A framework for data collection configuration enabling sample selection and range for artificial intelligence and machine learning (ai / ML)

Abstract

Methods, apparatuses, and systems provide recovery for artificial intelligence and machine learning (AI / ML) data collection. In the context of a method, the method includes receiving, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; selecting a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmitting a second message to the network node, the second message comprising second information indicative of the configuration.

Description

A FRAMEWORK FOR DATA COLLECTION CONFIGURATION ENABLING SAMPLESELECTION AND RANGE FOR ARTIFICIAL INTELLIGENCE AND MACHINELEARNING (AI / ML)RELATED APPLICATION

[0001] This application claims priority to US provisional Application No. 63 / 716575 filed November 5, 2024, which is incorporated herein by reference in its entirety.TECHNOLOGICAL FIELD

[0002] The present disclosure relates generally to techniques for artificial intelligence and machine learning (AI / ML) data collection and, more particularly, to a framework for data collection configuration enabling sample selection and range for AI / ML.BACKGROUND

[0003] Wireless communication networks may support one or more positioning protocols, such as a long-term evolution (LTE) positioning protocol (LPP) and / or a new radio (NR) positioning protocol A (NRPPa). LPP may be used, point-to-point, for communication between a location server and a target device to position the target device based on positioning-related information, such as position-related measurements. The NRPPa may be used for communication of positioning related information between a radio access network (RAN) node and a core network entity, such as a location management function (LMF). The LMF manages co-ordination and scheduling of resources used for determining a location of a target user equipment (UE) registered with or accessing the core network (e.g., a fifth generation core network (5GCN)). In some instances, the LMF calculates or verifies a location and velocity estimate for the target UE. Additionally, in some instances, the LMF may estimate an accuracy with which the location and / or velocity are determined. The LMF may receive location requests for a target UE from a serving access and mobility management function (AMF) using an Nlmf interface. The LMF may interact with a UE to exchange location information applicable to UE-assisted and UE-based position methods. In some instances, the LMF interacts with a RAN node, a non-3GPP interworking function (N3IWF), or a trusted non-3GPP access network (TNAN) to obtain location information. In some cases, the target device (e.g., a UE), the RAN node, and / or the core network entity may use AI / ML to improve positioning accuracy for the target device. Improvements in how positioning-related information is collected for AI / ML-based positioning are needed.BRIEF SUMMARY

[0004] Methods, apparatuses, and systems are disclosed for AI / ML data collection. In this regard, the methods, apparatuses, and systems are configured to support a framework for data collection configuration enabling sample selection and range for AI / ML. For example, the methods, apparatuses, and systems are configured to support a framework for data configuration enabling sample selection and range for AI / ML positioning so as to provide for improved positioning accuracy of AI / ML models. By providing for improved positioning accuracy, the methods, apparatuses, and systems may provide for reduced latency and improved performance of wireless communications in the network.

[0005] In at least one example embodiment, an apparatus is provided comprising at least one processor and at least one memory including computer program code (e.g., instructions) configured to, with the at least one processor, cause the apparatus at least to receive, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; select a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmit a second message to the network node, the second message comprising second information indicative of the configuration.

[0006] In at least one example embodiment, the configuration is indicative of at least one of the following: a granularity for the selection of channel measurement samples, one or more ranges for the selection of channel measurement samples, or one or more similarity metrics for the selection of channel measurement samples.

[0007] In at least one example embodiment, the configuration is indicative of the granularity by being indicative of at least one of the following: a range of samples, a starting point associated with the range of samples, one or more types of offsets associated with one or more samples, a quantity of samples, a quantity of consecutive samples, a quantity of non- consecutive samples, or a ratio of consecutive samples per non-consecutive sample.

[0008] In at least one example embodiment, the configuration is indicative of the one or more ranges by including an indication to determine a range associated with a sample based on an offset of the sample.

[0009] In at least one example embodiment, the configuration is indicative of the one or more similarity metrics by including an indication of a type of similarity metric.

[0010] In at least one example embodiment, the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: obtain a plurality of channel measurement samples based at least in part on one or more reference signals; and select one or more channel measurement samples from the plurality of channel measurement samples in accordance with the configuration, wherein the one or more channel measurement samples are associated with one or more offsets relative to a reference point, and wherein the second message comprises third information indicative of the one or more channel measurement samples and the one or more offsets.

[0011] In at least one example embodiment, the one or more channel measurement samples comprise at least a channel measurement sample having a value that is greater than one or more other values of one or more other channel measurement samples included in the plurality of channel measurement samples, wherein the channel measurement sample is associated with an offset of the one or more offsets, and wherein a quantity of channel measurement samples included in the one or more channel measurement samples is based at least in part on the offset.

[0012] In at least one example embodiment, the third information is indicative of the one or more channel measurement samples by including at least one of the following: an indication of a power level associated with the channel measurement sample and the quantity of channel measurement samples, or an indication of a respective power level associated with each channel measurement sample of the one or more channel measurement samples.

[0013] In at least one example embodiment, the third information is indicative of the quantity of channel measurement samples by including an indication of a range associated with the offset.

[0014] In at least one example embodiment, selecting the one or more channel measurement samples comprises: selecting a set of consecutive channel measurement samples from the plurality of channel measurement samples; identifying a first channel measurement sample of the set of consecutive channel measurement samples, wherein the first channel measurement sample is associated with a first offset relative to the reference point; and selecting a first subset of consecutive channel measurement samples based at least in part on the first offset, wherein the one or more channel measurement samples comprises at least the first subset of consecutive channel measurement samples.

[0015] In at least one example embodiment, selecting the one or more channel measurement samples further comprises: identifying a second channel measurement sample of the set of consecutive channel measurement samples, wherein the second channelmeasurement sample is associated with a second offset relative to the reference point, the second offset being greater than the first offset; and selecting a second subset of consecutive channel measurement samples based at least in part on the second offset, wherein a first quantity of samples included in the first subset is greater than a second quantity of samples included in the second subset based at least in part on the second offset being greater than the first offset.

[0016] In at least one example embodiment, the third information comprises an indication of the first offset and at least one of the following: an indication of the second offset or an indication of a difference between the first offset and the second offset.

[0017] In at least one example embodiment, in accordance with the configuration, the third information indicates the reference point.

[0018] In at least one example embodiment, the one or more offsets comprise at least one of the following: one or more time offsets, one or more phase offsets, or one or more angle offsets, and the reference point comprises at least one of the following: a reference time, a reference phase, or a reference angle.

[0019] In at least one example embodiment, the plurality of channel measurement samples is associated with a measurement time window, and wherein a quantity of channel measurement samples included in the one or more channel measurement samples is based at least in part on a level of similarity associated with the plurality of channel measurement samples.

[0020] In at least one example embodiment, selecting the one or more channel measurement samples comprises: identifying a first portion of the plurality of channel measurement samples that are associated with a first time-domain measurement window; determining a first level of similarity associated with the first portion based at least in part on one or more similarity metrics; and selecting a first subset of channel measurement samples from the first portion based at least in part on the first level of similarity, wherein the one or more channel measurement samples comprises at least the first subset of channel measurement samples.

[0021] In at least one example embodiment, selecting the one or more channel measurement samples further comprises: identifying a second portion of the plurality of channel measurement samples that are associated with a second time-domain measurement window; determining a second level of similarity associated with the second portion based at least in part on the one or more similarity metrics; and selecting a second subset of channel measurement samples from the second portion based at least in part on the second level ofsimilarity, wherein a first quantity of samples included in the first subset is greater than a second quantity of samples included in the second subset based at least in part on the second level of similarity being greater than the first level of similarity.

[0022] In at least one example embodiment, the plurality of channel measurement samples includes one or more types of channel measurements, and wherein the one or more types of channel measurements comprises at least one of the following: a received power, a phase, a time-domain channel impulse response (OR), a power delay profile (PDP), or a delay profile (DP).

[0023] In at least one example embodiment, the one or more characteristics comprise at least one of the following: one or more radio propagation conditions, one or more channel conditions, one or more capabilities of a target device, or one or more positioning accuracy constraints.

[0024] In at least one example embodiment, the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: receive, from the network node, a third message comprising information indicating activation of a subset of the set of configurations, wherein the configuration is selected from the subset.

[0025] In at least one example embodiment, the apparatus comprises a user equipment (UE) or a radio access network (RAN) node.

[0026] In at least one example embodiment, the apparatus comprises the UE, and wherein the one or more ML models are configured at the UE.

[0027] In at least one example embodiment, the first message and the second message are communicated in accordance with a long term evolution (LTE) positioning protocol (LPP).

[0028] In at least one example embodiment, the apparatus comprises the RAN node, and wherein the one or more ML models are configured at the RAN node or the network node.

[0029] In at least one example embodiment, the first message and the second message are communicated in accordance with a new radio (NR) positioning protocol.

[0030] In at least one example embodiment, the network node comprises at least one of the following: a radio access network (RAN) node, a location management function (LMF), an access and mobility management function (AMF), or a network data analytics function (NWDAF).

[0031] In at least one example embodiment, at least one of the first message or the second message are communicated in accordance with a radio resource control (RRC) protocol or a medium access control (MAC) protocol.

[0032] In at least one example embodiment, the channel measurement samples comprise at least one of the following: time-domain channel measurement samples, angle-domain channel measurement samples, or phase-domain channel measurement samples.

[0033] In at least one example embodiment, an apparatus is provided comprising at least one processor and at least one memory including computer program code (e.g., instructions) configured to, with the at least one processor, cause the apparatus at least to determine a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmit a first message to a node, the first message comprising first information indicative of the set of configurations; and receive, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

[0034] In at least one example embodiment, the node comprises a radio access network (RAN) node, and wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: transmit, to at least one other RAN node neighboring the RAN node, a third message indicative of the set of configurations; and receive, from the at least one other RAN node in response to the third message, a fourth message indicative of a second configuration of the set of configurations.

[0035] In at least one example embodiment, the second configuration is different from the configuration.

[0036] In at least one example embodiment, the node comprises a radio access network (RAN) node, and wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: transmit, to at least one other RAN node neighboring the RAN node in response to the second message, a third message indicative of the configuration.

[0037] In at least one example embodiment, determining the set of configurations is based at least in part on one or more positioning accuracy constraints.

[0038] In at least one example embodiment, the configuration is indicative of at least one of the following: a granularity for the selection of channel measurement samples, one or more ranges for the selection of channel measurement samples, or one or more similarity metrics for the selection of channel measurement samples.

[0039] In at least one example embodiment, the configuration is indicative of the granularity by being indicative of at least one of the following: a range of samples, a starting point associated with the range of samples, one or more types of offsets associated with oneor more samples, a quantity of samples, a quantity of consecutive samples, a quantity of non- consecutive samples, or a ratio of consecutive samples per non-consecutive sample.

[0040] In at least one example embodiment, the configuration is indicative of the one or more ranges by including an indication to determine a range associated with a sample based on an offset of the sample.

[0041] In at least one example embodiment, the configuration is indicative of the one or more similarity metrics by including an indication of a type of similarity metric.

[0042] In at least one example embodiment, the second message comprises third information indicative of one or more channel measurement samples and one or more offsets associated with the one or more channel measurement samples, wherein the one or more offsets are relative to a reference point, and wherein the one or more channel measurement samples are based at least in part on the configuration.

[0043] In at least one example embodiment, the one or more offsets comprise at least one of the following: one or more time offsets, one or more phase offsets, or one or more angle offsets, and the reference point comprises at least one of the following: a reference time, a reference phase, or a reference angle.

[0044] In at least one example embodiment, the third information is indicative of the one or more channel measurement samples by including at least one of the following: an indication of a power level associated with a channel measurement sample of the one or more channel measurement samples and a quantity of channel measurement samples in the one or more channel measurement samples, or an indication of a respective power level associated with each channel measurement sample of the one or more channel measurement samples.

[0045] In at least one example embodiment, the channel measurement sample is associated with an offset of the one or more offsets, and wherein the third information is indicative of the quantity of channel measurement samples by including an indication of a range associated with the offset.

[0046] In at least one example embodiment, the one or more channel measurement samples includes one or more types of channel measurements, and wherein the one or more types of channel measurements comprises at least one of the following: a received power, a phase, a time-domain channel impulse response (OR), a power delay profile (PDP), or a delay profile (DP).

[0047] In at least one example embodiment, the configuration is based at least in part on one or more characteristics of a wireless communication channel, and wherein the one or more characteristics comprise at least one of the following: one or more radio propagationconditions, one or more channel conditions, one or more capabilities of a target device, or one or more positioning accuracy constraints.

[0048] In at least one example embodiment, the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: transmit a third message comprising information indicating activation of a subset of the set of configurations, wherein the configuration is based at least in part on the subset.

[0049] In at least one example embodiment, the apparatus comprises a radio access network (RAN) node, a location management function (LMF), an access and mobility management function (AMF), or a network data analytics function (NWDAF).

[0050] In at least one example embodiment, a method is provided comprising receiving, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; selecting a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmitting a second message to the network node, the second message comprising second information indicative of the configuration.

[0051] In at least one example embodiment, a method is provided comprising determining a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmitting a first message to a node, the first message comprising first information indicative of the set of configurations; and receiving, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

[0052] In at least one example embodiment, a non-transitory computer readable storage medium is provided. The non-transitory computer readable storage medium comprises computer instructions that, when executed by an apparatus, cause the apparatus to receive, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; select a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmit a second message to the network node, the second message comprising second information indicative of the configuration.

[0053] In at least one example embodiment, a non-transitory computer readable storage medium is provided. The non-transitory computer readable storage medium comprises computer instructions that, when executed by an apparatus, cause the apparatus to determine a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmit a first message to a node, the first message comprising first information indicative of the set of configurations; and receive, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

[0054] In at least one example embodiment, an apparatus is provided that comprises means for receiving, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; selecting a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmitting a second message to the network node, the second message comprising second information indicative of the configuration.

