning model. The machine learning model may be referred to herein as a GLoc. For example, the machine learning model may include an artificial neural network (ANN). An example of an artificial neural network is shown in FIG. 10.

[0127] The first training data may include at least one of reference signal measurement information measured at one or more first network nodes from one or more received positioning reference signals (e.g., DL PRS, UL SRS, and / or SL PRS), emulated reference signal measurement information, or simulated reference signal measurement information related to the one or more first network nodes. For example, the reference signal measurement information may include at least channel impulse response (CIR) measurements that may be simulated or measured at the one or more first network nodes from the one or more received positioning reference signals. The emulated reference signal measurement information may be obtained, for example, by using an emulation tool such as ray tracing, a digital twin, etc.

[0128] The one or more first network nodes may comprise one or more types of network nodes, hi one example, the one or more first network nodes may comprise multiple network nodes of different types.

[0129] At block 402, the LMF defines error ranges for type x and type y. The error ranges may indicate internal timing errors caused during measurement collection due to RF imperfections or impairments. For example, a transmit timing error may indicate the time delay from when a digital signal is generated at baseband to when the RF signal is transmitted from the transmit antenna. A receive timing error may indicate the time delay from when an RF signal arrives at the receive antenna to when the signal is digitized and time-stamped at baseband. Error ranges may be defined by the LMF to classify network nodes into type x and type y. For example, a type x host may be associated with an error range −x of + / −a nanoseconds (e.g., a=5 nanoseconds).

[0130] In block 403, the LMF selects one or more third network nodes of type x (denoted as type x NR-HU) from which the LMF will request training data to update / refine the machine learning model (generic GLoc) to account for RF impairments specific to type x. The type of the one or more third network nodes may be different from the type of the one or more first network nodes.

[0131] The LMF may select one or more fourth network nodes of type y (denoted as type-y NR-HUs) from which the LMF will request training data to update / refine the machine learning model (generic GLoc) to account for RF defects specific to type y. Herein, type y may refer to a type different from type x.

[0132] As used herein, the term "network node" may refer to, for example, a target user device, a positioning reference unit, an anchor UE, a TRP, or an access node (e.g., gNB) of a radio access network.

[0133] As used herein, the term "type" may refer to, for example, a vendor-specific user device, a vendor-specific access node (e.g., gNB), a TRP with particular RF characteristics, a user device with a particular number of receive antennas, an Industrial Internet of Things (IIoT) device, a Low Power High Precision Positioning (LPHAP) device, a Reduced Capability (RedCap) device, a Handheld User Device, or a Roadside Unit (RSU). A type may be defined in terms of both the target positioning accuracy and the inherent error range that a given network node of that type is expected to introduce.

[0134] At block 404, the LMF sends a request to one or more third network nodes (type x NR-HUs) to provide second training data from the one or more third network nodes to the LMF to update the machine learning model at the LMF. For example, the request may be sent in an information element of an LPP request message. A given third network node may accept or reject the request from the LMF based on the load condition and / or hardware limitations of the third network node.

[0135] At block 405, the LMF sends a request to one or more fourth network nodes (type-y NR-HUs) to provide third training data from the one or more fourth network nodes to the LMF to update the machine learning model at the LMF. For example, the request may be sent in an information element of an LPP request message. A given fourth network node may accept or reject the request from the LMF based on the load condition and / or hardware limitations of the fourth network node.

[0136] At block 406, one or more third network nodes (Type x NR-HUs) send a response message to the LMF to accept the request. For example, the acceptance may be indicated in an information element of the LPP reply message that includes a yes or no flag.

[0137] In block 407, one or more fourth network nodes (type-y NR-HUs) send a response message to the LMF to accept the request. For example, the acceptance may be indicated in an information element of the LPP reply message that includes a yes or no flag.

[0138] At block 408, one or more third network nodes (type x NR-HUs) obtain second training data. For example, the one or more third network nodes may obtain the second training data by performing measurements, such as CIR measurements, on received positioning reference signals, such as DL PRS, UL SRS, or SL PRS. Alternatively, the one or more third network nodes may obtain the second training data by retrieving stored measurements from memory. The one or more third network nodes may clean the second training data prior to sending it to the LMF.

[0139] At block 409, one or more third network nodes (type x NR-HU) transmit second training data to the LMF. The LMF may store the received second training data in its memory. The one or more third network nodes may transmit information indicating a first set of constraints for updating the machine learning model to the LMF. For example, the first set of constraints may include a maximum supported depth (i.e., number of layers) of the artificial neural network, available activation functions, etc.

