Machine learning assisted location determination

Abstract

Methods, devices, and systems for machine learning (ML) assisted location determination are disclosed. Information indicative of an artificial intelligence / machine learning (AI / ML) model for determining a location is received. Information indicating a transmit reference point (TRP) (502, 504) associated with a corner is received. Information indicating a reference signal received power (RSRP) is received. The TRP associated with the corner includes a first TRP. It is determined that the WTRU is located at a corner based on the RSRP of a positioning reference signal (PRS) (506) received from the first TRP being above an RSRP threshold. Location information is determined based on the AI / ML location model and the determination that the WTRU is located at the corner. And transmitting information indicating the position of the WTRU.

Description

[0001] Related Applications

[0002] This application claims the benefit of and priority to U.S. Provisional Application No. 63 / 335,533, filed on April 27, 2022, entitled “Machine Learning Assisted Positioning”; U.S. Provisional Application No. 63 / 389,168, filed on July 14, 2022, entitled “Machine Learning Assisted Positioning”; and U.S. Provisional Application No. 63 / 445,595, filed on February 14, 2023, entitled “Machine Learning Assisted Positioning”, the entire contents of each of these U.S. provisional applications are incorporated herein by reference. Background Art

[0003] Downlink techniques, uplink techniques, and both downlink and uplink techniques may be used to determine the location of a device in a wireless network environment, sometimes referred to as "position determination," "position fix," "position determination," "positioning," or similar terms or combinations of such terms. Such techniques may use positioning reference signals and / or sounding reference signals for position determination purposes. In some implementations, the environment plays a role in the measurement of such reference signals, which may affect the positioning accuracy that can be achieved by such position determination techniques. Summary of the invention

[0004] Some specific implementations provide a method implemented in a wireless transmit / receive unit (WTRU). Receive information indicating an artificial intelligence / machine learning (AI / ML) model for determining a location. Receive information indicating a transmit reference point (TRP) associated with a corner and information indicating a reference signal received power (RSRP) threshold, the TRP associated with the corner including a first TRP. Determine that the WTRU is located at the corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being greater than the RSRP threshold. Determine location information of the WTRU based on the AI / ML location model and the determination that the WTRU is located at the corner. Send information indicating the location of the WTRU.

[0005] In some implementations, the AI / ML location model includes a single TRP fingerprinting AI / ML location model. In some implementations, the single TRP fingerprinting AI / ML location model includes weights for a neural network. In some implementations, the AI / ML model for determining location includes a support vector machine (SVM) or a k-nearest neighbor (KNN) model. Some implementations include sending a request for information indicating the AI / ML model for determining location. In some implementations, a TRP is associated with a corner based on its proximity to a corner of a deployment environment. Some implementations include sending a request for a PRS to one or more TRPs. Some implementations include receiving a PRS from each of a plurality of TRPs including a first TRP. In some implementations, the WTRU has a line of sight (LOS) path to the first TRP. In some implementations, the input to the AI / ML location model includes peak power or average power measurements of one or more transmissions from the first TRP.

[0006] Some specific implementations provide a WTRU. The WTRU includes a receiver circuit configured to receive information indicating an artificial intelligence / machine learning (AI / ML) model for determining a location. The receiver circuit is configured to receive information indicating a transmit reference point (TRP) associated with a corner. The receiver circuit is also configured to receive information indicating a reference signal received power (RSRP) threshold, the TRP associated with the corner including a first TRP. The WTRU also includes a processing circuit configured to determine that the WTRU is located at a corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being higher than the RSRP threshold. The processing circuit is also configured to determine location information based on the AI / ML location model and a determination that the WTRU is located at a corner. The WTRU also includes a transmitter circuit configured to send information indicating the location of the WTRU.

[0007] In some implementations, the AI / ML location model includes a single TRP fingerprinting AI / ML location model. In some implementations, the single TRP fingerprinting AI / ML location model includes weights for a neural network. In some implementations, the AI / ML model for determining location includes a support vector machine (SVM) or a k-nearest neighbor (KNN) model. In some implementations, the transmitter is further configured to send a request for information indicating the AI / ML model for determining location. In some implementations, a TRP is associated with a corner based on its proximity to a corner of a deployment environment. In some implementations, the transmitter is further configured to send a request for a PRS to one or more TRPs. In some implementations, the receiver is further configured to receive a PRS from each of a plurality of TRPs including a first TRP. In some implementations, the WTRU has a line of sight (LOS) path to the first TRP. In some implementations, the input to the AI / ML location model includes peak power or average power measurements of one or more transmissions from the first TRP.

[0008] Some specific implementations provide a method implemented in a WTRU. Receive information indicating an artificial intelligence / machine learning (AI / ML) model for determining location. Receive information indicating a transmit reference point (TRP) associated with line of sight (LOS) communication. Receive information indicating a reference signal received power (RSRP) threshold. Receive information indicating a threshold number of TRPs associated with LOS communication. Determine a total number of TRPs associated with LOS communication for which an RSRP measured on a received positioning reference signal (PRS) is less than the RSRP threshold. Determine that the WTRU is located in a non-line of sight (NLOS) environment based on the determined total number of TRPs being less than the threshold number of TRPs. Determine location information of the WTRU based on the AI / ML location model and the determination that the WTRU is located in a NLOS environment. Send information indicating the location of the WTRU.

[0009] In some implementations, the location model includes a multi-TRP fingerprinting AI / ML location model. In some implementations, the multi-TRP fingerprinting AI / ML location model includes weights for a neural network. In some implementations, the AI / ML model for determining location includes a support vector machine (SVM) or a k-nearest neighbor (KNN) model. Some implementations include sending a request for information indicating the AI / ML model for determining location. In some implementations, the indication of the TRPs associated with the LOS communication includes a list of TRP identifiers and an associated pre-position indicating LOS or non-line of sight (NLOS). Some implementations include sending a request for a PRS to each of the TRPs associated with the LOS communication. Some implementations include receiving a PRS from multiple TRPs associated with the LOS communication. In some implementations, the LOS path to at least one of the TRPs associated with the LOS communication is obscured. In some implementations, inputs to the AI / ML location model include peak power or average power measurements of one or more transmissions from each of a plurality of TRPs associated with LOS communications.

[0010] Some specific implementations provide a WTRU. The WTRU includes a receiver circuit configured to receive information indicating an artificial intelligence / machine learning (AI / ML) model for determining a location. The receiver circuit is further configured to receive information indicating a transmit reference point (TRP) associated with line of sight (LOS) communication, information indicating a reference signal received power (RSRP) threshold, and information indicating a threshold number of TRPs associated with LOS communication. The WTRU includes a processing circuit configured to determine a total number of TRPs associated with LOS communication for which the RSRP measured on a received positioning reference signal (PRS) is below the RSRP threshold. The processing circuit is further configured to determine that the WTRU is located in a non-line of sight (NLOS) environment based on the total number of TRPs determined being less than the threshold number of TRPs. The processing circuit is further configured to determine location information based on the AI / ML location model and the determination that the WTRU is located in a NLOS environment. The WTRU includes a transmitter circuit configured to send information indicating the location of the WTRU.

[0011] In some implementations, the location model includes a multi-TRP fingerprinting AI / ML location model. In some implementations, the multi-TRP fingerprinting AI / ML location model includes weights for a neural network. In some implementations, the AI / ML model for determining the location includes a support vector machine (SVM) or a k-nearest neighbor (KNN) model. In some implementations, the transmitter circuit is further configured to send a request for information indicating the AI / ML model for determining the location. In some implementations, the indication of the TRPs associated with the LOS communication includes a list of TRP identifiers and an associated predetermined position indicating LOS or non-line of sight (NLOS). In some implementations, the transmitter is further configured to send a request for a PRS to each of the TRPs associated with the LOS communication. In some implementations, the receiver is further configured to receive PRSs from multiple TRPs associated with the LOS communication. In some implementations, the LOS path to at least one of the TRPs associated with the LOS communication is shielded. In some implementations, input to the AI / ML position includes peak power or average power measurements of one or more transmissions from each of a plurality of TRPs associated with LOS communications.

[0012] Some specific implementations provide methods, devices, and systems for machine learning (ML) assisted location determination. A ML assisted non-line-of-sight (NLOS) identification request and capability information are sent to a network device. In response to the identification request and the capability information, a trained ML model and a first list of input features are received. A model for NLOS identification is selected based on a desired rate of NLOS prediction. A NLOS transmit receive point (TRP) is predicted based on the selected model. A location is determined based on the predicted NLOS TRP.

[0013] In a first aspect, the present disclosure relates to a method for machine learning assisted location determination implemented in a wireless transmit / receive unit (WTRU). The method includes receiving an artificial intelligence / machine learning (AI / ML) model for AI / ML assisted positioning. The method also includes receiving an indication of a transmit reference point (TRP) associated with a corner and an indication of a reference signal received power (RSRP) threshold, the TRP associated with the corner including a first TRP. The method also includes determining that the WTRU is located at a corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being higher than the RSRP threshold. The method also includes obtaining the location of the WTRU using single TRP fingerprinting AI / ML positioning by the WTRU based on the determination that the WTRU is located at the corner. The method also includes sending an indication of the location of the WTRU.

[0014] In a second aspect, the present disclosure relates to a device for machine learning assisted location determination, the device comprising a wireless transmit / receive unit (WTRU), the wireless transmit / receive unit comprising a receiver, a processor, and a transmitter. The receiver is configured to: receive an artificial intelligence / machine learning (AI / ML) model for AI / ML assisted positioning; and receive an indication of a transmit reference point (TRP) associated with a corner and an indication of a reference signal received power (RSRP) threshold, the TRP associated with the corner comprising a first TRP. The processor is configured to: determine that the WTRU is located at a corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being higher than the RSRP threshold; and obtain the location of the WTRU using single TRP fingerprinting AI / ML positioning based on the determination that the WTRU is located at the corner. The transmitter is configured to send an indication of the location of the WTRU.

[0015] In a third aspect, the present disclosure relates to a method for machine learning (ML) assisted location determination implemented in a wireless transmit / receive unit (WTRU). The method includes receiving an artificial intelligence / machine learning (AI / ML) model for AI / ML assisted positioning. The method also includes receiving an indication of a transmit reference point (TRP) associated with line of sight (LOS) communication, an indication of a reference signal received power (RSRP) threshold, and an indication of a threshold number of TRPs associated with LOS communication. The method also includes determining a total number of TRPs associated with LOS communication for which the RSRP measured on a received positioning reference signal (PRS) is below the RSRP threshold. The method also includes determining that the WTRU is located in a non-line of sight (NLOS) environment based on the total number of TRPs determined being less than the threshold number of TRPs. The method also includes obtaining the location of the WTRU using multi-TRP fingerprinting AI / ML positioning by the WTRU based on the determination that the WTRU is located in a NLOS environment. The method also includes sending an indication of the location of the WTRU.

[0016] In a fourth aspect, the present disclosure relates to a device for machine learning (ML) assisted location determination. The device includes a wireless transmit / receive unit (WTRU), which includes a receiver, a processor, and a transmitter. The receiver is configured to: receive an artificial intelligence / machine learning (AI / ML) model for AI / ML assisted positioning; and receive an indication of a transmit reference point (TRP) associated with line of sight (LOS) communication, an indication of a reference signal received power (RSRP) threshold, and an indication of a threshold number of TRPs associated with LOS communication. The processor is configured to: determine the total number of TRPs associated with LOS communication, for which the RSRP measured on the received positioning reference signal (PRS) is lower than the RSRP threshold; determine that the WTRU is located in a non-line of sight (NLOS) environment based on the total number of TRPs determined to be less than the threshold number of TRPs; and obtain the location of the WTRU using multi-TRP fingerprinting AI / ML positioning based on the determination that the WTRU is located in the NLOS environment. The transmitter is configured to send an indication of the location of the WTRU.

[0017] In a fifth aspect, the present disclosure relates to a method for machine learning assisted location determination. The method includes identifying, by a wireless transmit / receive unit (WTRU), a spatial configuration of the WTRU relative to one or more transmit receive points (TRPs). The method also includes selecting, by the WTRU, a location determination algorithm from a plurality of location determination algorithms based on the identified spatial configuration, the location determination algorithm including at least one trained ML model. The method also includes determining, by the WTRU, the location of the WTRU using the selected location determination algorithm. In some embodiments, the method includes determining, in response to detecting that a positioning reference signal (PRS) received by the WTRU from a TRP associated with a corner exceeds a threshold, that the WTRU is associated with a corner within an environment. In another embodiment, the method includes selecting a single TRP fingerprint recognition ML model in response to determining that the WTRU is associated with a corner within the environment. In some embodiments, the method includes determining that the WTRU is in a non-line of sight (NLOS) environment. In another embodiment, the method includes selecting a multi-TRP fingerprint recognition ML model in response to determining that the WTRU is in a NLOS environment.

[0018] In a sixth aspect, the present disclosure relates to a device for machine learning assisted location determination. The device includes a wireless transmit / receive unit (WTRU) including a processor. The processor is configured to: identify a spatial configuration of the WTRU relative to one or more transmit / receive points (TRPs); select a location determination algorithm from a plurality of location determination algorithms based on the identified spatial configuration, the location determination algorithm including at least one trained ML model; and determine the location of the WTRU using the selected location determination algorithm. In some embodiments, the processor is further configured to determine that the WTRU is associated with a corner within an environment in response to detecting that a positioning reference signal (PRS) received via a receiver from a TRP associated with a corner exceeds a threshold. In another embodiment, the processor is further configured to select a single TRP fingerprint recognition ML model in response to determining that the WTRU is associated with a corner within the environment. In some embodiments, the processor is further configured to determine that the WTRU is in a non-line of sight (NLOS) environment. In another embodiment, the processor is further configured to select a multi-TRP fingerprint recognition ML model in response to determining that the WTRU is in a NLOS environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] A more detailed understanding may be obtained from the following description given by way of example in conjunction with the accompanying drawings, in which like reference numerals indicate like elements, and in which:

[0020] Figure 1A is a system diagram illustrating an example communication system in which one or more disclosed embodiments may be implemented;

[0021] Figure 1B is an example of an embodiment in which Figure 1A A system diagram of an example wireless transmit / receive unit (WTRU) for use within the illustrated communication system;

[0022] Figure 1C is an example of an embodiment in which Figure 1A a system diagram of an example Radio Access Network (RAN) and an example Core Network (CN) used within the illustrated communication system;

[0023] Figure 1D is an example of an embodiment in which Figure 1A A system diagram of another example RAN and another example CN used within the illustrated communication system;

[0024] Figure 2 is a system diagram illustrating an example RSRP fingerprint;

[0025] Figure 3 is a system diagram illustrating an example NLOS propagation link between a WTRU and a TRP;

[0026] Figure 4 is a system diagram illustrating the effect of the reference TRP position on the achieved position determination accuracy;

[0027] Figure 5 is a system diagram illustrating an example WTRU in a situation where a limited number of PRSs can be received;

[0028] Figure 6 is a block diagram illustrating an example artificial neural network;

[0029] Figure 7 is a system diagram illustrating an example WTRU in a hybrid LOS / NLOS environment;

[0030] Figure 8 is a message sequence chart illustrating a sample model selection process;

[0031] Fig. 9 is a flow chart illustrating an example process for determining a multi-level ML model;

[0032] Fig.10 is a block diagram illustrating an example model deployment process for fingerprinting-based location determination;

[0033] Fig.11 is a block diagram illustrating an example model training process for fingerprint-based location determination;

[0034] Fig.12 are perspective and plan views of example deployments of WTRUs;

[0035] Fig.13 is another perspective and plan view of an example deployment of a WTRU;

[0036] Fig.14 is a block diagram illustrating inputs and outputs for an example ML-assisted position estimation implementation;

[0037] Fig.15 is a block diagram illustrating an example ML-assisted position estimation scenario;

[0038] Fig.16 is a block diagram illustrating an example training process for ML-assisted position estimation;

[0039] Fig.17 is a system diagram illustrating an example WTRU in a case where the example WTRU is located at a corner;

[0040] Fig.18 is a system diagram illustrating an example WTRU in an example scenario in which the WTRU is located in an example NLOS environment;

[0041] Fig.19 is a system diagram illustrating an example WTRU in an example scenario in which the WTRU is in a LOS environment; and

[0042] Fig. 20 is a flow chart illustrating an embodiment of a method for ML-assisted position estimation. DETAILED DESCRIPTION

[0043] Figure 1A 1 is a diagram illustrating an example communication system 100 in which one or more disclosed embodiments may be implemented. The communication system 100 may be a multiple access system that provides content such as voice, data, video, messaging, broadcast, etc. to multiple wireless users. The communication system 100 may enable multiple wireless users to access such content through sharing of system resources (including wireless bandwidth). For example, the communication system 100 may employ one or more channel access methods, such as code division multiple access (CDMA), time division multiple access (TDMA), frequency division multiple access (FDMA), orthogonal FDMA (OFDMA), single carrier FDMA (SC-FDMA), zero tail unique word discrete Fourier transform spread OFDM (ZT-UW-DFT-S-OFDM), unique word OFDM (UW-OFDM), resource block filtered OFDM, filter bank multi-carrier (FBMC), etc.

[0044] like Figure 1AAs shown, the communication system 100 may include wireless transmit / receive units (WTRUs) 102a, 102b, 102c, 102d, a radio access network (RAN) 104, a core network (CN) 106, a public switched telephone network (PSTN) 108, the Internet 110 and other networks 112, but it should be understood that the disclosed embodiments contemplate any number of WTRUs, base stations, networks, and / or network elements. Each of the WTRUs 102a, 102b, 102c, 102d may be any type of device configured to operate and / or communicate in a wireless environment. By way of example, the WTRUs 102a, 102b, 102c, 102d (any of which may be referred to as a station (STA)) may be configured to send and / or receive wireless signals and may include user equipment (UE), a mobile station, a fixed or mobile subscriber unit, a subscription-based unit, a pager, a cellular phone, a personal digital assistant (PDA), a smart phone, a laptop, a netbook, a personal computer, a wireless sensor, a hotspot or Mi-Fi device, an Internet of Things (IoT) device, a watch or other wearable device, a head-mounted display (HMD), a vehicle, a drone, medical devices and applications (e.g., remote surgery), industrial devices and applications (e.g., robots and / or other wireless devices operating in an industrial and / or automated process chain environment), a consumer electronic device, a device operating on a commercial and / or industrial wireless network, etc. Any of the WTRUs 102a, 102b, 102c, and 102d may be interchangeably referred to as a UE.

[0045] The communication system 100 may also include a base station 114a and / or a base station 114b. Each of the base stations 114a, 114b may be any type of device configured to wirelessly interface with at least one of the WTRUs 102a, 102b, 102c, 102d to facilitate access to one or more communication networks (such as the CN 106, the Internet 110, and / or other networks 112). By way of example, the base stations 114a, 114b may be a base transceiver station (BTS), a Node B, an evolved Node B (eNB), a Home Node B, a Home evolved Node B, a next generation Node B (such as a gNode B (gNB)), a new radio (NR) Node B, a site controller, an access point (AP), a wireless router, and the like. While the base stations 114a, 114b are each depicted as a single element, it will be appreciated that the base stations 114a, 114b may include any number of interconnected base stations and / or network elements.