[0055] In at least one example embodiment, an apparatus is provided that comprises means for determining a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmitting a first message to a node, the first message comprising first information indicative of the set of configurations; and receiving, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

[0056] The above summary is provided merely for purposes of summarizing at least some example embodiments to provide a basic understanding of some aspects of the disclosure. Accordingly, it will be appreciated that the above-described embodiments are merely examples and should not be construed to narrow the scope or spirit of the disclosure in any way. It will also be appreciated that the scope of the disclosure encompasses many potential embodiments in addition to those summarized here, some of which will be further described below.BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Having thus described certain example embodiments of the present disclosure in general terms, reference will hereinafter be made to the accompanying drawings, which are not necessarily drawn to scale, and wherein:

[0058] FIG. 1 illustrates an example diagram of a communication network to which one or more examples disclosed herein may be applied;

[0059] FIG. 2 illustrates an example diagram of time-domain channel samples and sub-samples to which one or more examples disclosed herein may be applied;

[0060] FIG. 3 illustrate an example diagram of time-domain channel information to which one or more examples disclosed herein may be applied;

[0061] FIG. 4 illustrates an example diagram of channel measurement sample selection to which one or more examples disclosed herein may be applied;

[0062] FIG. 5 illustrates an example signaling diagram to which one or more examples disclosed herein may be applied;

[0063] FIG. 6 illustrate an example signaling diagram to which one or more examples disclosed herein may be applied;

[0064] FIG. 7 illustrates an example flowchart of a method to which one or more examples disclosed herein may be applied;

[0065] FIG. 8 illustrates an example flowchart of a method to which one or more examples disclosed herein may be applied;

[0066] FIG. 9 illustrates an example block diagram of an apparatus to which one or more examples disclosed herein may be applied.DETAILED DESCRIPTION

[0067] The following embodiments are exemplary. Although the specification may refer to “an”, “one”, or “some” embodiment(s) in several locations of the text, this does not necessarily mean that each reference is made to the same embodiment(s), or that a particular feature only applies to a single embodiment. Single features of different embodiments may also be combined to provide other embodiments. Further, when a particular feature, structure, or characteristic is described in connection of an embodiment, it is within the knowledge of one skilled in the art to apply such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described. It shall be understood that although the terms “first,” “second” and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another.

[0068] For the purposes of the present disclosure, the phrases “at least one of A or B”, “at least one of A and B”, and “A and / or B” means (A), (B), or (A and B). For the purposes of the present disclosure, the phrase “A, B, and / or C” means (A), (B), (C), (A and B), (A and C), (B and C), or (A, B, and C).

[0069] Embodiments described may be implemented in a communication network, such as any of the following radio access technologies (RATs): Worldwide Interoperability for Micro-wave Access (WiMAX), Global System for Mobile communications (GSM, 2G), GSM EDGE radio access Network (GERAN), General Packet Radio Service (GRPS), Universal Mobile Telecommunication System (UMTS, 3G) based on basic wideband-code division multiple access (W-CDMA), high-speed packet access (HSPA), Long Term Evolution (LTE), LTE-Advanced, and enhanced LTE (eLTE), 5G (also called NR), or any future RAT such as 6G. Moreover, communication within the communication network may utilize any proper wireless communication technology, comprising but not limited to: Code Division Multiple Access (CDMA), Frequency Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), Frequency Division Duplex (FDD), Time Division Duplex (TDD), Multiple-Input Multiple-Output (MIMO), Orthogonal Frequency Division Multiple (OFDM), and / or Discrete Fourier Transform spread OFDM (DFT-s-OFDM).

[0070] As used herein, the term “network device” or “network node” refers to a node in a communication network via which user equipment may access the network and / or which is capable of controlling radio communication and managing radio resources within a cell. The network node or network device may be referred to as a base station (BS), an access point (AP) or an access node. The network device may be, depending on the applied technology, for example, a node B (NodeB or NB), an evolved NodeB (eNodeB or eNB), an NR NB (also referred to as a gNB), a Remote Radio Unit (RRU), a radio head (RH), a remote radio head (RRH), a relay, an Integrated Access and Backhaul (IAB) node, a low power node, a nonterrestrial network (NTN) or non-ground network device such as a satellite network device, a low earth orbit (LEO) satellite and a geostationary orbit (GEO) satellite, or an aircraft network device.

[0071] Moreover, in connection of split radio access network (RAN), the network device may refer to a centralized unit (CU) of a base station and / or a distributed unit (DU) of a base station. An interface between CU and DU may be referred to as an Fl interface in NR. In the split RAN architecture, node operations may be carried out, at least partly, in the central / centralized unit, CU, (e.g. server, host or node) operationally coupled to the DU, (e.g. a radio head / node). One CU may control one or more DUs, acting at least as transmit / receive(Tx / Rx) nodes. In at least one embodiment, the DUs may comprise e.g. a radio link control (RLC), medium access control (MAC) layer and a physical (PHY) layer, whereas the CU may comprise the layers above RLC layer, such as a packet data convergence protocol (PDCP) layer, a radio resource control (RRC) layer and an internet protocol (IP) layer. Other functional splits are possible too. In practice, any processing task may be performed in either the CU or the DU and the boundary where the responsibility is shifted between the CU and the DU may depend on the applied implementation.

[0072] The term “terminal device” refers to any end device that may be capable of wireless communication. By way of example, a terminal device may be referred to as a communication device, user equipment (UE), a Subscriber Station (SS), or a Mobile Station (MS). The terminal device may include a mobile phone, a cellular phone, a smart phone, voice over IP (VoIP) phones, wireless local loop phones a tablet, a wearable terminal device, a personal digital assistant (PDA), portable computers, desktop computer, image capture terminal devices such as digital cameras, gaming terminal devices, music storage and playback appliances, vehicle-mounted wireless terminal devices, USB dongles, an Internet of Things (loT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medical device and applications (e.g., remote surgery), an industrial device and applications (e.g., a robot and / or other wireless devices operating in an industrial and / or an automated processing chain contexts), a consumer electronics device, a device operating on commercial and / or industrial wireless networks, and the like.

[0073] A term “resource”, as used herein, may refer to radio resources in time domain, in frequency domain, in space domain, and / or in code domain. Some examples of resources include e.g. a physical resource block (PRB), a radio frame, a subframe, a time slot, a subband, a frequency region, a sub-carrier, a beam, etc. The term “transmission” and / or “reception” may refer to wirelessly transmitting and / or receiving via a wireless propagation channel on radio resources.

[0074] FIG. 1 illustrates an example of a communication network to which examples disclosed herein may be applied. The communication network (also referred to herein as a cellular communication network or system) may comprise a network node 110 providing one or more cells, such as source cell 100, and a network node 112 providing one or more other cells, such as target cell 102. Each cell may be, e.g., a macro cell, a micro cell, femto, or a pico cell, for example. The cell may define a coverage area or a service area of the corresponding access node.

[0075] The network node 110 may provide a user equipment (UE) 120 (one or more UEs) with wireless access to the communication network. The wireless access may comprise downlink (DL) communication from the network node to the UE 120 and uplink (UL) communication from the UE 120 to the network node. Examples of uplink channels comprise physical uplink control channel (PUCCH) for transmitting control information and physical uplink shared channel (PUSCH) for transmitting data towards the network. Examples of downlink channels comprise physical downlink control channel (PDCCH) for transmitting control information and physical downlink shared channel (PDSCH) for transmitting data towards the user equipment.

[0076] There may be a plurality of UEs 120, 122 in the system. Each of them may be served by the same or by different network nodes 110, 112. UE may be configured with dual connectivity (DC), wherein the UE, e.g. UE 120, may be connected to multiple network nodes 110, 112. The UEs 120, 122 may communicate with each other, in case device-to- device (D2D) communication interface is established between them via a so-called sidelink (SL). Such D2D communications may be referred to as machine-to-machine, peer-to-peer (P2P) communications, or vehicle-to-vehicle (V2V), for example.

[0077] In the case of multiple network nodes in the communication network, the network nodes may be connected to each other via an interface. LTE specifications call such an interface as X2 interface. An interface between an LTE node and a 5G node, or between two 5G nodes may be called Xn interface.

[0078] The network nodes 110 and 112 may be further connected via another interface to a core network 116 (also referred to herein as the core 116) of the communication network. The LTE specifications specify the core network as an evolved packet core (EPC), and the core network may comprise e.g. a mobility management entity (MME) and a gateway node. The MME may handle mobility of terminal devices in a tracking area encompassing a plurality of cells and handle signaling connections between the terminal devices and the core network. The gateway node may handle data routing in the core network and to / from the terminal devices. The 5G specifications specify the core network as a 5G core (5GC). The 5G core may comprise e.g. an access and mobility management function (AMF) and a user plane function / gateway (UPF) and other functions. The AMF may handle termination of non-access stratum (NAS) signaling, NAS ciphering & integrity protection, registration management, connection management, mobility management, access authentication and authorization, security context management. The UPF node may support packet routing and forwarding, packet inspection and quality of service (QoS) handling, for example.

[0079] In 5G New Radio, different kinds of data transfer services are offered by the medium access control (MAC) layer. To accommodate different kinds of data transfer services, multiple types of logical channels are defined. A MAC protocol data unit (MAC PDU) may consist of one or more MAC control elements (MAC CEs) corresponding to one or more features requiring the MAC CEs. According to the third generation partnership project (3GPP) technical specification (TS) 38.321, the MAC PDU includes a subheader with a logical channel identification (LCID) value or an extended LCID (eLCID) value. In some examples, the UEs 120, 122 may be configured to transmit information pertaining to a beam management procedure to the network nodes 110, 112 via one or more MAC CEs.

[0080] A MAC PDU is a bit string that is byte aligned (i.e., multiple of 8 bits) in length. The bit strings are represented by tables in which the most significant bit is the leftmost bit of the first line of the table, the least significant bit is the rightmost bit on the last line of the table, and more generally the bit string is to be read from left to right and then in the reading order of the lines. The bit order of each parameter field within a MAC PDU is represented with the first and most significant bit in the leftmost bit and the last and least significant bit in the rightmost bit.

[0081] A MAC SDU is a bit string that is byte aligned (i.e., multiple of 8 bits) in length. A MAC SDU is included into a MAC PDU from the first bit onward. A MAC CE is a bit string that is byte aligned (i.e., multiple of 8 bits) in length. A MAC subheader is a bit string that is byte aligned (i.e., multiple of 8 bits) in length. Each MAC subheader is placed immediately in front of the corresponding MAC SDU, MAC CE, or padding. The MAC entity shall ignore the value of the Reserved bits in downlink MAC PDUs. The MAC SDUs may have variable sizes. The MAC PDU may comprise one or more MAC subPDUs. Each MAC subPDU comprises one of the following: a MAC subheader only (including padding); a MAC subheader and a MAC SDU; a MAC subheader and a MAC CE; or a MAC subheader and padding. The MAC CEs may be placed together. DL MAC subPDU(s) with MAC CE(s) is placed before any MAC subPDU with MAC SDU and MAC subPDU with padding. UL MAC subPDU(s) with MAC CE(s) is placed after all the MAC subPDU(s) with MAC SDU and before the MAC subPDU with padding in the MAC PDU. The size of padding can be zero. A maximum of one MAC PDU can be transmitted per TB per MAC entity.

[0082] Referring back to FIG. 1 , the system may be configured to support multiple input multiple output (MIMO) operations, for example, at the UEs 120, 122 or the network nodes 110, 112. In some examples, the system may support UE event driven reporting for MIMO operations. For example, the system may support one or more features to facilitate UE-initiated / event-driven beam management for reducing overhead and / or latency. As used herein, UE-initiated / event-driven beam management refers to event-driven beam management that is initiated by a UE. In some examples, the term event-driven beam management refers to a beam management procedure that is triggered in response to one or more events. In some examples, UE-initiated / event-driven beam management may include the use of a unified transmission configuration indicator (TCI) and / or one or more CSI measurement and reporting configuration frameworks. Additionally, in some examples, UE- initiated / event-driven beam management may utilize one or more frequency ranges, such as FR2 (Frequency Range 2) or other frequency ranges that may include operational frequencies in the millimeter wave (mmWave) region (e.g., above 24 giga Hertz (GHz)). Additionally, in some examples, UE-initiated / event-driven beam management may include (or be otherwise associated with) a transmission reception point (TRP), such as a single transmission reception point (sTRP) with intra-cell beam management and / or inter-cell beam management. In some instances, uplink signaling (e.g., uplink signaling content(s) and / or uplink signaling procedure(s)) for UE-initiated / event-driven beam reporting facilitates relatively fast beam switching. Additionally, in some instances, the UE-initiated / event-driven nature of uplink transmissions, uplink signaling medium(s) / container(s) may be used for (e.g., designed for) the purpose of beam reporting.

[0083] The system of FIG. 1 may use AI / ML for the air interface including, for example, AI / ML-based positioning. In some examples, the system may support one or more AI / ML frameworks for one-sided AI / ML models, which may include signaling and protocol aspects of life cycle management (LCM) enabling functionality, model selection, activation, deactivation, and switching, as well as fallback options. Additionally, or alternatively, some such frameworks may include identification related signaling, signaling / mechanism(s) for LCM to facilitate model training, inference, performance monitoring, and data collection (e.g., except for the purpose of core network (CN) operations and management (0AM) or over-the-top (OTT) collection of UE-sided model training data) for both UE-sided and network-sided models, and / or signaling mechanism of applicable functionalities and models. In some examples, the system may employ AI / ML to improve the accuracy of UE positioning.

