Training machine learning positioning models in wireless communication networks

Abstract

A network node in a wireless communications network is provided, the network node comprising: a transmitter configured to transmit a positioning training dataset configuration to at least one data source of the wireless communications network, the positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on a location of the network node; the network node further comprising: a receiver configured to receive a response from the at least one data source, the response including the positioning training dataset in the required data format; and the network node further comprising a processor configured to train the machine learning positioning model using the positioning training dataset.

Description

[Technical Field]

[0001] The subject matter disclosed herein generally relates to the field of implementing training of machine learning positioning models in wireless communication networks. This document defines a network node in a wireless communication network and a method in the network node, the network node being in the wireless communication network. [Background technology]

[0002] Positioning requirements for 3GPP® New Radio (NR) air interface (Uu) signals and standalone (SA) architectures (e.g., beam-based transmission) are specified in 3GPP Release 16. Target use cases also included commercial and regulatory (emergency services) scenarios given in 3GPP Release 15. Furthermore, the current 3GPP Release 17 Positioning recently defines positioning performance requirements for commercial and industrial internet of things (IIoT) use cases.

[0003] Positioning techniques in 3GPP may currently be configured and implemented based on location management function (LMF) and UE capability requirements, which attempt to enable the calculation of a location estimate for the UE. Summary of the Invention [Means for solving the problem]

[0004] Artificial intelligence (AI) and / or machine learning (ML) for air interfaces corresponding to target use cases in the 3GPP framework relate to CSI feedback, beam management, and positioning accuracy improvement in terms of aspects such as performance, complexity, and potential impact of specifications. The use of AI / ML-based procedures to enhance different location procedures, including 3GPP positioning procedures, and / or signaling in the RAN and / or UE and / or location server (LMF), may be particularly beneficial and improvements in the aforementioned aspects may be required.

[0005] For example, AI / ML techniques can be implemented to optimize and predict various metrics in time and space (e.g., PRS / SRS configuration, final location accuracy) in a positioning session of a target UE. Therefore, assuming the described training and inference models are already deployed in the location server or NG-RAN / UE, there is a need to support efficient model training and inference signaling mechanisms for improved measurements and reporting in the positioning framework.

[0006] However, current positioning frameworks lack a known method for selecting, triggering, and configuring the collection of positioning data required to train a configured AI / ML positioning model. This includes both online and offline training of positioning data for a specific AI / ML model. Furthermore, the type of data source is important for the input training dataset to be statistically significant, which impacts the accuracy of the AI / ML model to be trained.

[0007] This disclosure presents apparatus and methods detailing support for various scenarios that can enable selection, triggering, and signaling of positioning training data to support an AI / ML framework for positioning. Disclosed herein are procedures for training machine learning positioning models in wireless communications networks. The procedures may be implemented by network nodes and methods in network nodes.

[0008] A network node in a wireless communications network is provided, the network node comprising: a transmitter configured to transmit a positioning training dataset configuration to at least one data source of the wireless communications network, the positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on a location of the network node; the network node further comprising: a receiver configured to receive a response from the at least one data source, the response including the positioning training dataset in the required data format; and the network node further comprising a processor configured to train the machine learning positioning model using the positioning training dataset.

[0009] A method in a network node is also provided, the network node being in a wireless communications network, the method including transmitting a positioning training dataset configuration to at least one data source of the wireless communications network, the positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on a location of the network node. The method further includes receiving a response from the at least one data source, the response including the positioning training dataset in the required data format. The method further includes training the machine learning positioning model using the positioning training dataset.

[0010] To describe the manner in which the advantages and features of the present disclosure may be obtained, the description of the present disclosure will be made by reference to several apparatus and methods that are illustrated in the accompanying drawings. Each of these drawings illustrates only certain aspects of the present disclosure and therefore should not be considered as limiting the scope of the disclosure. The drawings have been simplified for clarity and are not necessarily drawn to scale.

[0011] A method and apparatus for training a machine learning positioning model in a wireless communications network is now described, by way of example only, with reference to the accompanying drawings, in which: [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 illustrates an embodiment of a wireless communication system. [Figure 2] FIG. 1 illustrates an embodiment of a user equipment device. [Figure 3] FIG. 1 illustrates an embodiment of a network node. [Figure 4] FIG. 1 illustrates an embodiment of a system demonstrating NR beam-based positioning in a wireless communication network. [Figure 5]1 illustrates an embodiment of a multi-cell RTT positioning procedure for a wireless communication network. [Figure 6] FIG. 1 illustrates an embodiment of relative range estimation using an existing single g Node B RTT positioning framework. [Figure 7] FIG. 10 illustrates the RequestLocationInformation message body in an LPP message used by a location server to request positioning measurements. [Figure 8] FIG. 10 illustrates the ProvideLocationInformation message body in an LPP message used by a target device to provide positioning measurements. [Figure 9] FIG. 1 illustrates an embodiment of an AI / ML functional block diagram for RAN intelligence. [Figure 10] 1 illustrates an embodiment of a method in a network node, the network node being in a wireless communications network. [Figure 11] FIG. 1 illustrates an embodiment of a system illustrating entities / nodes capable of performing AI / ML training using positioning input data in a wireless communications network. [Figure 12] FIG. 1 illustrates an embodiment of LMF-based collection and training of positioning measurement data. [Figure 13] 1 illustrates an embodiment of a method and apparatus for NG-RAN-based collection and training of positioning data. [Figure 14] 1 illustrates an embodiment of a method and apparatus for UE-based collection and training of positioning data. [Figure 15] 1 illustrates an embodiment of a method and apparatus for request and response for positioning training dataset collection. [Figure 16] FIG. 1 illustrates an embodiment of an AI / ML direct positioning training database structure. [Figure 17]FIG. 1 illustrates an embodiment of an AI / ML-assisted positioning training database structure. DETAILED DESCRIPTION OF THE INVENTION

[0013] Direct artificial intelligence (AI)-based positioning and AI-assisted positioning methods can be leveraged to improve UE location accuracy performance within a 3GPP-defined positioning framework. Training AI / machine learning (ML) models using accurate and reliable positioning data is a key element of this process to achieve desired AI / ML positioning outputs. The disclosure herein details methods for supporting the selection, triggering, and signaling of positioning training data from configured data sources for both offline and online training of positioning data. In summary, these methods enable the selection and configuration of data sources depending on the network entity that performed the AI / ML training; the LMF configuring the collection of AI / ML training data for positioning purposes; generating training data sets depending on whether AI / ML direct (standalone) or AI / ML-assisted positioning is configured; or any combination thereof.

[0014] For purposes of this disclosure, positioning-related reference signals may be referred to as reference signals used in a positioning procedure / purpose, e.g., PRS, or may be based on existing reference signals, such as CSI-RS or SRS, to estimate the location of a target UE, and the target UE may be referred to as a device / entity to be located / positioned. In various embodiments, the term PRS may refer to any signal, such as a reference signal, that may or may not be used primarily for positioning.

[0015] For purposes of this disclosure, the target UE may be referred to as the subject UE, the location (absolute or relative) of which will be obtained by the network or by the UE itself.

[0016] For the purposes of this disclosure, the terms AI and ML are used interchangeably to refer to intelligent software components or systems.

[0017] As will be appreciated by those skilled in the art, aspects of the present disclosure may be embodied as a system, apparatus, method, or program product. Accordingly, the configurations described herein may be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or combining software and hardware aspects.

[0018] For example, the disclosed methods and apparatus may be implemented as a hardware circuit comprising custom very-large-scale integration ("VLSI") circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors, or other discrete components. The disclosed methods and apparatus may also be implemented within programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices, and the like. As another example, the disclosed methods and apparatus may include one or more physical or logical blocks of executable code, which may be organized as, for example, objects, procedures, or functions.

[0019] Furthermore, the methods and apparatus may take the form of a program product embodied in one or more computer-readable storage devices that store machine-readable code, computer-readable code, and / or program code, hereafter referred to as "code." The storage devices may be tangible, non-transitory, and / or non-transmittable. The storage devices may not embody signals. In some arrangements, the storage devices merely utilize signals to access the code.

[0020] Any combination of one or more computer-readable mediums may be used. The computer-readable medium may be a computer-readable storage medium. The computer-readable storage medium may be a storage device that stores the code. The storage device may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, holographic, micro-mechanical, or semiconductor system, apparatus, or device, or any suitable combination of the above.

[0021] More specific examples (a non-exhaustive list) of storage devices would include an electrical connection having one or more wires, a portable computer diskette, a hard disk, random-access memory ("RAM"), read-only memory ("ROM"), erasable programmable read-only memory ("EPROM" or flash memory), a portable compact disc read-only memory ("CD-ROM"), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the context of this document, a computer-readable storage medium may be any tangible medium that can contain or store a program for use by or associated with an instruction execution system, apparatus, or device.

[0022] Throughout this specification, a reference to a particular method or apparatus example, or similar language, means that the particular feature, structure, or characteristic described in connection with that example is included in at least one implementation of the methods and apparatus described herein. Thus, references to features of a particular method or apparatus example, or similar language, may, but do not necessarily, all refer to the same example and, unless otherwise specified, mean "one or more, but not all, examples." The terms "including," "comprising," and "having," and variations thereof, mean "including, but not limited to," unless otherwise specified. An enumerated list of items does not imply that any or all of the items are mutually exclusive unless otherwise specified. The terms "a," "an," and "the" also refer to "one or more," unless otherwise specified.

[0023] As used herein, a list with the conjunction "and / or" includes any single item in the list or a combination of items in the list. For example, a list of A, B, and / or C includes A only, B only, C only, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term "one or more of" includes any single item in the list or a combination of items in the list. For example, one or more of A, B, and C includes A only, B only, C only, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term "one of" includes one of, and only one of, any single item in the list. For example, "one of A, B, and C" includes A only, B only, or C only, and excludes A, B, and C. As used herein, "a member selected from the group consisting of A, B, and C" includes only one of A, B, or C, and excludes the combination of A, B, and C. As used herein, "a member selected from the group consisting of A, B, and C, and combinations thereof" includes A only, B only, C only, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C.