[0046] The base station 114a may be part of the RAN 104, which may also include other base stations and / or network elements (not shown), such as a base station controller (BSC), a radio network controller (RNC), a relay node, etc. The base station 114a and / or the base station 114b may be configured to send and / or receive wireless signals on one or more carrier frequencies, which may be referred to as a cell (not shown). These frequencies may be in a licensed spectrum, an unlicensed spectrum, or a combination of licensed and unlicensed spectrums. A cell may provide coverage of wireless services to a specific geographic area, which may be relatively fixed or may change over time. The cell may be further divided into cell sectors. For example, a cell associated with the base station 114a may be divided into three sectors. Therefore, in one embodiment, the base station 114a may include three transceivers, i.e., one transceiver for each sector of the cell. In an embodiment, the base station 114a may employ multiple-input multiple-output (MIMO) technology and may utilize multiple transceivers for each sector of the cell. For example, beamforming may be used to send and / or receive signals in a desired spatial direction.

[0047] The base stations 114a, 114b may communicate with one or more of the WTRUs 102a, 102b, 102c, 102d over an air interface 116, which may be any suitable wireless communication link (e.g., radio frequency (RF), microwave, centimeter wave, micrometer wave, infrared (IR), ultraviolet (UV), visible light, etc.). The air interface 116 may be established using any suitable radio access technology (RAT).

[0048] More specifically, as noted above, the communication system 100 may be a multiple access system and may employ one or more channel access schemes, such as CDMA, TDMA, FDMA, OFDMA, SC-FDMA, etc. For example, the base station 114a in the RAN 104 and the WTRUs 102a, 102b, 102c may implement a radio technology, such as Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access (UTRA), that may use Wideband CDMA (WCDMA) to establish the air interface 116. WCDMA may include communication protocols, such as High Speed ​​Packet Access (HSPA) and / or Evolved HSPA (HSPA+). HSPA may include High Speed ​​Downlink (DL) Packet Access (HSDPA) and / or High Speed ​​Uplink (UL) Packet Access (HSUPA).

[0049] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology such as Evolved UMTS Terrestrial Radio Access (E-UTRA), which may establish the air interface 116 using Long Term Evolution (LTE) and / or Advanced LTE (LTE-A) and / or Advanced LTE Pro (LTE-A Pro).

[0050] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology (such as NR radio access) that may establish the air interface 116 using NR.

[0051] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement multiple radio access technologies. For example, the base station 114a and the WTRUs 102a, 102b, 102c may together implement LTE radio access and NR radio access, for example using the dual connectivity (DC) principle. Thus, the air interface utilized by the WTRUs 102a, 102b, 102c may be characterized by multiple types of radio access technologies and / or transmissions transmitted to / from multiple types of base stations (e.g., eNBs and gNBs).

[0052] In other embodiments, the base station 114a and the WTRUs 102a, 102b, 102c may implement radio technologies such as IEEE 802.11 (i.e., Wireless Fidelity (WiFi)), IEEE 802.16 (i.e., Worldwide Interoperability for Microwave Access (WiMAX)), CDMA2000, CDMA2000 1X, CDMA2000 EV-DO, Interim Standard 2000 (IS-2000), Interim Standard 95 (IS-95), Interim Standard 856 (IS-856), Global System for Mobile Communications (GSM), Enhanced Data rates for GSM Evolution (EDGE), GSM EDGE (GERAN), etc.

[0053] Figure 1AThe base station 114b in the may be, for example, a wireless router, a Home NodeB, a Home eNodeB, or an access point, and may utilize any suitable RAT to facilitate wireless connectivity in local areas such as commercial venues, homes, vehicles, campuses, industrial facilities, sky corridors (e.g., for use by drones), roads, and the like. In one embodiment, the base station 114b and the WTRUs 102c, 102d may implement radio technologies such as IEEE 802.11 to establish a wireless local area network (WLAN). In an embodiment, the base station 114b and the WTRUs 102c, 102d may implement radio technologies such as IEEE 802.15 to establish a wireless personal area network (WPAN). In yet another embodiment, the base station 114b and the WTRUs 102c, 102d may utilize a cellular-based RAT (e.g., WCDMA, CDMA2000, GSM, LTE, LTE-A, LTE-A Pro, NR, etc.) to establish a picocell or a femtocell. As Figure 1A As shown, the base station 114b may have a direct connection to the Internet 110. Therefore, the base station 114b may not be required to access the Internet 110 via the CN 106.

[0054] The RAN 104 may be in communication with the CN 106, which may be any type of network configured to provide voice, data, applications, and / or Voice over Internet Protocol (VoIP) services to one or more of the WTRUs 102a, 102b, 102c, 102d. The data may have different quality of service (QoS) requirements, such as different throughput requirements, latency requirements, error tolerance requirements, reliability requirements, data throughput requirements, mobility requirements, etc. The CN 106 may provide call control, billing services, mobile location-based services, pre-paid calling, Internet connectivity, video distribution, etc., and / or perform high-level security functions, such as user authentication. Although not described in detail in the specification, the CN 106 may be configured to provide voice, data, applications, and / or Voice over Internet Protocol (VoIP) services to one or more of the WTRUs 102a, 102b, 102c, 102d. Figure 1A Although not shown in the figure, it will be appreciated that the RAN 104 and / or the CN 106 may be in direct or indirect communication with other RANs that employ the same RAT as the RAN 104 or a different RAT. For example, in addition to being connected to the RAN 104, which may utilize NR radio technology, the CN 106 may also be in communication with another RAN (not shown) that employs GSM, UMTS, CDMA 2000, WiMAX, E-UTRA, or WiFi radio technology.

[0055] The CN 106 may also act as a gateway for the WTRUs 102a, 102b, 102c, 102d to access the PSTN 108, the Internet 110, and / or other networks 112. The PSTN 108 may include a circuit-switched telephone network that provides plain old telephone service (POTS). The Internet 110 may include a global system of interconnected computer networks and devices that use common communication protocols, such as the Transmission Control Protocol (TCP), the User Datagram Protocol (UDP), and / or the Internet Protocol (IP) in the TCP / IP Internet protocol suite. The networks 112 may include wired communication networks and / or wireless communication networks owned and / or operated by other service providers. For example, the networks 112 may include another CN connected to one or more RANs, which may employ the same RAT as the RAN 104 or a different RAT.

[0056] Some or all of the WTRUs 102a, 102b, 102c, 102d in the communication system 100 may include multi-mode capabilities (e.g., the WTRUs 102a, 102b, 102c, 102d may include multiple transceivers for communicating with different wireless networks via different wireless links). Figure 1A The illustrated WTRU 102c may be configured to communicate with the base station 114a, which may employ a cellular-based radio technology, and with the base station 114b, which may employ an IEEE 802 radio technology.

[0057] Figure 1B is a system diagram illustrating an example WTRU 102. Figure 1B As shown, the WTRU 102 may include, among other things, a processor 118, a transceiver 120, a transmit / receive element 122, a speaker / microphone 124, a keypad 126, a display / touchpad 128, non-removable memory 130, removable memory 132, a power source 134, a global positioning system (GPS) chipset 136, and / or other peripherals 138. It will be appreciated that the WTRU 102 may include any sub-combination of the foregoing elements while remaining consistent with an embodiment.

[0058] The processor 118 may be a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor (DSP), a plurality of microprocessors, one or more microprocessors associated with a DSP core, a controller, a microcontroller, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), any other type of integrated circuit (IC), a state machine, etc. The processor 118 may perform signal decoding, data processing, power control, input / output processing, and / or any other functionality that enables the WTRU 102 to operate in a wireless environment. The processor 118 may be coupled to the transceiver 120, which may be coupled to the transmit / receive element 122. Although Figure 1B The processor 118 and the transceiver 120 are depicted as separate components, but it is understood that the processor 118 and the transceiver 120 may be integrated together in an electronic package or chip.

[0059] The send / receive element 122 may be configured to send a signal to a base station (e.g., base station 114a) or receive a signal from the base station via an air interface 116. For example, in one embodiment, the send / receive element 122 may be an antenna configured to send and / or receive an RF signal. In an embodiment, the send / receive element 122 may be a transmitter / detector configured to send and / or receive, for example, IR, UV or visible light signals. In another embodiment, the send / receive element 122 may be configured to send and / or receive both RF signals and optical signals. It should be understood that the send / receive element 122 may be configured to send and / or receive any combination of wireless signals.

[0060] Although the transmit / receive element 122 Figure 1B 1 as a single element, the WTRU 102 may include any number of transmit / receive elements 122. More specifically, the WTRU 102 may employ MIMO technology. Thus, in one embodiment, the WTRU 102 may include two or more transmit / receive elements 122 (e.g., multiple antennas) for transmitting and receiving wireless signals over the air interface 116.

[0061] The transceiver 120 may be configured to modulate signals to be transmitted by the transmit / receive element 122 and to demodulate signals received by the transmit / receive element 122. As noted above, the WTRU 102 may have multi-mode capabilities. For example, the transceiver 120 may include multiple transceivers for enabling the WTRU 102 to communicate via multiple RATs, such as NR and IEEE 802.11.

[0062] The processor 118 of the WTRU 102 may be coupled to and may receive user input data from a speaker / microphone 124, a keypad 126, and / or a display / touchpad 128 (e.g., a liquid crystal display (LCD) display unit or an organic light emitting diode (OLED) display unit). The processor 118 may also output user data to the speaker / microphone 124, the keypad 126, and / or the display / touchpad 128. In addition, the processor 118 may access information from and store data in any type of suitable memory, such as a non-removable memory 130 and / or a removable memory 132. The non-removable memory 130 may include a random access memory (RAM), a read-only memory (ROM), a hard disk, or any other type of memory storage device. The removable memory 132 may include a subscriber identity module (SIM) card, a memory stick, a secure digital (SD) memory card, and the like. In other embodiments, the processor 118 may access information from and store data in a memory that is not physically located on the WTRU 102, such as on a server or a home computer (not shown).

[0063] The processor 118 may receive power from the power source 134, and may be configured to distribute and / or control power to other components in the WTRU 102. The power source 134 may be any suitable device for powering the WTRU 102. For example, the power source 134 may include one or more dry cell batteries (e.g., nickel-cadmium (NiCd), nickel-zinc (NiZn), nickel-metal hydride (NiMH), lithium-ion (Li-ion), etc.), solar cells, fuel cells, etc.

[0064] The processor 118 may also be coupled to the GPS chipset 136, which may be configured to provide location information (e.g., longitude and latitude) regarding the current location of the WTRU 102. In addition to or in lieu of the information from the GPS chipset 136, the WTRU 102 may receive location information from a base station (e.g., base stations 114a, 114b) over the air interface 116 and / or determine its location based on the timing of signals received from two or more nearby base stations. It will be appreciated that the WTRU 102 may acquire location information by any suitable location-determination method while remaining consistent with an embodiment.

[0065] The processor 118 may also be coupled to other peripherals 138, which may include one or more software modules and / or hardware modules that provide additional features, functionality, and / or wired or wireless connectivity. For example, the peripherals 138 may include an accelerometer, an electronic compass, a satellite transceiver, a digital camera (for photos and / or video), a universal serial bus (USB) port, a vibration device, a television transceiver, a hands-free headset, module, a frequency modulation (FM) radio unit, a digital music player, a media player, a video game player module, an Internet browser, a virtual reality and / or augmented reality (VR / AR) device, an activity tracker, etc. The peripheral device 138 may include one or more sensors. The sensor may be one or more of the following: a gyroscope, an accelerometer, a Hall effect sensor, a magnetometer, an orientation sensor, a proximity sensor, a temperature sensor, a time sensor; a geographic location sensor, an altimeter, a light sensor, a touch sensor, a magnetometer, a barometer, a gesture sensor, a biometric sensor, a humidity sensor, etc.

[0066] The WTRU 102 may include a full-duplex radio for which transmission and reception of some or all of the signals (e.g., associated with specific subframes for both UL (e.g., for transmission) and DL (e.g., for reception)) may be concurrent and / or simultaneous. The full-duplex radio may include an interference management unit for reducing and / or substantially eliminating self-interference via signal processing performed by hardware (e.g., a choke) or via a processor (e.g., a separate processor (not shown) or via the processor 118). In an embodiment, the WTRU 102 may include a half-duplex radio for which transmission and reception of some or all of the signals (e.g., associated with specific subframes for both UL (e.g., for transmission) or DL ​​(e.g., for reception)) may be concurrent and / or simultaneous.

[0067] Figure 1C 1 is a system diagram illustrating the RAN 104 and the CN 106 according to one embodiment. As noted above, the RAN 104 may employ an E-UTRA radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0068] The RAN 104 may include evolved Node-Bs 160a, 160b, 160c, though it will be appreciated that the RAN 104 may include any number of evolved Node-Bs while remaining consistent with an embodiment. The evolved Node-Bs 160a, 160b, 160c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 116. In one embodiment, the evolved Node-Bs 160a, 160b, 160c may implement MIMO technology. Thus, the evolved Node-B 160a, for example, may use multiple antennas to transmit wireless signals to and / or receive wireless signals from the WTRU 102a.

[0069] Each of the evolved Node Bs 160a, 160b, 160c may be associated with a particular cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in the UL and / or DL, etc. Figure 1C As shown, the eNode-Bs 160a, 160b, 160c may communicate with one another via an X2 interface.

[0070] Figure 1C The illustrated CN 106 may include a mobility management entity (MME) 162, a serving gateway (SGW) 164, and a packet data network (PDN) gateway (PGW) 166. Although the foregoing elements are depicted as part of the CN 106, it should be understood that any of these elements may be owned and / or operated by an entity other than the CN operator.

[0071] The MME 162 may be connected to each of the evolved Node-Bs 162a, 162b, 162c in the RAN 104 via an S1 interface and may serve as a control node. For example, the MME 162 may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, bearer activation / deactivation, selecting a particular serving gateway during an initial attach of the WTRUs 102a, 102b, 102c, etc. The MME 162 may provide a control plane function for switching between the RAN 104 and other RANs (not shown) that employ other radio technologies, such as GSM and / or WCDMA.

[0072] The SGW 164 may be connected to each of the evolved Node-Bs 160a, 160b, 160c in the RAN 104 via an S1 interface. The SGW 164 may generally route and forward user data packets to / from the WTRUs 102a, 102b, 102c. The SGW 164 may perform other functions, such as anchoring the user plane during an inter-evolved Node-B handover, triggering paging when DL data is available for the WTRUs 102a, 102b, 102c, managing and storing the context of the WTRUs 102a, 102b, 102c, and the like.

[0073] The SGW 164 may be connected to the PGW 166, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices.

[0074] The CN 106 may facilitate communications with other networks. For example, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to circuit-switched networks, such as the PSTN 108, to facilitate communications between the WTRUs 102a, 102b, 102c and traditional land-line communications devices. For example, the CN 106 may include, or may be in communication with, an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) that acts as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to other networks 112, which may include other wired networks and / or wireless networks owned and / or operated by other service providers.

[0075] Although the WTRU Figures 1A to 1D Although described as wireless terminals, it is contemplated that in certain representative embodiments, such terminals may (eg, temporarily or permanently) use a wired communication interface with a communication network.

[0076] In a representative embodiment, the other network 112 may be a WLAN.

[0077] A WLAN in infrastructure basic service set (BSS) mode may have an access point (AP) for the BSS and one or more stations (STA) associated with the AP. The AP may have access or an interface to a distribution system (DS) or another type of wired / wireless network that carries traffic to and / or carries traffic away from the BSS. Traffic originating from outside the BSS and leading to the STA may be reached by the AP and may be delivered to the STA. Traffic originating from the STA and leading to a destination outside the BSS may be transmitted to the AP to be delivered to the corresponding destination. Traffic between STAs within the BSS may be transmitted by the AP, for example, wherein the source STA may transmit traffic to the AP, and the AP may deliver traffic to the target STA. Traffic between STAs within the BSS may be considered and / or referred to as point-to-point traffic. Direct link establishment (DLS) may be used to transmit point-to-point traffic between the source STA and the target STA (e.g., directly between them). In certain representative embodiments, the DLS may use 802.11e DLS or 802.11z tunnel DLS (TDLS). A WLAN using an independent BSS (IBSS) mode may not have an AP, and STAs within or using the IBSS (eg, all STAs in the STA) may communicate directly with each other. The IBSS communication mode may sometimes be referred to herein as an "ad-hoc" communication mode.

[0078] When using the 802.11ac infrastructure operating mode or a similar operating mode, the AP may send beacons on a fixed channel (such as a primary channel). The primary channel may be a fixed width (e.g., a 20 MHz wide bandwidth) or a dynamically set width. The primary channel may be a working channel of the BSS and may be used by the STA to establish a connection with the AP. In certain representative embodiments, carrier sense multiple access / collision avoidance (CSMA / CA) may be implemented, for example, in an 802.11 system. For CSMA / CA, a STA (e.g., each STA) (including the AP) may listen to the primary channel. If the primary channel is listened / detected and / or determined to be busy by a particular STA, the particular STA may back off. One STA (e.g., only one station) may transmit at any given time in a given BSS.

[0079] High throughput (HT) STAs may communicate using a 40 MHz wide channel, for example, via a combination of a primary 20 MHz channel and adjacent or non-adjacent 20 MHz channels to form a 40 MHz wide channel.

[0080] Very high throughput (VHT) STA can support 20MHz, 40MHz, 80MHz and / or 160MHz wide channels. 40MHz channels and / or 80MHz channels can be formed by combining continuous 20MHz channels. 160MHz channels can be formed by combining 8 continuous 20MHz channels, or by combining two non-continuous 80MHz channels (this can be called 80+80 configuration). For 80+80 configuration, after channel coding, the data can pass through a segment parser that can divide the data into two streams. Each stream can be processed by inverse fast Fourier transform (IFFT) and time domain processing separately. These streams can be mapped to two 80MHz channels, and data can be sent by sending STA. At the receiver of the receiving STA, the above-mentioned operation for 80+80 configuration can be reversed, and the combined data can be transmitted to the medium access control (MAC).

[0081] 802.11af and 802.11ah support operating modes below 1GHz. Channel operating bandwidths and carriers are reduced in 802.11af and 802.11ah relative to those used in 802.11n and 802.11ac. 802.11af supports 5MHz, 10MHz, and 20MHz bandwidths in the TV White Space (TVWS) spectrum, and 802.11ah supports 1MHz, 2MHz, 4MHz, 8MHz, and 16MHz bandwidths using non-TVWS spectrum. According to a representative embodiment, 802.11ah may support meter type control / machine type communication (MTC), such as MTC devices in macro coverage areas. MTC devices may have certain capabilities, such as limited capabilities, including support for (e.g., only support for) certain bandwidths and / or limited bandwidths. MTC devices may include batteries with battery life above a threshold (e.g., to maintain very long battery life).

