Method and apparatus for delayed reporting of performance monitoring for artificial intelligence / machine learning (ai / ML) model evaluation and training

Abstract

Methods, apparatuses, and systems are provided for radio measurements in mobile communications. User Equipment (UE) can perform radio measurements and infer other radio measurements using an artificial intelligence / machine learning (AI / ML) model functionality at the UE. The UE stores the radio measurements, the inferred radio measurements output from the AI / ML model functionality, and information about the AI / ML model or dataset on which it was trained in a circular buffer at the UE. Based upon a certain triggering event or upon receipt of a request from a network, the user equipment can provide a measurement report and / or monitoring report to the network that includes some or all of the logged radio measurements, inferred radio measurements, or AI / ML model information.

Description

METHOD AND APPARATUS FOR DELAYED REPORTING OF PERFORMANCE MONITORING FOR ARTIFICIAL INTELLIGENCE / MACHINE LEARNING (AI / ML) MODEL EVALUATION AND TRAININGTechnological Field

[0001] An example embodiment relates generally to wireless communications and, more particularly, but not exclusively, to a method, apparatus and computer program product for delayed reporting of performance monitoring for artificial intelligence / machine learning model evaluation and training.Background

[0002] The 3rd Generation Partnership Project (3GPP) is a standards organization which develops protocols for mobile telephony and is known for the development and maintenance of various standards including second generation (2G), third generation (3G), fourth generation (4G), Long Term Evolution (LTE), and fifth generation (5G) standards. The 5G network has been designed as a Service Based Architecture (SB A) or, in other words, a system architecture in which the system functionality is achieved by a set of network functions providing services to other authorized network functions to access their services.

[0003] The 5G network may comprise a plurality of base stations (e.g., Next generation NodeB (gNB), etc.) that serve multiple cells across a particular area. As a User Equipment (UE) moves through the particular area cell changes, known as handovers in the connected mode, occur to maintain connectivity between the UE and the serving Radio Access Network (RAN). Moreover, cells transmit and receive data via multiple beams. Handover procedures may be triggered as a result of rotation of the UE or as a result of obstructions between the UE and base stations (e.g., a wall, etc.).

[0004] A UE typically reports out to the network on cell-level signal or service quality to the network, or the cell can provide to the network or UE regular reports on cell quality for serving and candidate cells or reference signals (RS).

[0005] Wireless communication systems such as 5G systems and technologies such as new radio (NR), also referred to mobile communication systems and protocols, are under constant development. One core aspect in mobile communication systems is the provision of radio measurements from UE to the network for supporting communication operations.Such radio measurements may be channel state information (CSI) measurements that are provided in CSI-reports from the UE to the network at configured reporting instances based on the latest radio measurements relating to a channel state and quality.Brief Summary

[0006] According to some aspects, there is provided the subject matter of the independent claims. Some further aspects are defined in the dependent claims. The embodiments that do not fall under the scope of the claims are to be interpreted as examples useful for understanding the disclosure.

[0007] According to an embodiment, a method can be carried out by, e.g., a user equipment (UE). An example method can comprise: using a machine learning (ML) model to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network; maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements; and in response to receiving, at the UE, a UE measurement reporting request event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information.

[0008] In some embodiments, the inference output information corresponds to one or more outputs of the ML model. In some embodiments, the inference output information comprises one or more of: predicted channel state information, predicted beam index information, predicted beam layer 1 reference signal received power (Ll-RSRP) information, predicted UE latitude information, predicted UE longitude information, predicted line of sight information, or predicted non-line of sight information. In some embodiments, the UE measurement information comprises one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR values, one or more Ll-RSRP values, one or more filtered Layer 3 -RSRP (L3-RSRP) values, one or more Positioning Reference Signal-RSRP (PRS-RSRP) values, one or more PRS-Reference Signal Received Path Power (PRS-RSRPP) values, one or more power delay profiles, one or more power delay values, one or more channel impulse responses, one or more frequency error measurements, or one or more timing error measurements.