[0024] Furthermore, the illustrated features, structures, or characteristics described herein may be combined in any suitable manner. In the following description, numerous specific details are provided, such as examples of programming, software modules, user selections, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc., to provide a thorough understanding of the present disclosure. However, those skilled in the art will recognize that the disclosed methods and apparatuses may be practiced without one or more of the specific details, or with other methods, components, materials, etc. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0025] Aspects of the disclosed methods and apparatus are described below with reference to schematic flowchart illustrations and / or schematic block diagrams of methods, apparatus, systems, and program products. It will be understood that each block of the schematic flowchart illustrations and / or schematic block diagrams, and combinations of blocks in the schematic flowchart illustrations and / or schematic block diagrams, may be implemented by code. This code may be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce machines, whereby the instructions, executed by the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the schematic flowchart illustrations and / or schematic block diagrams.

[0026] The code may also be stored in a storage device that can instruct a computer, other programmable data processing apparatus, or other device to function in a particular manner, whereby the instructions stored on the storage device create an article of manufacture that includes instructions that implement the functions / acts specified in the schematic flowchart diagrams and / or schematic block diagrams.

[0027] The code may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus, or other device, resulting in a computer-implemented process such that the code running on the computer or other programmable apparatus provides a process for implementing the functions / acts specified in the schematic flowchart diagrams and / or schematic block diagrams.

[0028] The schematic flowchart diagrams and / or schematic block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, systems, methods, and program products. In this regard, each block in the schematic flowchart diagrams and / or schematic block diagrams may represent a module, segment, or portion of code, which contains one or more executable instructions of code for implementing the specified logical function(s).

[0029] It should also be noted that in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. Other steps and methods may be devised that are equivalent in function, logic, or effect to one or more blocks, or portions thereof, of the illustrated figures.

[0030] The description of an element in each figure may refer to the element in the preceding figure. Like numbers refer to the same element in all figures.

[0031] 1 illustrates an embodiment of a wireless communication system 100 for training a machine learning positioning model in a wireless communication network. In one embodiment, the wireless communication system 100 includes a remote unit 102 and a network unit 104. While a particular number of remote units 102 and network units 104 are shown in FIG. 1, those skilled in the art will understand that any number of remote units 102 and network units 104 may be included in the wireless communication system 100.

[0032] In one embodiment, the remote unit 102 may include a computing device such as a desktop computer, a laptop computer, a personal digital assistant (PDA), a tablet computer, a smartphone, a smart television (e.g., a television connected to the Internet), a set-top box, a game console, a security system (including security cameras), a vehicle-mounted computer, a network device (e.g., a router, a switch, a modem), an aircraft, a drone, etc. In some embodiments, the remote unit 102 includes a wearable device such as a smart watch, a fitness band, an optical head-mounted display, etc. Furthermore, the remote unit 102 may be referred to as a subscriber unit, a mobile, a mobile station, a user, a terminal, a mobile terminal, a fixed terminal, a subscriber station, a UE, a user terminal, a device, or by other terms used in the art. The remote unit 102 may communicate directly with one or more of the network units 104 via UL communication signals. In certain embodiments, the remote unit 102 may communicate directly with other remote units 102 via sidelink communication.

[0033] The network units 104 may be distributed throughout a geographic region. In certain embodiments, the network units 104 may include access points, access terminals, base stations, base stations, Node Bs, eNBs, gNBs, Home Node Bs, relay nodes, devices, core networks, air servers, radio access nodes, APs, NRs, network entities, Access and Mobility Management Functions (“AMFs”), Unified Data Management Functions (“UDMs”), Unified Data Repository (UDRs), UDM / UDRs, Policy Control Functions (“PCFs”), Location Management Functions (“LMFs”), Radio Access Networks (“RANs”), Network Slice Selection Functions (“NSSFs”), operations, administration, and management functions.and management ("OAM"), session management function ("SMF"), user plane function ("UPF"), application function, authentication server function ("AUSF"), security anchor functionality ("SEAF"), trusted non-3GPP gateway function ("TNGF"), application function, service enabler architecture layer ("SEAL") function, vertical application enabler server, edge enabler server, edge configuration server, mobile edge computing platform function, mobile edge computing application, application data analytics enabler server, SEAL data distribution server, middleware entity, network slice capability management server, or any other terminology used in the art. The network unit 104 is generally part of a radio access network including one or more controllers communicatively coupled to one or more corresponding network units 104. The radio access network is generally communicatively coupled to one or more core networks, which may be coupled to other networks such as the Internet and the public switched telephone network, among other networks. These and other elements of the radio access and core networks are not shown but are generally familiar to those skilled in the art.

[0034] In one implementation, the wireless communication system 100 conforms to the New Radio (NR) protocol standardized in 3GPP, with the network unit 104 transmitting on the downlink (DL) using an Orthogonal Frequency Division Multiplexing ("OFDM") modulation scheme and the remote unit 102 transmitting on the uplink (UL) using a Single Carrier Frequency Division Multiple Access ("SC-FDMA") scheme or an OFDM scheme. However, more generally, the wireless communication system 100 may implement any other open or proprietary communication protocol, such as WiMAX, IEEE 802.11 variants, GSM, GPRS, UMTS, LTE variants, CDMA2000, Bluetooth, ZigBee, Sigfox, among other protocols. This disclosure is not intended to be limited to any particular wireless communication system architecture or protocol implementation.

[0035] The network unit 104 may serve several remote units 102 within a serving area, e.g., a cell or a cell sector, via wireless communication links. The network unit 104 transmits DL communication signals to serve the remote units 102 in the time, frequency, and / or spatial domains.

[0036] FIG. 2 illustrates a user equipment device 200 that can be used to implement the methods described herein. The user equipment device 200 can be used to implement one or more of the solutions described herein. The user equipment device 200 conforms to one or more of the user equipment devices described in the embodiments herein. In particular, the user equipment device 200 can be the device 102 of FIG. 1, the device 450 of FIG. 4, the device 630 of FIG. 6, the device 1150 or 1160 of FIG. 11, or the devices 1410, 1420, and 1430 of FIG. 14; therefore, the reference numeral 200 will hereinafter be used to refer to the user equipment devices 102, 450, 630, 1150, 1160, 1410, 1420, and 1430. The user equipment device 200 includes a processor 205, a memory 210, an input device 215, an output device 220, and a transceiver 225.

[0037] The input device(s) 215 and the output device(s) 220 may be combined into one device, such as a touchscreen. In some implementations, the user equipment device 200 does not include any input device(s) 215 and / or output device(s) 220. The user equipment device 200 may include one or more of the processor 205, the memory 210, and the transceiver 225, and may not include the input device(s) 215 and / or the output device(s) 220.

[0038] As shown, the transceiver 225 includes at least one transmitter 230 and at least one receiver 235. The transceiver 225 may communicate with one or more cells (or wireless coverage areas) supported by one or more base units. The transceiver 225 may be capable of operating on an unlicensed spectrum. Moreover, the transceiver 225 may include multiple UE panels supporting one or more beams. Furthermore, the transceiver 225 may support at least one network interface 240 and / or application interface 245. The application interface 245 may support one or more APIs. The network interface 240 may support 3GPP reference points such as Uu, N1, PC5, etc. As will be appreciated by those skilled in the art, other network interfaces 240 may also be supported.

[0039] The processor 205 may include any known controller capable of executing computer-readable instructions and / or performing logical operations. For example, the processor 205 may be a microcontroller, microprocessor, central processing unit ("CPU"), graphics processing unit ("GPU"), auxiliary processing unit, field programmable gate array ("FPGA"), or similar programmable controller. The processor 205 may execute instructions stored in the memory 210 to implement the methods and routines described herein. The processor 205 is communicatively coupled to the memory 210, the input device 215, the output device 220, and the transceiver 225.

[0040] Processor 205 may control user equipment device 200 to implement the user equipment device behaviors described herein. Processor 205 may include an application processor (also known as a “main processor”) that manages application domain and operating system (“OS”) functions, and a baseband processor (also known as a “baseband radio processor”) that manages radio functions.

[0041] Memory 210 may be a computer-readable storage medium. Memory 210 may include a volatile computer storage medium. For example, memory 210 may include RAM, including dynamic RAM (“DRAM”), synchronous dynamic RAM (“SDRAM”), and / or static RAM (“SRAM”). Memory 210 may include a non-volatile computer storage medium. For example, memory 210 may include a hard disk drive, flash memory, or any other suitable non-volatile computer storage device. Memory 210 may include both volatile and non-volatile computer storage media.

[0042] The memory 210 may store relevant data for implementing the traffic category fields described herein. The memory 210 also stores program code and relevant data, such as an operating system or other controller algorithms running on the device 200.

[0043] The input device 215 may include any known computer input device, including a touch panel, buttons, a keyboard, a stylus, a microphone, etc. The input device 215 may be integrated with the output device 220, for example, as a touch screen or similar touch display. The input device 215 may include a touch screen such that text may be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. The input device 215 may include two or more different devices, such as a keyboard and a touch panel.

[0044] Output device 220 may be designed to output visual, audible, and / or tactile signals. Output device 220 may include an electronically controllable display or display device capable of outputting visual data to a user. For example, output device 220 may include, but is not limited to, a liquid crystal display ("LCD"), a light-emitting diode ("LED") display, an organic LED ("OLED") display, a projector, or similar display device capable of outputting images, text, etc. to a user. As another non-limiting example, output device 220 may include a wearable display that is separate from but communicatively coupled to the rest of user equipment device 200, such as a smartwatch, smart glasses, a head-up display, etc. Furthermore, output device 220 may be a component of a smartphone, a personal digital assistant, a television, a tablet computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, etc.

[0045] The output device 220 may include one or more speakers for producing sound. For example, the output device 220 may produce an audible alert or notification (e.g., a beep or chime). The output device 220 may include one or more haptic devices for producing vibration, movement, or other haptic feedback. All or some portions of the output device 220 may be integrated with the input device 215. For example, the input device 215 and the output device 220 may form a touchscreen or similar touch display. The output device 220 may be located near the input device 215.

[0046] The transceiver 225 communicates with one or more network functions of a mobile communications network via one or more access networks. The transceiver 225 operates under the control of the processor 205 to transmit messages, data, and other signals and to receive messages, data, and other signals. For example, the processor 205 may selectively activate the transceiver 225 (or portions thereof) at particular times to transmit and receive messages.