[0140] At block 410, one or more fourth network nodes (type-y NR-HUs) obtain third training data. For example, the one or more fourth network nodes may obtain the third training data by performing measurements, such as CIR measurements, on received positioning reference signals, such as DL PRS, UL SRS, or SL PRS. Alternatively, the one or more fourth network nodes may obtain the third training data by retrieving stored measurements from memory. The one or more fourth network nodes may clean the third training data prior to sending it to the LMF.

[0141] At block 411, one or more fourth network nodes (type y NR-HU) transmit third training data to the LMF. The LMF may store the received third training data in its memory. The one or more fourth network nodes may transmit information indicating a second set of constraints for updating the machine learning model to the LMF. For example, the second set of constraints may include a maximum supported depth (i.e., number of layers) of the artificial neural network, available activation functions, etc.

[0142] At block 412, the LMF obtains a first updated machine learning model (type-x-GLoc) by updating the machine learning model (original GLoc) based at least in part on second training data received from one or more third network nodes.

[0143] The update may include one or more of the following steps 1 to 6. 1) Initialize the machine learning model using the original GLoc. 2) Accessing second training data. 3) Cleaning the second training data and preparing the input to match the input format of GLoc. 4) Selecting cost functions specific to one or more third network nodes, for example, by using a variant of the GLoc generic cost function. 5) Selecting an activation function to generate the output format of the GLoc. 6) Performing updating / training under a set of constraints dictated by one or more third network nodes.

[0144] The LMF may obtain a second updated machine learning model (Type-y-GLoc) by updating the machine learning model (Original GLoc) based at least in part on third training data received from one or more fourth network nodes.

[0145] In block 413, the LMF transmits or distributes the first updated machine learning model (Type-x-GLoc) to one or more second network nodes of Type-x. In other words, the LMF distributes the first updated machine learning model to one or more third network nodes (Type-x NR-HUs) and other entities of the same type. The one or more second network nodes may be referred to as Type-x units in this specification. The difference between Type-x NR-HUs and Type-x units is that Type-x NR-HUs have the ability to train the updated machine learning model, i.e., the ability and computational resources to collect and label second training data, and are therefore selected to produce an updated machine learning model that works well for all Type-x units.

[0146] A given second network node (Type-x unit) may be configured to use the first updated machine learning model (Type-x-GLoc) to position a target UE. For example, if the Type-x unit is a gNB or anchor UE, the Type-x unit may measure reference signals received from the target UE, e.g., to obtain CIR measurements. Alternatively, if the Type-x unit is a target UE, the Type-x unit may measure reference signals received from the gNB or anchor UE, e.g., to obtain CIR measurements. The Type-x unit may then provide these measurements as input to the Type-x-GLoc. The output of the Type-x-GLoc may be a location estimate for the target UE. Alternatively, the output of the Type-x-GLoc may be some other useful positioning-related information or intermediate feature, such as time of arrival (TOA) and / or angle of arrival (AOA) of (possible) line-of-sight (LOS) and / or strong non-line-of-sight (NLOS) paths.

[0147] A type-x-GLoc may be fed the same input information and produce the same output type as the original GLoc. The difference involves the accuracy of these models; a type-x-GLoc model provides higher accuracy for a specific type of device (i.e., for a type-x device) compared to the original GLoc because the type-x-GLoc model is refined based on the specific RF defects of this device type.

[0148] It should be noted that the LMF may transmit the first updated machine learning model to other types of network nodes, for example, to one or more first network nodes.

[0149] At block 414, the LMF transmits or distributes the type-y-GLoc to one or more fifth network nodes of type y. In other words, the LMF distributes the second updated machine learning model to other entities of the same type as the fourth network node (type-y NR-HU). The one or more fifth network nodes may be referred to as type-y units in this specification.

[0150] It should be noted that the LMF may transmit the second updated machine learning model to other types of network nodes, for example, to one or more of the first network nodes.

[0151] 5 shows a flowchart according to an example embodiment of a method performed by an apparatus such as, comprising, or included in a Location Management Function (LMF) or a Network Data Analytics Function (NWDAF). For example, the apparatus may correspond to the LMF 112 of FIG. 1 or the LMF 212 of FIG. 2.

[0152] Referring to FIG. 5, at block 501, a machine learning model for positioning is obtained, and the machine learning model is trained based on first training data associated with one or more first network nodes.

[0153] At block 502, information including at least one of a machine learning model, a request to update the machine learning model based on the second training data, or a request to provide the second training data to update the machine learning model at the device is transmitted.

[0154] At block 503, at least one of a message indicating an updated machine learning model or second training data is received.

[0155] At block 504, the updated machine learning model is transmitted to at least one or more second network nodes.