[0009] In some embodiments, the method can further comprise: receiving, from the mobile network, an indication configuring the circular buffer in the UE to have a specified buffer depth. In some embodiments, the UE measurement report comprises only a portion of contents of the circular buffer. In some embodiments, the UE measurement report comprises an entire contents of the circular buffer. In some embodiments, the UE measurement reporting request event is based on at least one of the following occurring: the UE receiving, from the mobile network, a UE measurement reporting request; or an occurrence of a measurement reporting period, the measurement reporting period being indicated to the UE by the mobile network during prior signaling.

[0010] In some embodiments, the UE measurement reporting request includes one of: an indication to report the entire contents of the circular buffer; a time-range of samples to report from the contents of the circular buffer; or a selection of specific samples to report from the contents of the circular buffer. In some embodiments, the UE measurement report further comprises one or more identifiers associated with the ML model, and the method can further comprise: maintaining, in the circular buffer at the UE, with the inference output information, the one or more identifiers identifying the ML model used to generate the inference output information, wherein the ML model identifiers correspond to at least the inference output information in the UE measurement report.

[0011] In some embodiments, the one or more identifiers associated with the ML model comprise at least one of: a model ID associated with the ML model, an associated ID corresponding to a network configuration used during data collection, an identifier of a data collection configuration used to train the ML model, an identifier of a dataset used to train the ML model, a UE hardware ID, or a UE vendor ID. In some embodiments, the UE measurement report further comprises one or more timing-related information associated with the inference output information, and the method can further comprise: maintaining, in the circular buffer at the UE, the one or more timing-related information identifying a respective time at which respective of the inference output information was generated, wherein the timing-related information corresponds to at least the inference output information in the UE measurement report.

[0012] In some embodiments, the one or more timing-related information comprise at least one of: a system frame number, a slot number, an orthogonal frequency division multiplexing (OFDM) symbol number, a global navigation satellite system (GNSS) timestamp, a UE timestamp, or another timestamp.

[0013] According to another embodiment, a method can be carried out by, e.g, an element or function of a mobile network, network-side device, base station, or network node. An example method can be carried out by a base station, the method comprising at least: receiving, from a user equipment (UE), in response to a base station-initiated measurement reporting event, a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements of one or more cells in the mobile network, and the inference output information being inferred by one or more machine learning (ML) models in a ML functionality of the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network. In some embodiments, the method can further comprise: providing, to an element or function in the mobile network, a UE performance report comprising the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information. In some embodiments, the method can further comprise: storing, at the base station, the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information. In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to perform a lifecycle management (LCM) action with regard to using the one or more ML models of the ML functionality in the UE.

[0014] In some embodiments, the inference output information in the UE measurement report is associated with one or more inference outputs from an ML model in the ML functionality in the UE. In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to perform inference, using the ML model of the ML model functionality in the UE. In some embodiments, the method can further comprise: determining, based at least upon the UE measurement report, an ML model performance of the ML model used to generate the one or more inference outputs associated with the inference output information; and determining whether the ML model performance of the ML model is below a ML model performance threshold.

[0015] In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to maintain the UE measurement information and the inference output information in a circular buffer maintained at the UE. In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to generate and transmit the UE measurement report comprising the UEmeasurement information and the inference output information. In some embodiments, the UE measurement report further comprises one or more identifiers associated with the ML model. In some embodiments, the indication that the UE is to perform inference using the ML model comprises or is provided in one or more of: a CSLReportConfig message, an ReportConfigNR message, or an LPP RequestLocationlnformation message.

[0016] In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the circular buffer in the UE. In some embodiments, the inference output information comprises one or more of: predicted channel state information, predicted beam index information, predicted beam layer 1 reference signal received power (Ll-RSRP) information, predicted UE latitude information, predicted UE longitude information, predicted line of sight information, or predicted non-line of sight information. In some embodiments, the indication configuring the circular buffer in the UE further configures the UE to store the UE measurement information in the buffer in the form of one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR values, one or more Ll-RSRP values, one or more filtered Layer 3 -RSRP (L3-RSRP) values, one or more Positioning Reference Signal-RSRP (PRS-RSRP) values, one or more PRS-Reference Signal Received Path Power (PRS-RSRPP) values, one or more power delay profiles, one or more power delay values, one or more channel impulse responses, one or more frequency error measurements, or one or more timing error measurements. In some embodiments, the indication configuring the circular buffer in the UE further configures the circular buffer in the UE to have a specified buffer depth.