[0084] The system may support multiple use cases for AI / ML-based positioning. For example, the system may support direct AI / ML positioning in which one or more AI / ML models are used to directly estimate the location of a UE (e.g., the UE 120 and / or the UE 122). That is, for direct AI / ML positioning, an output of an AI / ML model may include aposition at which a UE is located. For example, for UE-based direct AI / ML positioning, one or more AI / ML models are configured at (e.g., reside on) the UE side and, as such, the UE may use the AI / ML models to estimate the location of the UE (e.g., the UE use the AI / ML models to infer the position at which the UE is located). In other words, the output of the one or more AI / ML models includes (or otherwise indicates) a position of the UE. For AI / ML assisted positioning, the output of one or more AI / ML models may include intermediate key performance indicators (KPIs) and / or other intermediate features, such as timing related information (e.g., line-of-sight (LOS) information, non-line-of-sight NLOS) information). In some examples of AI / ML assisted poisoning, these intermediate features (or KPIs) may be reported to the network and the network may estimate the position of the UE based on the reported intermediate features (or KPIs).

[0085] Direct AI / ML positioning may include multiple use cases, such as Case 1 in which UE-based positioning is performed with one or more UE-side AI / ML models, Case 2b in which UE-assisted / LMF-based positioning is performed with one or more LMF-side AI / ML models, and Case 3b in which RAN node (e.g., NG-RAN node) assisted positioning is performed with one or more LMF-side AI / ML models. AI / ML assisted positioning may also include multiple use cases, such as Case 2a in which UE-assisted / LMF-based positioning is performed with one or more UE-side models and Case 3a in which NG-RAN node assisted positioning is performed with one or more gNB-side models. In other words, Case 1 includes downlink-based positioning at the UE and Case 3 a and Case3b include uplink-based positioning at the network. In some examples, of Case 1, the UE may collect channel information (e.g., channel measurements, such as received power measurements of downlink reference signals) to use as input for one or more AI / ML models, which may infer a position of the UE based on the input. In some examples, the UE may collect channel information or measurements based on one or more downlink reference signals, such as one or more positioning reference signals (PRSs). As used herein, a channel measurement refers to a process of assessing the channel. For example, a device (e.g., the UE, the RAN node) may collect a channel measurement (e.g., data, metrics) through the process of assessing the channel. Additionally, as used herein, channel information refers to information (e.g., knowledge) derived from one or more channel measurements. In some examples, channel information may be used to improve (e.g., optimize) wireless communications. In some examples of Case 3a and Case 3b, a RAN node (e.g., gNB or TRP, such as the network node 110 and / or the network node 112) may report channel information (e.g., channel measurements, such as received power measurements of uplink reference signals) to the LMFand the LMF may use the channel information as input for one or more AI / ML models, which may infer (e.g., estimate) the position of the UE. In some examples, the RAN node may obtain (and report) channel information or measurements based on one or more uplink reference signals, such as one or more sounding reference signals (SRSs). In some examples, the UE may estimate channel measurements (e.g., channel impulse response (OR), power delay profile (PDP), delay profile (DP)) based on channel conditions and / or one or more channel characteristic (e.g., LOS / NLOS, average power, mean / access delay, mobility conditions). In some examples, the UE estimates channel measurements (e.g., CIR, PDP, DP) based on channel conditions, channel characteristic LOS / NLOS, average power, and / or access delay and mobility conditions.

[0086] In some examples, time-domain channel measurements may be sample-based and / or path-based. That is, for AI / ML-based positioning, sample-based measurements and / or path-based measurements may be used for AI / ML model input, for example, for inference, training, and or data collection purposes. For example, AI / ML-based positioning may use sample-based measurements in which timing information is an integer multiple of sampling periods and / or path-based measurements in which timing information is according to detected path timing and may not be an integer multiple of sampling periods. In some examples, a granularity of timing information associated with sample-based measurements may be different from a granularity of timing information associated with path-based measurements.

[0087] In some examples, for sample-based measurements, a measurement may include Nt’ samples of an estimated channel response in the time-domain. The timing information for the Nt’ samples may be reported with a timing granularity T, in which T = 2feX Tc, and in which k represents a timing reporting granularity factor and Tcis the basic time unit for NR. The timing information may be defined relative to a reference time. The corresponding measurement (e.g., power, if reported) corresponds to the measurement for the reported Nt’ samples. In some examples, however, the UE and / or the RAN node may lack a framework for determining Nt’ and k (e.g., a value range of Nt’ and / or a value range of integer k for the timing granularity T). Additionally, or alternatively, the UE and / or the RAN node may lack a framework for signaling Nt’ and k (e.g., between the UE, the RAN node, and / or LMF). In other words, a granularity of sample-based measurements may be based on one or more parameters, including Nt and Nt’, in which Nt’ samples may be selected from a set of Nt samples. For example, the input dimension for a channel measurement may be NTRP* port * Nt, in which NTRPis the number of TRPs, Nportis the number of transmit / receiveantenna port pairs, and Nt is the number of consecutive time domain samples. In some examples, if Nt’ samples (where Nt’ < Nt) are selected as model input, with remaining (Nt- Nt’) time domain samples set to zero, timing information for the Nt’ samples may also be provided as model input. In some examples, Nt may be referred to as a sampling parameter (e.g., a parameter for sampling a channel measurement) and Nt’ may be referred to as a subsampling parameter (e.g., a parameter for sub-sampling the channel measurement).

[0088] In some examples, for sample-based measurement, the UE may lack a framework for selecting Nt’ samples out of a list of Nt consecutive sample, which may have timing granularity T, for determining a starting time of the list of Nt samples, and for determining a value range of Nt. For example, the UE (or the RAN node) may be configured with one or more protocols that specify measurements, signaling, and / or mechanism(s) to facilitate LCM operations for various positioning accuracy use cases and provide for consistency between training and inference regarding network-side conditions (if identified) for inference at the UE for one or more (related) positioning sub-use cases. However, the UE (or the RAN node) may lack one or more measurement configurations (or mechanisms) to reduce reporting overhead, while maintaining (or improving) positioning accuracy, such as a sample selection rule that considers the UE capability, radio conditions, and accuracy constraints for samplebased measurements. In some examples, the RAN node may select Nt’ sub-samples from Nt samples based on received power (e.g., the Nt’ selected samples may be samples with the highest received power among the Nt samples). In some such examples, however, the selected Nt’ samples may provide insufficient channel information and predictions based on the selected Nt’ samples may be relatively inaccurate.

[0089] For example, signaling between the UE and RAN node may occur via multipath channels in which signals propagate along multiple paths between a transmitter and a receiver. That is, signals transmitted from a transmitting device may reach a receiving device via multiple paths. In some examples, the UE may be configured to measure and report (e.g., via higher layer parameter AdditionalPath-relativeTiming-Request and subject to UE capability), timing and the quality metrics of up to 8 additional detected paths (or another suitable number of additional detected paths), that are associated with each reference signal timing difference (RSTD) or UE Rx-Tx time difference (e.g., as indicated via nr-RSTD or nr- UE-RxTxTimeDiff parameters). The timing of each additional path may be reported relative to path timing used for determining the higher layer parameters nr-RSTD or nr-UE- RxTxTimeDiff. A used herein, higher layer parameters refer to RRC parameters (e.g., parameters configured via RRC signaling). For UE positioning measurement reporting inhigher layer parameters NR-DL-TDOA-SignalMeasurementlnformation or NR-Multi-RTT- SignalMeasurementlnformation, the UE may be configured to measure and report, subject to UE capability, the path downlink position reference signal (PRS) reference signal received power (RSRP) of the first path and the up to 8 additional paths that are associated with each RSTD or UE Rx-Tx time difference. By constraining the selection of Nt’ samples to samples with the highest power (e.g., relative to other samples within the set of Nt samples), the selected Nt’ samples may not be representative of the multiple paths. For example, the highest power samples may correspond to the strongest (e.g., most direct) paths and, as such, may not be representative of other (more indirect) paths. Thus, constraining the selection of Nt’ samples to samples with the highest power (e.g., relative to other samples within the set of Nt samples) may lead to insufficient channel information and the system of FIG. 1 may lack a framework for configuring the UE or RAN node to select suitable samples, which may provide sufficient positioning accuracy (e.g., the network may lack a framework for configuring the UE or RAN node with Nt, Nt’, a range of samples, or timing information that provides suitable positioning accuracy). In other words, methods for (sub-)sampling based on strongest power may be insufficient for AI / ML data collection. For example, such methods may lead to insufficient inference input for various positioning scenarios and the UE and / or RAN node may lack a framework for selecting a suitable Nt’ sample set that balances positioning accuracy and signaling overhead. That is, reporting all samples may lead to increased overhead and reporting samples selected based (only) on strongest power may lead to reduced positioning accuracy.

[0090] In some examples, the UE or RAN node may be configured with different types of measurements as well as associated configurations for AI / ML-based positioning. For example, a model input specification may specify for a UE to down select options for timedomain samples selection by comparing positioning accuracy performance and the signaling overhead. Such options may include a first option in which the UE may select Nt consecutive time domain samples as model input, or a second option in which the UE may select Nt’ time-domain samples with the strongest power (among samples obtained by the UE) as model input. Additionally, or alternatively, a model input specification may specify multiple values of Nt or Nt’ for configuration. Additionally, or alternatively, a model specification may consider the representation of timing information for model input in accordance with a power-per-sample-based approach in which a representation of timing information is obtained by taking a quantity (K_fine) of samples on the sampling grid of an estimated CIR around (e.g., on the left and right of, within the overall CIR window) N samples thatcorrespond to the N samples with the highest power values among samples within the estimated OR. Additionally, or alternatively, the model specification may consider the representation of timing information for model input in accordance with a path-based approach in which the representation of timing information is obtained by taking K (consecutive or non-consecutive) time instances based on a timing grid of an output of a pathdetection. In some examples, the timing information includes reference time, sampling period, a value of K, and K timing instance values.

[0091] In some examples, the reporting of an increased quantity of additional paths may be considered for examples in which the model is deployed at the LMF. In some examples in which a network defines a sample selection mechanism for overhead reduction, the selection mechanism may include a rule that specifies a threshold (e.g., maximum) number of non-zero samples within a channel response measurement that may be configured. In some examples, the network may support an LMF configuring the value of Nt for sample-based measurement, in which the starting point of Nt may be determined based on the measurement report of Nt samples, which may be determined based on the first detected path in time. In some such examples, the measurement report of Nt samples may be determined based on a path having a power that is stronger than a configured threshold. In some examples, the measurement report of Nt samples is configured by LMF. In some examples, the measurement report of Nt samples is determined by TRP.

[0092] While the UE or RAN node may be configured with different types of measurements as well as associated configurations for AFML-based positioning, the configurations may fail to support (e.g., specify) how the UE or RAN node is to select sample measurements for inference and / or training of the AI / ML models. That is, when the network configures a UE (or RAN node) for data collection for AI / ML inference and / or training (e.g., when LMF requests sample measurements), the configuration may fail specify a mechanism for selecting suitable samples (e.g., may fail to specify Nt / Nt’, a range of samples, timing information) that satisfy positioning constraints.

[0093] Various aspects of the present disclosure provide a framework for sampling that supports AI / ML-driven data collection, while avoiding unnecessary increases in system overhead and ensuring suitable positioning accuracy (e.g., required positioning accuracy). For example, the present disclosure may provide one or more methods for sample set collection to enhance the AI / ML inference inputs, while preemptively considering UE capabilities, radio conditions, and accuracy constraints. In some examples, in accordance with the framework of the present disclosure, a core network node (e.g., an LMF, AMF, and / ornetwork data analytics function (NWDAF), such as the core 116) may determine one or more criteria (e.g., configurations) for selecting samples (e.g., selecting a set of Nt samples and / or selecting Nt’ samples from the set of Nt samples) and may configure one or more UEs for data collection purposes. The sample selection criteria may be based on, for example, consecutive and / or non-consecutive samples. For instance, the core network node may configure the UE or RAN node to select a fraction of consecutive samples per detected non- consecutive peak, depending on the observed channel characteristics. In some such examples, Nt’ may be equal to Nt or a subset of Nt.

[0094] The UE or RAN node may select a configuration from the configurations indicated by the network node based on, for example, radio propagation conditions. The UE or RAN node may collect samples in accordance with the selected configuration and report, to the core network node, the selected configuration (e.g., and information associated with the collected samples). In some examples, in accordance with the configuration, the UE or RAN node may select samples based on an offset of one or more strongest samples relative to a reference point, and a variance (e.g., range) applied across those strongest samples. In other words, in accordance with the configuration, sample selection may be based on one or more offsets of one or more strongest samples relative to a reference point, and one or more variances applied across those strongest samples. As used herein, a strongest sample refers to a sample having a higher value, such as a higher received power value, than one or more other samples obtained by the UE or RAN node. Additionally, as used herein, a reference point refers to a reference time (t0), a reference angle (0O), or a reference phase (o). In some examples, in accordance with the configuration, the UE or gNB may select samples based on a level of correlation between samples associated with a particular measurement time window.

[0095] For example, the UE or RAN mode may receive, from the core network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more ML models. In such an example, the UE or RAN node may select a configuration from the set of configurations based on one or more characteristics of a wireless communication channel. The UE or RAN node may transmit a second message to the core network node, the second message include at least second information indicative of the configuration.

[0096] FIG. 2 illustrates an example diagram 200 of time-domain channel sample and subsample collection to which one or more examples disclosed herein may be applied. Thetime-domain channel sample and subsample collection illustrated in the example of FIG. 2 may be implemented at a device, which may also be referred to herein as a node. The device (or node) may include a UE or a RAN node (e.g., TRP / gNB) illustrated by and described with reference to FIG. 1.