[0156] 6 shows a flowchart according to an example embodiment of a method implemented by an apparatus such as a network node. The network node may refer to, for example, a user device, a positioning reference unit, or an access node of a radio access network. For example, the network node may correspond to UE 100, UE 102, or access node 104 of FIG. 1, or any of UE 200, PRU 202, 202A, and 202B, or access nodes 204, 204A, and 204B of FIG. 2.

[0157] Referring to FIG. 6, at block 601, information including at least one of a machine learning model for positioning, a request to update the machine learning model in the device based on second training data, or a request to provide second training data to update the machine learning model is received, where the machine learning model has been trained based on first training data associated with one or more first network nodes.

[0158] At block 602, second training data is obtained.

[0159] At block 603, at least one of a message indicating an updated machine learning model or second training data is sent.

[0160] As used herein, "at least one of the following: <list of two or more elements>" )" and "at least one of <list of two or more elements>" )" and similar phrases where a list of two or more elements is 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 of the elements.

[0161] The blocks, associated functions, and information exchanges (messages) described above with reference to Figures 3-6 are not in absolute chronological order, and some of them may be performed simultaneously or in a different order than described. Other functions may be performed between or among them as well, other information may be sent, and / or other rules may apply. Some of the blocks, or portions of the blocks, or one or more pieces of information may also be omitted or replaced by a corresponding block, portion of the block, or one or more pieces of information.

[0162] 7 illustrates an example of an apparatus 700 comprising means for implementing any of the methods of FIGS. 3-6 or any other exemplary embodiment described above. For example, the apparatus 700 may be an apparatus such as, comprising, or included in a user device. The user device may correspond to any of the user devices 100, 102 of FIG. 1, the user device 200 of FIG. 2, or any of the PRUs 202, 202A, 202B of FIG. 2. The user device may also be referred to as a subscriber unit, a mobile station, a remote terminal, an access terminal, a user terminal, a terminal device, user equipment (UE), a target UE, a target user device, an anchor UE, a positioning reference unit (PRU), an NR-HU, a host, or a network node.

[0163] The device 700 comprises at least one processor 710. The at least one processor 710 interprets instructions (e.g., computer program instructions) and processes data. The at least one processor 710 may comprise one or more programmable processors. The at least one processor 710 may comprise programmable hardware with embedded firmware, or alternatively or additionally, may comprise one or more application-specific integrated circuits (ASICs).

[0164] The at least one processor 710 is coupled to the at least one memory 720. The at least one processor is configured to write / read data to / from the at least one memory 720. The at least one memory 720 may comprise one or more memory units. The memory units may be volatile or nonvolatile. It is noted that there may be one or more memory units of nonvolatile memory and one or more memory units of volatile memory, or alternatively, one or more memory units of nonvolatile memory, or alternatively, one or more memory units of volatile memory. The volatile memory may be, for example, random access memory (RAM), dynamic random access memory (DRAM), or synchronous dynamic random access memory (SDRAM). The nonvolatile memory may be, for example, read-only memory (ROM), programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), flash memory, optical storage, or magnetic storage. In general, memory may be referred to as a non-transitory computer-readable medium. The term "non-transitory," as used herein, refers to the limitation of the medium itself (i.e., tangible, not a signal) as opposed to the limitation of data storage persistence (e.g., RAM vs. ROM). At least one memory 720 stores computer-readable instructions, which are executed by at least one processor 710 to implement one or more of the exemplary embodiments described above. For example, non-volatile memory stores computer-readable instructions, and at least one processor 710 executes the instructions using volatile memory for temporary storage of data and / or instructions. The computer-readable instructions may also be referred to as computer program code.

[0165] The computer-readable instructions may be pre-stored in at least one memory 720, or alternatively or additionally, the computer-readable instructions may be received by the apparatus via an electromagnetic carrier signal and / or copied from a physical entity such as a computer program product. Execution of the computer-readable instructions by the at least one processor 710 causes the apparatus 700 to implement one or more of the exemplary embodiments described above. That is, the at least one processor and at least one memory storing instructions may provide a means for providing or effecting the performance of any of the methods and / or blocks described above.

[0166] In the context of this document, "memory" or "computer-readable media" or "computer-readable medium" may be any non-transitory medium or vehicle or means that can contain, store, communicate, propagate, or transport instructions for use by or in connection with an instruction execution system, apparatus, or device such as a computer. 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 regarding data storage persistence (e.g., RAM vs. ROM).

[0167] The device 700 may further include or be connected to an input unit 730. The input unit 730 may include one or more interfaces for receiving input. The one or more interfaces may include, for example, one or more temperature, motion, and / or orientation sensors, one or more cameras, one or more accelerometers, one or more microphones, one or more buttons, and / or one or more touch detection units. Additionally, the input unit 730 may include an interface to which external devices may connect.