[0082] WLAN systems that can support multiple channels and channel bandwidths (such as 802.11n, 802.11ac, 802.11af, and 802.11ah) include channels that can be designated as primary channels. The primary channel may have a bandwidth equal to the maximum common operating bandwidth supported by all STAs in the BSS. The bandwidth of the primary channel may be set and / or limited by a STA (which supports the minimum bandwidth operation mode) from all STAs operating in the BSS. In the example of 802.11ah, for STAs (e.g., MTC-type devices) that support (e.g., only support) a 1MHz mode, the primary channel may be 1MHz wide, even if the AP and other STAs in the BSS support 2MHz, 4MHz, 8MHz, 16MHz, and / or other channel bandwidth operation modes. Carrier sensing and / or network allocation vector (NAV) settings may depend on the state of the primary channel. If the primary channel is busy, for example, because a STA (which only supports a 1MHz operation mode) is transmitting to the AP, all available bands may be considered busy even if most of the available bands remain idle.

[0083] In the United States, the available frequency band for 802.11ah is 902MHz to 928MHz. In South Korea, the available frequency band is 917.5MHz to 923.5MHz. In Japan, the available frequency band is 916.5MHz to 927.5MHz. The total bandwidth available for 802.11ah is 6MHz to 26MHz, depending on the country code.

[0084] Figure 1D1 is a system diagram illustrating the RAN 104 and the CN 106 according to one embodiment. As noted above, the RAN 104 may employ NR radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0085] The RAN 104 may include gNBs 180a, 180b, 180c, though it will be appreciated that the RAN 104 may include any number of gNBs while remaining consistent with an embodiment. The gNBs 180a, 180b, 180c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 116. In one embodiment, the gNBs 180a, 180b, 180c may implement MIMO technology. For example, the gNBs 180a, 180b may utilize beamforming to send signals to and / or receive signals from the gNBs 180a, 180b, 180c. Thus, the gNB 180a may, for example, use multiple antennas to send wireless signals to and / or receive wireless signals from the WTRU 102a. In an embodiment, the gNBs 180a, 180b, 180c may implement carrier aggregation technology. For example, gNB 180a may transmit multiple component carriers to WTRU 102a (not shown). A subset of these component carriers may be on unlicensed spectrum, while the remaining component carriers may be on licensed spectrum. In an embodiment, gNBs 180a, 180b, 180c may implement coordinated multi-point (CoMP) technology. For example, WTRU 102a may receive coordinated transmissions from gNB 180a and gNB 180b (and / or gNB 180c).

[0086] The WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c using transmissions associated with parameter sets that may be scalable. For example, OFDM symbol spacing and / or OFDM subcarrier spacing may vary for different transmissions, different cells, and / or different portions of the wireless transmission spectrum. The WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c using subframes or Transmission Time Intervals (TTIs) of varying or scalable lengths (e.g., containing varying numbers of OFDM symbols and / or varying absolute time lengths over time).

[0087] The gNBs 180a, 180b, 180c may be configured to communicate with the WTRUs 102a, 102b, 102c in a standalone configuration and / or a non-standalone configuration. In a standalone configuration, the WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c while not accessing other RANs (e.g., such as the eNodeBs 160a, 160b, 160c). In a standalone configuration, the WTRUs 102a, 102b, 102c may use one or more of the gNBs 180a, 180b, 180c as mobility anchors. In a standalone configuration, the WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c using signals in an unlicensed band. In a non-standalone configuration, the WTRUs 102a, 102b, 102c may communicate / connect with the gNB 180a, 180b, 180c while also communicating / connecting with another RAN, such as the eNode-B 160a, 160b, 160c. For example, the WTRUs 102a, 102b, 102c may implement the DC principle to communicate with one or more gNBs 180a, 180b, 180c and one or more eNode-Bs 160a, 160b, 160c substantially simultaneously. In a non-standalone configuration, the eNode-B 160a, 160b, 160c may act as a mobility anchor for the WTRUs 102a, 102b, 102c, and the gNB 180a, 180b, 180c may provide additional coverage and / or throughput for serving the WTRUs 102a, 102b, 102c.

[0088] Each of the gNBs 180a, 180b, 180c may be associated with a specific cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in UL and / or DL, support of network slicing, interworking between DC, NR, and E-UTRA, routing of user plane data towards a user plane function (UPF) 184a, 184b, routing of control plane information towards an access and mobility management function (AMF) 182a, 182b, etc. Figure 1D As shown, gNBs 180a, 180b, and 180c may communicate with each other via an Xn interface.

[0089] Figure 1D The illustrated CN 106 may include at least one AMF 182a, 182b, at least one UPF 184a, 184b, at least one session management function (SMF) 183a, 183b, and possible data networks (DNs) 185a, 185b. Although the aforementioned elements are depicted as part of the CN 106, it should be understood that any of these elements may be owned and / or operated by an entity other than the CN operator.

[0090] The AMF 182a, 182b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via the N2 interface and may act as a control node. For example, the AMF 182a, 182b may be responsible for authenticating users of the WTRU 102a, 102b, 102c, support of network slicing (e.g., handling of different protocol data unit (PDU) sessions with different requirements), selecting a specific SMF 183a, 183b, management of registration areas, termination of non-access stratum (NAS) signaling, mobility management, etc. The AMF 182a, 182b may use network slicing in order to customize CN support for the WTRU 102a, 102b, 102c based on the type of services utilized by the WTRU 102a, 102b, 102c. For example, different network slices may be established for different use cases, such as services relying on ultra-reliable low latency (URLLC) access, services relying on enhanced massive mobile broadband (eMBB) access, services for MTC access, etc. The AMF 182a, 182b may provide a control plane function for switching between the RAN 104 and other RANs (not shown) that employ other radio technologies, such as LTE, LTE-A, LTE-A Pro, and / or non-3GPP access technologies, such as WiFi.

[0091] The SMF 183a, 183b may be connected to the AMF 182a, 182b in the CN 106 via the N11 interface. The SMF 183a, 183b may also be connected to the UPF 184a, 184b in the CN 106 via the N4 interface. The SMF 183a, 183b may select and control the UPF 184a, 184b, and configure the traffic routing through the UPF 184a, 184b. The SMF 183a, 183b may perform other functions, such as managing and allocating UE IP addresses, managing PDU sessions, controlling policy implementation and QoS, providing DL data notification, etc. The PDU session type may be IP-based, non-IP-based, Ethernet-based, etc.

[0092] The UPF 184a, 184b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via the N3 interface, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices. The UPF 184, 184b may perform other functions, such as routing and forwarding packets, enforcing user plane policies, supporting multi-homed PDU sessions, handling user plane QoS, buffering DL packets, providing mobility anchoring, etc.

[0093] The CN 106 may facilitate communications with other networks. For example, the CN 106 may include or may communicate with an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) that acts as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to other networks 112, which may include other wired networks and / or wireless networks owned and / or operated by other service providers. In one embodiment, the WTRUs 102a, 102b, 102c may be connected to the DNs 185a, 185b via the UPFs 184a, 184b via the N3 interfaces to the UPFs 184a, 184b and the N6 interfaces between the UPFs 184a, 184b and the local DNs 185a, 185b.

[0094] Given that Figures 1A to 1D as well as Figures 1A to 1D Corresponding to the description of the present invention, one or more or all of the functions described herein with reference to one or more of the following items may be performed by one or more simulation devices (not shown): WTRU102a-102d, base station 114a-114b, evolved Node B 160a-160c, MME 162, SGW 164, PGW 166, gNB 180a-180c, AMF 182a-182b, UPF 184a-184b, SMF 183a-183b, DN 185a-185b and / or any other device described herein. The simulation device may be one or more devices configured to mimic one or more or all of the functions described herein. For example, the simulation device may be used to test other devices and / or simulate network and / or WTRU functions.

[0095] The simulation device may be designed to implement one or more tests of other devices in a laboratory environment and / or in an operator network environment. For example, one or more simulation devices may perform one or more functions or all functions while being fully or partially implemented and / or deployed as part of a wired and / or wireless communication network in order to test other devices within the communication network. One or more simulation devices may perform one or more functions or all functions while being temporarily implemented / deployed as part of a wired and / or wireless communication network. The simulation device may be directly coupled to another device for the purpose of testing and / or performing tests using over-the-air wireless communications.

[0096] One or more simulation devices can perform one or more (including all) functions without being implemented / deployed as part of a wired and / or wireless communication network. For example, the simulation device can be used in a test scenario in a test laboratory and / or a non-deployed (e.g., testing) wired and / or wireless communication network to implement testing of one or more components. One or more simulation devices can be test equipment. Direct RF coupling and / or wireless communication via RF circuits (e.g., which may include one or more antennas) can be used by the simulation device to send and / or receive data.

[0097] In Release 17 in 3GPP, downlink techniques, uplink techniques, and downlink and uplink techniques can be used for position determination. Such techniques can use positioning reference signals, sounding reference signals, and sounding reference signals for positioning purposes. In some implementations, the environment plays a role in the measurement of such reference signals, which can affect the positioning accuracy that can be achieved by such position determination techniques.

[0098] In NR positioning or location determination, the WTRU may not receive PRS from a sufficient or required number (and / or quality) of LOS TRPs, which may result in inaccurate positioning. In some cases, such as in an unknown environment or due to the mobility of the WTRU, the WTRU may not have environmental information available. In some implementations, models can be used to determine these environmental conditions and can be used to enhance positioning performance (e.g., by selecting TRPs, positioning techniques, and / or positioning based on ML models (fingerprinting)).

[0099] An example process for ML-assisted positioning or location determination is as follows. Based on a predefined condition (e.g., the achieved positioning accuracy is less than a threshold), the WTRU sends a ML-assisted NLOS identification request to the network. The WTRU sends its capability information to the network. For example, the capability information may indicate that the WTRU is capable of NLOS identification, such as using a support vector machine (SVM) algorithm. Alternatively, the capability information may indicate that the WTRU is capable of using more than one method (e.g., algorithms based on SVM and neural networks) for NLOS identification. The WTRU receives a list of trained models (e.g., support vector machines, K nearest neighbors, etc.) and input features (e.g., peak power, average power, etc.) for AI / ML-based NLOS identification from the network. The WTRU receives a PRS priority level from the TRP. The WTRU measures the PRS from the network. The WTRU determines a model to use for NLOS identification based on the number of NLOS estimates required per second. If the NLOS estimation time of any model exceeds the maximum allowed NLOS estimation time, the WTRU requests the network to provide a new set of models that require a shorter NLOS estimation time. If the NLOS estimation can be completed in time, the WTRU indicates the selected model to the network. The WTRU uses the AI / ML model to predict / identify NLOS and / or LOS TRPs. Based on the predefined requirements for the conventional positioning algorithm (e.g., minimum number of LOS TRPs > 3), the WTRU sends a request to the network as to whether the network should perform conventional positioning or positioning based on the ML model (e.g., fingerprinting). If the WTRU sends a request to the network to use conventional positioning techniques, it provides the network with a set of (LOS) TRP IDs (e.g., TRP ID-34, TRP ID-27, etc.) to perform conventional TRPs.

[0100] This article uses various abbreviations and acronyms, for example, the following:

[0101] ACK

[0102] AoA Angle of Arrival

[0103] AoD Angle of Departure

[0104] ARFCN Absolute Radio Frequency Channel Number

[0105] BLER Block Error Rate

[0106] BWP Bandwidth Part

[0107] CAP Channel Access Priority

[0108] CAPC Channel Access Priority Class

[0109] CCA Clear Channel Assessment

[0110] CCE Control Channel Element

[0111] CE Control Elements

[0112] CG configuration authorization or cell group

[0113] CORESET Control Resource Set

[0114] CP Cyclic Prefix

[0115] CP-OFDM Cyclic Prefix OFDM

[0116] CQI Channel Quality Indicator

[0117] CRC Cyclic Redundancy Check

[0118] CSI Channel State Information

[0119] CW Contention Window

[0120] CWS contention window size

[0121] CO channel occupancy

[0122] DAI Downlink Assignment Index

[0123] DCI Downlink Control Information

[0124] DFI Downlink Feedback Information

[0125] DG Dynamic Authorization

[0126] DL Downlink

[0127] DM-RS Demodulation Reference Signal

[0128] DRB Data Radio Bearer

[0129] DRX Discontinuous Reception

[0130] ECID Enhanced Cell ID

[0131] eLAA Enhanced License Assisted Access

[0132] eMBB Enhanced Mobile Broadband

[0133] FeLAA further enhanced license assisted access

[0134] HARQ Hybrid Automatic Repeat Request

[0135] IM Interference Measurements

[0136] KNN k nearest neighbor

[0137] LAA License Assisted Access

[0138] LBT Listen before you speak

[0139] LCH Logical Channel

[0140] LCP Logical Channel Priority

[0141] LBT Listen before you speak

[0142] LOS Line of Sight

[0143] NLOS Non Line of Sight

[0144] LMF session management capabilities

[0145] LPP LTE Positioning Protocol

[0146] LTE LTE Long Term Evolution, e.g. from 3GPP LTE Release 8 and higher

[0147] MAC CE MAC Control Element

[0148] MAC Media Access Control

[0149] MCS modulation and coding scheme

[0150] MIMO Multiple Input Multiple Output

[0151] NACK Negative ACK

[0152] NAS Non-Access Stratum

[0153] NR New Radio

[0154] OFDM Orthogonal Frequency Division Multiplexing

[0155] OTDOA Observed Time Difference of Arrival

[0156] PDCCH Physical Downlink Control Channel

[0157] PDSCH Physical Downlink Shared Channel

[0158] PDU Packet Data Unit

[0159] PHY Physical Layer

[0160] PID Process ID

[0161] PO Paging time

[0162] PRACH Physical Random Access Channel

[0163] PRS Positioning Reference Signal

[0164] PRU Positioning Reference Unit

[0165] PSS Primary Synchronization Signal

[0166] PTRS Phase Tracking Reference Signal

[0167] PUCCH Physical Uplink Control Channel

[0168] PUSCH Physical Uplink Shared Channel

[0169] RA Random Access (or Random Access Procedure)

[0170] RACH Random Access Channel

[0171] RAR Random Access Response

[0172] RCU Radio Access Network Central Unit

[0173] RF Radio Front End

[0174] RLF Radio Link Failure

[0175] RLM Radio Link Monitoring

[0176] RNTI Radio Network Identifier

[0177] RNA RAN Notification Area

[0178] RO RACH Timing

[0179] RRC Radio Resource Control

[0180] RRM Radio Resource Management

[0181] RTT Round Trip Time

[0182] RP Receiving Point

[0183] RS reference signal

[0184] RSRP Reference Signal Received Power

[0185] RSTD Reference Signal Time Difference

[0186] RTT Round Trip Time

[0187] RSSI Received Signal Strength Indicator

[0188] RTOA Relative time of arrival

[0189] SDAP Service Data Adaptation Protocol

[0190] SDU Service Data Unit

[0191] SRB Signalling Radio Bearer

[0192] SRS Sounding Reference Signal

[0193] SS Sync Signal

[0194] SSS Auxiliary Sync Signal

[0195] SWG Switching Gap (in an independent subframe)

[0196] SPS Semi-Persistent Scheduling

[0197] SUL Supplementary Uplink

[0198] SVM Support Vector Machine

[0199] TB Transfer Block

[0200] TBS Transport Block Size

[0201] TDoA Time Difference of Arrival

[0202] TRP Transmit Receive Point

[0203] TSC Time Sensitive Communications

[0204] TSN Time Sensitive Network

[0205] TTI Transmit Time Interval

[0206] UCI Uplink Control Information

[0207] UL Uplink

[0208] URLLC Ultra-Reliable Low Latency Communications

[0209] WBWP Wide Bandwidth Part

[0210] WLAN Wireless Local Area Network and related technologies (IEEE 802.xx domain)

[0211] Various techniques are used to determine the location of a device such as a WTRU. For example, downlink techniques, uplink techniques, and downlink and uplink techniques may be used for location determination. Some implementations include one or more of the following types of techniques: DL positioning; UL positioning; and DL and UL positioning.

[0212] DL positioning may refer to any location determination technique based on downlink reference signals such as PRS. For example, in some implementations, the WTRU receives multiple reference signals from the TP and measures DL RSTD and / or RSRP. Examples of DL positioning include DL-AoD and DL-TDOA positioning.

[0213] UL positioning may refer to any location determination technique based on an uplink reference signal such as an SRS. For example, in some implementations, the WTRU sends an SRS to multiple RPs, and the RPs measure UL RTOA and / or RSRP. Examples of UL positioning include UL-TDOA or UL-AoA positioning.

[0214] DL and UL positioning may refer to any position determination technique based on both uplink reference signals and downlink reference signals. For example, in some implementations, the WTRU sends SRS to multiple TRPs and the gNB measures the Rx-Tx time difference. In some implementations, the gNB may measure the RSRP of the received SRS. In some implementations, the WTRU measures the Rx-Tx time difference for PRS sent from multiple TRPs. In some implementations, the WTRU may measure the RSRP of the received PRS. In some implementations, the Rx-TX difference and / or RSRP measured at the WTRU and the gNB are used to calculate the round trip time. Here, the Rx-TX difference refers to the difference between the arrival time of the reference signal sent by the TRP and the transmission time of the reference signal sent from the WTRU. Examples of DL and UL positioning include multi-RTT positioning.

[0215] A "DL positioning method" may refer to any location determination method that uses a downlink reference signal such as a PRS. The WTRU receives multiple reference signals from the TP and measures the DL RSTD and / or RSRP. Examples of DL positioning methods are DL-AoD or DL-TDOA positioning.

[0216] "UL positioning method" may refer to any position determination method that uses an uplink reference signal (such as SRS) for positioning. The WTRU sends the SRS to multiple RPs, and the RPs measure UL RTOA and / or RSRP. Examples of UL positioning methods are UL-TDOA or UL-AoA positioning.

[0217] "DL and UL positioning method" may refer to any position determination method that uses both uplink reference signals and downlink reference signals for positioning. In one example, the WTRU sends an SRS to multiple TRPs and the gNB measures the Rx-Tx time difference. The gNB may measure the RSRP for the received SRS. The WTRU measures the Rx-Tx time difference for the PRS sent from multiple TRPs. The WTRU may measure the RSRP for the received PRS. The Rx-TX difference and the RSRP that may be measured at the WTRU and the gNB are used to calculate the round trip time. Here, the Rx and Tx difference refers to the difference between the arrival time of the reference signal sent by the TRP and the transmission time of the reference signal sent from the UE. An example of a DL and UL positioning method is multi-RTT positioning. In the following, in order to distinguish from ML-assisted positioning, we refer to "DL positioning method", "UL positioning method" and "DL and UL positioning method" as non-ML positioning methods.

[0218] "Machine learning (ML) assisted positioning methods" may refer to any positioning method that uses a trained ML model to transform predefined input measurements to directly predict the location of the WTRU as the output of the model or predict measurements or parameters (e.g., LOS / NLOS indicators, TDOA, PRS-RSRP, etc.) that may be used by non-ML positioning techniques at a later stage to obtain the location of the WTRU. In some cases, the term "artificial intelligence machine learning" (AIML) is used interchangeably with the term ML.

[0219] Some specific implementations involve RSRP fingerprinting. The RSRP fingerprint may refer to a combination of RSRP values ​​of a reference signal that may be sent from multiple TRPs, multiple beams of the same TRP, or a combination of both. The WTRU may generate an RSRP fingerprint based on one or more PRSs received from a reference TRP or multiple neighboring TRPs. WTRUs located within a certain distance (e.g., predefined or preconfigured) from each other may measure similar RSRPs (e.g., low, high) in PRS measurements, and these similar RSRPs may be RSRP fingerprints.