[0047] The transceiver 225 includes at least one transmitter 230 and at least one receiver 235. The one or more transmitters 230 may be used to provide uplink communication signals to a base unit of a wireless communication network. Similarly, the one or more receivers 235 may be used to receive downlink communication signals from the base unit. Although only one transmitter 230 and one receiver 235 are shown, the user equipment device 200 may have any suitable number of transmitters 230 and receivers 235. Furthermore, the transmitters 230 and receivers 235 may be any suitable type of transmitter and receiver. The transceiver 225 may include a first transmitter / receiver pair used to communicate with a mobile communication network over a licensed radio spectrum and a second transmitter / receiver pair used to communicate with a mobile communication network over an unlicensed radio spectrum.

[0048] A first transmitter / receiver pair used to communicate with a mobile communications network over a licensed radio spectrum and a second transmitter / receiver pair used to communicate with a mobile communications network over an unlicensed radio spectrum may be combined into a single transceiver unit, e.g., a single chip that performs functions for use with both licensed and unlicensed radio spectrum. The first transmitter / receiver pair and the second transmitter / receiver pair may share one or more hardware components. For example, some transceivers 225, transmitters 230, and receivers 235 may be implemented as physically separate components that access shared hardware and / or software resources, such as, for example, a network interface 240.

[0049] One or more transmitters 230 and / or one or more receivers 235 may be implemented and / or integrated in a single hardware component, such as a multi-transceiver chip, a system-on-chip, an application-specific integrated circuit ("ASIC"), or other type of hardware component. One or more transmitters 230 and / or one or more receivers 235 may be implemented and / or integrated in a multi-chip module. Other components, such as a network interface 240 or other hardware components / circuits, may be integrated into a single chip with any number of transmitters 230 and / or receivers 235. The transmitters 230 and receivers 235 may be logically configured as a transceiver 225 using one or more common control signals, or as modular transmitters 230 and receivers 235 implemented within the same hardware chip or multi-chip module.

[0050] Memory 210 of user equipment device 200 may contain or store a machine learning positioning model for training using a positioning training data set. The machine learning positioning model may be stored before and / or after training. Positioning training data sets received through receiver 235, from data sources local to user equipment device 200, or through measurements by user equipment device 200 may also be stored in memory 210. Processor 205, in some embodiments, is arranged to train the machine learning positioning model.

[0051] 3 shows further details of a network node 300 that may be used to implement the methods described herein. The network node 300 may be an implementation of an entity, such as 104 of FIG. 1, 420, 430, or 440 of FIG. 4, 610 or 620 of FIG. 6, 1135, 1141, 1150, 1160, or 1144 of FIG. 11, 1210 of FIG. 12, 1320 of FIG. 13, or 1410, 1420, or 1430 of FIG. 14, in a wireless communications network, e.g., in one or more of the wireless communications networks described herein. The network node 300 may be, for example, the UE 200 described above, or a network function (NF) or application function (AF), or another entity in one or more of the wireless communications networks of the embodiments described herein. The network node 300 includes a processor 305, a memory 310, an input device 315, an output device 320, and a transceiver 325.

[0052] The input device(s) 315 and the output device(s) 320 may be combined into one device, such as a touchscreen. In some implementations, the network node 300 does not include any input device(s) 315 and / or output device(s) 320. The network node 300 may include one or more of the processor 305, the memory 310, and the transceiver 325, and may not include the input device(s) 315 and / or the output device(s) 320.

[0053] As shown, the transceiver 325 includes at least one transmitter 330 and at least one receiver 335, where the transceiver 325 communicates with one or more remote units 300. Additionally, the transceiver 325 may support at least one network interface 340 and / or application interface 345. The application interface 345 may support one or more APIs. The network interface 340 may support 3GPP reference points such as Uu, N1, N2, and N3. As will be appreciated by those skilled in the art, other network interfaces 340 may also be supported.

[0054] The processor 305 may include any known controller capable of executing computer-readable instructions and / or performing logical operations. For example, the processor 305 may be a microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, or similar programmable controller. The processor 305 may execute instructions stored in the memory 310 to implement the methods and routines described herein. The processor 305 is communicatively coupled to the memory 310, the input device 315, the output device 320, and the transceiver 325.

[0055] The memory 310 may be a computer-readable storage medium. The memory 310 may include a volatile computer storage medium. For example, the memory 310 may include RAM, including dynamic RAM (“DRAM”), synchronous dynamic RAM (“SDRAM”), and / or static RAM (“SRAM”). The memory 310 may include a non-volatile computer storage medium. For example, the memory 310 may include a hard disk drive, a flash memory, or any other suitable non-volatile computer storage device. The memory 310 may include both volatile and non-volatile computer storage media.

[0056] The memory 310 may store data related to the establishment of multipath unicast links and / or mobile operations. For example, the memory 310 may store parameters, configurations, resource allocations, policies, etc., as described herein. The memory 310 also stores program code and associated data, such as operating systems or other controller algorithms, running on the network node 300.

[0057] The input device 315 may include any known computer input device, including a touch panel, buttons, a keyboard, a stylus, a microphone, etc. The input device 315 may be integrated with the output device 320, for example, as a touch screen or similar touch display. The input device 315 may include a touch screen such that text may be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. The input device 315 may include two or more different devices, such as a keyboard and a touch panel.

[0058] The output device 320 may be designed to output visual, audible, and / or tactile signals. The output device 320 may include an electronically controllable display or display device capable of outputting visual data to a user. For example, the output device 320 may include, but is not limited to, an LCD display, an LED display, an OLED display, a projector, or similar display device capable of outputting images, text, etc. to a user. As another non-limiting example, the output device 320 may include a wearable display that is separate from but communicatively coupled to the rest of the network node 300, such as a smartwatch, smart glasses, a head-up display, etc. Furthermore, the output device 320 may be a component of a smartphone, a personal digital assistant, a television, a tablet computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, etc.

[0059] The output device 320 may include one or more speakers for producing sound. For example, the output device 320 may produce an audible alert or notification (e.g., a beep or chime). The output device 320 may include one or more haptic devices for producing vibration, movement, or other haptic feedback. All or some portions of the output device 320 may be integrated with the input device 315. For example, the input device 315 and the output device 320 may form a touchscreen or similar touch display. The output device 320 may be located near the input device 315.

[0060] The transceiver 325 includes at least one transmitter 330 and at least one receiver 335. The one or more transmitters 330 may be used to communicate with a UE, as described herein. Similarly, the one or more receivers 335 may be used to communicate with a network function in a PLMN and / or RAN, as described herein. Although only one transmitter 330 and one receiver 335 are shown, the network node 300 may have any suitable number of transmitters 330 and receivers 335. Furthermore, the transmitters 330 and receivers 335 may be any suitable type of transmitter and receiver.

[0061] The memory 310 of the network node 300 may contain or store a machine learning positioning model for training using a positioning training dataset. The machine learning positioning model may be stored before and / or after training. Positioning training datasets received through the receiver 335, or from data sources local to the network node 300, or through measurements by the network node 300 may also be stored in the memory 310. The processor 305, in some embodiments, is arranged to train the machine learning positioning model.

[0062] Positioning requirements for 3GPP New Radio (NR) air interface (Uu) signals and standalone (SA) architectures (e.g., beam-based transmissions) are specified in 3GPP Release 16. Target use cases also included commercial and regulatory (emergency services) scenarios given in 3GPP Release 15. Performance requirements from Specification #38.855 are given in Table 1 and include a horizontal positioning error of less than 3 m for 80% of indoor UEs and less than 10 m for 80% of outdoor UEs. For vertical positioning, performance requirements include a vertical positioning error of less than 3 m for 80% of UEs and less than 3 m for 80% of outdoor UEs.

[0063] [Table 1]

[0064] Additionally, the current 3GPP Release 17 Positioning recently defined positioning performance requirements for commercial and industrial Internet of Things (IIoT) use cases. These performance requirements from Specification #38.857 are given in Table 2 and, for commercial use cases, include a positioning error of less than 1 m for horizontal positioning for 90% of UEs, less than 3 m for vertical positioning for 90% of UEs, less than 10 ms for physical layer latency for UE location estimation, and less than 100 ms for end-to-end latency for UE location estimation. For IIoT use cases, the performance requirements include a positioning error of less than 0.2 m for horizontal positioning for 90% of UEs, less than 1 m for vertical positioning for 90% of UEs, less than 10 ms for physical layer latency for UE location estimation, and less than 100 ms for end-to-end latency for UE location estimation (although 10 ms is preferred).

[0065] [Table 2]

[0066] The distinct positioning techniques in 3GPP Release 16 from specification #38.305 are shown in Table 3. These positioning techniques can now be configured and implemented based on the requirements of the Location Management Function (LMF) and UE capabilities.

[0067] [Table 3]

[0068] The transmission of a positioning reference signal (PRS) enables a UE to perform UE positioning-related measurements to enable calculation of a UE location estimate. The PRS is configured per transmission reception point (TRP), and the TRP may transmit one or more beams. FIG. 4 illustrates an embodiment of a system 400 demonstrating NR beam-based positioning available from 3GPP Release 16 onward. The system 400 includes a location server or LMF 410, a first g Node B TRP 420, a second g Node B TRP 430, a third g Node B TRP 440, and a UE 450. The PRS can be transmitted by different base stations 420, 430, 440 (serving and neighboring) using beams spanning frequency ranges / bands FR1 and FR2, which is significantly different compared to LTE, where the PRS was transmitted throughout the cell. The PRS may be locally associated with a PRS resource ID and resource set ID for the base station (TRP). Similarly, UE positioning measurements such as reference signal time difference (RSTD) and PRS reference signal received power (RSRP) measurements are made on a per-beam basis (e.g., based on downlink (DL) PRS resources or DL ​​PRS resource sets) as opposed to different cells as is the case in LTE. Furthermore, there are additional UL positioning methods for the network to leverage to calculate the location of the target UE. Tables 4 and 5 show the reference signal-to-measurement mappings required for each supported radio access technology (RAT)-dependent positioning technique at the UE and gNodeB, respectively.RAT-dependent positioning techniques involve 3GPP RATs and core network entities to perform UE position estimation, and these techniques are distinguished from RAT-independent positioning techniques that rely on GNSS, IMU sensors, WLAN, and Bluetooth technologies to perform target device (UE) positioning.

[0069] [Table 4]

[0070] [Table 5]

[0071] The RAT dependent positioning techniques supported since 3GPP Release 16 (specification #38.305) are briefly introduced in the following paragraphs.