[0168] The device 700 may comprise an output unit 740. The output unit may comprise or be connected to one or more displays capable of rendering visual content, such as a light emitting diode (LED) display, a liquid crystal display (LCD), and / or a liquid crystal on silicon (LCoS) display. The output unit 740 may further comprise one or more audio outputs. The one or more audio outputs may be, for example, loudspeakers.

[0169] The device 700 further comprises a connectivity unit 750. The connectivity unit 750 enables wireless connectivity to one or more external devices. The connectivity unit 750 comprises at least one transmitter and at least one receiver, which may be integrated into the device 700 or to which the device 700 may be connected. The at least one transmitter comprises at least one transmitting antenna, and the at least one receiver comprises at least one receiving antenna. The connectivity unit 750 may comprise an integrated circuit or a set of integrated circuits that provide wireless communication capabilities for the device 700. Alternatively, the wireless connectivity may be a hardwired application-specific integrated circuit (ASIC). The connectivity unit 750 may provide means for implementing at least some of the blocks of one or more exemplary embodiments described above. The connectivity unit 750 may include one or more components, such as a power amplifier, a digital front end (DFE), an analog-to-digital converter (ADC), a digital-to-analog converter (DAC), a frequency converter, a (de)modulator, and / or encoder / decoder circuitry, controlled by a corresponding control unit.

[0170] It is noted that the apparatus 700 may further comprise various components not shown in Figure 7. The various components may be hardware components and / or software components.

[0171] FIG. 8 illustrates an example of an apparatus 800 comprising means for implementing any of the methods of FIGS. 3-6 or any other exemplary embodiment described above. For example, the apparatus 800 may be an apparatus such as, comprising, or included in an access node of a radio access network. The access node may correspond to any of the access nodes 104 of FIG. 1 or 204, 204A, and 204B of FIG. 2. The apparatus 800 may also be referred to as, for example, a network node, a radio access network (RAN) node, a next generation radio access network (NG-RAN) node, a Node B, an eNB, a gNB, a base transceiver station (BTS), a base station, an NR base station, a 5G base station, an access point (AP), a relay node, a repeater, an integrated access and backhaul (IAB) node, an IAB donor node, a distributed unit (DU), a central unit (CU), a baseband unit (BBU), a radio unit (RU), a radio head, a remote radio head (RRH), or a transmit and receive point (TRP).

[0172] The apparatus 800 may include, for example, circuitry or a chipset applicable to implementing one or more of the exemplary embodiments described above. The apparatus 800 may also be an electronic device including one or more electronic circuitry. The apparatus 800 may include communication control circuitry 810, such as at least one processor, and at least one memory 820 storing instructions that, when executed by the at least one processor, cause the apparatus 800 to implement one or more of the exemplary embodiments described above. Such instructions 822 may include, for example, computer program code (software), where the at least one memory and the computer program code (software) are configured to, with the at least one processor, cause the apparatus 800 to implement one or more of the exemplary embodiments described above. Herein, computer program code may then refer to instructions that, when executed by the at least one processor, cause the apparatus 800 to implement one or more of the exemplary embodiments described above. That is, at least one processor and at least one memory storing instructions may provide the means for providing or effecting the performance of any of the methods and / or blocks described above.

[0173] The processor is coupled to the memory 820. The processor is configured to write / read data to / from the memory 820. The memory 820 may comprise one or more memory units. The memory units may be volatile or nonvolatile. It is noted that there may be one or more units of nonvolatile memory and one or more units of volatile memory, or alternatively, one or more units of nonvolatile memory, or alternatively, one or more units of volatile memory. The volatile memory may be, for example, random access memory (RAM), dynamic random access memory (DRAM), or synchronous dynamic random access memory (SDRAM). The nonvolatile memory may be, for example, read-only memory (ROM), programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), flash memory, optical storage, or magnetic storage. In general, memory may be referred to as a non-transitory computer-readable medium. The term "non-transitory," as used herein, refers to the medium itself (i.e., tangible, not signal) as opposed to the data storage persistence (e.g., RAM vs. ROM). The memory 820 stores computer-readable instructions that are executed by the processor. For example, non-volatile memory stores the computer-readable instructions, and the processor executes the instructions using volatile memory for temporary storage of data and / or instructions.

[0174] The computer-readable instructions may be pre-stored in memory 820, or alternatively or additionally, the computer-readable instructions may be received by the device via an electromagnetic carrier signal and / or copied from a physical entity such as a computer program product. Execution of the computer-readable instructions causes the device 800 to perform one or more of the functions described above.