[0097] In some examples, the device may receive a reference signal, such as a PRS (in the case of a UE) or an SRS (in the case of the RAN node), across one or more resource elements (REs) of one or more orthogonal frequency division multiplexing (OFDM) symbol. As illustrated in the example of FIG. 2, at 202, the device may receive (and measure) a reference signal in REs 206 (e.g., 4096 REs) of an OFDM symbol 204. As such, the device may obtain a measurement of the channel response (e.g., a received power measurement, such as reference signal received power (RSRP) measurement) for 4096 REs of the OFDM symbol 204. An RE may include one subcarrier in the frequency domain and one OFDM symbol in the time domain. Thus, based on the received power measurement at 202, the device may obtain a frequency-domain representation of a measured channel response (e.g., the measured received power) during the OFDM symbol 204. At 208, the device may apply an inverse fast Fourier transform (IFF!) to obtain a time-domain channel measurement 210 for the REs 206. That is, based on the IFFT at 208, the device may obtain a time-domain representation of the channel during the OFDM symbol 204. The time-domain representation of the channel may include (e.g., show) individual delayed replicas (e.g., 4096 samples) of the reference signal received at the device due to reflections and multipath propagation of the reference signal within the channel. In other words, the time-domain channel measurement 210 includes 4096 samples (also referred to as taps), in which each sample is representative of the received power of a version of the reference signal received at the device. The version of the reference signal may correspond to the reference signal or a reflection of the reference signal that propagated along a particular path between a transmitter (e.g., the RAN node) and a receiver (e.g., the UE). As such, each sample is associated with a particular delay (also referred to as an offset) relative to a reference point. In the example of FIG. 2, the samples may be obtained in accordance with a sampling period 212. In some examples, the sampling period 212 may be determined in accordance with the following Equation 1:Sampling period = l / Nf x A ) (1) where Nf correspond to the quantity of REs and A corresponds to the subcarrier spacing. For example, for Nf =4096 and A =30 kHz, the sampling period may be about 8.14 (ns). Insome other examples, a value of the sampling period (e.g., 4.069 ns) may be configured at the device.

[0098] As illustrated in the example of FIG. 2, the device may be configured to select samples 214 (e.g., Nt samples) from the 4096 samples of the time-domain channel measurement 210. For example, the device may select a configuration which indicates for the device to select 256 consecutive samples starting from sample 6. That is, the configuration may indicate that Nt is equal to 265 and may indicate a starting time for the Nt samples, which corresponds to the sixth sample of the time-domain channel measurement 210. Accordingly, the samples 214 may include samples 6 through 261 of the time-domain channel measurement 210. As illustrated in the example of FIG. 2, the device may be configured to select samples 216 (e.g., Nt’ samples) from the samples 214. In some examples, the device may select the samples 216 based on the configuration.

[0099] FIG. 3 illustrate an example diagram 300 of time-domain channel information to which one or more examples disclosed herein may be applied. The time-domain channel information illustrated in the example of FIG. 3 may be obtained at a device, which may also be referred to herein as a node. The device (or node) may include a UE or a RAN node (e.g., TRP / gNB) illustrated by and described with reference to FIGs. 1 and 2.

[0100] A number and resolution of time-domain samples obtained from a time-domain channel measurement may impact a granularity of a channel estimation based on the timedomain channel measurement. Thus, the selection of samples to collect and report may impact AI / ML inference input and, in some examples, positioning information obtained from the AI / ML. In some instances, such as instances in which time-domain samples are used as model inputs and sampling and / or sub-sampling is applied, the selection of sample measurement may be based on the received signal power of detected paths (in time). For example, the device may be configured to select the strongest samples (e.g., one or more samples having a higher received power than one or more other samples obtained by the device and / or one or more samples having a received power greater than a threshold). However, this approach may fail to provide suitable channel information (e.g., a suitable range of samples) and may lead to increases in sampling overhead. Moreover, such an approach may necessitate transferring at least the channel tap powers for numerous time instances (e.g., even if many of these powers are relatively small or negligible). For example, in accordance with such approaches, a device may select samples randomly or, for example, based on a received power threshold. As illustrated in the example of FIG. 3, these approaches may lead to reporting of channel tap received powers for samples within sampleset 302 and / or sample set 306, which may be less useful compared to samples within sample set 304. Accordingly, random (sub)-sampling and / or (sub)-sampling based (only) on the samples having the strongest power (e.g., received powers greater than one or more other samples or received powers greater than a threshold) may be insufficient for meeting AI / ML data collection constraints (e.g., may be insufficient for determining a suitable size of samples to input and / or report for various positioning scenarios).

[0101] In some other examples, in accordance with the framework for data collection configuration enabling sample selection and range for AI / ML presented herein, the device may be configured with multiple configurations for sample selection and may select a configuration (e.g., a suitable value for Nt’) that balances positioning accuracy and signaling overhead. In other words, the present disclosure provides an efficient sampling and / or subsampling strategy that supports AI / ML-driven data collection while avoiding unnecessary increases in system overhead and, in some examples, supports positioning constraints. That is the systems, methods, and apparatuses of the present disclosure provide for sample set collection, which enhances the AI / ML inference inputs, while preemptively considering UE capability, radio conditions, and accuracy constraints.

[0102] In some examples, a core network node (e.g., LMF) may determines one or more criteria (e.g., configurations) for selecting samples and configures one or more devices for data collection purposes. The sample selection criteria / configurations may be based on, for example, consecutive and / or non-consecutive samples. For instance, a possible use case may include a device selecting a configuration that indicates a fraction of consecutive samples per detected non-consecutive peak, based on the channel characteristics.

[0103] In some examples, a device may select one suitable configuration (e.g., including one or more criterion) from a set of configuration options indicated by the core network node based on, for example, radio propagation conditions. The device may then collect samples in accordance with the selected configuration (e.g., the fraction of consecutive samples per detected non-consecutive peak) and report the selected configuration, as well as information pertaining to the collected samples, to the core network node. In some examples, in accordance with the selected configuration, sample selection may be based on a time offset of one or more samples (e.g., one or more strongest samples) relative to a reference point and a variance applied across those samples.

[0104] FIG. 4 illustrates an example diagram 400 of channel measurement sample selection to which one or more examples disclosed herein may be applied. The channel measurement sample selection illustrated in the example of FIG. 4 may be implemented at adevice, which may also be referred to herein as a node. The device (or node) may include a UE or a RAN node (e.g., TRP / gNB) illustrated by and described with reference to FIGs. 1-3.

[0105] In some examples, the device may obtain a time-domain channel measurement401 including multiple samples. The device may apply sample selection criteria (e.g., a criterion or multiple criteria included in, or otherwise indicated via, a configuration provided by a core network node) to select one or more samples from the multiple samples included in the time-domain channel measurement 401.

[0106] For example, the network node (e.g., EMF) may define (e.g., in one or more configurations) criteria for selecting samples. In some examples, the criteria selected by the device for sample selection (e.g., one or more parameters included in a configuration selected by the device) may indicate for the device to select samples based on one or more time offsets of one or more samples relative to a reference point (e.g., a reference time (t0), a reference angle (0O), or a reference phase (o)) and based on one or more variances (e.g., ranges) applied across those samples. For example, in accordance with the selected criteria, the device may increase a variance applied to a sample as the time offset decreases. In other words, when the time offset of the samples increases, the variance for selecting samples narrows, leading to a more focused sample selection. Conversely, as the time offset decreases, the variance broadens, allowing for a wider selection of samples.

[0107] In some examples, based on the configuration, the device may identify one or more strongest samples among the samples included in the time-domain channel measurement 401. As used herein, a strongest sample in a channel measurement means a sample having a higher value (e.g., received power value) than one or more other samples in the channel measurement. In some examples, the strongest sample may have a higher value than a threshold.

[0108] As illustrated in the example of FIG. 4, the device may determine that a sample 406 is a strongest sample among the samples included in the time-domain channel measurement 401 (e.g., may determine that the sample 406 has a received power that is higher than a respective received power of one or more other samples included in the timedomain channel measurement 401). In such an example, to determine one or more samples to select for inference input corresponding to one or more ML models, the device may apply a variance 402 (e.g., <5lz) to the sample 406. For example, the device may select a subset 418 of samples associated with the sample 406 for the inference input and may apply the variance402 to determine a quantity of samples to select for the subset 418. In other words, the devicemay apply the variance 402 (e.g., a range) to the sample 406 and select the samples within the variance 402 for the subset 418, which may be used as input for one or more AI / ML models.

[0109] Additionally, or alternatively, the device may determine that a sample 408 is another strongest sample among the samples included in the time-domain channel measurement 401. For example, the device may be configured to identify the strongest non- consecutive detected paths (also referred to herein as non-consecutive detected peaks). In such an example, the device may determine that the sample 406 is a first strongest sample (corresponding to a first detected path) and the sample 408 is the next strongest sample, corresponding to the next non-consecutive detected path. In such an example, the device may select a subset 420 of samples associated with the sample 408 for the inference input and may apply a variance 404 (e.g., <52) to determine a quantity of samples to select for the subset 420. In other words, the device may apply the variance 404 (e.g., a range) to the sample 408 and select the samples within the variance 404 for the subset 418, which may be used as input for one or more AI / ML models. In some examples, as illustrated in the example of FIG. 4, the subset 418 and the subset 420 may include consecutive samples. As used herein, a consecutive sample means a sample that is neighboring to (e.g., on either side of) another sample. For example, consecutive samples associated with the sample 406 include (among other samples included in the subset 418) a sample 411-a and a sample 411-b.

[0110] In some examples, a variance applied to a strongest sample may determine a range (e.g., and quantity) of samples selected by the device. For example, based on the variance 404, the subset 420 may include the sample 408, as well as (neighboring) sample 410-a and (neighboring) sample 410-b. In some other examples, the subset 418 and the subset 420 may include non-consecutive samples. In some examples, the criteria may include an indication of one or more ratios of consecutive samples per non-consecutive peaks. In such an example, the device may determine the variance 402 for the sample 406 and / or the variance 404 for the sample 408 based on the one or more ratios.

[0111] In some other examples, the network node (e.g., LMF, AMF, or NWDAF) may define (in one or more configurations) a metric for measuring a similarity between two or more sample measurement instances over a particular measurements time window. In such an example, the network node may trigger the device to report (e.g., filter-out) samples based on the metric. For instance, if the samples are highly correlated, the device may be configured to select less samples, and, if the samples are less correlated or un-correlated, the device may be configured to select more samples. In some examples, the metric may include a dynamic time warping (DTW) distance (or similar metric), a gaussian similarity metric, or a cosinesimilarity metric. In other words, the device may be configured to use an auto- or crosscorrelation among multiple samples within a time measurement window as one criterion for sample selection in a region or multiple regions of the time-domain channel measurement 401. Although the example of FIG. 4 is described in the context of a time-domain channel measurement, it is to be understood that the device may apply the framework for data collection configuration enabling sample selection and range for AI / ML, as described herein, to one or more other types of channel measurements, such as angle-domain measurements and / or phase-domain measurements.

[0112] FIG. 5 illustrates an example signaling diagram 500 to which one or more examples disclosed herein may be applied. The signaling diagram 500 illustrates operations performed, such as within the system of FIG. 1, by the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 in accordance with one or more aspects of the present disclosure. The network node 110, the network node 112, and the network node 114 may be examples of RAN nodes, such as gNBs or TRPs. As illustrated in the signaling diagram 500, the network node 110 may be a serving node for the UE 120 (e.g., may have established an RRC connection with the UE 120), while the network node 112 and the network node 114 may be neighboring nodes (e.g., network nodes located within a proximity of the UE 120, such that the network node 112 and / or the network node 114 may be capable of establishing an RRC connection with the UE 120). In the example of FIG. 5, the core 116 may be an example of an LMF, AMF, or NWDAF. One or more operations performed at the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may be performed in a different order than the example order shown. Additionally, or alternatively, one or more operations performed at the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may be omitted and / or one or more other operations may be added. The signaling diagram 500 may support one or more direct AI / ML positioning use cases, such as Case 1 in which UE-based positioning is performed with one or more UE-side models.

[0113] In some examples, the UE 120, the network node 110, the network node 112, the network node 114, and the core 116 may support a framework for data collection configuration enabling sample selection and range for AI / ML as described herein. For example, the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may support a framework for AI / ML sample selection for one or more UE- side ML models (e.g., one or more ML models executed at the UE 120). In some examples, the UE 120 may use the one or more UE-side ML models for determining a position of theUE 120 (e.g., a target device). For example, the one or more UE-side ML models may be trained to output positioning information associated with the UE 120. That is, in some examples, the UE 120 (e.g., and the network node 110, the network node 112, and a network node 114, and the core 116) may support a framework for AI / ML sample selection for direct AI / ML positioning. In other words, one or more entities illustrated in the example of FIG. 5 may be configured to support UE-based positioning with one or more UE-side models.

[0114] In some examples, at 502, the UE 120 may perform data collection for inference and / or training of one or more ML models configured at the UE 120. The one or more ML models may be used to support one or more network functionalities, such as positioning and / or beam management.

[0115] In some examples, at 504, the UE may perform an LPP capability transfer procedure with the network (e.g., one or more of the network node 110, the network node 112, the network node 114, or the core 116). For example, the one or more ML models may be trained for positioning of a target device. In such an example, the UE 120 may use the one or more ML models to support (e.g., improve a performance of) a positioning procedure, such as an LPP procedure. Accordingly, the UE 120 may initiate an LPP session with the network and may perform the LPP capability transfer in accordance with the LPP session (or to otherwise support the LPP session). That is, the UE 120 may perform a capability transfer procedure for the LPP procedure. The LPP capability transfer may enable the network (e.g., an LMF) to provide assistance data to the UE 120 (e.g., as part of the LPP procedure). The LPP capability procedure may additionally, or alternatively, enable the UE 120 to request assistance data from the LMF (e.g., as part the LPP procedure).

[0116] In some examples, at 506, the network may perform one or more NRPPa procedures. For example, to support the LPP session or other signaling associated with positioning of the UE 120, the network node 110 may communicate with the core 116 in accordance with an NRPPa procedure (or another suitable positioning procedure that support signaling between the network nodes and core network functions). In some examples, the network node 112 and / or the network node 114 may communicate with the LMF in accordance with (respective) NRPPa procedures.