[0072] The DL-TDOA positioning method utilizes DL RSTD (and optionally DL PRS RSRP) of downlink signals received at a UE from multiple TPs. The UE measures the DL RSTD (and optionally DL PRS RSRP) of the received signals using assistance data received from a positioning server, and the resulting measurements, together with other configuration information, are used to locate the UE relative to neighboring TPs.

[0073] The DL-AoD positioning method utilizes measured DL PRS RSRPs of downlink signals received at a UE from multiple TPs. The UE measures the DL PRS RSRPs of the received signals using assistance data received from a positioning server, and the resulting measurements, together with other configuration information, are used to locate the UE relative to neighboring TPs.

[0074] The multi-cell RTT positioning procedure 500 is shown in FIG. 5 and utilizes UE Rx-Tx measurements and DL PRS RSRPs of downlink signals received from multiple TRPs measured by the UE, and gNB Rx-Tx measurements and UL SRS-RSRPs of uplink signals transmitted from the UE measured at multiple TRPs. The UE measures the UE Rx-Tx measurements (and optionally the DL PRS RSRPs of the received signals) using assistance data received from the positioning server, and the TRPs measure the gNB Rx-Tx measurements (and optionally the UL SRS-RSRPs of the received signals) using assistance data received from the positioning server. The measurements are used to determine the RTT at the positioning server, which are used to estimate the UE's location. Multi-RTT is only supported for UE-assisted / NG-RAN-assisted positioning techniques, as described in Table 3. FIG. 6 provides an example 600 of relative range estimation using the existing single gNodeB RTT positioning framework. Shown in the figure are a positioning server 610 (LMF), a gNodeB 620, and multiple target UEs 630. Relative ranges 640 can be calculated between the UEs 630, and RTTs can be calculated at 650 to obtain absolute locations. Signals 660 are shown for UL-SRS 661, DL-SRS 662, as well as relative UE-to-UE distance / orientation 670.

[0075] In the enhanced Cell ID (CID) positioning method, the UE's location is estimated with knowledge of its serving ng-eNB, gNB, and cells and is based on LTE signals. Information about the serving ng-eNB, gNB, and cells can be obtained by paging, registration, or other methods. NR Enhanced Cell ID (NR E-CID) positioning refers to a technique that uses additional UE measurements and / or NR radio resources and other measurements to improve the UE location estimate using NR signals. NR E-CID positioning may use some of the same measurements as the measurement control system in the RRC protocol, but the UE is generally not expected to perform additional measurements for the sole purpose of positioning; i.e., the positioning procedure does not provide measurement configuration or measurement control messages, and the UE reports the measurements it has prepared rather than being required to take additional measurement actions.

[0076] The UL TDOA positioning method utilizes UL TDOA (and optionally UL SRS-RSRP) at multiple RPs of uplink signals transmitted from a UE. The RPs measure the UL TDOA (and optionally UL SRS-RSRP) of the received signals using assistance data received from a positioning server, and the resulting measurements are used together with other configuration information to estimate the location of the UE.

[0077] The UL AoA positioning method utilizes measured azimuth and zenith angles of arrival at multiple RPs of uplink signals transmitted from a UE. The RPs measure the A-AoA and Z-AoA of the received signals using assistance data received from a positioning server, and the resulting measurements are used, along with other configuration information, to estimate the location of the UE.

[0078] According to 3GPP Release 16, the PRS can be transmitted by different base stations (serving and neighboring) using narrow beams across frequency bands FR1 and FR2, which is significantly different compared to LTE, where the PRS was transmitted across the entire cell. The PRS may be locally associated with a PRS resource ID and resource set ID for the base station (TRP). Similarly, UE positioning measurements, such as reference signal time difference (RSTD) and PRS RSRP measurements, are performed across beams (e.g., between different pairs of DL PRS resources or DL ​​PRS resource sets), as opposed to different cells as in LTE. Furthermore, there are additional UL positioning methods for the network to leverage to calculate the location of the target UE.

[0079] The different DL measurements required for supported RAT dependent positioning techniques, including DL PRS-RSRP, DL RSTD, and UE Rx-Tx time difference, are shown in Table 6. The measurement configuration specified in 3GPP specification #38.215 includes that four pairs of DL RSTD measurements may be performed per pair of cells, each measurement between a different pair of DL PRS resources / resource sets with a single reference timing, and eight DL PRS RSRP measurements may be performed on different DL PRS resources from the same cell.

[0080] [Table 6A]

[0081] [Table 6B]

[0082] The following paragraphs briefly introduce a RAT-independent positioning technique that relies on GNSS, IMU sensors, WLAN and Bluetooth technologies to perform target device (UE) positioning (3GPP specification #38.305).

[0083] Network-assisted GNSS methods utilize a UE equipped with a radio receiver capable of receiving GNSS signals. In 3GPP specifications, the term GNSS encompasses both terrestrial and regional / augmented navigation satellite systems. Examples of global navigation satellite systems include GPS, modernized GPS, Galileo, GLONASS, and the BeiDou Navigation Satellite System (BDS). Regional navigation satellite systems include the Quasi-Zenith Satellite System (QZSS), and many augmentation systems are categorized under the collective term Space Based Augmentation Systems (SBAS) and provide regional augmentation services. Different GNSS (e.g., GPS, Galileo, etc.) can be used separately or in combination to determine the location of a UE.

[0084] Barometric pressure sensor positioning utilizes an atmospheric pressure sensor to determine the vertical component of the UE's position. The UE measures the barometric pressure, optionally aided by assistance data, to calculate the vertical component of its location or transmits the measurements to a positioning server for position calculation. This method should be combined with other positioning methods to determine the UE's 3D position.

[0085] The WLAN positioning method utilizes WLAN measurements (AP identifiers and optionally other measurements) and a database to determine the location of the UE. The UE, optionally aided by assistance data, measures received signals from WLAN access points and sends the measurements to a positioning server for position calculation. Using the measurement results and a reference database, the UE's location is calculated. Alternatively, the UE utilizes WLAN measurements and optionally WLAN AP assistance data provided by the positioning server to determine its location.

[0086] The Bluetooth positioning method utilizes Bluetooth measurements (beacon identifiers and optionally other measurements) to determine the location of a UE. The UE measures the received signals from Bluetooth beacons. Using the measurements and a reference database, the location of the UE is calculated. The Bluetooth method can be combined with other positioning methods (e.g., WLAN) to improve the accuracy of UE positioning.

[0087] A Terrestrial Beacon System (TBS) consists of a network of ground-based transmitters that broadcast signals for positioning purposes only. Current types of TBS positioning signals are MBS (Metropolitan Beacon System) signals and Positioning Reference Signals (PRS) (3GPP Specification #36.211). A UE measures received TBS signals, optionally aided by assistance data, to calculate its location or transmits the measurements to a positioning server for position calculation.

[0088] The motion sensor positioning method uses different sensors, such as accelerometers, gyros, and magnetometers, to calculate the displacement of the UE. The UE estimates the relative displacement based on a reference position and / or a reference time. The UE transmits a report containing the determined relative displacement, which can be used to determine the absolute position. This method should be used together with other positioning methods for hybrid positioning.

[0089] Measurements and reporting are performed for each configured RAT-dependent / RAT-independent positioning method. The overall measurement configuration and reporting is shown in Figures 7 and 8. In particular, Figure 7 provides an example 700 of a RequestLocationInformation message body in an LPP message used by a location server to request positioning measurements or position estimates from a target device, such as the UE 102. Figure 8 provides an example 800 of an instantiation of a ProvideLocationInformation message body in an LPP message used by a target device (such as the UE 102) to provide positioning measurements or position estimates to a location server.

[0090] AI / ML offers opportunities for improving positioning accuracy within wireless communication networks. Some opportunities and use cases are briefly introduced here as follows:

[0091] Direct AI / ML positioning where the output of the AI / ML model inference (e.g., fingerprinting based on channel observations as input to the AI / ML model; details of channel observations as input to the AI / ML model, such as CIR, RSRP and / or other types of channel observations, require further study; applicable scenarios and generalization aspects of the AI / ML model require further study) is the UE location.

[0092] AI / ML-aided positioning where the output of the AI / ML model inference is a new measurement and / or an improvement of an existing measurement (e.g., LOS / NLOS discrimination, timing and / or angle or measurement, likelihood of measurement; details of inputs and outputs for the corresponding AI / ML model require further study; applicable scenarios and generalization aspects of the AI / ML model require further study).

[0093] Additionally, there is potential impact of 3GPP specifications on aspects of AI / ML techniques for improving positioning accuracy, including, but not limited to, AI / ML model training (training data type / size, training data source determination, e.g., UE / PRU / TRP, supporting signaling and procedures for training data collection), AI / ML model instruction / configuration (e.g., model configuration, model activation / deactivation, model restoration / termination, supporting signaling and procedures for model selection), AI / ML model monitoring and update (e.g., model performance monitoring, supporting signaling and procedures for model update / tuning), AI / ML model inference input (reporting / feedback of model input for inference, e.g., network-side model inference, model input acquisition and preprocessing, UE feedback as input for model input type / definition), AI / ML model inference output (reporting / feedback of model inference output, post-processing of model inference output), UE capabilities for AI / ML models (e.g., for model training, model inference, and model monitoring). However, not all of these aspects may be applicable to all AI / ML techniques in a particular use case.

[0094] The opportunities and use cases discussed herein for improving positioning in wireless communication networks require the development of several aspects of the positioning framework. In the current and described positioning framework, there is no known method for selecting, triggering, and configuring the collection of positioning data required to train a configured AI / ML positioning model. This includes both online and offline training of positioning data for a particular AI / ML model. Furthermore, the type of data source is important for the input training dataset to be statistically significant, which impacts the accuracy of the AI / ML model to be trained. This application presents a solution to this problem.

[0095] 9 provides an illustration of an embodiment of an AI / ML functional block diagram 900 within a functional framework for RAN intelligence, which incorporates common terminology associated with the functional framework for RAN intelligence. Functional block diagram 900 includes data collection 910, model training 920, model inference 930, and actors 940.