[0175] The memory 820 may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, flash memory, magnetic memory devices and systems, optical memory devices and systems, fixed memory, and / or removable memory. The memory may comprise a configuration database for storing configuration data. For example, the configuration database may store a current neighbor cell list and, in some exemplary embodiments, the frame structure used in detected neighbor cells.

[0176] The device 800 may further comprise a communication interface 830 comprising hardware and / or software for achieving communication connectivity according to one or more communication protocols. The communication interface 830 comprises at least one transmitter (Tx) and at least one receiver (Rx), which may be integrated into the device 800 or to which the device 800 may be connected. The communication interface 830 may provide means for implementing some of the blocks of one or more exemplary embodiments described above. The communication interface 830 may comprise one or more components, such as a power amplifier, a digital front end (DFE), an analog-to-digital converter (ADC), a digital-to-analog converter (DAC), a frequency converter, a (de)modulator, and / or encoder / decoder circuitry, controlled by a corresponding control unit.

[0177] The communication interface 830 provides the device with wireless communication capabilities for communicating in a cellular communication system. The communication interface may, for example, provide a radio interface to one or more user devices. The device 800 may further comprise another interface towards a core network, such as a network coordinator device or AMF, and / or to an access node of the cellular communication system.

[0178] The apparatus 800 may further comprise a scheduler 840 configured to allocate radio resources. The scheduler 840 may be configured together with the communication control circuitry 810, or the scheduler 840 may be configured separately.

[0179] It is noted that the apparatus 800 may further comprise various components not shown in Figure 8. The various components may be hardware components and / or software components.

[0180] 9 illustrates an example of an apparatus 900 comprising means for implementing any of the methods of FIGS. 3-6 or any other exemplary embodiment described above. For example, the apparatus 900 may be an apparatus such as, comprising, or included in a central ML unit. The central ML unit may be referred to as, for example, a location management function (LMF), a location server, or a network data analytics function (NWDAF). For example, the central ML unit may correspond to the LMF 112 of FIG. 1 or the LMF 212 of FIG. 2.

[0181] The apparatus 900 may include, for example, circuitry or a chipset applicable to implementing one or more of the exemplary embodiments described above. The apparatus 900 may also be an electronic device including one or more electronic circuitry. The apparatus 900 may include communication control circuitry 910, such as at least one processor, and at least one memory 920 storing instructions that, when executed by the at least one processor, cause the apparatus 900 to implement one or more of the exemplary embodiments described above. Such instructions 922 may include, for example, computer program code (software), where the at least one memory and the computer program code (software) are configured to, with the at least one processor, cause the apparatus 900 to implement one or more of the exemplary embodiments described above. Herein, computer program code may then refer to instructions that, when executed by the at least one processor, cause the apparatus 900 to implement one or more of the exemplary embodiments described above. That is, at least one processor and at least one memory storing instructions may provide the means for providing or effecting the performance of any of the methods and / or blocks described above.

[0182] The processor is coupled to the memory 920. The processor is configured to write / read data to / from the memory 920. The memory 920 may comprise one or more memory units. The memory units may be volatile or nonvolatile. It is noted that there may be one or more units of nonvolatile memory and one or more units of volatile memory, or alternatively, one or more units of nonvolatile memory, or alternatively, one or more units of volatile memory. The volatile memory may be, for example, random access memory (RAM), dynamic random access memory (DRAM), or synchronous dynamic random access memory (SDRAM). The nonvolatile memory may be, for example, read-only memory (ROM), programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), flash memory, optical storage, or magnetic storage. In general, memory may be referred to as a non-transitory computer-readable medium. The term "non-transitory," as used herein, refers to the medium itself (i.e., tangible, not signal) as opposed to the data storage persistence (e.g., RAM vs. ROM). The memory 920 stores computer-readable instructions that are executed by the processor. For example, non-volatile memory stores the computer-readable instructions, and the processor executes the instructions using volatile memory for temporary storage of data and / or instructions.

[0183] The computer-readable instructions may be pre-stored in memory 920, or alternatively or additionally, the computer-readable instructions may be received by the device via an electromagnetic carrier signal and / or copied from a physical entity such as a computer program product. Execution of the computer-readable instructions causes the device 900 to perform one or more of the functions described above.

[0184] The memory 920 may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, flash memory, magnetic memory devices and systems, optical memory devices and systems, fixed memory, and / or removable memory. The memory may include a configuration database for storing configuration data. For example, the configuration database may store a current neighbor cell list and, in some exemplary embodiments, the frame structure used in detected neighbor cells.