[0117] In some examples, at 508, the core 116 may determine a set of configurations for selection of channel measurement samples for inference (or training) input corresponding to one or more ML models (e.g., at the UE 120). In some such examples, a configuration (e.g., each configuration) in the determined set may include one or more criteria for channel measurement sample selection. That is, a configuration (e.g., each configuration) of the setmay include one or more parameters for the selection of channel measurement samples for inference (or training) input for the one or more ML models. The core 116 may be unaware of channel conditions at the UE 120. Accordingly, the core 116 may determine multiple configurations (e.g., options) for channel measurement sample selection at the UE 120, such that the UE 120 may select a suitable configuration based on characteristics of the channel as observed by the UE 120. In some examples, the core 116 may determine one or more parameters of a configuration based on one or more positioning accuracy constraints. In other words, the LMF (or another network entity) may determine criteria for selecting samples (e.g., selecting a subset of Nt samples per each Nt’ sample) to provide efficient AI / ML-based channel inference with sufficient positioning accuracy. The criteria may indicate a granularity for the selection of channel measurement samples, one or more ranges (e.g., variances) for the selection of channel measurement samples, and / or one or more similarity metrics for the selection of channel measurement samples.

[0118] In some examples in which the criteria (e.g., one or more parameters in a configuration) indicate a granularity for the selection of channel measurement samples, the criteria may include consecutive samples, non-consecutive samples, and / or a fraction of consecutive samples per non-consecutive detected peak to be selected by the UE 120. In other words, the criteria may indicate at least one of the following: a range of samples, a starting point associated with the range of samples, one or more types of offsets associated with one or more samples, a quantity of samples, a quantity of consecutive samples, a quantity of non-consecutive samples, or a ratio of consecutive samples per non-consecutive sample. The starting point may include a starting time, a starting angle, or a starting phase. The range of samples may correspond to a range of Nt samples from which the UE 120 may select Nt’ channel measurement samples. In the example of FIG. 2, the range of samples may correspond to samples 6 through 261. Additionally, in some such examples, the starting time associated with a range of samples may be a time from which the UE 120 is to obtain the range of samples. In other words, the starting time of a list of Nt samples may be based on (e.g., defined by) reference timing information, such as an uplink (UL) relative time of arrival (RTOA) reference time. In the example of FIG. 2, the starting time may correspond to timing information associated with sample 6.

[0119] In some examples, a configuration (e.g., the criteria included in a configuration) may indicate for the UE 120 to select channel measurement samples based on a timing of the first detected path, a fixed offset, and / or a configured offset. In such an example, the UE 120 may report the timing, for example, in addition to a respective power level corresponding tothe offsets and / or detected paths. For example, a configuration may indicate one or more types of offsets associated to be reported by the UE 120 for one or more samples. In such an example, the one or more types of offsets may include an offset of a first detected path. That is, the configuration may indicate for the UE 120 to select channel measurement samples according to (and report) an offset of the first path detected by the UE 120. As used herein, the first detected path of a channel measurement obtained by a UE means a sample (e.g., tap) of the channel measurement with the highest value among samples included in the channel measurement. For example, the channel measurement may include a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the first detected path may include the sample with the highest received power among samples included in the time-domain channel measurement obtained at the UE 120. That is, in the example of FIG. 4, the first detected path may correspond to sample 406. That is, in the example of FIG. 4, the received power of sample 406 is greater than a respective received power of other samples (e.g., all other samples) included in time-domain channel measurement 401. Accordingly, the UE 120 may determine that the sample 406 is the first detected path. In such an example, the UE 120 may select one or more samples based on the first detected path. For example, the first detected path may be associated with an offset and the UE 120 may select one or more samples (e.g., a subset of Nt samples) based on the offset associated with the first detected path. In some examples, the offset is relative to a reference point.

[0120] In some other examples, the configuration may indicate a fixed or configured offset. That is, in some examples, the configuration may indicate for the UE 120 to select channel measurement samples according to (and report) a fixed or configured offset relative to a reference point. In such an example, the UE 120 may select one or more samples (e.g., a subset of Nt samples) based on the fixed or configured offset.

[0121] In some examples, the one or more types of offset may indicate for the UE 120 to select channel measurement samples according to (and report) one or multiple offsets associated with one or multiple samples (e.g., one or multiple paths) relative to a reference point. In other words, the configuration may indicate for the UE 120 to select channel measurement samples according to a respective offset of one or more paths, in which each offset is relative to the reference point. For example, the channel measurement may be an example of a time-domain channel measurement illustrate by and described with reference to FIG. 4. In such an example, the UE 120 may select the subset 418 according to the offset 412 of the sample 406 (the first detected path) and the subset 420 according to the offset 414 ofthe sample 408 (the next non-consecutive detected path), in which the offset 412 and the offset 414 are relative to the same reference point. In some other examples, the one or more types of offset may indicate for the UE 120 to select channel measurement samples according to (and report) a first offset of the first detected path relative to a reference point and one or more other offsets of one or more other detected paths relative to the first offset. In the example of FIG. 4, the UE 120 may select the subset 418 according to the offset 412 of the sample 406 (the first detected path) and the subset 420 according to the offset 416 of the sample 408 (the next non-consecutive detected path), in which the offset 412 is relative to a reference point and the offset 416 is relative to the first detected path.

[0122] In some examples, the configuration may indicate a quantity of samples to be reported by the UE 120 (e.g., a threshold quantity of samples to be selected and reported by the UE 120). For example, the configuration may indicate a threshold quantity of samples to be reported by the UE 120. Additionally, or alternatively, the configuration may indicate a quantity of consecutive samples to be selected by the UE 120. For example, the configuration may indicate a threshold quantity of consecutive samples to be selected by the UE 120. The quantity of consecutive samples may be a quantity of consecutive samples from which the UE 120 may select channel measurement samples to report to the network. In other words, the quantity of consecutive samples may indicate a value of Nt. In some examples, the configuration may indicate a quantity of non-consecutive samples to be selected by the UE 120 (e.g., a value of Nt’). Additionally, or alternatively, the configuration may indicate a ratio of consecutive samples per non-consecutive sample to be selected by the UE 120 (e.g., a value of Nt / Nt’). For example, the channel measurement may be an example of a timedomain channel measurement illustrated by and described with reference to FIG. 2. In such an example, the quantity of consecutive samples may be 256 and the quantity of non- consecutive samples may be 3. In some other examples, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the configuration may indicate multiple ratios of consecutive samples per non-consecutive sample. For example, the configuration may indicate a first ratio of consecutive samples per non-consecutive sample for the first detected path (e.g., the sample 406) and a second ratio of consecutive samples per non-consecutive sample for the next (non-consecutive) detected path (e.g., the sample 408). In such an example, the first ratio may indicate a quantity of samples to be included in the subset 418 and the second ratio may indicate a quantity of samples to be included in the subset 420.

[0123] In some examples, the configuration may indicate one or more ranges (e.g., variances) for the selection of channel measurement samples. For example, the configuration may indicate a degree to which samples within a subset may deviate (e.g., spread out) from a detected path. In other words, the configuration may indicate (e.g., via one or more variances) a threshold distance (e.g., temporal distance, phase distance, angle distance) that selected samples may be from a detected path. For example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the variance 402 (e.g., a first range) may indicate a threshold (e.g., maximum) distance from which samples included in the subset 418 may deviate from the sample 406 (e.g., the first detected path). Additionally, the variance 404 (e.g., a second range) may indicate a threshold (e.g., maximum) distance from which samples included in the subset 420 may deviate from the sample 408 (e.g., the next non-consecutive detected path). In some such examples, the configuration (e.g., criteria included in the configuration) may indicate for the UE 120 to determine a variance associated with a sample based on an offset of the sample. For example, as illustrated in FIG. 4, the UE 120 may select more samples for the subset 418 and fewer samples for the subset 420 based on the offset 412 being shorter than the offset 416.

[0124] In some examples, the configuration may indicate one or more similarity metrics for the selection of channel measurement samples. For example, the configuration may indicate one or more types of similarity metric for the UE 120 to use for the selection of channel measurement samples. In some examples, the one or more types of similarity metrics may include a DTW distance (or similar metric), a gaussian similarity metric, or a cosine similarity metric. Additionally, or alternatively, the one or more similarity metrics may indicate for the UE 120 to perform an auto-correlation or a cross-correlation among multiple samples within a time measurement window for sample selection in a region or multiple regions of the channel measurement. For example, the configuration may indicate (e.g., via one or more similarity metrics) for the UE 120 to determine, for a region of the channel measurement, a quantity of samples to include in a subset associated with the region based on a level of similarity among samples within the region. For example, a region of a channel measurement (e.g., a time-domain channel measurement) may include a relatively large quantity of samples corresponding to reflections. In such an example, the region may be associated with a lower level of similarity than another region of the channel measurement that includes a relatively small quantity of samples corresponding to reflections. Thus, by selecting samples based on one or more similarity metrics, the UE 120 may select moresamples in regions with significant multipath reflections and less samples in regions with insignificant multipath reflections, which may lead to improved accuracy of predictions based on the selected samples.

[0125] At 510, the UE 120 may receive a first message including first information indicative of a set of one or more configurations. That is, at 510, the network may configure (e.g., and activate) one or more configurations for channel measurement sample selection at the UE 120. In other words, the network (e.g., LMF) may configure the UE 120 (e.g., and one or more other UEs) with network determined criteria / configuration for sample collection purposes. The set of one or more configurations (e.g., each configuration included in the set) may include one or more parameters for selection of channel measurement samples for inference (or training) input corresponding to one or more ML models (e.g., one or more ML models at the UE 120). In some examples, such as examples in which the UE 120 uses the one or more ML models for positioning, the UE 120 may receive the first message in accordance with the LPP protocol. In other words, the first message may be an example of an LPP message.

[0126] In some examples, the network may use multiple messages to configure and activate configurations at the UE. In some such examples, after receiving the first message, the UE 120 may receive another message from the core network 116, which may include information indicating activation of a subset of the set of configurations. For example, the core 116 may use the first message to configure the UE 120 with the set of configurations and may use one or more other message to activate a subset of the set of configurations. In some examples, the activated subset is based on one or more capabilities of the UE 120 and / or one or more network constraints, such as one or more beam management constraints or positioning constraints.

[0127] In some examples, at 512, the UE 120 may receive one or more reference signals (e.g., PRSs) from the network node 110, the network node 112, and / or the network node 114. For example, the UE 120 may receive one or more reference signals in accordance with (or to otherwise support) the LPP procedure. In some examples, the UE 120 may obtain one or more received power measurements of the one or more PRSs. For example, the UE 120 may perform one or more received power measurements of the one or more PRSs to obtain a channel measurement, which may include multiple channel measurement samples (e.g., taps). In some examples, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIGs. 2-4. In some otherexamples, the channel measurement may be an example of a phase-domain channel measurement or an angle-domain channel measurement.

[0128] At 514, the UE 120 may select a configuration from the set of configurations (e.g., received at 510) based on one or more characteristics of a wireless communication channel observed at the UE 120. In other words, the UE 120 may evaluate the options provided by the network to select one suitable configuration. The one or more characteristics may include one or more radio propagation conditions, one or more channel conditions, one or more capabilities of a target device, and / or one or more positioning accuracy constraints. In other words, the evaluation by the UE 120 may be based on radio propagation conditions, a PDP of one or more CSI measurements at UE 120, a DP of one or more CSI measurements at the UE 120, one or more UE capabilities (e.g., transmit capabilities of the UE, receive capabilities of the UE, one or more capabilities indicated via the LPP capability transfer, a capability of the UE to perform the evaluation, a PRS processing capability of the UE, a maximum quantity of frequency layers supported at the UE, a bandwidth supported by the UE, an antenna configuration at the UE), a network load, and / or a level of positioning accuracy, among other types of channel characteristics. For example, the channel measurement may be a example of a OR, PDP, or DP, and the UE 120 may select a configuration for sampling the CIR, PDP, or DP based on characteristics of the CIR, PDP, or DP itself and / or one or more other factors, such as one or more previous (e.g., historic) channel measurements, one or more capabilities of the UE 120, a network load, and / or a level of positioning accuracy.

[0129] In some examples, at 516, the UE 120 may select (e.g., collect) one or more samples in accordance with the configuration selected at 514. In other words, the UE 120 may collect samples and / or sub-samples (e.g., may select Nt consecutive samples per each Nt’ non-consecutive samples) based on the selected configuration. In some examples, the UE 120 may select the one or more samples from among the samples obtained based on the PRS transmissions at 512. The one or more channel measurement samples may be associated with one or more offsets relative to a reference point.

[0130] The one or more channel measurement samples selected at the UE 120 may include one or more detected paths. A detected path may also be referred to herein as a strongest sample. For example, a sample of a channel measurement corresponding to a detected path may have a greater value (e.g., greater received power) than a respective value of one or more other samples included in the channel measurement. Thus, the one or more channel measurement samples may include one or more strongest samples (e.g., one or more samples having a greater value than one or more other samples obtained based on the PRStransmissions at 512). For example, the channel measurement may be an example of a timedomain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the strongest (non-consecutive) samples may include the sample 406 and the sample 408.

[0131] In some examples, the one or more strongest samples may include the first detected path and one or more other detected paths. In such an example, the first strongest sample may correspond to the first detected path and the second strongest (non-consecutive) sample may correspond to the next detected path. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the first strongest sample may include the sample 406 and the second strongest (non-consecutive) sample may include the sample 408. The sample 406 may be the first strongest sample (e.g., the first detected path) based on the sample 406 having a greater received power value than other samples (e.g., all other samples) included in the time-domain channel measurement 401. Additionally, the sample 408 may be the second strongest (non-consecutive) sample based on the sample 408 having a greater value than other remaining samples of the time-domain channel measurement 401, in which the other remaining samples exclude the sample 406 and samples that are consecutive with (e.g., neighboring to) the sample 406. That is, the sample 408 may be the second strongest (non-consecutive) sample based on the sample 408 having a greater value than samples other than the sample 406, the sample 411-a, and the sample 411-b.