[0096] Data collection 910 is a function that provides input data to the model training 920 and model inference 930 functions. AI / ML algorithm-specific data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) is not performed in the data collection 910 function. Examples of input data can be measurements from a UE or different network entities, feedback 941 from actors 940, and output from an AI / ML model. Training data 911 is data required as input for the AI / ML model training 920 function. Inference data 912 is data required as input for the AI / ML model inference 930 function.

[0097] Model Training 920 is a function that performs ML model training, validation, and testing, which may generate model performance metrics as part of the model checking procedure. The Model Training 920 function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and conversion), if necessary, based on the training data 911 provided by the Data Collection 910 function. Model Deployment / Update 921 is used to initially deploy a trained, validated, and tested AI / ML model to the Model Inference 930 function, or to provide an updated model to the Model Inference 930 function.

[0098] Model inference 930 is a function that provides an AI / ML model inference output 931 (e.g., a prediction or decision). It has yet to be determined whether to provide model performance feedback 932 to the model training 920 function. The model inference 930 function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation), if necessary, based on the inference data 912 provided by the data collection 910 function. The details of the inference output 931 are use case specific. It has also yet to be determined whether model performance feedback 932 is applied if particular information derived from the model inference 930 function is suitable for improving the AI / ML model trained in the model training 920 function. Feedback 941 from actors 940 or other network entities (via the data collection 910 function) may be required in the model inference 930 function to create the model performance feedback 932.

[0099] Actors 940 are functions that receive output 931 from the model inference 930 function and trigger or perform corresponding actions. Actors 940 may trigger actions directed at other entities or at themselves. Feedback 941 is training or inference data or information that may be needed to derive performance feedback.

[0100] The network nodes and methods described herein allow for the selection and configuration of data sources depending on the type of training of the positioning data and which network entity is performing the training of the AI / ML positioning model.

[0101] Described herein is a network node in a wireless communications network, comprising: a transmitter configured to transmit a positioning training dataset configuration to at least one data source of the wireless communications network, the positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on a location of the network node. The network node further comprises a receiver configured to receive a response from the at least one data source, the response including the positioning training dataset in the required data format. The network node further comprises a processor configured to train the machine learning positioning model using the positioning training dataset.

[0102] In some embodiments, the at least one data source comprises a data source selected from a group of data sources consisting of a serving or neighboring gNodeB, a serving or neighboring transmitting reception point, a positioning reference unit user equipment, a positioning reference unit transmitting reception point, a reference user equipment, a target user equipment, an assisting user equipment, a third party user equipment, a network data analysis function, and a location management function.

[0103] In some embodiments, the network node is selected from the group of network nodes consisting of a location server node, an advanced radio access network node, a positioning reference unit g Node B, a positioning reference unit transmission reception point, a positioning reference unit user equipment, a reference user equipment, an assisting user equipment, a third party user equipment, and a target user equipment. The advanced radio access network node may include a serving or neighboring g Node B or a serving or neighboring transmission reception point. The serving or neighboring g Node B may be applicable to terrestrial and / or non-terrestrial based networks.

[0104] In some embodiments, the positioning training data set configuration includes a reporting configuration that is instantaneous, periodic, event-based, or a combination thereof.

[0105] In some embodiments, the machine learning positioning model includes a machine learning direct positioning model (e.g., input positioning training data is provided to the machine learning model to output a final location estimate of the target UE) or a machine learning assisted positioning model (e.g., input positioning training data is provided to the machine learning model to output enhanced positioning-related measurements or metrics that are used to ultimately calculate a location estimate of the target UE). Some machine learning positioning models that are capable of both types of training (direct and assisted) may be used. However, the trained data used to develop the machine learning positioning model may differ depending on whether the model is direct or assisted.

[0106] In some embodiments, the required data format is structured according to the type of machine learning positioning model.

[0107] In some embodiments, the required data format includes a fingerprint number, a positioning measurement, configurable location information for the location where the positioning measurement is performed, time information for the positioning measurement, quality of the positioning measurement, or any combination thereof. The configurable location information may include at least one of 2D or 3D coordinates, a zone identifier, a grid identifier, a speed or velocity, an orientation, a bearing, a height, a length, and a rectangular grid having a width. The coordinates may include geodetic or spherical coordinates. The length and width may be configured depending on the location accuracy granularity.

[0108] In some embodiments, the receiver is further arranged to receive an activation request to activate transmission of a positioning training dataset configuration, reception of a positioning training dataset and / or training of a machine learning positioning model.

[0109] In some embodiments, the receiver is further arranged to receive a deactivation request to deactivate the transmission of the positioning training dataset configuration, the reception of the positioning training dataset and / or the training of the machine learning positioning model.

[0110] In some embodiments, the transmitter is further arranged to transmit an output of the trained machine learning positioning model to the first network node.

[0111] In some embodiments, the processor is further arranged to train the machine learning positioning model using a positioning training dataset received from a local data source.

[0112] In some embodiments, the training dataset configuration further includes an indication as to whether the positioning training dataset is intended for online and / or offline training of a machine learning positioning model.

[0113] 10 illustrates an embodiment of a method 1000 in a network node in a wireless communications network. In a first step 1010, a positioning training dataset configuration is sent to at least one data source of the wireless communications network, the positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on the location of the network node. In a second step 1020, a response is received from the at least one data source, the response including the positioning training dataset in the required data format. In a third step 1030, the machine learning positioning model is trained using the positioning training dataset.

[0114] In some embodiments, the at least one data source comprises a data source selected from a group of data sources consisting of a serving or neighboring gNodeB, a serving or neighboring transmitting reception point, a positioning reference unit user equipment, a positioning reference unit transmitting reception point, a reference user equipment, a target user equipment, an assisting user equipment, a third party user equipment, a network data analysis function, and a location management function.

[0115] In some embodiments, the network node is selected from the group of network nodes consisting of a location server node, a next generation radio access network node, a positioning reference unit g Node B, a positioning reference unit transmission reception point, a positioning reference unit user equipment, an assisting user equipment, a third party user equipment, a reference user equipment, and a target user equipment. The next generation radio access network node may include a serving or neighboring g Node B or a serving or neighboring transmission reception point.

[0116] In some embodiments, the positioning training data set configuration includes a reporting configuration that is instantaneous, periodic, event-based, or a combination thereof.

[0117] In some embodiments, the machine learning positioning model is a machine learning direct positioning model or a machine learning assisted positioning model.

[0118] In some embodiments, the required data format is structured according to the type of machine learning positioning model.

[0119] In some embodiments, the required data format includes a fingerprint number, a positioning measurement, configurable location information of the location where the positioning measurement is performed, time information of the positioning measurement, quality of the positioning measurement, or any combination thereof.

[0120] In some embodiments, the configurable location information includes at least one of 2D or 3D coordinates, a zone identifier, a grid identifier, a speed or velocity, an orientation, a heading, a rectangular grid having a height, a length, and a width.

[0121] Some embodiments further include transmitting a positioning training dataset configuration, receiving a positioning training dataset and / or receiving an activation request to activate training of a machine learning positioning model.

[0122] Some embodiments further include receiving a deactivation request to deactivate the transmission of the positioning training dataset configuration, the reception of the positioning training dataset and / or the training of the machine learning positioning model.

[0123] Some embodiments further include transmitting an output of the trained machine learning positioning model to the first network node.

[0124] Some embodiments further include training the machine learning positioning model using a positioning training dataset received from a local data source, which may be a data source at the network node, such as stored position measurement information, or the local data source may include acquisition of position information and / or measurements by the network node (e.g., when the network node is a UE or a gNodeB).

[0125] In some embodiments, the training dataset configuration further includes an indication as to whether the positioning training dataset is intended for online and / or offline training of a machine learning positioning model.

[0126] FIG. 11 illustrates an embodiment of a system 1100 illustrating entities / nodes within a positioning network setup capable of performing AI / ML training using positioning input data in a wireless communications network. The entities / nodes may also serve as configured data sources for providing positioning training datasets to other entities / nodes. The system 1100 includes an LCS client 1110 and application function 1120, a 5G core network 1130, a radio access network (NG-RAN) 1140, a target UE 1150 and LCS client 1151, and a PRU-UE 1160. The 5G core network 1130 includes a GMLC / LRF 1131, a UDM 1132, a NEF 1133, an AMF 1134, and an LMF 1135. The radio access network 1140 includes a first neighboring g Node B 1141, a second neighboring g Node B 1142, a serving g Node B 1143, and a PRU 1144 as a TRP. The LMF 1135 of the 5G core network 1130 may perform LMF training using the positioning data set. The gNodeBs 1141, 1142, 1143 may perform training using the positioning data set. The target UE 1150 may perform training using the positioning data set. The PRU UE 1160 may perform training using the positioning data set. The PRU TRP 1144 may perform training using the positioning data set.

[0127] An embodiment in which the LMF 1135 uses the positioning training dataset to train the machine learning positioning model will now be described in more detail. The LMF 1135 is supported to request training datasets from one or more configured data sources. In another implementation, the LMF 1135 may request a partial training dataset from data source A and another partial training dataset from data source B. The configured data sources may be one or more of the following entities, including serving and / or neighboring Node Bs / TRPs, such as 1141, 1142, 1143, 1144, positioning reference units, reference UEs, or NWDAFs as UEs and / or TRPs, such as 1160 and 1144.

[0128] Serving and / or neighboring gNBs / TRPs such as 1141, 1142, 1143, 1144 may act as data sources and provide a training dataset in the form of UL-based positioning measurements consisting of UL-RTOA, gNB Rx-Tx time difference measurements, UL-AoA (including LCS-GCS translation information), SRS-RSRP, and SRS-RSRPP (per-route RSRP). The training dataset should consist of ground truth measurements including at least location information associated with each of the measurements, and the quality of the measurements may vary depending on the type of training; for example, online positioning training data may require different location granularity and update intervals compared to positioning data trained in an offline manner.

[0129] In another implementation option, PRU TRPs such as 1144 may also act as data sources and be uniformly distributed across different indoor / outdoor scenarios such that ground truth measurements for each area of ​​the indoor / outdoor scenario are captured based on the fixed location of the deployed PRU TRPs such as 1144; for example, an example of such a scenario could be an IIoT indoor factory scenario where TRPs are deployed in a distributed manner across a work site.