[0185] The device 900 may further comprise a communication interface 930 comprising hardware and / or software for achieving communication connectivity according to one or more communication protocols. The communication interface 930 comprises at least one transmitter (Tx) and at least one receiver (Rx), which may be integrated into the device 900 or to which the device 900 may be connected. The communication interface 930 may provide means for implementing some of the blocks of one or more exemplary embodiments described above. The communication interface 930 may comprise one or more components, such as a power amplifier, a digital front end (DFE), an analog-to-digital converter (ADC), a digital-to-analog converter (DAC), a frequency converter, a (de)modulator, and / or encoder / decoder circuitry, controlled by a corresponding control unit.

[0186] The communication interface 930 provides the device with wireless communication capabilities for communicating in a cellular communication system. The communication interface may, for example, provide a radio interface to one or more user devices. The device 900 may further comprise another interface towards a core network, such as a network coordinator device or AMF, and / or to an access node of the cellular communication system.

[0187] It is noted that the apparatus 900 may further comprise various components not shown in Figure 9. The various components may be hardware components and / or software components.

[0188] As used in this application, the term "circuitry" may refer to one or more or all of the following: a) hardware-only circuit implementations (such as implementations with only analog and / or digital circuitry), and b) combinations of hardware circuitry and software, for example, (where applicable): i) combinations of analog hardware circuitry and / or digital hardware circuitry with software / firmware, and ii) any portion of a hardware processor with software (including digital signal processors, software, and memory that work together to cause a device such as a mobile phone to perform various functions), and c) hardware circuitry and / or processors, such as a microprocessor or portion of a microprocessor, that require software (e.g., firmware) to operate, but the software may be absent when not required for operation.

[0189] This definition of circuitry applies to all uses of the term in this application, including in any claim. As a further example, as used in this application, the term circuitry also covers merely a hardware circuit or processor (or processors), or certain portions of a hardware circuit or processor and its (or their) associated software and / or firmware implementations. The term circuitry also covers, for example, a baseband integrated circuit, or a processor integrated circuit for a mobile device, or similar integrated circuit in a server, cellular network device, or other computing or network device, if applicable to particular claim elements.

[0190] Figure 10 shows an example of an artificial neural network 1030 with one hidden layer 1002, and Figure 11 shows an example of a computational node 1004. However, it should be noted that the artificial neural network 1030 may comprise two or more hidden layers 1002.

[0191] An artificial neural network (ANN) 1030 contains a set of rules designed to perform tasks such as regression, classification, clustering, and pattern recognition. ANNs may achieve such goals through a learning / training procedure, where the goal is represented by various examples of input data along with the desired output. In this way, the ANN learns to identify the appropriate output for any input within the training data manifold. Learning / training by using labels is called supervised learning, while learning without labels is called unsupervised learning. Deep learning may require a large amount of input data.

[0192] Deep learning (also known as deep structured learning or hierarchical learning) is part of a broader family of layer-based machine learning models used in artificial neural networks. A deep neural network (DNN) 1030 is an artificial neural network with multiple hidden layers 1002 between an input layer 1000 and an output layer 1014. Training a DNN enables it to find the correct mathematical operations that transform inputs into appropriate outputs, even when the relationships are highly nonlinear and / or complex.

[0193] A given hidden layer 1002 comprises nodes 1004, 1006, 1008, 1010, and 1012 where computations occur. As shown in FIG. 11 , a given node 1004 combines input data 1000 with a set of coefficients or weights 1100 that amplify or attenuate that input 1000, thereby assigning significance to the input 1000 with respect to the task the algorithm is attempting to learn. The input-weight products are summed 1102, and the sum, through an activation function 1104, determines whether and how far the signal should proceed further through the neural network 1030 to affect the final outcome, such as the act of classification. In the process, the neural network learns to recognize correlations between certain relevant features and optimal results.

[0194] In the case of classification, the output of the DNN 1030 may be considered as the probability of a particular outcome. In this case, the number of layers 1002 may vary proportionally to the number of input data 1000 used. However, when the number of input data 1000 is large, the accuracy of the outcome 1014 is more reliable. On the other hand, when there are fewer layers 1002, the calculation may take less time, thus reducing latency. However, this is highly dependent on the particular DNN architecture and / or available computational resources.

[0195] The initial weights 1100 of the model can be set in a variety of alternative ways. During the training phase, the weights may be adapted to improve the accuracy of the process based on analytical errors in decision-making. Training a model is essentially a trial and error activity. In principle, a given node 1004, 1006, 1008, 1010, 1012 of the neural network 1030 will make decisions (input * The node makes a decision (weights) and then compares this decision with the collected data to find the difference relative to the collected data. In other words, the node determines the error based on which the weights 1100 are adjusted. Training a model may therefore be thought of as a corrective feedback loop.