[0220] In some implementations, the RSRP fingerprint can be generated based on an alternative measurement representing the signal strength of the signal. Examples of alternative RSRP include: RSRQ; SINR; QI; and / or a difference in RSRP values.

[0221] Figure 2is a system diagram 200 illustrating example RSRP fingerprints in a wireless environment. In this example, each RSRP fingerprint is generated by measuring the PRS-RSRP transmitted from 12 Tx beams (each beam having a unique Tx angle) of a single TRP 206. WTRU UE 1 and UE 2 are located within a certain distance from each other (indicated as group 204A), so we observe that they exhibit similar RSRP fingerprints (as shown in curve graph 202A, where the upper line corresponds to UE1 and the lower line corresponds to UE2). We observe the same phenomenon for WTRU UE 3 and UE 4 (indicated as group 204B and fingerprint 202B, where the lower line corresponds to UE3 and the upper line corresponds to UE4). In some specific implementations, in the case of NLOS environment between the WTRU and the TRP, the correlation of the fingerprint with the geographic location remains valid. For example, in Figure 2 In the example, UE1 and UE2 experience NLOS, while UE3 and UE4 experience LOS.

[0222] In some cases, environmental conditions may affect positioning accuracy. For example, for timing-based positioning techniques (e.g., DL positioning, UL positioning, and DL and UL positioning), environmental conditions may affect the achievability of positioning accuracy. Examples of such environmental factors include: one or more LOS / NLOS conditions between the WTRU and the TRP link; the location of one or more reference TRPs that send PRSs to the WTRU; and / or the number of TRPs from which the WTRU can detect PRSs.

[0223] An example environmental parameter that may affect the achievable positioning accuracy is the NLOS condition between the WTRU and the TRP link. For example, in an NLOS environment, the WTRU may receive a PRS with multipath reflections.

[0224] Figure 3 300 is a system diagram illustrating an example NLOS propagation link between a WTRU 304 and a TRP 302. In this example, a wall 306A is located between the WTRU 304 and the TRP 302, which blocks line-of-sight propagation (marked by dashed line 308). To overcome this blocking condition, the WTRU 304 receives a reflected signal 310 from the wall 2 306B. It is observed that the reflected signal 310 travels a distance greater than the assumed LOS propagation 308.

[0225] In some implementations, such reflection measurements may include positive deviations in addition to the actual time difference that may have been measured under LOS conditions between the WTRU 304 and the pair of TRPs 302. Such deviations to the timing difference measurements may be propagated to the timing difference measurements. In such cases, if a timing-based positioning technique is applied to such deviated timing difference measurements, it may provide a positioning output indicating a greater distance than the correct distance, which may result in inaccuracies.

[0226] Another factor that may limit the achievable positioning accuracy is the location of the TRP. In some implementations, an error is associated with each propagation delay measurement. For example, if the TRP is located in a row or cluster, the positioning accuracy may be reduced due to the lack of sufficient TRP resolution from the WTRU's perspective.

[0227] Figure 4 400A) and rows / groups of TRPs 402C, 402D with example measurement ranges 404C, 404D (bottom 400B) are shown with uncertainty regions shaded (406, in hash marks). In some implementations, the same phenomenon applies to timing difference based positioning techniques.

[0228] For timing difference based positioning techniques, in some implementations, the PRS coverage of the WTRU is based on at least 4 non-collocated TRPs. However, in some cases, the WTRU may receive PRS from less than 4 reference TRPs; for example, due to the WTRU location (e.g., where the WTRU is located at the corner of a building), lack of PRS resources from certain TRPs, or a combination of both.

[0229] Figure 5 5 is a system diagram 500 illustrating an example WTRU 508 (UE1 in the figure) in a situation where a limited number of PRS can be received. In this example, the WTRU 508 receives PRS from only one TRP; that is, the WTRU 508 may receive PRS 506 only from TRP6 504, with PRS from the remaining TRPs 502 being out of range, blocked, or otherwise unavailable.

[0230] Some specific implementations include artificial intelligence (AI). Artificial intelligence can broadly refer to the behavior exhibited by machines that mimic cognitive functions of perception, reasoning, adaptation, and action. Some specific implementations include machine learning (ML). Machine learning can broadly refer to algorithms that solve problems based on learning through experience (e.g., "data") without explicit programming (e.g., "configuration rule sets"). Machine learning can be considered a subset of AI. Based on the nature or feedback of the data available for learning algorithms, different machine learning paradigms can be envisioned. For example, supervised learning methods can include learning a function that maps inputs to outputs based on labeled training examples, where each training example can include a pair consisting of an input and a corresponding output. Unsupervised learning methods can include detecting patterns in data without pre-existing labels. Reinforcement learning methods can include performing a series of actions in an environment to maximize cumulative rewards. In some specific implementations, it is possible to use a combination of the above-mentioned methods or interpolation to apply machine learning algorithms. For example, a semi-supervised learning method can use a combination of a small amount of labeled data and a large amount of unlabeled data during training. In this regard, semi-supervised learning can be considered to fall between unsupervised learning (no labeled training data) and supervised learning (only labeled training data).

[0231] Some specific implementations include artificial neural networks (ANNs). Figure 6 is a block diagram illustrating an example artificial neural network 600 .

[0232] In some implementations, the purpose of training is to apply inputs and adjust weights, indicated in the figure as w and x (which may be referred to as neuron weights or link weights), so that the output from the neural network approaches the desired target value associated with the input value. In this example, the artificial neural network includes three layers 602, 604, 606, although other numbers of layers may be utilized in various embodiments, wherein each layer includes one or more nodes 608A-608N connected by one or more edges or connectors 610. In some implementations, during training of the artificial neural network, for a given input, the difference between the output and the desired value is calculated, and the difference is used to update the weights in the neural network. If a relatively large difference between the output and the desired value is observed, a relatively large change in the weight may be made, while if a relatively small difference between the output and the desired value is observed, a relatively small change in the weight may be made.

[0233] In some implementations, for positioning, the input may include reference signal parameters, and the output may be an estimated position. In some implementations, the expected value may include position information acquired by a global navigation satellite system (GNSS) (e.g., with high accuracy) or another suitable reference source.

[0234] In some implementations, after the artificial neural network completes its training (e.g., after the difference between the output and the expected value is below a threshold), the artificial neural network can be applied to the positioning task by taking the input and using the output as the expected result for the associated input. The output may be an estimated position or location of the WTRU.

[0235] The following information can be used to train a neural network: inputs to the neural network; expected outputs associated with the inputs; and actual outputs from the neural network, which are compared to the expected outputs. An example neural network model can be characterized by the following parameters: the number of weights; and / or the number of layers in the neural network

[0236] Deep learning (DL) may refer to a class of machine learning algorithms that employ artificial neural networks, particularly deep neural networks (DNNs), which are loosely inspired from biological systems and include at least one hidden layer. DNNs are a class of machine learning models inspired by the human brain, in which the input is linearly transformed and passed through a nonlinear activation function multiple times. DNNs typically include multiple layers, each of which includes a linear transformation function and a nonlinear activation function. In some implementations, DNNs may be trained via back propagation using training data. DNNs may provide high performance in multiple fields (such as speech, vision, natural language, etc.) and for various machine learning settings (e.g., supervised, unsupervised, and semi-supervised).

[0237] As used herein, “network” may include, for example, an AMF, an LMF, a gNB, or an NG-RAN. “Preconfiguration” and “configuration” may be used interchangeably. “Non-serving gNB” and “neighboring gNB” may be used interchangeably. “gNB” and “TRP” may be used interchangeably. “PRS” or “PRS resource” may be used interchangeably. “PRS” or “PRS resource” may be used interchangeably. The aforementioned terms “PRS” or “PRS resource” may belong to different PRS resource sets. “PRS” or “DL-PRS” or “DL PRS” may be used interchangeably. “Measurement gap” and “measurement gap pattern” may be used interchangeably. “Measurement gap pattern” may include parameters such as measurement gap duration or measurement gap repetition period or measurement gap periodicity. The terms “event” or “opportunity” may be used interchangeably.

[0238] A PRU may include any suitable device, such as a WTRU or a TRP, whose location (e.g., altitude, latitude, geographic coordinates, or local coordinates) is known to the network (e.g., gNB, LMF, etc.). The capabilities of a PRU may be the same as a WTRU or a TRP, for example, being able to receive PRS or send SRS or SRS for positioning, return measurements, or send PRS. A WTRU acting as a PRU may be used by the network for calibration purposes (e.g., correcting unknown timing offsets and / or correcting unknown angle offsets).

[0239] The LMF is a non-limiting example of a node or entity (eg, a network node or entity) that may be used for or support positioning. Any other node or entity may replace the LMF and still be consistent with the techniques described herein.

[0240] As discussed previously, artificial intelligence (AI) is a branch of science that solves problems by enabling machines to mimic human behavior. Machine learning (ML) is a subset of artificial intelligence that focuses on solving problems by learning from data and making predictions. The terms AI and ML are used interchangeably in this article.

[0241] A mathematical model is a mathematical formula that approximates the relationship between one or more variables (inputs) and another variable (output). For example, a model can be created based on ML techniques.

[0242] Model training may include the process of providing input variables and output variables to a model to learn from. Model deployment may include the process of deploying a trained model in the real world to make predictions or estimate outputs. Model inference is the process of providing input variables to a model and predicting / calculating outputs. For the scope of this disclosure, the terms "model deployment" and "model inference" may be used interchangeably. Model monitoring may include the process of monitoring the accuracy of output predictions of a particular model. Model accuracy may refer to a metric that compares different classification models, defined as the number of correct predictions made by a model versus the total sample size. Model updating may refer to a process in which one model is replaced by another model for prediction.

[0243] In some cases, positioning accuracy may be degraded, for example, due to poor environmental conditions. For example, in NR positioning, the WTRU may not be able to receive PRS (e.g., 4 for DL-TDOA) from the minimum number of TRPs required for a particular positioning method (e.g., as in the example of a WTRU located at the corner of a building), resulting in a decrease in positioning accuracy. Examples of poor environmental conditions include, for example, the WTRU being in NLOS, where all TRPs, TRPs concentrated in one area, where the WTRU may receive PRS from only one or two TRPs, etc.

[0244] Some implementations include machine learning assisted NLOS identification. In some implementations, the WTRU may be configured to train a machine learning model based on a preconfigured set of inputs and a preconfigured set of outputs. For example, the input set and the output set may be based on one or more of the following: WTRU measurements, sensor data, imaging / video data, positioning information from GNSS and another suitable reference source, input from the network (such as data and / or assistance information), output of a positioning method, etc.

[0245] In some implementations, the NLOS reference TRP may affect positioning accuracy. For example, the WTRU may be in a mixed LOS / NLOS environment such that the WTRU has more than one TRP in the LOS case and the rest of the TRPs from which the WTRU can detect PRS in the NLOS case.

[0246] Figure 7 700 is a system diagram illustrating an example WTRU 708 (UE1 in the figure) in a hybrid LOS / NLOS environment. In this example, 12 TRPs 702A-702C are deployed in an indoor scenario. Based on the location and radio conditions of the WTRU, it can detect PRS from 8 different TRPs (within the dashed oval 706). Of these 8 TRPs, 4 TRPs are LOS 702B and 4 are in NLOS 702C conditions with the WTRU. The LOS TRPs 702B are TRP3, TRP8, TRP10, TRP11 (shown as outlines within the dashed oval 706), and the NLOS TRPs 702C are TRP2, TRP4, TRP5, TRP9 (shown as solid lines within the dashed oval 706). In some implementations, if a subset of neighboring TRPs is selected for PRS transmission and subsequent TDOA measurements (e.g., RSTD) are performed without considering NLOS conditions, the measured TDOA measurement results may include positive measurement biases, resulting in inaccurate positioning.

[0247] In some implementations, the WTRU may receive a predefined positioning accuracy threshold from the network (e.g., from the LMF or gNB). The WTRU may evaluate the positioning accuracy as follows. Based on the positioning method configured by the network (e.g., WTRU-based DL-TDOA, WTRU-based DL-AoD), the WTRU may obtain estimated location information. The WTRU may obtain reference location information via a RAT independent positioning method (e.g., GNSS). The WTRU may evaluate the positioning accuracy by comparing the estimated location information and the reference location information. If the evaluated positioning accuracy is lower than the threshold, the WTRU may initiate an ML model-assisted NLOS identification.

[0248] In some implementations, the WTRU may receive an indication from the network to initiate NLOS identification based on the ML model. After the WTRU receives the indication, the WTRU may receive one or more configurations related to training of the ML model.

[0249] In some implementations, during the ANN training phase, the WTRU may obtain a reference LOS / NLOS condition relative to each TRP (e.g., the reference LOS / NLOS condition refers to an indication of whether the channel is LOS or NLOS, e.g., represented by 0 or 1. The condition may be estimated or generated by the network.). The reference LOS / NLOS condition may be indicated to the WTRU using a LOS / NLOS indicator. In some implementations, the LOS / NLOS indicator may be a "hard" indicator (e.g., 0 or 1) indicating a full LOS or full NLOS condition, or a "soft" indicator (e.g., a value between 0 and 1, inclusive) that may indicate a value between full LOS and full NLOS.

[0250] In some implementations, during the training phase, the WTRU trains the model based on preconfigured inputs and associated outputs. Example preconfigured inputs may include channel impulse response data (e.g., a plurality of taps, each of which is associated with a relative delay relative to a first tap and a relative power relative to a tap with the highest power); peak power (e.g., peak RSRP) data; average power (e.g., average RSRP) data; peak-to-average power ratio data; skew of the channel impulse response data; and / or spike value of the channel impulse response data.

[0251] The outputs may include a marker output (whose "actual value" is known). For example, the actual output may include a LOS indicator. Figure 7 The hard LOS indicators (labeled as outputs) of the illustrated WTRU 708 (UE1) may be as follows: TRP2-0, TRP3-1, TRP4-0, TRP5-0, TRP8-1, TRP9-0, TRP10-1, TRP11-1. In this example, 0 indicates a NLOS condition and 1 indicates a LOS condition.

[0252] As another example, the WTRU 708 may use RSRP measurements made on the configured PRS resources and associate them with the LOS indicator. For example, after the training completion criteria are met, the WTRU stops training and returns the trained model to the network. The completion criteria may include one or more of the following: the difference between the estimated LOS indicator and the reference LOS indicator is below a preconfigured threshold; the change between the estimated LOS indicator for consecutive occasions is below a preconfigured threshold; a threshold number of measurements have been collected (e.g., preconfigured inputs and associated outputs); a threshold number of measurements have been collected per category (e.g., LOS measurements); and / or measurements have been collected for a preconfigured duration (e.g., training is performed for 100 minutes).

[0253] An example of a trained model returned to the network may include: Parameters of trained ML model 1: model type - (e.g., support vector machine); input features - (e.g., peak power, average power); output - (e.g., hard LOS value); kernel function - (e.g., Gaussian).

[0254] In some implementations, the WTRU may indicate to the network whether it is able to train a particular ML model (e.g., SVM). In such implementations, the network may select the model and define the input measurements for the WTRU to perform training (e.g., peak power, average power).

[0255] In some implementations, the WTRU may indicate to the network that it is able to train an ML model without specifying model specific capabilities. In some such implementations, the WTRU and the network interact as follows. The network may send more than one model (e.g., SVM, KNN, and random forest) for NLOS identification. Based on hardware / implementation availability, the WTRU may select an SVM for NLOS identification. The network may provide the WTRU with model details (e.g., input features, outputs, kernel type, kernel scale) of a trained SVM for NLOS identification.

[0256] In some implementations, the WTRU may receive a trained model from the network. Examples of trained models may include parameters for SVM. In some implementations, the WTRU measures channel impulse responses from neighboring TRPs and applies these measurements to the trained model. In some implementations, the WTRU obtains the LOS probability for each neighboring TRP. In some implementations, based on the LOS probabilities of different TRPs, the WTRU selects a neighboring TRP to send PRS and measure TDOA. In some implementations, TDOA measurements performed with the LOS TRP minimize positioning errors.

[0257] In some implementations, the WTRU may indicate to the network that it is capable of ML-based model training. In some such implementations, the WTRU receives a model, a type of input to the model, and an output of the model (e.g., an expected output). In some implementations, the WTRU may select one of the following as an input type to classify the LOS / NLOS condition (obtained as an output). Example pre-configured inputs may include channel impulse response data (e.g., a plurality of taps, each of which is associated with a relative delay relative to a first tap and a relative power relative to a tap with the highest power); peak power (e.g., peak RSRP) data; average power (e.g., average RSRP) data; peak-to-average power ratio data; skew of the channel impulse response data; and / or spike value of the channel impulse response data.

[0258] In some implementations, the WTRU may select an ML model to classify the LOS / NLOS condition. In some implementations, the ML model may include one of the following ML models: support vector machine; logistic regression; K-nearest neighbor; random forest; decision tree; and / or naive Bayes.

[0259] Some implementations include model selection for NLOS identification.

[0260] Figure 8 is a message sequence chart 800 illustrating an example model selection process. Figure 8 In an example of FIG. 8 , a WTRU or UE 802 (generally referred to as WTRU 802) receives a location information request 808. The WTRU 802 receives a PRS configuration 810, for example, from an LMF 806. The WTRU 802 receives PRSs from a plurality of TRPs (not shown). Based on predefined conditions, the WTRU sends a ML assisted positioning request 812 to a network device 804. The WTRU receives a ML model for training and / or deployment 814, for example, from the network device 804. The WTRU performs validation and / or selection 816 to validate the ML model and / or selects a ML model to perform training and / or deployment based on coverage of the model. Examples of model coverage include the WTRU's geographic location, a reference TRP, a reference beam ID, a NLOS-only environment, or an indoor environment. In some implementations, the WTRU sends a notification 818 indicating a validation status and / or model selection to the network device 804 at step 818, and performs training and / or deployment 820 of the model to determine location information, or for training purposes. In some implementations, the WTRU returns location information or an indication 822 of the trained model to the network (e.g., to the LMF 806).

[0261] Examples of configuration aspects associated with ML models, such as SVM models, are described herein. The term "ML model" is used to describe AI and / or ML models. The various examples and implementations described herein are not limited to these exemplary configuration aspects and / or SVM models, and are applicable to any attributes, properties, and / or characteristics associated with any ML model. For example, configuration aspects of a model may include one or more of: the dimensionality of the model; the kernel type or its configuration; and / or regularization parameters.

[0262] With respect to the model dimension configuration, for example, the dimension configuration may include the input shape and / or input dimension of the model, the number of features, the number of categories for classification, category labels, category weights, etc.

[0263] For kernel type configuration, for example, a kernel may refer to any function that takes as input in a low dimensional space and transforms it to a higher dimensional space. In some solutions, kernels may be applied to classifications involving nonlinear decision boundaries. For example, a kernel configuration may indicate the type of kernel, such as a linear kernel, a polynomial kernel, a Gaussian kernel, a radial basis kernel, etc. In some implementations, each kernel may include additional parameterizations based on the type of kernel.

[0264] For regularization parameter configuration, for example, the regularization parameter can control the tradeoff between overfitting (e.g., good model performance on the training dataset but poor performance during testing and / or deployment) and generalization (e.g., good model performance during testing and / or deployment). A relatively low value of the regularization parameter can result in reduced classification error (e.g., during training) at the expense of poor generalization (e.g., during deployment). A relatively high value of the regularization parameter can result in better generalization at the expense of some misclassification.