[0130] The PRU 1160 as a UE may also act as a data source for collecting ground truth measurements, including DL-based measurements such as RSTD, UE Rx-Tx time difference measurements, PRS RSRP, and PRS RSRPP (path-specific RSRP), and UEs based on different configured location granularity may be applicable to both offline and online training of AI / ML models. Similarly, a reference UE may also collect information related to training datasets with respect to pre-classifying measurements as LOS and / or NLOS. This may be applicable to both offline and online training of AI / ML models.

[0131] A normal positioning UE (also referred to as a target UE) may also be configured to provide positioning training data on a best-effort basis to update an offline training database of measurements. This may be applicable when an AI / ML model may need to be updated based on real-time environmental changes to measurements not captured during the offline training phase.

[0132] The type of location information associated with each positioning measurement in the training data set may consist of one or more of the following: UE 2D (x,y) or 3D (x,y,z) coordinates, including geodetic or spherical or latitude / longitude coordinates; Zone ID / Grid ID; UE speed / velocity; orientation; heading; height; and antenna array location information.

[0133] In an enhanced implementation, the configured data source may be provided with location information based on RAT-independent positioning measurements / methods such as, for example, GNSS, Bluetooth, WiFi, Inertial Measurement Units (IMUs), e.g., accelerometers, gyroscopes, etc.

[0134] 12 illustrates an embodiment 1200 of location server (LMF)-based collection and training of positioning measurement data, including a systematic signaling procedure for LMF-based training. The configured data source may be one or more of the following entities: a serving and / or neighboring gNB / TRP 1220, a positioning reference unit 1230 as a TRP, a positioning reference unit 1240 as a UE, a reference UE 1250, and a target UE 1260. In some implementations, separate measurement configurations for AI / ML positioning purposes may be configured for the aforementioned data sources, which may differ from measurement configurations for non-AI / ML positioning methods. These may include the type of positioning method, different PRS configurations including different comb patterns, number of symbols, repetition, QCL assumptions, muting configurations, periodicity, or a combination thereof.

[0135] The steps in an embodiment 1200 of LMF-based collection and training of positioning measurement data will now be described.

[0136] In a first step 1201, a request for training data (ground truth data collection) is made. In this step 1201, the location server node 1210 (e.g., LMF) requests a training data set from the NG-RAN nodes 1220 and 1230, including an indication whether the requested training data (measurement data) is to be used for offline and / or online training and a location for each measurement. The request may include a request for RAT-independent and UL-based positioning measurements. The request may be signaled using the NRPPa interface. In a further step 1202, the location server node 1210 (e.g., LMF) requests a training data set from the UE node (PRU UE 1240, reference UE 1250, and / or target UE 1260), including an indication whether the requested training data (measurement data) is to be used for offline and / or online training and a location for each measurement. The request may include RAT-independent and DL-based positioning measurements. This request may be signaled using LPP.

[0137] In a further step 1203, the UE nodes 1240, 1250, 1260 perform measurements over a period of time, providing each location with statistical significance or providing the stored measurements with the required timestamp and location of each stored location.

[0138] In a further step 1204, the gNB nodes 1220 and 1230 perform measurements over a period of time, giving each location statistical significance or being prepared to provide stored measurements with a requested timestamp and location for each stored location.

[0139] In a further step 1205, the UE nodes 1240, 1250, 1260 respond with a set of structured training data consisting of measurements and the location of each measurement. This response may be signaled using LPP.

[0140] In a further step 1206, the gNB nodes 1220 and 1230 respond with a set of structured training data consisting of measurements and the timestamp / location of each measurement. This response may be signaled using NRPPa.

[0141] In a further step 1207, the LMF 1210 performs training using the collected data from the data sources.

[0142] In an alternative extended implementation, the type of learning and the AI / ML model for positioning may also be signaled along with the data collection configuration. These may include the type of learning model, such as unsupervised or supervised learning, clustering, classification, dimensionality reduction, regression, etc. These may further include the type of AI / ML model, such as deep neural network, support vector machine (SVM), k-nearest neighbor (KNN) classifier, etc.

[0143] According to some embodiments, the NG-RAN node and / or PRU as a TRP is supported to request training datasets from one or more of the configured data sources. Training at the NG-RAN node and / or PRU as a TRP provides the opportunity to use both RAT-independent and UL-based positioning measurement data to train AI / ML models at the gNB side, which can save signaling overhead for training datasets in some scenarios. Figure 13 shows an embodiment 1300 of NG-RAN node-based training and signaling.

[0144] The configured data source may be one or more of the following entities including a positioning reference unit as the UE 1330, a reference UE 1340, and a target UE 1350. In another implementation, the NG-RAN node 1320 may separately request a positioning dataset from the LMF 1310 that includes measurements / data received from other cells as additional information.

[0145] The steps in the NG-RAN node-based training embodiment 1300 are as follows:

[0146] In a first step 1301, the LMF 1310 may activate the collection and training of positioning AI / ML models in the desired NG-RAN node / PRU TRP 1320. This may include requesting the output of the trained models. This step 1301 may be signaled using NRPPa.

[0147] In a further step 1302, the NG-RAN node 1320 may request collection of positioning training data from configured data sources (such as UEs 1330, 1340, 1350), where the data may include RAT-independent positioning measurements / location estimates and SRS configurations for UEs to transmit SRS for measurement collection and training purposes at the NG-RAN node 1320. This request may use RRC.

[0148] In a further step 1303, a response to the NG-RAN 1320 request is given by the UE (e.g., 1330, 1340, 1350) comprising reporting RAT independent positioning measurements / location estimates, e.g., RAT independent location information reported using GNSS reported using MDT, RRC CommonLocationInfo message, as well as transmission of SRS for positioning. In another implementation, timing information, e.g., in the form of a timestamp, may also be reported with the data point comprising the location information in addition to the location information where the SRS was transmitted.

[0149] In a further step 1304, the NG-RAN node 1320 performs measurements and collects the measurements.

[0150] In a further step 1305, the NG-RAN node 1320 performs training on the collected measurements.

[0151] In a further step 1306, the NG-RAN node 1320 may forward the output of the trained model depending on the type of AI / ML positioning task, e.g., for direct AI / ML positioning or AI / ML-assisted positioning. This output may be a location estimate or an enhanced measurement result. This output may use NRPPa.

[0152] In a further step 1307, the LMF 1310 may deactivate the collection and training of AI / ML models for positioning. This deactivation may use NRPPa.

[0153] In some embodiments, the PRU-UE and target UE nodes are supported to request positioning training data from one or more of the configured data sources. Training at the PRU UE as a node provides an opportunity to perform extensive ground truth collection, including measurements and associated locations. In another aspect, performing training for a UE-based positioning method supported by the target UE may help update the ground truth positioning database at several locations where the UE reports measurements. Figure 14 provides an illustration of an embodiment 1400 of UE-based collection and training of positioning data and a systematic signaling procedure for UE node-based training of positioning. The configured data sources may be one or more of the following entities: the target UE 1410, the PRU UE 1420, other UEs 1440, including the reference UE 1430, and a location server (LMF) 1450. The steps shown in embodiment 1400 are as follows:

[0154] In a first step 1401, the UE 1410, 1420, 1430 may perform configured SL(PC5) / DL PRS measurements over a specific period and location and / or use stored SL(PC5) / DL PRS measurements from the past (the period of past historical measurements may also be further configured during the training phase based on the historical measurements).

[0155] In a further step 1402, the UEs 1410, 1420, 1430 may use the PC5 interface to request positioning training data (consisting of SL or DL ​​positioning measurements or a combination thereof) from other UEs / PRU UEs 1440. The positioning training data set may be based on online / offline training RAT-independent and DL-based positioning measurements.

[0156] In a further step 1403, the UEs 1410, 1420, 1430 may use the LPP interface to request positioning training data from the LMF 1450. The positioning training data may include DL and / or SL measurements, if available. The request for the positioning training data set may be based on online / offline training.

[0157] In a further step 1404, the UEs 1410, 1420, 1430 receive responses with the requested measurement / positioning data from other surrounding / neighboring UEs 1440. Among the different responses, the UEs 1410, 1420, 1430 may receive an indication that the requested training data set is unavailable or will become available at a future time period / moment. The responses may use a PC5 interface.

[0158] In a further step 1405, the UE 1410, 1420, 1430 receives a response with the requested measurement / positioning data from the LMF 1450. Among the different responses, the UE 1410, 1420, 1430 may receive an indication that the requested training data set is unavailable or will become available at a future time period / moment. The response may use the LPP interface.

[0159] In a further step 1406, the UEs 1410, 1420, 1430 perform training based on collected data from configured data sources.

[0160] In other implementations, the positioning data to be collected may consist of DL-based or SL-based measurements or a combination thereof.

[0161] In some embodiments, a service request trigger is used to perform positioning data collection. Because the LMF / NG-RAN node / UE is the consumer of measurement data for AI / ML training purposes, a separate AI / ML positioning data collection service request should be managed by the LMF. A common framework for data collection should be established by the variable nodes where training may be performed, where the LMF may act as a centralized coordination entity. According to FIG. 13, in the case of training of an NG-RAN node 1320, the LMF 1310 may activate 1301 and deactivate 1307 the collection of positioning data and the training of AI / ML models in the NG-RAN node 1320. Similarly, for all embodiments described herein, the LMF may initiate the configuration of data sources for training data collection. Data sources may be identified by separate capabilities that verify the UE's ability to collect positioning data / measurements for AI / ML model training purposes. FIG. 15 shows an example description 1500 of such a request and response for positioning training dataset collection.

[0162] In a first step 1501, a location server (LMF) 1510 requests positioning training dataset collection capability from at least one data source, which may include a PRU UE 1520, a reference UE 1530, and a normal UE 1540. The request may use an LPP interface.

[0163] In a further step 1502, the UEs 1520, 1530, 1540 provide a response to the LMF 1510 with their positioning training data set collection capabilities. The response may use the LPP interface.

[0164] In an enhanced implementation, capability request 1501 and response 1502 distinguish between the ability to collect training data for online and / or offline training, support for direct AI / ML methods (e.g., fingerprinting), AI / ML-assisted methods (AI / ML-assisted DL-TDOA, AoD, RTT), and providing a known location and associated location determination source. In another implementation, these capabilities may be requested and responded for each positioning method for AI / ML and non-AI / ML positioning method. Non-AI / ML methods refer to positioning methods that do not require AI / ML support.