[0196] For example, the neural network model may be trained using a stochastic gradient descent optimization algorithm, for which gradients are calculated using a backpropagation algorithm. The gradient descent algorithm attempts to change the weights 1100 so that the next evaluation reduces the error, meaning that the optimization algorithm navigates down the gradient (or slope) of the error. Any other suitable optimization algorithm may also be used if it provides sufficiently accurate weights 1100. As a result, the trained parameters of the neural network 1030 may include the weights 1100.

[0197] In the context of optimization algorithms, the function used to evaluate candidate solutions (i.e., sets of weights) is called the objective function. For neural networks, where the target is to minimize error, the objective function may also be called a cost function or loss function. Any suitable method may be used as the loss function when adjusting the weights 1100. Some examples of loss functions are mean squared error (MSE), maximum likelihood estimation (MLE), and cross entropy.

[0198] Regarding the activation function 1104 of a node 1004, the activation function defines the output 1014 of that node 1004 given an input or set of inputs 1000. The node 1004 calculates a weighted sum of the inputs, possibly adding a bias, and then makes a decision as to "activate" or "not activate" based on a decision threshold as a binary activation or using an activation function 1104 that provides a nonlinear decision function. Any suitable activation function 1104 may be used, such as sigmoid, rectified linear unit (ReLU), rectified exponential function (softmax), softplus, tanh, etc. In deep learning, the activation function 1104 may be set at the layer level and applied to all neurons (nodes) in that layer. The output 1014 is then used as the input for the next node, and so on, until the desired solution to the original problem is found.

[0199] The techniques and methods described herein may be implemented by various means. For example, these techniques may be implemented in hardware (one or more devices), firmware (one or more devices), software (one or more modules), or a combination thereof. In a hardware implementation, the apparatus of the exemplary embodiments may be implemented within one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), graphics processing units (GPUs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. In a firmware or software implementation, the implementation may be implemented through modules (e.g., procedures, functions, etc.) of at least one chipset that perform the functions described herein. Software code may be stored in a memory unit and executed by a processor. The memory unit may execute within the processor or external to the processor. In the latter case, the memory unit may be communicatively coupled to the processor by various means, as is known in the art. Furthermore, the components of the systems described herein may be rearranged and / or supplemented by additional components to facilitate accomplishing various aspects, etc., described with respect to the components, and the components are not limited to the precise configurations set forth in the given figures, as will be recognized by those skilled in the art.

[0200] It will be obvious to those skilled in the art that as technology advances, the concept of the present invention may be implemented in various ways. The embodiments are not limited to the exemplary embodiments described above, but may vary within the scope of the claims. Therefore, all words and expressions should be interpreted broadly and are intended to illustrate, not limit, exemplary embodiments.

Claims

1. 1. 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 to perform at least: obtaining a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmitting information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide the second training data for updating the machine learning model at the device; receiving at least one of a message indicating an updated machine learning model or the second training data; transmitting the updated machine learning model to at least one or more second network nodes; A device that performs the following.

2. further selecting a third network node, said information being transmitted to said third network node; 2. The apparatus of claim 1, wherein a type of the third network node is different from a type of the one or more first network nodes, and a type of the one or more second network nodes corresponds to the type of the third network node.

3. 3. The apparatus of claim 1, further comprising: causing the updated machine learning model to be further validated or modified prior to transmitting the updated machine learning model to the one or more second network nodes.

4. The apparatus according to claim 1 , further comprising: a set of constraints for updating the machine learning model.

5. The device of any one of claims 1 to 4, further comprising transmitting information about a reference training procedure to update the machine learning model, wherein the information about the reference training procedure includes at least a set of parameters used to construct the reference training procedure.

6. The apparatus of claim 1 , wherein the message includes the updated machine learning model.

7. The apparatus of claim 1 , wherein the message includes an updated set of weights and biases associated with the updated machine learning model.

8. 8. The apparatus of claim 1, further comprising: updating the machine learning model based at least in part on the second training data, thereby obtaining the updated machine learning model.

9. The apparatus of claim 1 , further configured to receive information indicating a set of constraints for updating the machine learning model at the apparatus.

10. The apparatus of claim 1 , further causing the one or more first network nodes to transmit the updated machine learning model.

11. The apparatus of any preceding claim, wherein the one or more first network nodes comprise a plurality of network nodes of different types.