[0265] In some implementations, the WTRU may be configured with one or more attributes associated with the ML model, and the WTRU may train the ML model based on one or more of the configured attributes.

[0266] For example, in some implementations, the WTRU may receive an RRC configuration including at least one configuration aspect of an ML model associated with NLOS identification. For example, the configuration may include the type of ML model used for NLOS identification. Example types of ML models include SVM, logistic regression, k-nearest neighbor, random forest, decision tree, naive Bayes, etc. For example, the WTRU may be configured to apply an ML model (e.g., an SVM ML model) to NLOS identification. In some implementations, the RRC configuration may include configuration aspects of the model. In some implementations, the WTRU may apply the received configuration when training the model for NLOS identification. In some implementations, the WTRU may receive a range of values ​​(e.g., minimum and maximum values) for a specific parameter associated with the model as a configuration. In some implementations, the WTRU may select a specific value within a pre-configured range. For example, the WTRU may be configured to input one or more aspects related to channel measurements, including but not limited to channel impulse response, peak power, average power, peak-to-average power ratio, skew of the channel impulse response, spike value of the channel impulse response, etc.

[0267] In some implementations, the WTRU may be configured to determine one or more attributes and / or parameters associated with an ML model that meets preconfigured criteria (e.g., preconfigured training and / or deployment performance metrics), and the WTRU may be configured to report one or more of the determined attributes / parameters associated with the ML model.

[0268] For example, the WTRU may determine at least one configuration aspect of a model (e.g., an SVM model) trained for NLOS identification such that a pre-configured condition is satisfied, and may report the determined configuration aspect of the model. In some implementations, the pre-configured condition may include one or more of the following: a parameter (or parameter set) that minimizes the training time for a given training data set, a parameter (or parameter set) that maximizes the classification accuracy, a parameter (or parameter set) that minimizes the PRS overhead (or equivalently, the size of the training data), a parameter (or parameter set) that minimizes the inference time (e.g., maximizing the number of inferences per second), and / or a parameter (or parameter set) that minimizes the input size, etc.

[0269] In some implementations, the WTRU may be configured with two (or more) ML models. In some implementations, the WTRU may use the output of one model to determine one or more aspects related to the selection and / or configuration of another ML model. In some implementations, the WTRU may apply the other ML model to obtain the WTRU position and / or its parameters.

[0270] In some implementations, the WTRU may be configured with a first ML model and a set of ML models and conditions for determining a ML model from the set of ML models that may be used as a second ML model. For example, the first ML model may be used for NLOS identification. Based on the output of the first ML model, for example, if a NLOS condition is detected and if at least one ML model is configured, the WTRU may determine the suitability of an ML model from the set of ML models to be used as a second ML model. If an applicable second ML model exists, the WTRU may apply the second ML model to perform fingerprint-based positioning. If an applicable second ML model does not exist (e.g., such a model does not exist, or the model does not meet the suitability criteria), the WTRU may use a traditional positioning method. In some implementations, the suitability criteria may include one or more of the following: whether the ML model is valid for the current geographic location, whether the performance of the ML model is above a pre-configured threshold, whether the ML model inference latency is above a threshold, etc. In some implementations, the WTRU may be configured with a first ML model, wherein the output of the first ML model can be used to select input features and / or pre-process inputs for the second ML model.

[0271] Fig. 9 900 is a flow chart illustrating an example process for determining a WTRU location using more than one ML model. Fig. 9 In the example of , the WTRU receives an indication identifying a LOS / NLOS state for a TRP. The WTRU is configured with a positioning method (e.g., DL-TDOA). At 902, the WTRU uses a LOS / NLOS identification ML model to determine a LOS / NLOS condition. Under the condition 904 that the path corresponding to the TRP is NLOS (904, yes), the WTRU determines at 906 from one or more models 908 whether there is an applicable ML model. Under the condition 910 that an applicable ML model is available at the WTRU (910, yes), the WTRU selects the model at 914 and applies the ML model for positioning at 916. Under the condition 910 that an applicable ML model is not available at the WTRU (910, no), the WTRU determines at 912 to use the configured positioning method. Under the condition 904 that the path corresponding to the TPR is LOS (904, no), the WTRU determines at 912 to use the configured positioning method (e.g., DL-TDOA).

[0272] Some implementations include WTRU-based training for NLOS identification. In some implementations, the WTRU may receive a model for training and / or parameters associated with the model. The WTRU may determine to train the model at the WTRU. The WTRU may receive a target output type (e.g., hard or soft LOS indicator) and an input type corresponding to the model (e.g., RSRP, channel impulse response measurement, RSTD), and may perform training when the WTRU receives and measures PRS. In some implementations, the WTRU may receive a PRS priority level from the network. In some implementations, the WTRU may receive a duration for training from the network. In some implementations, the WTRU may determine, for example, based on the PRS priority level, that the training cannot be completed at the end of the duration. In some implementations, the WTRU may request a training model that requires a shorter training duration, for example, in response to determining that the training cannot be completed within a specific time (e.g., at the end of the duration).

[0273] The following example process is illustrative: The WTRU may receive a list of models (e.g., SVM, KNN) and training completion criteria for ML (e.g., minimum validation accuracy) from the network. The WTRU may receive a PRS priority level from the TRP. The WTRU may receive a timing of a first LOS / NLOS report (or LOS / NLOS reporting frequency) from the network. The WTRU may measure the PRS from the network. The WTRU may determine a model for LOS / NLOS classification based on the PRS priority and the training time. In some implementations, if the priority level of the PRS is low, the WTRU may not be able to process the PRS as scheduled. If the training time of any model exceeds the timing of the first LOS / NLOS report and / or the PRS priority level is low, the WTRU requests a new set of models that require a shorter training time from the network. If the training can be completed before the timing of the first report, the WTRU indicates the selected model. After the training is completed based on the predefined criteria (e.g., after achieving a predefined validation accuracy), the WTRU may send a training completion notification and / or details of the trained model (e.g., kernel function) to the network.

[0274] Some implementations include WTRU-assisted training for NLOS identification. In some implementations, the WTRU may return measurements made on the PRS to the network, which are used by the network for training performed by the network. The WTRU may receive an indication from the network to perform a position estimate, such as so that the WTRU may report the measurement results and the associated position estimate to the network. In some implementations, the WTRU may assist the network in collecting measurements for NLOS identification. In some implementations, the WTRU may log these measurements.

[0275] Some implementations include WTRU-based deployments for NLOS identification. For example, in some implementations, the WTRU may estimate the LOS indicator using an ML model. In some implementations, the WTRU may receive an indication from the network to estimate the LOS indicator for the indicated PRS resource and / or TRP using an ML model. In some implementations, the WTRU may receive parameters related to the ML model (e.g., dimensions of the SVM model).

[0276] The following example process is illustrative: Based on a predefined condition (e.g., the achieved positioning accuracy is less than a threshold), the WTRU may send a request for ML-assisted NLOS identification to the network. The WTRU may send its capability information to the network. The WTRU may receive a list of trained models (e.g., SVM, KNN) and input features (e.g., peak power, average power) for AI / ML-based NLOS identification from the network. The WTRU may receive a PRS priority level from the TRP. The WTRU may measure the PRS from the network. The WTRU may determine a model to use for NLOS identification based on the number of NLOS estimates required per second. If the NLOS estimation time of any model exceeds the maximum allowed NLOS estimation time, the WTRU may request a new set of models that require a shorter NLOS estimation time from the network. If the NLOS estimation can be completed in time, the WTRU may indicate the selected model to the network. The WTRU may use the AI / ML model to predict and / or identify NLOS and / or LOS TRP. Based on predefined requirements for conventional positioning algorithms (e.g., minimum number of LOS TRPs > 3), the WTRU may send a request to the network as to whether conventional positioning or positioning based on an ML model (e.g., fingerprinting) should be performed. If the WTRU sends a request to the network to use conventional positioning techniques, it may provide the network with a set of (LOS) TRP IDs for performing conventional positioning.

[0277] Some implementations provide for WTRU assisted deployment for NLOS identification. In some implementations, the WTRU may provide assistance to the network to estimate the LOS indicator for the PRS resource or TRP. In some implementations, the WTRU may return the measurement results to the network so that the network can use the measurement results to derive the LOS indicator.

[0278] Some implementations provide machine learning assisted positioning with fingerprinting. For example, the WTRU may be configured to deploy a machine learning model that can convert predefined inputs into positioning information (as the output of the model). In some implementations, the set of inputs and the set of outputs may be based on one or more of the following: WTRU RSRP measurement results, WTRU SINR measurement results, sensor data, imaging / video data, positioning information (e.g., GNSS, etc.), input from the network (including data and / or assistance information), output of the positioning method, etc.

[0279] In some implementations, the RSRP fingerprint is a unique sequence of measured RSRP values ​​from different transmission sources (e.g., unique, locally unique, or otherwise identifiable). In some implementations, the transmission sources can be different TRPs or different beams from the same TRP. In some implementations, the RSRP fingerprint is unique for each geographic location, and WTRUs located in the same geographic location (e.g., within a threshold distance of each other, or within a defined boundary) record similar RSRP patterns. In some implementations, the machine learning model learns the RSRP fingerprint pattern for each location during a training phase. In some implementations, during a deployment phase, the model receives the RSRP fingerprint as input and estimates and / or predicts the location as output.

[0280] In some implementations, the fingerprint may include RSRP measurements based on PRS sent from one or more TRPs. Figure 7 For a WTRU in a mixed LOS / NLOS environment, an example RSRP fingerprint may have the following form: Per-TRP RSRP fingerprint for UE1: {TRP2: -90dbm, TRP3: -72dbm, TRP4: -68dbm, TRP9: -75dbm} Per-beam per-TRP RSRP fingerprint for UE1: {TRP2_beam5: -99dbm, TRP3_beam2: -80dbm, TRP4_beam6: -82dbm, TRP9_beam1: -75dbm} Per-beam RSRP fingerprint for a single TRP: {TRP4_beam1: -99dbm, TRP4_beam2: -80dbm, TRP4_beam3: -72dbm, TRP4_beam4: -85dbm}.

[0281] Some implementations include fingerprint-based positioning in deployments. For example, Fig.101 is a block diagram 1000 illustrating an example model deployment process for fingerprint-based positioning. In some implementations, the model 1002 may have limited coverage (e.g., may be limited to a specific TRP, AMF, and / or LMF) or limited to a specific environment (e.g., specific to a NLOS environment or NLOSTRP). The model 1002 may include a trained neural network (NN), such as the one described above in conjunction with Figure 6 The described networks, or any other type or form of machine learning model, classifier, or predictor.

[0282] Examples of RSRP fingerprinting inputs may include one or more of the following: RSRP for each PRS resource; RSRP for each TRP; average RSRP for each PRS resource; statistical measures of RSRP for PRS resources (e.g., standard deviation, variance); PRS resource ID; PRS resource set ID; location of the TRP; line-of-sight angle information for the PRS resources; PRS resource ID corresponding to the maximum or minimum RSRP for the TRP; and / or a LOS indicator for each TRP.

[0283] In some implementations, in addition to or in lieu of RSRP-based fingerprinting, fingerprinting may be based on one or more of: TDOA per PRS resource; TDOA per TRP; TOA per PRS resource; TOA per TRP; RSRP per path per TRP; and / or RSRP per path per PRS resource. Fingerprint input may be provided to the model in any suitable form, such as an array, a string, a concatenated value with a predetermined length, a variable length value with a delimiter, etc.

[0284] In some implementations, the WTRU positioning information output may be expressed, for example, in one or more of the following forms: Cartesian coordinates; spherical coordinates; GPS coordinates; and / or distance to a reference object.

[0285] In some implementations, the WTRU determines neighboring LOS TRPs. For example, if the available LOS TRPs are less than a preconfigured threshold in number (e.g., <3), the WTRU may send a request to the network to initiate fingerprint-based positioning. In some implementations, the WTRU determines that it can receive PRS from only one TRP. In some such cases, the WTRU sends a request to the network to initiate fingerprint-based positioning based on a single TRP.

[0286] Some implementations include model training for fingerprint-based positioning. For example, in some implementations, the WTRU may indicate to the network that it is capable of training an ML model for fingerprint-based positioning.

[0287] Fig.111 is a block diagram 1100 illustrating an example model training process for fingerprint-based positioning. Fig.11 In the example of , the WTRU may receive a model and an RSRP fingerprint configuration as input. The model 1102 may include a trained neural network, such as the one described above in conjunction with Fig.10 The network described herein, or any other type or form of machine learning model, classifier, or predictor. For example, the WTRU may receive Fig.10 The model and fingerprint configuration utilized in the implementation of . In this example, the WTRU is able to determine its location (e.g., using GNSS-based location determination or sensor-based location determination). As another example, the WTRU may obtain RAT-dependent positioning, for example, during a training phase, such as described below. In some specific implementations, for example, based on a positioning method configured by the network (e.g., WTRU-based DL-TDOA, WTRU-based DL-AoD), the WTRU may obtain estimated location information. In some specific implementations, the WTRU may evaluate the positioning accuracy by comparing the estimated location information with the reference location information. After achieving the desired model accuracy, the WTRU may terminate the training and return the model to the network.

[0288] In some implementations, the WTRU may receive location points to cover for training purposes. For example, in some implementations, the WTRU moves to each location point (e.g., received from the network) in a deployment area (e.g., 50m radius from TRP1 with a measurement density of 10 measurements / m2) and provides RSRP fingerprint measurements to the model during the training process.

[0289] Fingerprinting may include single TRP fingerprinting and / or multi-TRP fingerprinting. For example, in some implementations, the WTRU requests training and deploys single TRP training. In some implementations, in single TRP training, all RSRP measurements are recorded from a single TRP. The WTRU may receive a TRP-related configuration, such as a TRP ID, from the network. For example, based on the TRP ID, the WTRU may measure the PRS sent from the TRP with the TRP ID and train an ML model for fingerprinting.

[0290] In some implementations, in multi-TRP fingerprinting, RSRP fingerprints are recorded from multiple TRPs. In some implementations, multi-TRP fingerprinting provides a higher resolution of the input data, which may result in a relatively complex machine learning model. In some implementations, the WTRU may receive a configuration related to the TRP, such as a TRP ID, from the network. For example, based on the TRP ID, the WTRU may measure the PRS sent from the TRP with the TRP ID and train the ML model for fingerprinting.

[0291] Various criteria can be used to select the type of fingerprinting. For example, in some implementations, if the WTRU identifies that it can receive PRS from only one TRP, then the WTRU requests single-TRP training. In some implementations, if the WTRU has limited computing power and cannot handle complex machine learning models, for example due to hardware resources or processing latency limitations, then the WTRU requests single-TRP training. In some implementations, if the WTRU identifies that a limited number of RSRP measurements are available from a single TRP, then the WTRU requests multi-TRP training.

[0292] Some implementations include exit conditions for fingerprinting. For example, in some implementations, the WTRU may fall back to conventional positioning based on predefined fallback criteria (e.g., if the achieved positioning accuracy is less than a threshold, if the number of visible LOS TRPs is less than a threshold, etc.).

[0293] In some implementations, conventional positioning refers to positioning that does not rely on models that require training. In some implementations, conventional positioning may include measurement results (e.g., RSRP, RSTD, AoA, AoD), and the WTRU and / or the network may determine location information based on linear or nonlinear processing of the measurement results. Examples of conventional positioning include DL-TDOA, DL-AoD, multi-RTT, UL-TDOA, UL-AoA.

[0294] Some implementations include WTRU-based training for positioning using fingerprinting. For example, in some implementations, the WTRU may determine validity conditions for the trained ML model during training. In some implementations, before initiating the training phase, the WTRU determines a method for determining its position. In some implementations, the WTRU may be configured with a positioning method that is RAT-dependent (e.g., DL-TDOA) or RAT-independent (e.g., GNSS).

[0295] In some implementations, the number of TRPs configured for the WTRU may vary depending on the configured positioning method. For example, in some implementations, if the WTRU is configured with a conventional RAT-dependent positioning method (e.g., DL-TDOA), the WTRU may be configured with more than one TRP. In some implementations, if the WTRU is configured with a RAT-independent positioning method, the WTRU may be configured with one TRP for training the ML model. In some implementations, during training, the WTRU may use measurements of PRS sent from the configured TRP as input to the ML model and use the position determined from the RAT-dependent / RAT-independent positioning method as the target. In some implementations, after training is completed (e.g., after a timer reaches the training duration, or after a time window in which the WTRU performs training times out), the WTRU determines the validity for the ML model. The validity for the ML model may be expressed in terms of, for example, area validity, time validity, and / or accuracy validity.

[0296] In some implementations, for example, for area validity, the WTRU may determine to associate a TRP ID with an ML model so that the WTRU or the network may use the ML model when the WTRU is located near the TRP associated with the TRP ID. In some implementations, for example, for time validity, the WTRU may determine a duration for which the WTRU may use the AI / ML model for position estimation. The WTRU may determine the time validity for the ML model based on the training duration or measurement conditions (e.g., RSRP of the PRS). In some implementations for accuracy validity, for example, the WTRU may associate a confidence level or indicator indicating the credibility of the trained ML model. The WTRU may be pre-configured with a table associating the confidence level of the PRS with the RSRP. For example, the WTRU may determine that a low RSRP is associated with a low confidence level.

[0297] In some implementations, if the WTRU is configured with more than one TRP, the WTRU may determine to associate a TRP with an ML model from which the WTRU receives a PRS with an RSRP above a preconfigured threshold. In some implementations, if the WTRU is configured with one TRP, the WTRU may determine to associate a TRP with an ML model.

[0298] The following example procedures are illustrative: The WTRU may determine a method to derive its location (e.g., WTRU-based RAT-dependent method, GNSS, WiFi) based on availability and / or WTRU capabilities. If the WTRU has an available method to determine its location, the WTRU may indicate to the NW (e.g., via signaling, WTRU capabilities) that the WTRU has the capability to train an ML model. The WTRU may receive a PRS configuration from the NW (e.g., a list of TRPs from which the WTRU receives PRSs) and may receive an indication to perform ML training. If the WTRU indicates a RAT-dependent positioning method to derive its location, more than one TRP may be configured. If the WTRU indicates a RAT-independent positioning method, one TRP may be configured by the TRP. The WTRU may request a training window (e.g., MG). The WTRU may receive a window configuration from the NW. During the window, the WTRU may receive PRSs and may process measurements (e.g., RSRP, RSTD, AoA, AoD, ToA) obtained from the received PRSs and train the ML model using the determined WTRU location. After the window times out, the WTRU may determine a validity condition for the trained ML model based on the measurement conditions (e.g., duration of the authorized window, corrected measurement quantity, RSRP statistics such as average RSRP, frequency information (e.g., frequency layer, FR1 / FR2, BWP, band, method used to determine its location), where the validity condition may be expressed in terms of time and / or associated TRP. If the WTRU is configured with multiple TRPs, the WTRU may determine the TRP associated with the trained ML model based on the number of beams with RSRP above a threshold. If the WTRU is configured with one TRP, the WTRU may determine to associate the configured TRP with the ML model.) The WTRU may return the weights for the ML model to the NW, along with, for example, the associated validity, the associated TRP, and the method used to train the ML model, confidence intervals (and possibly also a timestamp when the training was completed or a timestamp indicating when the weights for the ML model were sent), etc.