[0132] Additionally, in some examples, the one or more channel measurement samples selected by the UE 120 may include one or more other samples associated with the one or more strongest samples (e.g., the one or more detected paths). For example, the one or more channel measurement samples may include the first strongest sample (e.g., first detected path) and one or more channel measurement associated with the first strongest sample. Additionally, in some examples, the one or more channel measurement samples may include one or more next strongest samples (e.g., one or more next detected paths) and one or more channel measurements associated with the one or more next strongest samples. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the one or more channel measurement samples selected by the UE 120 may include the subset 418 (e.g., the first strongest sample and one or more (consecutive) samples associated with the first strongest sample). Additionally, the one or more channel measurements may includethe subset 420 (e.g., the next (non-consecutive) strongest sample and one or more (consecutive) samples associated with the next strongest sample).

[0133] In some examples, in accordance with the one or more parameters included in the selected configuration, the UE 120 may select the one or more samples based on one or more time offsets of the one or more strongest samples relative to a reference point and based on one or more variances applied across the one or more strongest samples. For example, the core 116 may define (e.g., in the configuration) criteria for selecting samples based on one or more time offsets of the one or more strongest samples relative to a reference point, and a respective variance applied across those strongest samples. In other words, the configuration may indicate for the UE 120 to determine a range (e.g., variance) associated with a sample (e.g., a strongest sample) based on an offset of the sample. In some examples, the criteria may indicate for the UE 120 to decrease the variance as the offset increases. For example, as the offset of a strongest sample increases, the variance for selecting samples associated with the strongest sample narrows, leading to a more focused selection. Conversely, as the offset of a strongest sample decreases, the variance for selecting samples associated with the strongest sample broadens, leading to a wider selection of samples.

[0134] For example, the UE 120 may select a set of consecutive channel measurement samples from the channel measurement samples obtained by the UE 120 based on the PRS transmissions at 512. The UE 120 may then identify a first (strongest) channel measurement sample of the set of consecutive channel measurement samples. That is, the UE 120 may identify a sample from among the set of consecutive channel measurement samples, which corresponds to the first detected path. The first channel measurement sample may be associated with a first offset relative to a reference point. The UE 120 may then select a first subset of consecutive channel measurement samples based on the first offset. In such an example, the one or more channel measurement samples include at least the first subset of consecutive channel measurement samples. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the set of consecutive channel measurement samples may include the selected and non-selected samples of the time-domain channel measurement 401. The first channel measurement sample may include the sample 406 having the offset 412. Accordingly, the UE 120 may select samples to include in the subset 418 based on the offset 412.

[0135] The UE 120 may then identify a second (next strongest) channel measurement sample of the set of consecutive channel measurement samples. That is, the UE 120 mayidentify the next detected path (the next strongest sample). The second channel measurement sample may be associated with a second offset relative to the reference point. The UE 120 may then select a second subset of consecutive channel measurement samples based on the second offset. In some examples, the second offset is greater than the first offset. In some such examples, a first quantity of samples included in the first subset may be greater than a second quantity of samples included in the second subset based on the second offset being greater than the first offset. That is, in accordance with the configuration, the UE 120 may apply a larger variance (e.g., larger range) to the first channel measurement sample and a smaller variance (e.g., smaller range) to the second channel measurement sample based on the first channel measurement sample having a smaller offset than the second channel measurement sample. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the second (next strongest) channel measurement sample may include the sample 408 having the offset 414. Accordingly, the UE 120 may select samples to include in the subset 420 based on the offset 414. Additionally, the UE 120 may apply a larger variance to the sample 406 and a smaller variance to the sample 408 based on the offset 412 being smaller than the offset 414. Thus, the subset 418 may include less samples than the subset 420.

[0136] In some other examples, the UE 120 may select the one or more samples based on a level of similarity associated with the samples. For example, the channel measurement samples obtained by the UE 120 may be associated with one or more measurement time windows (e.g., regions) and a quantity of channel measurement samples included in the one or more channel measurement samples may be based on a level of similarity associated with channel measurement samples included in the one or more measurement time windows. For example, the core 116 may define (e.g., in the configuration) a metric for measuring similarity between two or more sample measurement instances over a particular measurement time window. In such an example, the configuration may trigger the UE 120 (or the UE 120 may be otherwise triggered) to select samples based on a level of similarity determined for the measurement time window based on the metric. In other words, the UE 120 may be triggered to filter-out (and report) samples (e.g., Nt consecutive samples per non-consecutive sample) based on a metric for measuring similarity between two or more sample measurement instances over a particular measurement time window. The metric may include a DTW distance (or similar metric), gaussian similarity, or cosine similarity, among other examples of metrics that may be used to determine a level of similarity between two or moresamples. In other words, the network may trigger the UE 120 to use an auto-correlation or cross-correlation among multiple samples as one criterion to select suitable samples within a region (e.g., each measurement time window) of the time-domain channel measurement. In some such examples, if the samples in a region have a relatively high level of similarity (e.g., are highly correlated), the UE 120 may be configured to select less samples. Conversely, if samples in a region have a relatively low level of similarity (e.g., are less correlated or are uncorrelated), the UE 120 may be configured to select more samples.

[0137] For example, the UE 120 may select a set of consecutive channel measurement samples from the channel measurement samples obtained by the UE 120 based on the PRS transmissions at 512. The UE 120 may identify a first portion of the channel measurement samples obtained by the UE 120. The first portion may be associated with a first time-domain measurement window. The UE 120 may then determine a first level of similarity associated with the first portion based on one or more similarity metrics (e.g., the similarity metric indicated via the configuration). The UE 120 may then select a first subset of channel measurement samples from the first portion based on the first level of similarity. In some examples, the UE 120 may identify a second portion of the channel measurement samples obtained by the UE 120. The second portion may be associated with a second time-domain measurement window. The UE 120 may determine a second level of similarity associated with the second portion based on the one or more similarity metrics. The UE 120 may select a second subset of channel measurement samples from the second portion based on the second level of similarity. In some examples, the samples within the second region are more correlated than the samples within the first region. In such examples, the second level of similarity is greater than the first level of similarity and, as such, a first quantity of samples included in the first subset is greater than a second quantity of samples included in the second subset. That is, the UE 120 may select more samples for the first subset and less samples for the second subset based on the second level of similarity being greater than the first level of similarity.

[0138] At 518, the UE 120 may transmit a second message to the core 116. The second message may include at least second information indicative of the configuration. In some examples, the second message may also include third information indicative of the one or more channel measurement samples (e.g., indicative of a received power value of the one or more channel measurement samples) and the one or more offsets. That is, at 518, the UE 120 may transmit a sample information report for the collected samples to the core 116 and the report may indicate the selected configuration for sample selection. In some examples, basedon the selected configuration, the report may also indicate the collected samples, the offsets (e.g., timing offsets) corresponding to the collected samples, and / or the reference point (e.g., if requested). As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the UE 120 may report, to the network, a configuration used to obtain (e.g., collect) the selected samples, the offset 412, the offset 414 (or the offset 416), the reference point (e.g., reference time) from which the offset 412 and the offset 414 are measured, and / or a received power associated with one or more of the selected samples.

[0139] In some examples, the third information is indicative of the one or more channel measurement samples by including at least one of the following: an indication of a power level associated with the channel measurement sample and the quantity of channel measurement samples, or an indication of a respective power level associated with each channel measurement sample of the one or more channel measurement samples. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the UE 120 may report the selected samples by reporting a respective value for each of the selected samples. Alternatively, to reduce reporting overhead, the UE 120 may report the selected samples by reporting a first received power level associated with the sample 406 and a first quantity of samples included in the subset 418, as well as a second received power level associated with the sample 408 and a second quantity of samples included in the subset 420.

[0140] In some examples, the third information is indicative of the quantity of channel measurement samples by including an indication of a range associated with the offset. For example, the UE 120 may report (e.g., preserve) the strongest power samples (e.g., a set of “h” samples) and may dynamically adjust (and report to the network) a respective range applied across each strongest sample in the set of “h” strongest samples. In some examples, by utilizing the reported ranges (e.g., and timing information), the network may identify the contributions of each multi-path component (e.g., each sampled within the range) and may thus reconstruct the channel profile (e.g., as part of data collection) based on the reported ranges. As an illustrative example, the channel measurement may be an example of a timedomain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the UE 120 may report the selected samples by reporting a first received power level (and offset) associated with the sample 406 and a first range (e.g., the variance 402) associated with the subset 418, as well as a second received power level (and offset)associated with the sample 408 and a second range (e.g., the variance 404) associated with the subset 420. In such an example, the core 116 may reconstruct the channel profile based on the reported information.

[0141] In some examples, the variance and offset information may be considered to indicate the starting time of the list of Nt samples. That is, by mapping the variance and the offset, the network (or the UE 120) may determine the starting point of the list of Nt samples (e.g., the starting point of Nt consecutive samples from which Nt’ samples are selected and / or a starting point of a list of Nt samples associated with one of the Nt’ samples). For example, if a reported variance is relatively large and comparable to the corresponding offset, the starting point of the list of Nt samples associated with the offset may be defined based on the variance value. In some other examples, if the variance is relatively small compared to the corresponding offset, the starting point of Nt samples associated with the offset may be defined based on the offset value. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the variance 402 may be relatively large and comparable to the offset 412. Accordingly, the network (or the UE 120) may determine the starting point for the list of Nt samples included in the subset 418 based on the value of the variance 402. Additionally, the variance 404 may be relatively small compared to the offset 414. Accordingly, the network (or the UE 120) may determine the starting point for the list of Nt samples included in the subset 420 based on the value of the offset 414 (or the offset 416). In some examples, the starting time of a list of Nt samples may be based on (e.g., defined by) reference timing information, such as an uplink (UL) relative time of arrival (RTOA) reference time.

[0142] In some examples, the third information may be indicative of the one or more offsets by include an indication of a respective offset associated with the first channel measurement sample and the second channel measurement sample, in which the respective received power offset for the first channel measurement and the second channel measurement are relative to the reference point. As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the UE 120 may report the offset 412 for the sample 406 and the offset 414 for the sample 408. Alternatively, in some examples, the third information may be indicative of the one or more offsets by including an indication of the first offset and an indication of a difference between the first offset and the second offset.That is, in the example of FIG. 4, the UE may report the offset 412 for the sample 406 and the offset 416 for the sample 408.

[0143] Time information, phase information, and / or angle information associated with a sample may be indicative of a direction (e.g., propagation path) associated with the sample. Accordingly, in some examples, the UE 120 may report time-domain channel information, phase-domain channel information, and / or angle-domain channel information. That is, in some examples, the UE 120 may report time-domain channel information to the network. In such examples, the offsets may include time offsets relative to a reference time (or another time offset). That is, in some examples, the reference point may be a reference time. Additionally, or alternatively, the UE 120 may report phase-domain information to the network. In such an example, the offsets may include phase offsets relative to a reference phase (or another phase offset). That is, in some examples, the reference point may be a reference phase (o). Additionally, or alternatively, the UE 120 may report angle-domain information to the network. In such an example, the offsets may include angle offsets relative to a reference angle (or another angle offset). That is, in some examples, the reference point may be a reference angle (0O).

[0144] Although the example of FIG. 5 illustrates the signaling at 510 and 518 between the UE 120 and the core 116, it is to be understood that the signaling at 510 and 518 may additionally, or alternatively, occur between the UE 120 and the network node 110. In some examples in which the signaling at 510 and 510 occurs between the UE 120 and the network node 110, the signaling at 510 and 518 may occur via an RRC protocol or a MAC protocol.

[0145] FIG. 6 illustrates an example signaling diagram 600 to which one or more examples disclosed herein may be applied. The signaling diagram 600 illustrates operations performed, such as within the system of FIG. 1, by the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 in accordance with one or more aspects of the present disclosure. The signaling diagram 600 may implement one or more aspects of the signaling diagram 500. For example, in the signaling diagram 500, the UE 120 may perform one or more operations to support sample selection based on downlink PRSs. In the signaling diagram 600, the network node 110 may perform the same (or similar) one or more operations to support sample selection based on uplink sounding reference signals (SRSs). In other words, one or more operations illustrated by and described with reference to FIG. 5 for directed AI / ML positioning may be applied to the signaling diagram 600 for assisted-AI / ML positioning that considers uplink SRSs. The signaling diagram 600 may support one or more use cases for direct AI / ML positioning, such as Case 3b in which NG-RAN node assisted positioning is performed with one or more LMF-side models. The signaling diagram 600 may additionally, or alternatively, support one or more use cases for AI / ML assisted positioning, such as Case 3a in which NG-RAN node assisted positioning is performed with one or more gNB-side models.

[0146] In the example of FIG. 6, the network node 110, the network node 112, and the network node 114 may be examples of RAN nodes, such as gNBs or TRPs. As illustrated in the signaling diagram 600, the network node 110 may be a serving node for the UE 120 (e.g., may have established an RRC connection with the UE 120), while the network node 112 and the network node 114 may be neighboring nodes (e.g., network nodes located within a proximity of the UE 120, such that the network node 112 and / or the network node 114 may be capable of establishing an RRC connection with the UE 120). In the example of FIG. 6, the core 116 may be an example of an LMF, AMF, or NWDAF. One or more operations performed at the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may be performed in a different order than the example order shown. Additionally, or alternatively, one or more operations performed at the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may be omitted and / or one or more other operations may be added.

[0147] In some examples, the UE 120, the network node 110, the network node 112, the network node 114, and the core 116 may support a framework for data collection configuration enabling sample selection and range for AI / ML as described herein. For example, the UE 120, the network node 110, the network node 112, and a network node 114, and the core 116 may support a framework for AI / ML sample selection for one or more network-side ML models (e.g., one or more ML models executed at the network node 110 or the core network 116). In some examples, the network may use the one or more network-side ML models (e.g., gNB-side or LMF-side modes) for determining a position of the UE 120 (e.g., a target device). For example, the one or more network-side ML models may be trained to output positioning information associated with the UE 120. That is, in some examples, the network node 110 (e.g., and the UE 120, the network node 112, and a network node 114, and the core 116) may support a framework for AI / ML sample selection for assisted AI / ML positioning. In other words, one or more entities illustrated in the example of FIG.6 may be configured to support UE-based positioning with one or more network-side models. In some examples in which the one or more ML models are executed at the network node 110 (e.g., a gNB), the positioning information output by an ML model may include intermediate positioning information (e.g., key performance indicators (KPIs)), which the network node110 may provide to the core 116 for determining a position of the target device. In some other examples, the positioning information output by an ML model may include may indicate (e.g., predict) the position of the target device.