12. 12. The apparatus of claim 1, wherein the first training data comprises at least one of reference signal measurement information measured at the one or more first network nodes, emulated reference signal measurement information, or simulated reference signal measurement information associated with the one or more first network nodes.

13. The apparatus according to any one of claims 1 to 12, wherein the second training data includes reference signal measurement information measured at a third network node different from the one or more first network nodes.

14. 14. The apparatus of claim 12 or 13, wherein the reference signal measurement information comprises at least channel impulse response measurements measured from one or more received positioning reference signals.

15. 1. 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 to perform at least: receiving information including at least one of a machine learning model for positioning, a request to update the machine learning model at the device based on second training data, or a request to provide the second training data for updating the machine learning model, wherein the machine learning model has been trained based on first training data associated with one or more first network nodes; obtaining the second training data; sending at least one of a message indicating an updated machine learning model or the second training data; A device that performs the following.

16. The apparatus of claim 15 , further comprising: updating the machine learning model based on the second training data to obtain the updated machine learning model.

17. The apparatus of claim 15 or 16, wherein the message is sent based on an estimated performance improvement of the updated machine learning model exceeding a threshold.

18. The apparatus of any one of claims 15 to 17, further configured to receive information indicating a set of constraints for updating the machine learning model at the apparatus.

19. The apparatus of any one of claims 15 to 18, further comprising: a processor configured to receive information about a reference training procedure for updating the machine learning model in the apparatus; and the information about the reference training procedure includes at least a set of parameters used to construct the reference training procedure.

20. The apparatus of any one of claims 15 to 19, wherein the message includes the updated machine learning model.

21. The apparatus of any one of claims 15 to 19, wherein the message includes an updated set of weights and biases associated with the updated machine learning model.

22. The apparatus of any one of claims 15 to 21, further configured to transmit a set of constraints for updating the machine learning model.

23. 23. The apparatus of claim 15, wherein the first training data comprises at least one of reference signal measurement information measured at the one or more first network nodes, emulated reference signal measurement information, or simulated reference signal measurement information associated with the one or more first network nodes.

24. The device according to any one of claims 15 to 23, wherein the second training data includes reference signal measurement information measured by the device.

25. 25. The apparatus of claim 23 or 24, wherein the reference signal measurement information comprises at least channel impulse response measurements measured from one or more received positioning reference signals.

26. 1. An apparatus comprising: means for obtaining a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; means for transmitting information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide the second training data for updating the machine learning model at the device; means for receiving at least one of a message indicating an updated machine learning model or the second training data; means for transmitting the updated machine learning model to at least one or more second network nodes; An apparatus comprising:

27. 1. An apparatus comprising: means for receiving information including at least one of a machine learning model for positioning, a request to update the machine learning model in the device based on second training data, or a request to provide the second training data for updating the machine learning model, wherein the machine learning model has been trained based on first training data associated with one or more first network nodes; and means for acquiring the second training data; means for transmitting at least one of a message indicating an updated machine learning model or the second training data; An apparatus comprising:

28. 1. A method comprising: obtaining, by an apparatus, a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmitting, by the device, information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide the second training data for updating the machine learning model at the device; receiving, by the device, at least one of a message indicating an updated machine learning model or the second training data; transmitting, by the device, the updated machine learning model to at least one or more second network nodes; A method comprising:

29. 1. A method comprising: receiving, by the device, information including at least one of a machine learning model for positioning, a request to update the machine learning model at the device based on second training data, or a request to provide the second training data for updating the machine learning model, wherein the machine learning model has been trained based on first training data associated with one or more first network nodes; acquiring the second training data by the device; transmitting, by the device, at least one of a message indicating an updated machine learning model or the second training data; A method comprising:

30. A non-transitory computer-readable medium containing program instructions that, when executed by an apparatus, cause the apparatus to perform at least: obtaining a machine learning model for positioning, the machine learning model being trained based on first training data associated with one or more first network nodes; transmitting information including at least one of the machine learning model, a request to update the machine learning model based on second training data, or a request to provide the second training data for updating the machine learning model at the device; receiving at least one of a message indicating an updated machine learning model or the second training data; transmitting the updated machine learning model to at least one or more second network nodes; A non-transitory computer-readable medium for implementing the above.

31. A non-transitory computer-readable medium containing program instructions that, when executed by an apparatus, cause the apparatus to perform at least: receiving information including at least one of a machine learning model for positioning, a request to update the machine learning model at the device based on second training data, or a request to provide the second training data for updating the machine learning model, wherein the machine learning model has been trained based on first training data associated with one or more first network nodes; obtaining the second training data; sending at least one of a message indicating an updated machine learning model or the second training data; A non-transitory computer-readable medium for implementing the above.

Images