[0299] The following example process for multi-TRP training based on WTRU location determination is illustrative: The WTRU may receive a PRS configuration from the NW. The WTRU may request a window. The WTRU may receive a window configuration from the NW. The WTRU may determine the validity area to which the WTRU belongs based on the determined location. The WTRU may train the ML model using measurements from the TRPs in the area validity. The WTRU may associate the validity area with the trained ML model. The WTRU may return the trained ML model to the NW and may return the associated validity and / or method for training the ML model.

[0300] Some implementations include WTRU-assisted training for positioning using fingerprinting. For example, in some implementations, the WTRU may return measurements made on the PRS to the network for WTRU-assisted training for positioning based on fingerprinting. Examples of measurement results may include RSRP, RSTD, AoA, etc.

[0301] Some implementations include WTRU-based deployments for positioning using fingerprinting. For example, in some implementations, the WTRU may determine an ML model based on a target TRP. The WTRU may determine the target TRP based on measurements of a PRS (e.g., RSRP). The WTRU may determine whether an authorized model or an available ML model at the WTRU is associated with the target TRP. If at least one of the ML models is associated with the target TRP, the WTRU may determine to use the ML model to estimate its position.

[0302] The following example process is illustrative: The WTRU may receive a PRS configuration from the NW. The WTRU may request a ML model corresponding to the configured TRP from the NW. The NW may grant the WTRU the trained ML model and its validity condition. The WTRU may determine a target TRP based on the RSRP of the PRS sent from the TRP (e.g., the strongest RSRP, RSRP above a threshold). Based on the validity condition for the trained ML model: If the WTRU has a valid ML model for the target TRP, the WTRU may use the ML model to determine its position. If the WTRU does not have a valid ML model for the target TRP, the WTRU may determine to send a request to the NW to configure a positioning method (e.g., DL-TDOA, DL-AoD) and initiate the positioning method).

[0303] If one or more pre-configured ML models or authorized ML models are associated with more than one TRP, the WTRU may determine which zone the WTRU is located in. In some implementations, the WTRU may determine the zone based on RSRP measurements of the configured PRS.

[0304] The following example process is illustrative: The WTRU may receive a PRS configuration from the NW. The WTRU may request a ML model, for example, with associated zones, and for example, relationships between zones and TRPs. The NW may grant the WTRU a trained ML model and may provide its validity conditions. The WTRU may determine a target zone based on the RSRP of the PRS sent from the TRP (e.g., strongest RSRP, RSRP above a threshold). Based on the validity conditions for the trained NN: If the WTRU has a valid NN for the target zone, the WTRU may use the NN to determine its position. If the WTRU does not have a valid NN for the target TRP, the WTRU may determine to send a request to the NW to configure a positioning method (e.g., DL-TDOA, DL-AoD) and may initiate a positioning process.

[0305] Some implementations provide for WTRU-assisted deployment for positioning using fingerprinting. For example, in some implementations, in a WTRU-assisted deployment for fingerprinting-based positioning, the WTRU may provide a target TRP to the network so that the network may determine which ML model to use. The WTRU may determine the target TRP based on measurements made on the PRS.

[0306] The following process is exemplary: The WTRU may receive a PRS configuration from the NW. The WTRU may determine a target TRP based on the RSRP of the PRS sent from the TRP (eg, the strongest RSRP, RSRP above a threshold). The WTRU may send metadata for model selection to the network (e.g., the metadata may include a data set that provides additional information about the WTRU related to the model to make a decision), which may include one or more of the following: auxiliary information: (e.g., TRP Id, LOS / NLOS conditions with TRP, speed) and / or model selection criteria (e.g., power requirements (low complexity model for low power WTRU), criticality, required accuracy, previous positioning methods, etc.). The NW may grant the trained ML model and its validity condition to the WTRU. Based on the validity condition for the trained NN: if the WTRU has a valid ML model for the target TRP, the WTRU may use the NN to determine its position, or if the WTRU does not have a valid ML model for the target TRP, the WTRU may determine to send a request to the NW to configure a positioning method (e.g., DL-TDOA, DL-AoD) and initiate the positioning method.

[0307] Some implementations include model monitoring and updating for positioning enhancement. For example, in some implementations, multiple ML models (e.g., for NLOS identification and / or fingerprint-based positioning) may have been trained, where each model is optimized for a particular condition or use case (e.g., a model specific to a TRP, a group of TRPs, a beam set, an NLOS environment, an indoor factory environment, coverage of a single TRP, WTRU computing power, etc.). The WTRU may receive a validity condition for the model after downloading the model. When in the deployment phase, the WTRU may periodically monitor the validity condition associated with the model.

[0308] Some implementations relate to model validity. For example, in some implementations, the WTRU may receive a set of ML models with trained models having optimization criteria for each model (e.g., for NLOS identification and / or fingerprint-based positioning). The WTRU may identify the appropriate criteria and request the corresponding model (e.g., a model optimized for NLOS environments) from the network. In some such cases, the WTRU receives a single trained model with the requested criteria. In some implementations, after receiving the trained model, the WTRU may use the trained model until the validity criteria for the model are met or the model is no longer valid.

[0309] Some implementations involve model monitoring. For example, in some implementations, the WTRU may receive an ML model and an associated timing-based validity criterion. The WTRU may receive a model with a validity based on future timestamp information. In some implementations, if the validity timestamp is exceeded (or the expected time elapses), the WTRU receives a new model or switches to conventional positioning techniques.

[0310] In some implementations, the WTRU may receive an ML model and associated condition-based validity criteria. The WTRU may periodically monitor the conditions associated with the model (e.g., reference TRP ID, beam ID, number of LOS TRPs, model accuracy, etc.), and the WTRU may periodically monitor the validity criteria. If the selected model is no longer valid, the WTRU may notify the network.

[0311] Some implementations involve model updates. For example, in some implementations, the WTRU may update and / or replace the ML model and may download the new ML model and train and / or deploy specific performance metrics. In some implementations, the model update may be followed by a model monitoring process. In some implementations, the WTRU may receive updates from the TRP, LMF, or a local (within the WTRU) trained model, depending on the location of the trained model.

[0312] Some implementations include an indication of a trained model upon or upon entering or leaving an area. For example, in some implementations, the WTRU may send an indication of an ML model and associated parameters upon detecting that it is entering (or leaving) an area on which such a model was trained. For example, the detection may be based on determining that a measurement metric (such as RSRP) is above a threshold for a particular cell or for a particular set of cells. As another example, the detection may be based on the identity of a serving cell in connected mode, inactive mode, or idle mode. In such cases, the WTRU may initiate the sending of an indication after a handover (RRC configuration with mobility) and / or a cell reselection from / to a serving cell. The identity of the serving cell may include at least one of a universal cell identifier, a physical cell identity, and / or a carrier frequency. The WTRU may have determined the measurement threshold and cell set during previous use of the model.

[0313] In some implementations, the WTRU may include in the indication one or more of: an identifier for the trained model; an identifier for the WTRU; and / or at least one indication of a positioning measurement configuration that may be used with the trained model.

[0314] In some implementations, the indication of the positioning measurement configuration may include at least one attribute for at least one positioning resource, such as a carrier frequency, a periodicity, a frequency range, and / or a bandwidth portion. In some implementations, the indication may include a periodicity for measurement reporting. In some implementations, the indication may include a separate indication for each such attribute, or a separate indication of a single label indicating one of a set of possible positioning measurement configurations. In some implementations, the WTRU may have previously received, via signaling (e.g., RRC and / or LPP), a configuration for an identifier for the trained model and at least one label for at least one positioning measurement configuration, such as during earlier use of the trained model.

[0315] Some implementations involve adaptive hybrid positioning (e.g., positioning technology selection). In some cases, it may be more practical for the WTRU to apply an ML-assisted positioning method rather than a non-ML positioning method, for example, due to the positioning accuracy that each method can achieve. Some example cases where it may be more practical for the WTRU to apply an ML-assisted positioning method include when the WTRU is at the edge of a room (e.g., at the edge of a factory hall) and it is in LOS with 2 TRPs (e.g., as in Fig.12 ) and / or the WTRU is at the edge of a room (e.g., at the edge of a factory hall) and it is in LOS with 1 TRP (e.g., as shown in Fig.13 In some implementations, if non-ML positioning techniques are applied to two WTRUs (e.g., Fig.12 and Fig.13 ), the positioning accuracy may be reduced (e.g., relative to corresponding ML-based techniques), for example, due to a lack of a sufficient number of TRPs available to locate the WTRU in the corner of the room, or due to the presence of obstacles (e.g., blocking one or more TRPs).

[0316] In some cases (e.g., due to different environmental factors and / or TRP locations relative to the WTRU), the WTRU may experience a reduction in positioning accuracy after applying non-ML positioning techniques. However, in some cases, such as due to practical limitations (e.g., unavailability of a trained ML model, power consumption limitations of the WTRU, hardware limitations of the WTRU, etc.), it may not be practical to implement ML-assisted positioning methods. It should be noted that non-ML-assisted positioning methods and ML-assisted positioning methods may include more than one method. For example, non-ML positioning includes DL-TDOA, AoD, etc., and ML-assisted positioning techniques include single TRP positioning, multi-TRP positioning, etc.

[0317] In some cases, for example, due to the location of the WTRU, different positioning techniques and / or sub-techniques (e.g., single TRP positioning versus multiple TRP positioning in the case of ML-assisted positioning, or timing-based positioning versus angle-based positioning for non-ML positioning) may result in different achievable positioning accuracy. Therefore, in some implementations, it may be desirable to select a preferred or optimal positioning method.

[0318] Fig.12 A perspective view 1200A and a plan view 1200B of an example deployment of a WTRU 1204 at a corner of a deployment area 1208 with 2 LOS TRPs 1202A-1202B. The environment or deployment area may include obstacles 1206 that may not necessarily interfere with the signals from the TRPs 1202A-1202B such that they may provide LOS signals. Fig.13 A perspective view 1300A and a plan view 1300B of an example deployment of a WTRU 1304 at a corner of a deployment area 1308 with one LOS TRP 1302A. A non-LOS TRP 1302B may be blocked by an obstacle such as obstacle 1310, such that the TRP 1302 cannot provide a LOS signal (although a non-LOS path may exist).

[0319] In some implementations, for example, to address the aforementioned issues, a hybrid approach (e.g., a two-step adaptive hybrid positioning approach between non-ML positioning techniques and ML-assisted positioning techniques) may be implemented. In some implementations, the hybrid approach includes a position estimate of the WTRU and a selection of a positioning technique by the WTRU. For example, the position estimate may be performed by the WTRU based on measurements made by the WTRU.

[0320] In some cases, the techniques described herein may be referred to as "adaptive hybrid" because they may use ML-assisted positioning or non-ML-assisted positioning depending on the situation. In some implementations, whether to use ML-assisted positioning or non-ML-assisted positioning may be based on an estimate of the WTRU's position (e.g., in combination with other predefined conditions) to obtain the WTRU's position.

[0321] In some implementations, a position estimate (eg, an initial estimated position or determination of a coarse position of the WTRU) may be derived using non-ML-assisted position estimates or ML-assisted position estimates.

[0322] In some implementations, the positioning technology used by the WTRU to determine its position is selected based on predefined conditions associated with each positioning method received by the WTRU from the network (e.g., from the LMF) (e.g., a threshold number of LOS TRPs, a threshold position of the TRP relative to the WTRU, and / or availability of a trained ML model for the WTRU for coarse position, etc.). In some implementations, the WTRU applies and / or selects a corresponding positioning technology to determine its position based on satisfying certain conditions and position estimation criteria (e.g., obtained during position estimation).

[0323] In some implementations, the WTRU may be configured to apply location estimation, for example, to determine location estimate information based on predefined input criteria. Example inputs may include one or more of the following examples: WTRU RSRP measurement results, WTRU SINR measurement results, TOA measurement results, TDOA measurement results, sensor data (e.g., from a radar sensor, from a LIDAR sensor), imaging data / video data, positioning information from GNSS, etc. The output of the location estimation technique may include an estimated location, and may include, for example, absolute Cartesian coordinates, relative Cartesian coordinates, polar coordinates, absolute Cartesian coordinates with error bounds, zone IDs (e.g., where the network divides a deployment area into zones, such that each sub-deployment area, also referred to as a zone, has a unique (or locally unique) identifier, which may be referred to as a zone ID, and where the WTRU has a pre-configured mapping between zone IDs and geographic areas), etc.

[0324] In some implementations, the WTRU may be configured to apply ML-assisted position estimation. For example, in some implementations, the WTRU may be configured to deploy an ML model. In some implementations, the ML model may be configured to convert a predefined input (e.g., RSRP fingerprint) into positioning estimate information (e.g., zone ID).

[0325] Fig.14 1400 is a block diagram illustrating the input and output of an example ML-assisted position estimation implementation using zone IDs and a model 1402. The model 1402 may include a trained neural network, such as the one described above in conjunction with Figure 6 The described networks, or any other type or form of machine learning model, classifier, or predictor.

[0326] Fig.15 1500 is a block diagram illustrating an example ML-assisted location estimation scenario. In this example, the deployment area is divided into five zones 1502A-1502E, where each zone has a unique (e.g., locally unique) zone ID (e.g., zone ID-1, zone ID-2, ..., zone ID-5). In some implementations, for each location in the deployment area, there is at least one associated zone. In some implementations, the WTRUs 1504A (UE1 in the figure), 1504B (UE2 in the figure) receive a trained ML model (e.g., Fig.14 The trained ML model may convert the RSRP fingerprint measurement result into the region ID of the region where the WTRU is located. In this example, the RSRP fingerprint measurement result is the input to the model and the region ID is the output of the model. In some implementations, for Fig.15 The ML-assisted position estimate of the WTRU in the example of may be as follows:

[0327] - RSRP fingerprint for UE1 1504A: {TRP9: -90dbm, TRP10: -91dbm},

[0328] The zone Id estimation for UE1 is: 2 (1502B).

[0329] - RSRP fingerprint for UE2 1504B: {TRP12: -90dbm, TRP6: -95dbm},

[0330] The zone ID estimation for UE2: 4 (1502D).

[0331] Some implementations include WTRU-based training with coarse location information for ML-assisted location estimation. For example, the WTRU may indicate to the network that it has the capability to train an ML model for fingerprinting-based location estimation.

[0332] Fig.16 1600 is a block diagram illustrating an example training process for ML-assisted position estimation. In this example, the WTRU receives a model and an RSRP fingerprint configuration. The WTRU also receives a mapping of a first coarse position to a region ID conversion.

[0333] In some implementations, the WTRU may be able to determine its first coarse position (e.g., using GNSS-based position determination or sensor-based position determination), or the WTRU may obtain a RAT-dependent position during a training phase, e.g., as described below. In some implementations, the WTRU may obtain estimated position information, for example, based on a positioning method configured by the network (e.g., WTRU-based DL-TDOA, WTRU-based DL-AoD, etc.). The estimated position information may be expressed in terms of coordinates (e.g., in terms of global coordinate system coordinates or local coordinate system). In the examples described herein, the term "coarse position" may be used interchangeably with "position information".

[0334] In some implementations, for example, for WTRU-based training, the WTRU receives a mapping between zones and locations from the network. By applying the configured mapping between the coarse location and the zone ID, the WTRU may obtain its zone ID and provide it as a target for the model during training. For example, the WTRU may derive a coarse location based on measurements (e.g., RSRP) using a positioning method (e.g., DL-TDOA, DL-AoD). Based on the configured mapping, the WTRU may determine the zone ID corresponding to the coarse location. The zone ID may be marked as a target for training the ML model. Thus, for each set of measurements (e.g., RSRP, RSTD, RTT, AoA, AoD), there may be a corresponding zone ID as a target for training. For example, an ML model may be trained on N sets of measurements with corresponding N zone IDs. The ML model may be trained such that given a set of measurements, the ML model generates a zone ID corresponding to the set of measurements. For example, if the ML model is given a set of RSRP measurements corresponding to RSRP for 3 beams transmitted from a TRP of [-50dBm, -40dBm, -30dBm], and the corresponding zone ID is "3" (target), then if RSRP measurements close to [-50dBm, -40dBm, -30dBm] are input to the trained ML model, then the WTRU may expect the trained ML model to generate an ID number of "3". Alternatively, if the WTRU inputs [-50dBm, -50dBm, -50dBm] and a target zone ID of "2" to the ML model during training, then when RSRP measurements close to [-50dBm, -50dBm, -50dBm] are presented to the trained ML model, then the WTRU may expect the trained ML model to generate an ID number of "2". Training an ML model may require adjusting the internal parameters of the model (e.g., adjusting the weights used in the model). The measurement results may include information related to the multiple paths (e.g., RSRP for each path, RSTD for each path, time difference for each path, AoA for each path, AoD for each path, etc.). In some implementations, the WTRU may also or alternatively use the coarse position itself as a target to train the ML model for ML-assisted position estimation. In some implementations, if the WTRU does not receive a mapping from the network, the WTRU determines to use the coarse position (e.g., obtained from RAT-dependent positioning, GNSS, sensors) as a target for training.

[0335] In some implementations, the WTRU terminates training (eg, after a desired model accuracy is achieved) and may return the model to the network, eg, for use by other WTRUs.

[0336] Some implementations include network-based training with coarse location information for ML-assisted location estimation. In some implementations, the WTRU receives a trained ML model from the network. In some implementations, the WTRU receives information related to a configuration for the trained ML model, which may include information indicating whether the ML model is trained with a region ID or a coarse location as a target.

[0337] Some implementations include trigger conditions for the WTRU to initiate adaptive hybrid positioning. In some implementations, the WTRU may send a request to the network to enable adaptive hybrid positioning (e.g., to the LMF). In some implementations, the request is sent based on one or more of the following example trigger conditions: the WTRU is capable of performing both non-ML positioning and ML-assisted positioning; the non-ML positioning does not meet the accuracy requirement; and / or the WTRU determines that the measurement results for the non-ML positioning include errors.

[0338] In some embodiments, the WTRU may monitor fluctuations in PRS-RSRP and / or RSTD measurement results from each TRP, and may identify fluctuations in those measurement results (e.g., fluctuations in measurement results from the same TRP) that are above or below a preconfigured threshold (e.g., threshold standard deviation, variance, and / or maximum or minimum value of the measurement results) as measurement results with errors.

[0339] In some implementations, the WTRU may determine whether it is at a corner of a deployment area (e.g., based on a zone ID associated with a location in the corner) and / or may take action based on whether it is at a corner of a deployment area (e.g., based on a zone ID associated with a location in the corner).

[0340] In some implementations, the WTRU may determine whether the computational complexity of machine learning-assisted positioning is above a predefined threshold and / or may take action based on whether the computational complexity of machine learning-assisted positioning is above a predefined threshold.

[0341] In some implementations, the WTRU may be configured with a maximum allowed ML model complexity (e.g., a maximum number of FLOPs (floating operations)) required to perform positioning, and the WTRU may determine whether a trained ML model received by the WTRU has a computational complexity above the threshold. In some implementations, the WTRU may initiate adaptive hybrid positioning based on a trained ML model received by the WTRU having a computational complexity above the threshold.