[0148] In some examples, at 602, the network node 110 may perform data collection for inference and / or training of one or more ML models configured at the network node 110. The one or more ML models may be used to support one or more network functionalities, such as positioning and / or beam management.

[0149] Additionally, or alternatively, at 604, the core 116 may perform data collection for inference and / or training of one or more ML models configured at the core 116. The one or more ML models may be used to support one or more network functionalities, such as positioning and / or beam management.

[0150] In some examples, at 606, the network may perform one or more NRPPa procedures. For example, to support signaling associated with positioning of the UE 120, the network node 110 may communicate with the core 116 in accordance with an NRPPa procedure (or another suitable positioning procedure that support signaling between the network nodes and core network functions). In some examples, the network node 112 and / or the network node 114 may communicate with the LMF in accordance with (respective) NRPPa procedures.

[0151] In some examples, at 608, the core 116 may determine a set of configurations for selection of channel measurement samples for inference (or training) input corresponding to one or more ML models (e.g., at the network node 110). In some such examples, a configuration (e.g., each configuration) in the determined set may include one or more criteria for channel measurement sample selection. That is, a configuration (e.g., each configuration) of the set may include one or more parameters for the selection of channel measurement samples for inference (or training) input for the one or more ML models. The core 116 may be unaware of channel conditions at the network node 110. Accordingly, the core 116 may determine multiple configurations (e.g., options) for channel measurement sample selection at the network node 110, such that the network node 110 may select a suitable configuration based on characteristics of the channel as observed by the network node 110. In some examples, the core 116 may determine one or more parameters of a configuration based on one or more positioning accuracy constraints. In other words, the LMF (or another network entity) may determine criteria for selecting samples (e.g., selecting a subset of Nt samples per each Nt’ sample) to provide efficient AI / ML-based channel inference with sufficient positioning accuracy. The criteria may be an example of criteriadescribed with reference to FIG. 5. For example, the criteria may indicate a granularity for the selection of channel measurement samples, one or more ranges (e.g., variances) for the selection of channel measurement samples, and / or one or more similarity metrics for the selection of channel measurement samples.

[0152] At 610, the network node 110 may receive a first message including first information indicative of a set of one or more configurations. That is, at 610, the core 116 may configure (e.g., and activate) one or more configurations for channel measurement sample selection at the network node 110. For example, the EMF may configure the network node 110 (e.g., and one or more other network nodes, such as the network node 112 and / or the network node 114) with network determined criteria / configurations for sample collection purposes. The set of one or more configurations may be an example of a set of configurations illustrated by and described with reference to FIG. 5. For example, the set of one or more configurations (e.g., each configuration included in the set) include one or more parameters for selection of channel measurement samples for inference (or training) input corresponding to one or more ML models (e.g., one or more ML models at the network node 110). In some examples, such as examples in which the network node 110 uses the one or more ML models for positioning, the network node 110 may receive the first message in accordance with the NRPPa procedure. In other words, the first message may be an example of an NRPPa message. In some examples, the core 116 may also configure the network node 112 and / or the network node 114 with the set of one or more configurations. In some such examples, the core 116 may transmit the first message to the network node 112 and / or the network node 114 (e.g., in accordance with respective NRPPa procedures).

[0153] In some examples, the core 116 may use multiple messages to configure and activate configurations at the network node 110. In some such examples, after receiving the first message, the network node 110 may receive another message from the core network 116, which may include information indicating activation of a subset of the set of configurations. For example, the core 116 may use the first message to configure the network node 110 with the set of configurations and may use one or more other message to activate a subset of the set of configurations. In some examples, the activated subset may be based on one or more capabilities of the network node 10 (or the UE 120) and / or one or more network constraints, such as one or more beam management constraints or positioning constraints.

[0154] In some examples, the UE 120 may transmit one or more reference signals (e.g.,SRSs) to the network node 110, the network node 112, and / or the network node 114. For example, the UE 120 may transmit one or more reference signals in accordance with (or tootherwise support) one or more UE positioning procedures. In some such examples, at 611, the network node 110 (e.g., the serving node) may configure / activate SRS transmissions at the UE 120. For example, the network node 110 may configure the UE 120 to transmit one or more SRSs to the network node 110, the network node 112 and / or the network node 114.

[0155] In some such examples, at 612, the network node 110 (e.g., and one or both of the neighboring nodes) may receive SRS transmissions from the UE 120. In some such examples, the network node 110 may obtain one or more received power measurements of the one or more SRSs. For example, the network node 110 may perform one or more received power measurements of the one or more SRSs to obtain a channel measurement, which may include multiple channel measurement samples (e.g., taps). In some examples, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIGs. 2-4. In some other examples, the channel measurement may be an example of a phase-domain channel measurement or an angle-domain channel measurement.

[0156] At 614, the network node 110 may select a configuration from the set of configurations (e.g., received at 610) based on one or more characteristics of a wireless communication channel observed at the network node 110. In other words, the network node 110 may evaluate the options provided by the core 116 to select one suitable configuration. The one or more characteristics may include one or more radio propagation conditions, one or more channel conditions, one or more capabilities of a target device (e.g., the UE 120), and / or one or more positioning accuracy constraints. For example, the evaluation by the network node 110 may be based on radio propagation conditions, a PDP of one or more CSI measurements at UE 120, a DP of one or more CSI measurements at the UE 120, one or more UE capabilities (e.g., transmit capabilities of the UE, receive capabilities of the UE, one or more capabilities indicated via an LPP capability transfer, a capability of the network node 110 to perform the evaluation), a network load, and / or a level of positioning accuracy, among other types of channel characteristics. For example, the channel measurement may be an example of a CIR, PDP, or DP, and the network node 110 may select a configuration for sampling the CIR, PDP, or DP based on characteristics of the CIR, PDP, or DP itself and / or one or more other factors, such as one or more previous (e.g., historic) channel measurements, one or more capabilities of the UE 120, a network load, and / or a level of positioning accuracy.

[0157] In some examples, at 616, the network node 110 may select (e.g., collect) one or more samples in accordance with the configuration selected at 614. In other words, thenetwork node 110 may collect samples and / or sub-samples (e.g., may select Nt consecutive samples per each Nt’ non-consecutive sample) based on the selected configuration. In some examples, the network node 110 may select the one or more samples from among the samples obtained based on the SRS transmissions at 612. The one or more channel measurement samples may be associated with one or more offsets relative to a reference point.

[0158] At 618, the network node 110 may transmit a second message to the core 116. The second message may include second information indicative of the configuration. In some examples, the second message may also include third information indicative of the one or more channel measurement samples (e.g., indicative of a received power value of the one or more channel measurement samples) and the one or more offsets. That is, at 618, the network node may transmit a sample information report for the collected samples to the core 116 and the report may indicate the selected configuration for sample selection. The sample information report may be an example of a sample information report illustrated by and described with reference to FIG. 5. For example, based on the selected configuration, the report may also indicate the collected samples, the timing offsets corresponding to the collected samples, and / or the reference time (e.g., if requested). As an illustrative example, the channel measurement may be an example of a time-domain channel measurement illustrated by and described with reference to FIG. 4. In such an example, the network node 110 may report, to the core 116, a configuration used to obtain (e.g., collect) the selected samples, the offset 412, the offset 414 (or the offset 416), the reference time from which the offset 412 and the offset 414 are measured, and / or a received power associated with one or more of the selected samples.

[0159] In some examples, such as examples in which the core 116 configures the network node 112 and / or the network node 114 with the set of configurations at 610, the core 116 may receive one or more messages from the network node 112 and / or the network node 114, in which the one or more messages are indicative of one or more configurations selected by the network node 112 and / or the network node 114. The one or more configurations may be the same as (or different from) the configuration selected by the network node 110. For example, the channel conditions observed at the network node 112 and / or the network node 114 may be different from the channel conditions observed at the network node 110 and, as such the network node 112 and / or the network node 114 may select a different configuration than the network node 110.

[0160] Additionally, or alternatively, at 620 in response to the sample information report received from the network node 110 at 618, the core 116 may transmit one or more messagesto the network node 112 and / or the network node 114. The one or more messages may be indicative of the configuration selected by the network node 110 at 614. In some such examples, the network node 110, the network node 112, and / or the network node 114 may use the sample configuration for sample selection.

[0161] FIG. 7 illustrates an example flowchart 700 of a method to which one or more examples disclosed herein may be applied. The method may be computer-implemented. The method may be performed by a node, such as a UE or a RAN node illustrated by and described with reference to FIGs. 1-6. In some examples, the node may be an example of an apparatus 10 illustrated by and described with reference to FIG. 9.

[0162] As shown in FIG. 7, the node at block 710 receives, from a core network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more ME models. For example, the node may include the means (e.g., a processor 12, a memory 14, a radio interface 16) for receiving, from the core network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more ML models. The first message may be an example of a first message illustrated by and described with reference to FIGs. 5 and 6. For example, the first message may indicate configurations for sample selection.

[0163] As shown in FIG. 7, the node at block 712 selects a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel. For example, the node may include the means (e.g., a processor 12, a memory 14) for selecting a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel. The configuration may be an example of a configuration illustrated by and described with reference to FIGs. 1- 6. For example, the configuration may indicate for the node to select samples in accordance with an offset and variance and / or a level of similarity associated with one or more samples obtained at the node.

[0164] As shown in FIG. 7, the node at block 714 transmits a second message to the core network node, the second message comprising second information indicative of the configuration. For example, the node may include the means (e.g., a processor 12, a memory 14, a radio interface 16) for transmitting a second message to the core network node, the second message comprising second information indicative of the configuration. The secondmessage may be an example of a second message illustrated by and described with reference to FIGs. 5 and 6. For example, the second message may indicate the selected configuration, as well as one or more channel measurement samples selected in accordance with the configuration, one or more offsets associated with the one or more channel measurement samples, and / or a reference point from which the one or more offsets may be measured.

[0165] FIG. 8 illustrates an example flowchart 800 of a method to which one or more examples disclosed herein may be applied. The method may be computer-implemented. The method may be performed by a network node, such as a RAN node or a core network node illustrated by and described with reference to FIGs. 1-6. In some examples, the network node may be an example of an apparatus 10 illustrated by and described with reference to FIG. 9.

[0166] As shown in FIG. 8, the node at block 810 determines a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more ML models. For example, the network node may include the means (e.g., a processor 12, a memory 14) for determining a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more ML models. The set of configurations may be an example of a set of configurations illustrated by and described with reference to FIGs. 5 and 6. For example, the set of configurations may include one or more configurations for sample selection.

[0167] As shown in FIG. 8, the node at block 812 transmits a first message to a device, the first message comprising first information indicative of the set of configurations. For example, the network node may include the means (e.g., a processor 12, a memory 14, a radio interface 16) for transmitting a first message to a device, the first message comprising first information indicative of the set of configurations. The first message may be an example of a first message illustrated by and described with reference to FIGs. 1-6. For example, the network node may transmit the first message via an LPP procedure or an NRPPa procedure.

[0168] As shown in FIG. 8, the node at block 814 receives, from the device in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations. For example, the network node may include the means (e.g., a processor 12, a memory 14, a radio interface 16) for receiving, from the device in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations. The second message may be an example of a second message illustrated by and described with reference to FIGs. 5 and 6. For example, the second message may indicate the selected configuration, as well as one or more channelmeasurement samples selected in accordance with the configuration, one or more offsets associated with the one or more channel measurement samples, and / or a reference point from which the one or more offsets may be measured.

[0169] FIG. 9 illustrates an example block diagram 900 of an apparatus to which one or more examples disclosed herein may be applied. FIG. 9 shows, by way of example, a block diagram of an apparatus 10. The apparatus 10 comprises, for example, at least one processor 12 and at least one memory 14 storing instructions 15 that, when executed by the at least one processor, cause the apparatus 10 at least to perform the method or methods as disclosed herein, and any of the embodiments thereof. In an example, the at least one memory and the instructions (e.g. a computer program code, software), are configured, with the at least one processor, to cause the apparatus 10 to perform the method or methods as disclosed herein, and any of the embodiments thereof.

[0170] A processor 12 may comprise circuitry, or be constituted as circuitry or circuitries, the circuitry or circuitries being configured to perform phases of methods in accordance with one or more example embodiments described herein. As used in this application, the term “circuitry” may refer to one or more or all of the following: (a) hardware-only circuit implementations, such as implementations in only analog and / or digital circuitry, and (b) combinations of hardware circuits and software, such as, as applicable: (i) a combination of analog and / or digital hardware circuit(s) with software / firmware and (ii) any portions of hardware processor(s) with software (including digital signal processor(s)), software, and memory(ies) that work together to cause an apparatus, such as a user equipment, to perform various functions) and (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessor(s), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation. This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in server, a cellular network device, or other computing or network device.

[0171] The memory 14 may be implemented using any suitable data storage technology. The memory may comprise a database for storing data. The memory 14 may be at least in part external to apparatus 10 but accessible to apparatus 10.

[0172] The instructions 15 may be comprised in a computer readable medium or a non- transitory computer readable medium. A term non-transitory, as used herein, is a limitation of the medium itself (i.e. tangible, not a signal) as opposed to a limitation on data storage persistency (e.g. random access memory, RAM, vs. read only memory, ROM).

[0173] For example, the apparatus 10 may be a terminal device, such as the UE of FIGs. 1 5, and 6. As another example, the apparatus may be comprised in such a terminal device, e.g. as a chipset configured to control the terminal device. The apparatus 10 may be caused or configured to perform at least the method of FIGs. 7 and / or any one or more of the embodiments described.