[0165] The data collection reporting configuration may be designed according to the following types of reports (which may also be requested by the LMF 1510): Real-time (instantaneous) positioning data collection reports based on an immediate request from the LMF 1510 with a configured response time, or the LMF 1510 may configure the UEs 1520, 1530, 1540 to store all measurements and ground truth for a specific time period (e.g., 24 hours), and the UEs 1520, 1530, 1540 may then report all data to the LMF 1510 upon request; Periodic data collection reports based on reporting interval and reporting volume parameters, or the periods for measuring / logging and reporting may be the same or different; Event-based data collection reports based on event change scenarios, for example, when the UE 1540 moves into a new positioning area with a different area ID, or when the PRU UE 1520 moves another 10 cm or 1 cm away from the last position reported to the LMF 1510.

[0166] The structure of an embodiment of a positioning training dataset is described herein, which may facilitate various AI / ML-based positioning schemes, including, for example, fingerprinting and / or AI / ML-assisted positioning, e.g., direct AI / ML positioning using intermediate positioning measurement optimization. Furthermore, the input training dataset structure varies depending on the AI / ML model type and task (classification, prediction (regression algorithm)). The training dataset structure may impact the size of the training dataset (e.g., payload) and the type of signaling, e.g., user plane or control plane signaling, to be configured to transfer such information. For this solution, the user plane (new LCS user plane protocol) or control plane signaling (e.g., LPP) may be configured to transfer the training dataset configuration.

[0167] The training data set can be structured as an N-dimensional database of features. In the case of direct AI / ML positioning and / or AI / ML-assisted positioning, the fingerprint database may be utilized to directly determine the location of the UE.

[0168] For example, for AI / ML direct positioning, the training database (dataset structure / configuration / required data format) may be exemplified as shown in Figure 16. The AI / ML directional positioning training database structure 1600 is an N-dimensional database / configuration, where N is the number of column-wise features / labels associated with fingerprints including multiple fingerprints 1610, and for each fingerprint 1610, location information 1620 (at configurable granularity), positioning measurements 1630, timestamps 1640, and quality 1650 are provided.

[0169] The positioning measurements 1630 may include at least one of DL RSTD (DL-based measurement), DL PRS RSRP (DL-based measurement), DL PRS RSRPP (DL-based measurement), UE Rx-Tx Time Difference (DL-based measurement), SS-RSRP (RSRP for RRM), SS-RSRQ (for RRM), CSI-RSRP (for RRM), CSI-RSRQ (for RRM), SS-RSRPB (for RRM) (DL-based measurement), UL RTOA (UL-based measurement), UL SRS-RSRP (UL-based measurement), UL SRS-RSRPP (UL-based measurement), UL RTOA (UL-based measurement), gNB Rx-Tx Time Difference UL RTOA (UL-based measurement), UL-AoA and UL-ZoA UL RTOA (UL-based measurement) (depending on whether the UE or the NG-RAN node performs the measurements).

[0170] In the case of UL-based measurements, the UE location of the SRS transmission may also be given. The timestamp 1640 may be implemented using a time base defined in units of milliseconds, seconds, minutes, hours, etc. Additionally, in other implementations, absolute time units may also be used, such as, for example, UTC time.

[0171] The database 1600 may be signaled using either control plane or user plane signaling. For control plane signaling, an example signaling mechanism may include LPP signaling, and for user plane signaling, a new protocol, such as LCUP (LCS User plane), may be used.

[0172] As a further example, for an AI / ML-assisted positioning training database structure, an illustration is shown in Figure 17. Dataset structure / configuration / required data format 1700 includes multiple fingerprints 1710 and, for each fingerprint, location information (at configurable granularity) 1720, channel impulse response peak power (CIR) 1730, number of paths 1740, and time (window) range 1750. Features of the channel impulse response 1730 are extracted with respect to generating ground truth measurements.

[0173] The locations 1720 may be specific points in 2D or 3D space, or in another implementation, the locations 1720 may consist of a 2D or 3D rectangular grid with length L, width W, and height H, where L, W, and H may be configured depending on the location precision granularity.

[0174] It is desirable to consider a 3GPP framework for AI / ML for the air interface that addresses each target use case, including CSI feedback, beam management, and positioning accuracy improvement, with respect to aspects such as performance, complexity, and the impact of possible specifications. In the current positioning framework, there is no known way to select, trigger, or configure the collection of positioning data required to train a configured AI / ML positioning model. This includes both online and offline training of positioning data for a specific AI / ML model, which depends on the network entity performing the training. Furthermore, the type of data source and the method of collection are important components for the statistical significance of the input training dataset, which impacts the accuracy of the AI / ML model.

[0175] Training of AI / ML positioning models can be supported in various scenarios where different network entities may request positioning datasets, which then subsequently train the AI / ML models. We describe procedures by which the network entity or node performing the training may configure and request a positioning dataset depending on a defined set of criteria, i.e., based on AI / ML direct and AI / ML-assisted positioning methods. We also present ways in which the LMF may be aware of nodes / entities capable of performing positioning data collection, as well as how such data is reported. The next aspect of the invention describes a general structure by which positioning training datasets should be trained and signaled via different interfaces, which are differentiated based on AI / ML direct and AI / ML-assisted positioning methods.

[0176] In DOI:10.1109 / OJVT.2021.3110134, several architectural enhancements related to data collection were discussed for AI / ML-assisted positioning methods, but this relied on existing MDT data collection procedures. MDT data collection procedures are designed with different objectives for network operation and maintenance and may not be suitable or optimized for the type of AI / ML tasks required for UE positioning. In SMM920210198-US-PSPF, a measurement and reporting framework for 3GPP positioning is discussed, but it does not address the configuration of data sources by training location and training dataset content. Currently, training and inference model deployment are described in TR37.817 in the 3GPP context for use cases such as mobility optimization and load balancing, without considering any positioning-specific use cases for AI / ML model training and data collection.

[0177] As described herein, in one embodiment, methods are described for data source selection and configuration depending on which network entity performed the AI / ML training. These include AI / ML positioning training performed at a location server (LMF), an NG-RAN node, a PRU as a UE, or a conventional positioning UE, and an indication of whether the training is online or offline, which impacts the configuration and reporting of the positioning dataset collection. In another embodiment, methods are detailed for the LMF to configure the collection of AI / ML training data for positioning purposes using various reporting methods, which can be configured based on the data collection scenario. In a final embodiment, a method is proposed for generating a training dataset depending on whether it is AI / ML direct (standalone) or AI / ML-assisted positioning, which impacts the structure of the training dataset to be signaled.

[0178] Aspects described herein relate to a method in a wireless communications network, in which a network entity that performs training of a desired AI / ML positioning model configures at least one data source for providing a positioning training dataset, the positioning training dataset configuration may include at least a request to the data source and a format of the training dataset based on a location of the network entity that performs training of the AI / ML model for positioning purposes, the data source responds to the dataset configuration by providing the required training dataset according to the desired data format, and the network entity performs training of the desired AI / ML model based on the collected training dataset.

[0179] In some embodiments, the data source may consist of a serving or neighboring gNB / TRP, a positioning reference unit UE or positioning reference unit TRP, a reference UE, a target UE, another UE, or a combination thereof.

[0180] In some embodiments, the training data set may be provided based on whether training is performed in at least one of a serving or neighboring gNB / TRP, a positioning reference unit UE or positioning reference unit TRP, a reference UE, a target UE, or another UE.

[0181] In some embodiments, the AI / ML positioning model may perform AI / ML direct positioning or AI / ML-assisted positioning.

[0182] In some embodiments, the positioning training data set configuration includes a reporting configuration that can be either instantaneous, periodic, event-based, or a combination thereof.

[0183] In some embodiments, the training dataset structure or database may be formatted according to the type of AI / ML positioning model, including AI / ML direct positioning, AI / ML assisted positioning, or a combination thereof.

[0184] In some embodiments, the network entity may be a positioning reference unit UE, a reference UE, a target UE, a reference UE, an NG-RAN node, a positioning reference unit gNB, a location server, or the like.

[0185] In some embodiments, the NG-RAN nodes may consist of a serving gNB or TRP, a neighboring gNB or TRP.

[0186] In some embodiments, the data format may include a fingerprint number, a configurable location where the positioning measurement was performed, the positioning measurement, time information regarding the occurrence of the measurement, the quality of the positioning measurement performed, or a combination thereof.

[0187] In some embodiments, the training dataset configuration may include an indication of whether the data collection is intended for online and / or offline training of a machine learning positioning model.

[0188] It should be noted that the above-described methods and apparatus illustrate rather than limit the present invention, and that those skilled in the art can design many alternative arrangements without departing from the scope of the appended claims. The word "comprising" does not exclude the presence of elements or steps other than those listed in a claim, and "a" or "an" does not exclude a plurality; a single processor or other unit may fulfill the functions of several units recited in a claim. Any reference signs in the claims shall not be intended to limit their scope.

[0189] Furthermore, while examples have been provided in the context of particular communication standards, these examples are not intended to be the limitations on communication standards to which the disclosed methods and apparatus may be applied. For example, while specific examples are provided in the context of 3GPP, the principles disclosed herein may also be applied to other wireless communication systems, and indeed any communication system that uses routing rules.

[0190] The method may also be embodied in a set of instructions stored on a computer-readable medium, which when loaded into a computer processor, digital signal processor (DSP) or the like, cause the processor to perform the method described above.

[0191] The described method and apparatus may be embodied in other specific forms. The described method and apparatus are to be considered in all respects as illustrative only and not restrictive. The scope of the invention is, therefore, indicated by the appended claims, rather than by the foregoing description. All changes that come within the meaning and range of equivalency of the claims are to be embraced within their scope.