[0342] In some implementations, the WTRU may determine whether it has been performing machine learning assisted positioning for a duration longer than a predefined threshold and / or may take action based on whether it has been performing machine learning assisted positioning for a duration longer than a predefined threshold.

[0343] In some implementations, the WTRU may be configured to perform ML-assisted positioning (with a fixed or maximum computational complexity) for a predefined maximum duration. In some implementations, the WTRU may initiate adaptive hybrid positioning if the inference (e.g., estimated position) generated by the ML model takes longer than a predefined threshold.

[0344] Some implementations include conditions associated with ML-assisted positioning. In some implementations, the WTRU may apply one or more criteria to select a specific ML-assisted positioning technique. For example, in some implementations, the WTRU may use a specific ML-assisted positioning technique based on whether the LOS probability of each TRP is at or above (or below) a threshold (e.g., as configured by the LMF). In some implementations, the WTRU may use a specific ML-assisted positioning technique based on whether the RSRP value of each TRP is at or above (or below) a threshold (e.g., as configured by the LMF).

[0345] In some implementations, the WTRU may use a specific ML-assisted positioning technique based on whether all affected TRPs are in one direction with reference to the WTRU's estimated location. In some implementations, the WTRU is configured or pre-configured with valid zone IDs and assistance information for each zone ID (e.g., indicating whether the zone is located at a corner of a deployment area, whether a trained ML model is available for the zone, etc.). In some implementations, if the WTRU maps its coarse or estimated location to a zone ID that is located at a corner, it may determine that all TRPs are in one direction from its location.

[0346] based on Fig.15 , one such example scenario may include where both WTRUs receive a corner zone ID -{4,5}. After obtaining a position estimate (i.e., position estimate) from stage 1, the first WTRU (UE1) may estimate that it is in zone 2, and that zone 2 is not at the corner of the deployment area. However, the second WTRU (UE2) may recognize that its position estimate corresponds to zone ID 4, and therefore recognize that it is located in zone 4. Since zone 4 is a corner zone ID, the second WTRU may recognize that all TRPs are located in one direction (e.g., within a 180 degree arc) relative to its position.

[0347] In some implementations, the WTRU may receive information indicating corner coordinates and corner detection distances of a deployment area. If the difference between the WTRU's position estimate and one of the corner coordinates is less than the corner detection distance, the WTRU may identify that it is located at a corner of the deployment area.

[0348] In some implementations, the trained ML model may be used for the rough location of the WTRU to perform ML-assisted positioning. Fig.15 An example deployment of is used to illustrate that in some implementations, trained ML models are available for zone ID-1, zone ID-2, zone ID-3, and zone ID-4. In this example, the trained ML model for zone ID-5 is not available (e.g., due to insufficient training). In some implementations, the WTRU is pre-configured with a model availability flag or other indication. In this example, the model availability flag for zone ID-1, zone ID-2, zone ID-3, and zone ID-4 may be set to true (e.g., set to binary 1), and the availability flag for zone ID-5 may be set to false (e.g., set to binary 0).

[0349] The following example embodiments illustrate the technology described herein. The order of the various steps is exemplary and can be rearranged in any suitable manner, or can include any suitable additional steps, and / or some steps can be omitted in any suitable manner.

[0350] An example implementation includes WTRU-based zone ID-aware adaptive hybrid positioning. In some implementations, the WTRU is configured by the network to perform adaptive hybrid positioning. In some implementations, the WTRU receives a trained ML model for ML-assisted position estimation, a trained ML model for ML-assisted positioning, and a TRP location and associated LOS / NLOS indicator from the network. In some implementations, the WTRU receives a threshold for the number of LOSTRPs required for multi-TRP ML-assisted positioning. For example, in some implementations, the threshold for the number of LOS TRPs required for multi-TRP ML-assisted positioning = 2. In some implementations, the WTRU receives an indication of all valid zone IDs and assistance information for each zone (e.g., an indication of the ID of the zone for the corner can be represented as corner zone ID-{4,5}). In some implementations, the WTRU is configured with a RAT-dependent non-ML positioning method (e.g., DL-TDOA). In some implementations, the WTRU measures the PRS. In some implementations, the WTRU provides one or more of the measurements (eg, RSRP) to the ML-assisted position estimation model and obtains the region ID from the model based on the one or more measurements.

[0351] In some implementations, the WTRU makes the following determinations related to positioning: If the WTRU determines that it is in a corner, for example, based on its estimated zone ID and assistance information, the WTRU determines to use ML-assisted positioning (e.g., if the WTRU determines that it is in zone ID-4). If the WTRU determines that it is in a corner, for example, based on its zone ID and assistance information, and the number of TRPs with LOS indicators is equal to or above a threshold, the WTRU determines to use multi-TRP fingerprinting and returns an estimated position. (e.g., if the number of LOS TRPs visible to the WTRU is greater than or equal to a threshold number of LOS TRPs for multi-TRP ML-assisted positioning). If the WTRU determines that it is in a corner, for example, based on its associated zone ID and assistance information, and the number of TRPs with LOS indicators is below a threshold, the WTRU determines to use a single TRP fingerprinting and returns its estimated position (e.g., the number of LOS TRPs visible to the WTRU is less than a threshold number of LOS TRPs for multi-TRP ML-assisted positioning). If the WTRU determines that it is not located in a corner, for example based on the associated zone ID and assistance information, then the WTRU determines to use the configured RAT-dependent non-ML positioning method. In some implementations, the WTRU returns its estimated position to the LMF.

[0352] An example implementation includes WTRU-based ML-assisted position estimation-aware adaptive hybrid positioning. In some implementations, the WTRU is configured by the network to perform adaptive hybrid positioning. In some implementations, the WTRU receives a trained ML model for ML-assisted position estimation, ML-assisted positioning, and TRP locations and associated LOS / NLOS indicators from the network. In some implementations, the WTRU receives a threshold for the number of LOS TRPs required for multi-TRP ML-assisted positioning. For example, in some implementations, the threshold for the number of LOS TRPs required for multi-TRP ML-assisted positioning = 2. In some implementations, the WTRU receives coordinates of corners of a deployment area and an indication of a threshold distance from the corner for corner detection. In some implementations, the WTRU is configured with a RAT-dependent non-ML positioning method (e.g., DL-TDOA). In some implementations, the WTRU measures the PRS.

[0353] In some implementations, the WTRU provides one or more of the measurements (e.g., RSRP) to the ML-assisted position estimation model and obtains its position estimate from the model based on the one or more measurements. In some implementations, the ML model deployed to obtain a provisional or intermediate position estimate provides a significantly lower target positioning accuracy (e.g., approximately 10 m) than the ML model deployed to estimate the final position of the WTRU (e.g., approximately 0.1 m).

[0354] In some implementations, the WTRU makes the following determinations related to positioning: If the WTRU determines that it is at a corner, for example, based on the difference between the WTRU position estimate and one of the corner coordinates being below a threshold corner detection distance, the WTRU determines to use an ML-assisted positioning method. If the WTRU determines that it is at a corner and the number of TRPs with an LOS indicator is equal to or above a threshold, the WTRU determines to use a multi-TRP fingerprinting technique and returns an estimated position. If the WTRU determines that it is at a corner and the number of TRPs with an LOS indicator is below a threshold, the WTRU determines to use a single TRP fingerprinting technique and returns an estimated position. If the WTRU determines that it is not at a corner, for example, based on the minimum value of the difference between the WTRU position estimate and all corner coordinates being above a threshold, the WTRU determines to use a configured RAT-dependent non-ML positioning method.

[0355] In some implementations, the WTRU returns its estimated position to the LMF. For example, the WTRU returns its estimated position and reports the positioning method used to determine the estimated position to the LMF.

[0356] Some implementations include WTRU-based non-ML position estimate-aware adaptive hybrid positioning. In some implementations, the WTRU is configured by the network to perform adaptive hybrid positioning. In some implementations, the WTRU receives a trained ML model, TRP location, and associated LOS / NLOS indicator for ML-assisted positioning from the network. In some implementations, the WTRU receives a threshold value for the number of LOS TRPs required for multi-TRP ML-assisted positioning. For example, in some implementations, the threshold value for the number of LOS TRPs required for multi-TRP ML-assisted positioning = 2. In some implementations, the WTRU receives the coordinates of the corners of the deployment area and the threshold distance from the corners for corner detection. In some implementations, the WTRU is configured with a RAT-dependent non-ML positioning method (e.g., DL-TDOA). In some implementations, the WTRU measures the PRS. In some implementations, the WTRU provides one or more measurement results (e.g., RSRP) to a RAT-dependent non-ML positioning (e.g., DL-AoD), and obtains its position and uses the position as a position estimate.

[0357] In some implementations, the WTRU makes the following determinations related to positioning: If the WTRU determines that it is at a corner based on the difference between the WTRU position estimate and one of the corner coordinates being below a threshold corner detection distance, the WTRU determines to use an ML-assisted positioning method. If the WTRU determines that it is at a corner and the number of TRPs with LOS indicators is equal to or above a threshold, the WTRU determines to use multi-TRP fingerprinting and returns an estimated position. If the WTRU determines that it is at a corner and the number of TRPs with LOS indicators is below a threshold, the WTRU determines to use single TRP fingerprinting and returns an estimated position. If the WTRU determines that it is not at a corner based on the minimum value of the difference between the WTRU position estimate and all corner coordinates being above a threshold distance, the WTRU returns a previously obtained position.

[0358] In some implementations, the WTRU returns its estimated position to the LMF. For example, the WTRU returns its estimated position and reports the positioning method used to determine the estimated position to the LMF.

[0359] Some implementations include WTRU-based adaptive hybrid positioning. In some implementations, the WTRU receives a TRP location and an associated LOS / NLOS indicator from the network. The WTRU receives a LOS threshold and a LOS TRP threshold. In some implementations, the WTRU is configured with a RAT-dependent positioning method (e.g., DL-TDOA). The WTRU receives a configuration for a multi-RTT positioning method. In some implementations, the WTRU measures the PRS. In some implementations, the WTRU obtains a position estimate from an ML model-based positioning (e.g., a measurement based on multi-TRP fingerprinting) based on the measurement results.

[0360] In some implementations, the WTRU makes the following determinations related to subsequent positioning: If the WTRU determines that it is located at a corner (e.g., based on the difference between the WTRU position and the corner coordinates being below a threshold) and the number of TRPs with LOS indicators being above the LOS threshold and greater than or equal to the LOS TRP threshold, the WTRU determines to use multi-RTT. If the WTRU determines that it is located at a corner but the number of TRPs with LOS indicators being above the LOS threshold is less than the LOS TRP threshold, the WTRU returns the position estimate from the ML. If the WTRU is not located at a corner, the WTRU determines to use the configured RAT-dependent positioning method. In some implementations, the WTRU returns its estimated position to the LMF.

[0361] Some implementations involve positioning method selection. For example, in some cases, the WTRU may be located at the corner of a deployment area. In some cases, if a WTRU located at the corner of a deployment area were to apply a conventional positioning method (e.g., DL-TDOA), the positioning accuracy may be reduced (e.g., severely), for example, because all TRPs are located in the same direction relative to the target WTRU. In some implementations, if the WTRU identifies that it is located at the corner of a deployment area, the WTRU may use a single TRP AIML positioning method to obtain a more accurate position (e.g., compared to conventional positioning). In some implementations, using a single TRP AIML positioning method, the WTRU may determine its position by providing the AIML model with one or more measurement results from the same TRP (e.g., an RSRP fingerprint created by RSRP measurements of multiple beams from a unified TRP).

[0362] In some implementations, the WTRU may, for example, receive from the network one or more TRP IDs associated with a corner and conditions for corner association (e.g., the WTRU is within a preconfigured distance threshold from the TRP associated with the corner, or the WTRU is within a preconfigured distance threshold from one of the corners, etc.). For example, in some implementations, the WTRU may determine the coordinates of the corners of the room based on auxiliary data sent from the network. In some implementations, the WTRU may be configured with a TRP ID associated with each corner (e.g., the configured coordinates). In some implementations, if the TRP is within a preconfigured threshold distance from the coordinates of the corner, the WTRU may determine that the TRP is associated with the corner. In some implementations, after the conditions for corner association are met (e.g., if the WTRU is within a preconfigured threshold from the TRP associated with the corner, the WTRU is within a preconfigured threshold from one of the corners), the WTRU may determine that it is located at a corner and may use single TRP AIML positioning to obtain its position.

[0363] Fig.171700 is a system diagram of an example WTRU 1702 in an example scenario, wherein the WTRU 1702 is located at a corner of an environment (e.g., bounded by a wall, building or other structure or signal blocking element, or otherwise separated from other transmitters or other devices on two or more sides). The WTRU 1702 is labeled "UE" in the figure, however any suitable WTRU may be used. In this example, the network indicates to the WTRU 1702 that TRP7 1706 and TRP1 1708 are associated with the upper left corner (TRP 1704 is not associated with the corner). The network (e.g., LMF, gNB) also indicates to the WTRU 1702 the conditions for the corner association. Note that in some implementations, the condition may be received by the WTRU 1702 from another device (e.g., a sensor) and / or in any other suitable manner, and / or may be pre-configured. In this example, the condition for corner association indicates that if the RSRP from any of the TRPs (1706, 1708) associated with the corner is above a threshold, the WTRU should be associated with the corner. As another example, the condition for corner association may indicate that if the RSRP from any of the beam IDs associated with the corner is above a threshold, the WTRU should be associated with the corner. If the WTRU 1702 satisfies the condition (e.g., if the RSRP from either or both of TRP 7 and TRP1 is above a threshold), it associates itself with the corner (e.g., the WTRU determines that it is at the corner). In some embodiments, if the WTRU 1702 associates itself with the corner, the WTRU may determine a positioning method that is configured to position the WTRU at the corner (e.g., AIML positioning based on a single TRP). In this example, the WTRU 1702 receives an RSRP threshold for corner association from the network. In this example, the WTRU receives PRS from TRP1, TRP2, TRP7, and TRP8, and determines that the measured RSRP of TRP7 is above the RSRP threshold for corner association (in this example, the measured RSRP of TRP1, TRP2, and TRP8 may be above or below the threshold). Since TRP7 is associated with the upper left corner, the WTRU 1702 determines that it is located at the upper left corner. Based on the association with the corner, the WTRU 1702 uses a single TRP AIML positioning method to obtain its position. In one example, based on the configuration, the WTRU 1702 may determine to use a single TRP AIML positioning method that is applicable to any corner of the room. For another example, the WTRU 1702 may be configured with a single TRP AIML positioning method associated with each corner of the room. In this case, the WTRU 1702 may determine which corner the WTRU is associated with, and determine which single TRP AIML positioning method the WTRU should use.

[0364] In some cases, the WTRU may be located in a NLOS heavy environment; for example, where the WTRU is in a NLOS condition with a majority of the TRPs from which it receives PRSs. In such cases, if the WTRU were to apply conventional positioning methods, the WTRU may observe a reduction in accuracy, for example due to inaccurate positioning measurements from the NLOS TRPs. In some implementations, if the WTRU detects a NLOS heavy environment, the WTRU may apply a multi-TRP AIML positioning method and obtain its position, for example to determine its position with a higher positioning accuracy than conventional positioning. In some implementations, the network provides a LOS indicator associated with each TRP (e.g., for a NLOS TRP, a LOS indicator value = 0; for a LOS TRP, a LOS indicator value = 1) so that the WTRU can determine whether it is in a NLOS heavy environment. The WTRU receives a LOSTRP threshold for NLOS environment detection from the network. If the WTRU determines that the number of observed LOS TRPs is less than the LOS TRP threshold, the WTRU may determine that it is in a NLOS heavy environment and may determine to use a multi-TRP positioning method to obtain its position.

[0365] Fig.18 1800 is a system diagram of a WTRU 1802 in an example scenario in which the example WTRU 1802 is located in an example NLOS environment. The WTRU 1802 is labeled "UE" in the figure, however any suitable WTRU may be used. In this example, the WTRU 1802 receives a LOS TRP threshold for NLOS environment detection, where Threshold = 3 (e.g., 3 TRPs in LOS) (other thresholds may be used in other embodiments, such as 2, 4, 5, 10, or any other such value). The WTRU also receives PRS and LOS indicators from TRP2, TRP3, TRP9, and TRP8. Since the WTRU 1802 is in a LOS state with only 1 TRP (TRP9 1806), which is below the threshold for NLOS environment detection in this example scenario (where the NLOS TRP 1808 is obscured or blocked from the LOS), the WTRU 1802 determines that it is in a NLOS heavy environment and applies a multi-TRP AIML positioning method to estimate its position. In some embodiments, a comparison of the LOS indicator (or NLOS indicator) to the LOS TRP threshold may be referred to as a desired rate for NLOS prediction, such that if the desired rate for NLOS prediction is exceeded, a multi-TRP AI / ML positioning approach may be utilized.

[0366] In some cases, the WTRU may be located in a portion of its deployment area that is neither in a corner nor in a NLOS environment. In some such cases, the WTRU may use (or continue to use) conventional positioning methods (e.g., DL-TDOA).

[0367] Fig.19 1900 is a system diagram of an example WTRU 1904 in an example scenario in which the WTRU 1902 is in a LOS environment (e.g., where the WTRU is located in a portion of its deployment area that is neither in a corner nor in a NLOS environment because it has LOS connections to TRPs 2, 3, 8, and 9 (identified as LOS TRPs 1906), in which case TRP 1902 is out of range). WTRU 1902 is labeled as "UE" in the figure, however any suitable WTRU may be used. In this scenario, the WTRU 1902 may use (or continue to use) conventional positioning methods, for example, because AIML positioning methods (e.g., single TRP positioning or multi-TRP positioning) typically require more overhead for transfer training and models, and may not provide any additional accuracy gain compared to conventional methods in a LOS environment.

[0368] In such Figures 17 to 19 In some embodiments of the example, the physical environment or location of the WTRU may be referred to as a spatial configuration relative to the TRP. Fig.17 In , the spatial configuration of the WTRU may be defined as being at a corner or associated with a corner (or close to or associated with a TRP that is itself associated with a corner). Fig.18 In the NLOS environment, the spatial configuration of the WTRU can be defined. Fig.19 , the spatial configuration of a WTRU may be defined as being in a LOS environment. Thus, the spatial configuration of a WTRU or other device may broadly refer to the relative positioning of a WTRU or other device relative to one or more TRPs, corners, and / or other objects that block LOS transmissions.

[0369] Similarly, in some embodiments, various positioning models or algorithms may be referred to as position determination algorithms. For example, the position determination algorithm may include a single TRP AI / ML position prediction model, a multi-TRP AI / ML position prediction model, or a non-AI / ML positioning method.