[0174] As another example, the apparatus 10 is a network node (e.g. a RAN node) or core of FIGs. 1, 5 and 6. In another embodiment, the apparatus is comprised in such a network node, e.g. as a chipset configured to control the network node. The apparatus 10 may be caused or configured to perform at least the method of FIGs. 7 and 8 and / or any one or more of the embodiments described.

[0175] The apparatus may comprise one or more entities of any of protocol layers, such as a MAC entity, an RRC entity, an REC entity, a PDCP entity or a PHY entity. In at least one embodiment, the entity is configured to perform at least the method of Figs. 7 and 8, and / or any one or more of the embodiments described herein.

[0176] The apparatus 10 comprises a radio interface 16. The radio interface 16 may provide the apparatus 10 with communication capabilities. The radio interface 16 may comprise a receiver configured to receive information in accordance with at least one cellular or non-cellular standard. The radio interface 16 may comprise a transmitter configured to transmit information in accordance with at least one cellular or non-cellular standard. The receiver may comprise more than one receiver. The transmitter may comprise more than one transmitter. The radio interface 16 may comprise a transceiver configured to receive and transmit information in accordance with at least one cellular or non-cellular standard. The transceiver may comprise more than one transceiver.

[0177] The apparatus 10 may comprise a user interface 18 comprising, for example, at least one of a keypad, a microphone, a touch display, a display, a speaker, etc. The user interface 18 may be used to control the apparatus by the user. The user interface 18 may be external to the apparatus 10. For example, the apparatus 10 may be connected to another device, such as a computer, either via wireless or wired connection, and the apparatus 10 is controlled by the user via the computer.

[0178] In at least one embodiment, at least some of the processes described herein may be carried out by an apparatus comprising means for carrying out at least some of the described processes. Means for performing methods as disclosed herein may include software and / or hardware components of the apparatus 10. For example, the at least one processor 12, the memory 14, and the computer program code form means for carrying out the method or methods as disclosed herein, and any of the embodiments thereof. As used herein the term “means” is to be construed in singular form, i.e. referring to a single element, or in plural form, i.e. referring to a combination of single elements. Therefore, terminology “means for [performing A, B, C]”, is to be interpreted to cover an apparatus in which there is only one means for performing A, B and C, or where there are separate means for performing A, B and C, or partially or fully overlapping means for performing A, B, C. Further, terminology “means for performing A, means for performing B, means for performing C” is to be interpreted to cover an apparatus in which there is only one means for performing A, B and C, or where there are separate means for performing A, B and C, or partially or fully overlapping means for performing A, B, C.

[0179] Even though the present disclosure has been described above with reference to an example according to the accompanying drawings, it is clear that the present disclosure is not restricted thereto but can be modified in several ways within the scope of the appended claims. Therefore, all words and expressions should be interpreted broadly, and they are intended to illustrate, not to restrict, the embodiment. It will be obvious to a person skilled in the art that, as technology advances, the inventive concept can be implemented in various ways. Further, it is clear to a person skilled in the art that the described embodiments may, but are not required to, be combined with other embodiments in various ways.

Claims

What is claimed is:

1. An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: receive, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; select a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmit a second message to the network node, the second message comprising second information indicative of the configuration.

2. An apparatus according to claim 1, wherein the configuration is indicative of at least one of the following: a granularity for the selection of channel measurement samples, one or more ranges for the selection of channel measurement samples, or one or more similarity metrics for the selection of channel measurement samples.

3. An apparatus according to claim 2, wherein the configuration is indicative of the granularity by being indicative of at least one of the following: a range of samples, a starting point associated with the range of samples, one or more types of offsets associated with one or more samples, a quantity of samples, a quantity of consecutive samples, a quantity of non-consecutive samples, or a ratio of consecutive samples per non-consecutive sample.

4. An apparatus according to claim 2 or 3, wherein the configuration is indicative of the one or more ranges by including an indication to determine a range associated with a sample based on an offset of the sample.

5. An apparatus according to any one of claims 2-4, wherein the configuration is indicative of the one or more similarity metrics by including an indication of a type of similarity metric.

6. An apparatus according to any one of claims 1-5, wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: obtain a plurality of channel measurement samples based at least in part on one or more reference signals; and select one or more channel measurement samples from the plurality of channel measurement samples in accordance with the configuration, wherein the one or more channel measurement samples are associated with one or more offsets relative to a reference point, and wherein the second message comprises third information indicative of the one or more channel measurement samples and the one or more offsets.

7. An apparatus according to claim 6, wherein the one or more channel measurement samples comprise at least a channel measurement sample having a value that is greater than one or more other values of one or more other channel measurement samples included in the plurality of channel measurement samples, wherein the channel measurement sample is associated with an offset of the one or more offsets, and wherein a quantity of channel measurement samples included in the one or more channel measurement samples is based at least in part on the offset.

8. An apparatus according to claim 7, wherein the third information is indicative of the one or more channel measurement samples by including at least one of the following: an indication of a power level associated with the channel measurement sample and the quantity of channel measurement samples, or an indication of a respective power level associated with each channel measurement sample of the one or more channel measurement samples.

9. An apparatus according to claim 8, wherein the third information is indicative of the quantity of channel measurement samples by including an indication of a range associated with the offset.

10. An apparatus according to claim 6, wherein selecting the one or more channel measurement samples comprises: selecting a set of consecutive channel measurement samples from the plurality ofchannel measurement samples; identifying a first channel measurement sample of the set of consecutive channel measurement samples, wherein the first channel measurement sample is associated with a first offset relative to the reference point; and selecting a first subset of consecutive channel measurement samples based at least in part on the first offset, wherein the one or more channel measurement samples comprises at least the first subset of consecutive channel measurement samples.

11. An apparatus according to claim 10, wherein selecting the one or more channel measurement samples further comprises: identifying a second channel measurement sample of the set of consecutive channel measurement samples, wherein the second channel measurement sample is associated with a second offset relative to the reference point, the second offset being greater than the first offset; and selecting a second subset of consecutive channel measurement samples based at least in part on the second offset, wherein a first quantity of samples included in the first subset is greater than a second quantity of samples included in the second subset based at least in part on the second offset being greater than the first offset.

12. The apparatus according to claim 11, wherein the third information comprises an indication of the first offset and at least one of the following: an indication of the second offset or an indication of a difference between the first offset and the second offset.

13. An apparatus according to claim 12, wherein, in accordance with the configuration, the third information indicates the reference point.

14. An apparatus according to claim 13, wherein: the one or more offsets comprise at least one of the following: one or more time offsets, one or more phase offsets, or one or more angle offsets, and the reference point comprises at least one of the following: a reference time, a reference phase, or a reference angle.

15. An apparatus according to claim 6, wherein the plurality of channel measurement samples is associated with a measurement time window, and wherein a quantityof channel measurement samples included in the one or more channel measurement samples is based at least in part on a level of similarity associated with the plurality of channel measurement samples.

16. An apparatus according to claim 6, wherein selecting the one or more channel measurement samples comprises: identifying a first portion of the plurality of channel measurement samples that are associated with a first time-domain measurement window; determining a first level of similarity associated with the first portion based at least in part on one or more similarity metrics; and selecting a first subset of channel measurement samples from the first portion based at least in part on the first level of similarity, wherein the one or more channel measurement samples comprises at least the first subset of channel measurement samples.

17. An apparatus according to claim 16, wherein selecting the one or more channel measurement samples further comprises: identifying a second portion of the plurality of channel measurement samples that are associated with a second time-domain measurement window; determining a second level of similarity associated with the second portion based at least in part on the one or more similarity metrics; and selecting a second subset of channel measurement samples from the second portion based at least in part on the second level of similarity, wherein a first quantity of samples included in the first subset is greater than a second quantity of samples included in the second subset based at least in part on the second level of similarity being greater than the first level of similarity.

18. An apparatus according to claim 6, wherein the plurality of channel measurement samples includes one or more types of channel measurements, and wherein the one or more types of channel measurements comprises at least one of the following: a received power, a phase, a time-domain channel impulse response (OR), a power delay profile (PDP), or a delay profile (DP).

19. An apparatus according to any one of claims 1-18, wherein the one or more characteristics comprise at least one of the following: one or more radio propagationconditions, one or more channel conditions, one or more capabilities of a target device, or one or more positioning accuracy constraints.

20. An apparatus according to any one of claims 1-19, wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: receive, from the network node, a third message comprising information indicating activation of a subset of the set of configurations, wherein the configuration is selected from the subset.

21. An apparatus according to any one of claims 1-20, wherein the apparatus comprises a user equipment (UE) or a radio access network (RAN) node.

22. An apparatus according to claim 21, wherein the apparatus comprises the UE, and wherein the one or more ML models are configured at the UE.

23. An apparatus according to claim 22, wherein the first message and the second message are communicated in accordance with a long term evolution (LTE) positioning protocol (LPP).

24. An apparatus according to claim 21, wherein the apparatus comprises the RAN node, and wherein the one or more ML models are configured at the RAN node or the network node.

25. An apparatus according to claim 24, wherein the first message and the second message are communicated in accordance with a new radio (NR) positioning protocol.

26. An apparatus according to any one of claims 1-25, wherein the network node comprises at least one of the following: a radio access network (RAN) node, a location management function (LMF), an access and mobility management function (AMF), or a network data analytics function (NWDAF).

27. An apparatus according to any one of claims 1-26, wherein at least one of the first message or the second message are communicated in accordance with a radio resource control (RRC) protocol or a medium access control (MAC) protocol.

28. An apparatus according to any one of claims 1-27, wherein the channel measurement samples comprise at least one of the following: time-domain channel measurement samples, angle-domain channel measurement samples, or phase-domain channel measurement samples.

29. An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: determine a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmit a first message to a node, the first message comprising first information indicative of the set of configurations; and receive, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

30. An apparatus according to claim 29, wherein the node comprises a radio access network (RAN) node, and wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: transmit, to at least one other RAN node neighboring the RAN node, a third message indicative of the set of configurations; and receive, from the at least one other RAN node in response to the third message, a fourth message indicative of a second configuration of the set of configurations.

31. An apparatus according to claim 30, wherein the second configuration is different from the configuration.

32. An apparatus according to claim 29, wherein the node comprises a radio access network (RAN) node, and wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to: transmit, to at least one other RAN node neighboring the RAN node in response to the second message, a third message indicative of the configuration.

33. An apparatus according to any one of claims 29-32, wherein determining the set of configurations is based at least in part on one or more positioning accuracy constraints.

34. An apparatus according to any one of claims 29-33, wherein the configuration is indicative of at least one of the following: a granularity for the selection of channel measurement samples, one or more ranges for the selection of channel measurement samples, or one or more similarity metrics for the selection of channel measurement samples.

35. An apparatus according to claim 34, wherein the configuration is indicative of the granularity by being indicative of at least one of the following: a range of samples, a starting point associated with the range of samples, one or more types of offsets associated with one or more samples, a quantity of samples, a quantity of consecutive samples, a quantity of non-consecutive samples, or a ratio of consecutive samples per non-consecutive sample.

36. An apparatus according to claim 34, wherein the configuration is indicative of the one or more ranges by including an indication to determine a range associated with a sample based on an offset of the sample.

37. An apparatus according to claim 34, wherein the configuration is indicative of the one or more similarity metrics by including an indication of a type of similarity metric.

38. An apparatus according any one of claims 29-37, wherein the second message comprises third information indicative of one or more channel measurement samples and one or more offsets associated with the one or more channel measurement samples, wherein the one or more offsets are relative to a reference point, and wherein the one or more channel measurement samples are based at least in part on the configuration.

39. An apparatus according to claim 38, wherein: the one or more offsets comprise at least one of the following: one or more time offsets, one or more phase offsets, or one or more angle offsets, and the reference point comprises at least one of the following: a reference time, a reference phase, or a reference angle.

40. An apparatus according to claim 38, wherein the third information is indicative of the one or more channel measurement samples by including at least one of the following: an indication of a power level associated with a channel measurement sample of the one or more channel measurement samples and a quantity of channel measurement samples in the one or more channel measurement samples, or an indication of a respective power level associated with each channel measurement sample of the one or more channel measurement samples.

41. An apparatus according to claim 40, wherein the channel measurement sample is associated with an offset of the one or more offsets, and wherein the third information is indicative of the quantity of channel measurement samples by including an indication of a range associated with the offset.

42. An apparatus according to claim 38, wherein the one or more channel measurement samples includes one or more types of channel measurements, and wherein the one or more types of channel measurements comprises at least one of the following: a received power, a phase, a time-domain channel impulse response (OR), a power delay profile (PDP), or a delay profile (DP).

43. An apparatus according to any one of claims 29-42, wherein the configuration is based at least in part on one or more characteristics of a wireless communication channel, and wherein the one or more characteristics comprise at least one of the following: one or more radio propagation conditions, one or more channel conditions, one or more capabilities of a target device, or one or more positioning accuracy constraints.

44. An apparatus according to any one of claims 29-43, wherein the at least one memory and the instructions, when executed by the at least one processor, cause the apparatus to:transmit a third message comprising information indicating activation of a subset of the set of configurations, wherein the configuration is based at least in part on the subset.

45. An apparatus according to any one of claims 29-44, wherein the apparatus comprises a radio access network (RAN) node, a location management function (LMF), an access and mobility management function (AMF), or a network data analytics function (NWDAF).

46. A method, comprising: receiving, from a network node, a first message comprising first information indicative of a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; selecting a configuration from the set of configurations based at least in part on one or more characteristics of a wireless communication channel; and transmitting a second message to the network node, the second message comprising second information indicative of the configuration.

47. A method, comprising: determining a set of configurations comprising at least one or more parameters for selection of channel measurement samples for inference input corresponding to one or more machine learning (ML) models; transmitting a first message to a node, the first message comprising first information indicative of the set of configurations; and receiving, from the node in response to the first message, a second message comprising second information indicative of a configuration of the set of configurations.

Images