[0192] The following abbreviations used herein will be clear to those skilled in the art: ADR Accumulated Delta Range A GNSS Assisted GNSS AI artificial intelligence AP Access point ARFCN Absolute Radio Frequency Channel Number ARP Antenna Reference Point BFD Beam Failure Detection BSSID Basic Service Set Identifier BTS base transceiver station (GERAN) BWP Bandwidth Part CBR Channel Busy Rate CG Configured Grant CID Cell ID (positioning method) CRS cell-specific reference signal CSI Channel State Information CSI-RS Channel State Information Reference Signal DCI Downlink Control Information DL Downlink DL-AoD Downlink launch angle DL-TDOA Downlink Time Difference of Arrival DM-RS demodulation reference signal DS-TWR (Bi-directional ranging) ECEF Earth-centered, Earth-fixed ECGI Evolved Cell Global Identifier E CID Extended Cell ID (positioning method) E-SMLC Extended Serving Mobile Location Center E-UTRAN Evolved Universal Terrestrial Radio Access Network EOP Earth Orientation Parameter EPDU External Protocol Data Unit FDMA Frequency Division Multiple Access FEC Forward Error Correction FTA Precision Time Assistance GAGAN GPS Assisted Geo-Augmented Navigation GNSS Global Navigation Satellite System GPS Global Positioning System HA GNSS High precision GNSS (RTK, PPP) IMU Inertial Measurement Unit IS interface specifications LMC Location Management Components LMF Location Management Function LMU Location Measurement Unit LOS Line of Sight LPP LTE Positioning Protocol LPPa LTE Positioning Protocol Annex LSB least significant bit MAC main supplementary concept Media Access Control MAC CE Medium Access Control Element MBS Urban Beacon System ML Machine Learning MO-LR Mobile Originated Location Request MSB Most Significant Bit MT-LR Mobile Terminated Location Request Multi-RTT Multiple Round Trip Time NAV Navigation NB-IoT Narrowband Internet of Things NCGI NR Cell Global Identifier NI-LR Network Induced Location Request NLOS Non-Line-of-Sight NPRS narrowband positioning reference signal NR NR Radio Access NRPPa NR Positioning Protocol Annex NRSRP narrowband reference signal received power NRSRQ Narrowband Reference Signal Reception Quality NWDAF Network Data Analysis Function OSR Observation Space Representation OTDOA Observed Time Difference of Arrival PDU Protocol Data Unit PDCP Packet Data Convergence Protocol PDCCH Physical Downlink Control Channel PDSCH Physical Downlink Shared Channel PHY Physical Layer PSCCH Physical Sidelink Control Channel PSSCH Physical Sidelink Shared Channel PSBCH Physical Sidelink Broadcast Channel PPP Precise Point Positioning PRB Physical Resource Block PRC pseudorange correction PRS Positioning Reference Signal posSIB Positioning System Information Block P-RNTI Paging Radio Network Temporary Identifier PT-RS Phase Tracking Reference Signal PUCCH Physical Uplink Control Channel PUSCH Physical Uplink Shared Channel QCL pseudo-collocation RAT Radio Access Technology RF radio frequency RLC Radio Link Control RRC Range Rate Correction Radio Resource Control RRM Radio Resource Management RS reference signal RSRP reference signal received power RSRPP Reference Signal Received Power by Path RSRQ Reference Signal Reception Quality RSTD reference signal time difference RSU Roadside Unit RTK Real-time Kinematics RTT Round Trip Time SBAS Space-Based Augmentation System SBCH Sidelink Broadcast Channel SCCH Sidelink Control Channel SCI Sidelink Control Information SET SUPL compatible devices SFN System Frame Number SL Side Link SL-PRS Sidelink Positioning Reference Signal SLP SUPL Location Platform SPS semi-permanent SS / PBCH Synchronization Signal / Physical Broadcast Channel SSBRI SS / PBCH Block Resource Index SSID Service Set Identifier SSR State Space Representation SS-TWR Unilateral bidirectional ranging STCH Sidelink Transport Channel SUPL Secure User Plane Location TB terrestrial beacon TBS terrestrial beacon system TCI Transmit Configuration Indicator TECU TEC unit TLM Telemetry TOA arrival time TOF Time of Flight TP Transmission Point TRP Transmit Receiving Point UE User Equipment UDRE User Differential Range Error ULP User Plane Location Protocol URA User Range Accuracy UTC Coordinated Universal Time WGS 84 World Geodetic System 1984 WLAN Wireless Local Area Network [Explanation of symbols]

[0193] 100 Wireless Communication System 102 Remote Unit, Device, UE 104 Network Unit 200 User Equipment Device, Device, UE 205 processors 210 memory 215 Input Devices 220 output devices 225 Transceiver 230 Transmitter 235 receiver 240 network interfaces 245 Application Interface 300 Network Nodes, Remote Units 305 processor 310 memory 315 Input Devices 320 output device 325 Transceiver 330 Transmitter 335 Receiver 340 Network Interface 345 Application Interface 400 System 410 Location Server, LMF 420 First gNodeB TRP, base station 430 Second gNodeB TRP, base station 440 Third gNodeB TRP, base station 450 devices, UE 610 Positioning Server (LMF) 620g Node B 630 Device, Target UE 910 Data Collection 911 Training Data 920 Model Training 921 Model Expansion / Update 912 Inference Data 930 Model Inference 931 Inference Output 932 Model Performance Feedback 940 Actors 941 Feedback 1100 System 1110 LCS Client 1120 Application Features 1130 5G Core Network 1131 GMLC / LRF 1132 UDM 1133 NEF 1134 AMF 1135 LMF 1140 Radio Access Network (NG-RAN) 1141 first neighbor gNodeB, gNodeB, serving and / or neighbor gNodeB / TRP, serving and / or neighbor gNB / TRP 1142 second neighbor gNodeB, gNodeB, serving and / or neighbor gNodeB / TRP, serving and / or neighbor gNB / TRP 1143 Serving gNodeB, gNodeB, Serving and / or Neighboring gNodeB / TRP, Serving and / or Neighboring gNB / TRP 1144 PRU, UE, PRU TRP 1150 Device, Target UE, UE, TRP 1151 LCS Client 1160 Device, PRU-UE, PRU 1210 Location Server, Location Server Node 1220 Serving and / or neighboring gNB / TRP, NG-RAN node, gNB node 1230 Positioning Reference Unit, NG-RAN node, gNB node 1240 Positioning Reference Unit, PRU UE, UE Node 1250 Reference UE, UE node 1260 Target UE, UE node 1310 LMF 1320 NG-RAN Node, PRU TRP 1330 UE 1340 Reference UE, UE 1350 Target UE, UE 1410 Device, Target UE, UE 1420 Device, PRU UE, UE 1430 Device, Reference UE, UE 1440 Other UE, PRU UE 1450 Location Server (LMF) 1510 Location Server (LMF) 1520 PRU UE, UE 1530 Reference UE, UE 1540 Normal UE, UE

Claims

1. A network node for wireless communication, comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor causing the network node to: transmitting a positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on locations of the network nodes; receiving a response including the positioning training data set in the required data format; configured to perform Network node.

2. The at least one processor in the network node: Serving or neighboring gNode B, a serving transmit-receive point or a neighboring transmit-receive point; Positioning Reference Unit User Equipment, Positioning reference unit transmitting receiving point, Reference User Equipment, Target user equipment, Supporting user equipment, Third Party User Equipment; Network data analysis functions, or Location Management Function and receiving the response from at least one of the 2. The network node of claim 1.

3. The network node Location Server Node, Next-generation radio access network nodes, Positioning Reference Unit g Node B, Positioning reference unit transmitting receiving point, Positioning Reference Unit User Equipment, Reference User Equipment, Supporting user equipment, Third-party user equipment, or Target User Device At least one of 2. The network node of claim 1.

4. The next generation radio access network node includes a serving gNodeB or a neighboring gNodeB, or a serving transmission reception point or a neighboring transmission reception point.

4. The network node of claim 3.

5. The positioning training data set configuration includes a reporting configuration that is at least one of instantaneous, periodic, or event-based.

2. The network node of claim 1.

6. The network node of claim 1 , wherein the machine learning positioning model comprises a machine learning direct positioning model or a machine learning assisted positioning model.

7. The required data format is structured according to the type of the machine learning positioning model.

2. The network node of claim 1.

8. The required data format includes at least one of a positioning measurement, configurable location information of a location where the positioning measurement is performed, time information of the positioning measurement, or quality of the positioning measurement.

2. The network node of claim 1.

9. The configurable location information comprises: 2D or 3D coordinates, Zone identifier, Grid identifier, Speed ​​or velocity, Orientation, direction, Height, or A rectangular grid with a length and a width Contains at least one of 9. A network node according to claim 8.

10. The at least one processor is further configured to cause the network node to receive an activation request for the network node to activate at least one of transmitting the positioning training data set configuration, receiving the positioning training data set, or training the machine learning positioning model.

2. The network node of claim 1.

11. The at least one processor is further configured to cause the network node to receive a deactivation request to the network node for at least one of not transmitting the positioning training data set configuration, not receiving the positioning training data set, or not training the machine learning positioning model.

2. The network node of claim 1.

12. The at least one processor is further configured to train the machine learning positioning model using a positioning training data set received from a local data source.

2. The network node of claim 1.

13. The positioning training dataset configuration further includes an indication as to whether the positioning training dataset is intended for online and / or offline training of the machine learning positioning model.

2. The network node of claim 1.

14. The at least one processor is further configured to cause the network node to train the machine learning positioning model using the positioning training data set.

2. The network node of claim 1.

15. The at least one processor is further configured to cause the network node to transmit an output of the trained machine learning positioning model to other network nodes.

15. A network node according to claim 14.

16. 1. A method performed by a network node, comprising: transmitting a positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on locations of the network nodes; receiving a response including the positioning training data set in the required data format; training the machine learning positioning model using the positioning training dataset; A method comprising:

17. A network node for wireless communications, comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor causing the network node to: receiving a positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on locations of the network nodes; transmitting a response including the positioning training data set in the required data format; configured to perform Network node.

18. The at least one processor in the network node: Serving gNodeB or neighboring gNodeB, a serving transmit-receive point or a neighboring transmit-receive point; Positioning Reference Unit User Equipment, Positioning reference unit transmitting receiving point, Reference User Equipment, Target user equipment, Supporting user equipment, Third Party User Equipment; Network data analysis functions, or Location Management Function and receiving the response from at least one of the 18. A network node according to claim 17.

19. The required data format includes at least one of a positioning measurement, configurable location information of a location where the positioning measurement is performed, time information of the positioning measurement, or quality of the positioning measurement.

18. A network node according to claim 17.

20. A method performed by a network node, comprising: receiving a positioning training dataset configuration including a request for a positioning training dataset for training a machine learning positioning model and a required data format of the positioning training dataset based on locations of the network nodes; transmitting a response including the positioning training data set in the required data format; A method comprising:

Images