[0370] Fig. 202000 is a flow chart illustrating an embodiment of a method 2000 for ML-assisted position estimation. In brief, in some embodiments, at 2005, a device may receive a model, a TRP position, and / or a LOS indicator. At 2010, a device may receive a corner identifier and / or a threshold. At 2015, a device may receive one or more PRS configuration identifiers. At 2020, a device may measure a PRS signal. At 2025, a device may determine whether it is at a corner. If so, at 2030, the device may select a single TRP fingerprint recognition model. If not, at 2035, the device may determine whether it is in a NLOS environment. If so, at 2040, the device may select a multi-TRP fingerprint recognition model. If not, at 2045, the device may select a RAT-dependent positioning algorithm. At 2050, the device may determine its position using the selected model or algorithm. At 2055, the device may report its position to the network. It should be noted that in some implementations, some or all of the actions and / or conditions may be omitted or rearranged. For example, in some implementations, condition 2035 and action 2040 are omitted, or condition 2025 and action 2030 are omitted.

[0371] Still reference Fig. 20 And in more detail, at 2005, the device may receive a model, a TRP position, and / or a LOS indicator. The device may be or include a WTRU, a UE, or other wireless device. In some embodiments, the device may receive one or more models for AI / ML assisted positioning, such as a single TRP fingerprint recognition model or a multi-TRP fingerprint recognition model, as discussed above. In some embodiments, the model may include a single unified model. In some embodiments, the model may be separate. In the presence of multiple models, the models may be homogeneous (e.g., similar neural networks with different weights or biases) or heterogeneous (e.g., DNNs and Bayesian classifiers, or any other combination of different AI models or algorithms). For example, in some specific implementations, the device may include hardware for executing a neural network, such as a TPU coprocessor, an ASIC, or other circuit. The weights of the selected model may be loaded into the circuit as firmware, allowing the model to be executed on hardware.

[0372] In some embodiments, a device may receive TRP location information. The TRP location information may be in any suitable format, such as rectangular or polar coordinates from a given origin; in relative or absolute positions; or any other type and form of location information. In many embodiments, the TRP location information may be associated with each TRP based on a TRP identifier.

[0373] In some embodiments, the device may receive a LOS indicator associated with each TRP identifier. As discussed above, the LOS indicator may be a binary flag (e.g., 1 or 0), a data string, or other predetermined signal. In some embodiments, 2005 may include one or more additional sub-processes; for example, in some embodiments, the device or WTRU may send a request for TRP location information and / or a LOS indicator, and the TRP or another network device (e.g., LMF, gNB, etc.) may respond with the requested information. In some embodiments, the model may be preloaded into the memory of the device. Thus, in some such embodiments, 2005 may be considered to cover a considerable period of time (e.g., loading the model during an overnight system update, and sending a request for TRP location and / or LOS indicator when accessing the network at some later time). In some embodiments, preloading of the model may also be considered to be a separate predecessor process (not shown).

[0374] In some embodiments, at 2010, the device may receive a corner identifier and / or threshold. In some embodiments, the corner identifier may include a list of TRPs whose coverage includes corners (e.g., corners of a wireless environment or boundaries beyond which other TRPs are unavailable; corners where a device may be blocked from receiving signals on two or more sides by walls or environmental objects; etc.). Identifying a TRP as being associated with a corner may be determined by a system administrator, installer, engineer, or other entity, or may be determined dynamically (e.g., by the TRP executing a discovery protocol to identify other accessible TRPs, by measuring broadcast signals in various directions to detect walls based on reflections or returns, or in any other such method). In some cases, no TRP may be associated with a corner, and therefore, in some such embodiments, no corner identifier may be provided to the device (e.g., no response to a query for a corner identifier, or a blank response, etc.); in other embodiments, an explicit negative response may be provided to the device (identifying that no TRP is associated with the corner). In some embodiments, each TRP identifier may be associated with a flag or string or predetermined position to indicate whether it is associated with a corner; thus, in some such embodiments, the TRP identifier and the corner identifier may be provided together (e.g., at 2005).

[0375] In some embodiments, the device may receive one or more threshold values. For example, in some embodiments, the device may receive a RSRP threshold value for measuring whether the device is close to a corner TRP or within the coverage of a corner TRP. In some embodiments, the device may receive a threshold number of TRPs for indicating a LOS or NLOS environment. The threshold value may be provided together or via a separate message. In some embodiments, 2010 may include one or more additional subprocesses. For example, in some embodiments, the device may send one or more requests for a threshold value and / or a corner identifier, and may receive one or more corresponding responses. Although shown as being separated from 2005, in some embodiments, 2005 and 2010 may be combined (for example, the device may send a request for TRP location information, a LOS indicator, and a corner identifier; and may receive corresponding information). Similarly, as discussed above, in some embodiments, the threshold value may be loaded into the memory of the device at an earlier time (for example, during a periodic system update). Therefore, part of 2010 may occur simultaneously with 2010, before it, or after it.

[0376] At 2015, the device may receive a PRS configuration. As discussed above, the PRS configuration may identify one or more TRPs from which the device may receive PRS. The PRS configuration may be sent by a TRP or any suitable network device (e.g., LMF, gNB, etc.), and in some embodiments may be provided to the device as a response to a sent request for a PRS configuration.

[0377] Although shown separately, in some embodiments, 2015 may appear as part of 2005 or 2010. For example, in one such embodiment, the device may send a request for a TRP location, LOS indicator, corner identifier, and PRS configuration, and may receive one or more corresponding responses including the requested information.

[0378] At 2020, the device may measure the PRS signal. The device may measure received signal strength, signal loss relative to an indicated transmit strength, flight time based on a synchronized clock and transmit timestamp, and / or any other such measurement. Although shown as a single process, in many embodiments, the device may measure PRS signals from multiple TRPs, and thus in various embodiments may perform 2020 serially or in parallel for each TRP.

[0379] At 2025, the device may determine whether it is at a corner or associated with a corner. In some embodiments, the device may determine whether the RSRP from a TRP identified as associated with a corner is above a threshold. This may indicate that the device is adjacent to the TRP associated with the corner, and is therefore similarly associated with the corner. Therefore, in many embodiments, the device may not directly determine whether it is at a corner, but may determine that it is near the TRP associated with the corner and is therefore also associated with the corner. In some embodiments, other measurements may be utilized instead of RSRP or in addition to RSRP, such as the flight time of a signal from a TRP. In some cases, as discussed above, measurements from multiple TRPs associated with a corner may exceed a threshold; such measurements may indicate that multiple corners are close to each other in the environment, or that multiple TRPs are associated with the same corner (which may be advantageous in cases where there are other obstacles in the environment near the corner that may affect RSRP or other measurements).

[0380] If the device determines that it is associated with a corner based on the measurements of the TRP associated with the corner, then in some embodiments, the device may select a single TRP fingerprinting model at 2030. As discussed above, the single TRP fingerprinting model may include any type and form of AI / ML model, such as weights and biases for a neural network or other classifier, that is trained to determine a location within an environment associated with a corner based on measurements associated with a single TRP.

[0381] If the device determines that it is not associated with a corner based on the measurement of the TRP associated with the corner, then in some embodiments, at 2035, the device can determine whether it is in a NLOS environment. As discussed above, in some embodiments, the device can determine whether the number of LOS indicators received from the network and / or TRP is below a LOS TRP threshold. If so, the device can determine that it is in a NLOS environment. In some other embodiments, the device can determine whether the number of NLOS indicators received from the network and / or TRP is above a NLOS TRP threshold. If so, the device can determine that it is in a NLOS environment.

[0382] If the device determines that it is in a NLOS environment, then in some embodiments, the device may select a multi-TRP fingerprinting model at 2040. As discussed above, the multi-TRP fingerprinting model may include any type and form of AI / ML model, such as weights and biases for a neural network or other classifier, that is trained to determine a location within an environment based on measurements associated with a TRP having LOS to the device.

[0383] If the device determines that it is not in a NLOS environment, then in some embodiments, the device may select a RAT-dependent positioning algorithm or other positioning method, such as DL-TDOA, DL-AoD, multi-RTT, UL-TDOA, UL-AoA, or any other suitable method at 2045. As discussed above, in many embodiments, since the device may have LOS to a sufficient number of TRPs, a RAT-dependent positioning algorithm may have higher accuracy than a fingerprinting model, may be faster, may require less power or computing resources, or may otherwise be desirable in such instances.

[0384] At 2050, in some embodiments, the device may use the selected model or algorithm to determine its location (e.g., a single TRP fingerprinting ML model, a multi-TRP fingerprinting ML model, other positioning algorithms, etc.). As discussed above, using the model may include loading the model as firmware into a TPU or other coprocessor or hardware accelerator, loading the model into RAM or other memory, applying weights and biases to the nodes and edges of the model, and providing input data to the model and reading output data. In some embodiments, providing input data to the model may include normalizing input data such as RSRP measurements or other input data, transcoding the data into vectors or arrays, or otherwise preprocessing the data. In some embodiments, the input data may be filtered, truncated, scaled, or otherwise preprocessed. Similarly, the output data may be processed, transcoded, filtered, truncated, or otherwise configured as location data for sending to the network.

[0385] At 2055, in some embodiments, the device may report its location to the network based on the output of the model or algorithm. The location report may be sent via any suitable type and form of transmission and may be sent independently as part of another transmission such as a data uplink and / or in response to a request from another device (e.g., LMF, gNB, TRP, etc.).

[0386] Accordingly, the present disclosure relates to methods, devices, and systems for machine learning (ML) assisted positioning. A ML assisted non-line-of-sight (NLOS) recognition request and capability information are sent to a network device. In response to the recognition request and capability information, a trained ML model and a first list of input features are received. A model for NLOS recognition is selected based on a desired rate of NLOS prediction. A NLOS transmit receive point (TRP) is predicted based on the selected model. A location is determined based on the predicted NLOS TRP.

[0387] In some implementations, the WTRU may enhance its positioning accuracy and select the most efficient positioning technology using a method that includes some or all of the following procedures, in the following order or in a different order:

[0388] In some implementations, the WTRU may receive an AI / ML model, a TRP location, and an associated LOS indicator (e.g., 1 or 0) for AI / ML assisted positioning from the network (e.g., from an LMF or gNB). In some implementations, the WTRU may, for example, receive a list of TRP IDs associated with corners (e.g., TRPs whose coverage includes corners), a first threshold (e.g., RSRP), and a second threshold (e.g., the number of LOS TRPs) from the network. In some implementations, the WTRU may receive a PRS configuration from the network (e.g., LMF, gNB). In some implementations, the WTRU may perform PRS measurements. In some implementations, if the RSRP measured from at least one of the corner TRPs is higher than the first threshold, the WTRU may determine that it is located at a corner associated with the corner TRP. In some implementations, if the number of LOS TRPs is less than the second threshold, the WTRU may determine that it is in a NLOS environment.

[0389] In some implementations, the WTRU may select a positioning method, wherein if the WTRU is located at a corner, the WTRU selects a single TRP fingerprinting AI / ML positioning method. If the WTRU is not located at a corner but is in a NLOS environment, the WTRU selects a multi-TRP fingerprinting AI / ML positioning method; otherwise, if the WTRU is not located at a corner and is not in a NLOS environment, the WTRU selects a configured RAT-dependent non-AI / ML positioning method (e.g., DL-TDOA). Note that in some implementations, when the WTRU determines that it is located in a corner, the WTRU determines a single TRP fingerprinting AI / ML positioning method as the positioning method. Since the fingerprint of the measurement results can capture both the LOS and / or NLOS characteristics in the measurement results, the single TRP fingerprinting AI / ML positioning method can be adapted to a variety of different scenarios (e.g., corners with only LOS, corners with LOS and NLOS, corners with only NLOS).

[0390] In some implementations, the WTRU may use the selected positioning method to obtain its position. In some implementations, the WTRU may report its position to the network.

[0391] Although features and elements are described above in specific combinations, it will be understood by those of ordinary skill in the art that each feature or element may be used alone or in any combination with other features and elements. In addition, the methods described herein may be implemented in a computer program, software, or firmware incorporated in a computer-readable medium for execution by a computer or processor. Examples of computer-readable media include electronic signals (sent via a wired or wireless connection) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, read-only memory (ROM), random access memory (RAM), registers, cache memory, semiconductor memory devices, magnetic media (such as built-in hard disks and removable disks), magneto-optical media, and optical media (such as CD-ROM disks and digital versatile disks (DVDs)). A processor associated with software may be used to implement a radio frequency transceiver for a WTRU, UE, terminal, base station, RNC, or any host computer.

Claims

1. A method implemented in a wireless transmit / receive unit (WTRU), the method comprising: receiving information indicating an artificial intelligence / machine learning (AI / ML) model used to determine the location; receiving information indicating a transmit reference point (TRP) associated with a corner and information indicating a reference signal received power (RSRP) threshold; the TRP associated with the corner comprising a first TRP; determining that the WTRU is located at a corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being greater than the RSRP threshold; determining, by the WTRU, location information based on an AI / ML location model and the determination that the WTRU is located at the corner; as well as Sending information indicating the location of the WTRU.

2. The method of claim 1, wherein the AI / ML location model comprises a single TRP fingerprinting AI / ML location model.

3. The method of claim 2, wherein the single TRP fingerprinting AI / ML location model comprises weights for a neural network.

4. The method of claim 1, wherein the AI / ML model used to determine location comprises a support vector machine (SVM) or a k-nearest neighbor (KNN) model.

5. The method of claim 1, further comprising sending a request for the information indicative of an AI / ML model used to determine location.

6. The method of claim 1, wherein the TRP is associated with a corner based on its proximity to the corner of the deployment environment.

7. The method according to claim 1 also includes sending a request for PRS to one or more TRPs.

8. The method of claim 1, further comprising receiving a PRS from each of a plurality of TRPs including the first TRP.

9. The method of claim 1, wherein the WTRU has a line of sight (LOS) path to the first TRP.

10. The method of claim 1, wherein inputs to the AI / ML location model include peak power or average power measurements of one or more transmissions from the first TRP.

11. A wireless transmit / receive unit (WTRU), comprising: a receiver circuit configured to receive information indicative of an artificial intelligence / machine learning (AI / ML) model used to determine location; and The receiver circuit is further configured to receive information indicating a transmit reference point (TRP) associated with a corner and information indicating a reference signal received power (RSRP) threshold, the TRP associated with the corner comprising a first TRP; processing circuitry configured to determine that the WTRU is located at a corner based on an RSRP of a positioning reference signal (PRS) received from the first TRP being greater than the RSRP threshold; The processing circuit is further configured to determine location information based on an AI / ML location model and the determination that the WTRU is located at the corner; and Transmitter circuitry configured to transmit information indicative of the location of the WTRU.

12. The WTRU of claim 11, wherein the AI / ML location model comprises a single TRP fingerprinting AI / ML location model.

13. The WTRU of claim 12, wherein the single TRP fingerprinting AI / ML location model comprises weights for a neural network.

14. The WTRU of claim 11, wherein the AI / ML model used to determine location comprises a support vector machine (SVM) or a k-nearest neighbor (KNN) model.

15. The WTRU of claim 11, wherein the transmitter is further configured to send a request for the information indicative of an AI / ML model used to determine location.

16. The WTRU of claim 11, wherein the TRP is associated with a corner based on its proximity to a corner of a deployment environment.

17. The WTRU of claim 11, wherein the transmitter is further configured to send a request for a PRS to one or more TRPs.

18. The WTRU of claim 11, wherein the receiver is further configured to receive a PRS from each of a plurality of TRPs including the first TRP.

19. The WTRU of claim 11, wherein the WTRU has a line of sight (LOS) path to the first TRP.

20. The WTRU of claim 11, wherein inputs to the AI / ML location model include peak power or average power measurements of one or more transmissions from the first TRP.

21. A method implemented in a wireless transmit / receive unit (WTRU), the method comprising: receiving information indicating an artificial intelligence / machine learning (AI / ML) model used to determine the location; receiving information indicating a transmit reference point (TRP) associated with line of sight (LOS) communication, information indicating a reference signal received power (RSRP) threshold, and information indicating a threshold number of TRPs associated with the LOS communication; determining a total number of TRPs associated with LOS communications for which an RSRP measured on a received positioning reference signal (PRS) is below the RSRP threshold; determining that the WTRU is in a non-line-of-sight (NLOS) environment based on the determined total number of TRPs being less than the threshold number of TRPs; determining, by the WTRU, location information based on an AI / ML location model and the determination that the WTRU is in the NLOS environment; as well as Sending information indicating the location of the WTRU.

22. The method of claim 21, wherein the location model comprises a multi-TRP fingerprinting AI / ML location model.

23. The method of claim 22, wherein the multi-TRP fingerprinting AI / ML location model comprises weights for a neural network.

24. The method of claim 21, wherein the AI / ML model used to determine location comprises a support vector machine (SVM) or a k-nearest neighbor (KNN) model.

25. The method of claim 21, further comprising sending a request for the information indicative of an AI / ML model used to determine location.

26. The method of claim 21, wherein the indication of TRPs associated with LOS communications comprises a list of TRP identifiers and associated predetermined bits indicating LOS or non-line of sight (NLOS).

27. The method of claim 21 further comprising sending a request for a PRS to each of the TRPs associated with LOS communications.

28. The method of claim 21, further comprising receiving PRSs from a plurality of TRPs associated with LOS communications.

29. The method of claim 21, wherein a LOS path to at least one of the TRPs associated with LOS communications is obscured.

30. The method of claim 21, wherein inputs to the AI / ML location model include one or more transmitted peak power or average power measurements from each of a plurality of TRPs associated with LOS communications.

31. A wireless transmit / receive unit (WTRU): a receiver circuit configured to receive information indicative of an artificial intelligence / machine learning (AI / ML) model used to determine location, and The receiver circuit is further configured to receive information indicating a transmit reference point (TRP) associated with line of sight (LOS) communications, information indicating a reference signal received power (RSRP) threshold, and information indicating a threshold number of TRPs associated with LOS communications; processing circuitry configured to determine a total number of TRPs associated with LOS communications for which an RSRP measured on a received positioning reference signal (PRS) is below the RSRP threshold; The processing circuit is further configured to determine that the WTRU is in a non-line-of-sight (NLOS) environment based on the determined total number of TRPs being less than the threshold number of TRPs, and The processing circuit is further configured to determine location information based on the AI / ML location model and the determination that the WTRU is located in the NLOS environment; and Transmitter circuitry configured to transmit information indicative of the location of the WTRU.

32. The WTRU of claim 31 , wherein the location model comprises a multi-TRP fingerprinting AI / ML location model.

33. The WTRU of claim 32, wherein the multi-TRP fingerprinting AI / ML location model comprises weights for a neural network.

34. The WTRU of claim 31 , wherein the AI / ML model used to determine location comprises a support vector machine (SVM) or a k-nearest neighbor (KNN) model.

35. The WTRU of claim 31 , wherein the transmitter circuit is further configured to send a request for the information indicative of an AI / ML model that determines location.

36. The WTRU of claim 31, wherein the indication of a TRP associated with LOS communications comprises a list of TRP identifiers and associated predetermined bits indicating LOS or non-line of sight (NLOS).

37. The WTRU of claim 31, wherein the transmitter is further configured to send a request for a PRS to each of the TRPs associated with the LOS communication.

38. The WTRU of claim 31 , wherein the receiver is further configured to receive PRSs from a plurality of TRPs associated with LOS communications.

39. The WTRU of claim 31, wherein a LOS path to at least one of the TRPs associated with LOS communications is obscured.

40. The WTRU of claim 31 , wherein the input to the AI / ML position comprises one or more transmitted peak power or average power measurements from each of a plurality of TRPs associated with LOS communications.

Images