[0017] In some embodiments, the base station-initiated measurement reporting event comprises one or more of: the base station providing, to the UE, a UE measurement reporting request, or an occurrence of a measurement reporting period, the measurement reporting period being indicated to the UE by the mobile network during prior signaling. In some embodiments, the UE measurement reporting request comprises one or more of: an indication to report all contents of the circular buffer, a time-range of samples to report from the contents of the circular buffer, or a selection of specific samples to report from the contents of the circular buffer. In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to include one or more ML model identifiers in the UE measurement report.

[0018] In some embodiments, the UE measurement report comprises the one or moreML model identifiers associated with the one or more ML models used by the UE to generatethe inference output information in the UE measurement report. In some embodiments, the method can further comprise: determining, based upon the one or more ML model identifiers in the UE measurement report, the one or more ML models used by the UE to generate the inference output information. In some embodiments, the one or more identifiers comprise at least one of: one or more model IDs associated with the one or more ML models, one or more associated IDs corresponding to one or more network configurations used during data collection, one or more identifiers of one or more data collection configurations used to train the one or more ML models, one or more identifiers of one or more datasets used to train the one or more ML models, a UE hardware ID, or a UE vendor ID.

[0019] In some embodiments, the method can further comprise: providing, to the UE, an indication configuring the UE to include timing-related information in the UE measurement report, the timing-related information being associated with the ML model used to generate the inference output information. In some embodiments, the timing-related information comprise at least one of: a system frame number, a slot number, an orthogonal frequency division multiplexing (OFDM) symbol number, a global navigation satellite system (GNSS) timestamp, a UE timestamp, or another timestamp. In some embodiments, the UE measurement information comprises one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR values, one or more LI -RSRP values, one or more filtered Layer 3 -RSRP (L3-RSRP) values, one or more Positioning Reference Signal -RSRP (PRS-RSRP) values, one or more PRS-Reference Signal Received Path Power (PRS- RSRPP) values, one or more power delay profiles, one or more power delay values, one or more channel impulse responses, one or more frequency error measurements, or one or more timing error measurements.

[0020] The above-noted aspects and features may be implemented in systems, apparatuses, methods, articles and non-transitory computer-readable media depending on the desired configuration. The subject disclosure may be implemented in and used with a number of different types of devices, including but not limited to cellular phones, tablet computers, wearable computing devices, portable media players, and any of various other computing devices. For example, a user equipment (UE) can be provided that comprises at least one processor and at least one memory that stores thereon instructions which, when executed by the at least one processor, cause the UE to perform some or all of the elements of the above-described method, according to various embodiments. Alternatively, a networkside device such as a network node or base station can be provided that comprises at leastone processor and at least one memory that stores thereon instructions which, when executed by the at least one processor, cause the network-side device to perform some or all of the elements of the above-described method, according to various embodiments. In other examples, a computer program product, such as a non-transitory computer-readable storage medium can be provided that comprises instructions stored thereon that, when executed by at least one processor of an apparatus, cause the apparatus to perform some or all elements of a method such as that descried above, according to some embodiments. In other examples, an apparatus can be provided that comprises means for carrying out a method - such means can include, e.g., a processor and a memory storing computer-executable instructions or computer codes thereon that, when executed by the processor, cause the apparatus to perform some or all of a method such as one of the methods described herein.

[0021] This summary is intended to provide a brief overview of some of the aspects and features according to the subject disclosure. Accordingly, it will be appreciated that the above-described features are merely examples and should not be construed to narrow the scope of the subject disclosure in any way. Other features, aspects, and advantages of the subject disclosure will become apparent from the following detailed description, drawings and claims.List of the Drawings

[0022] A better understanding of the subject disclosure may be obtained when the following detailed description of various embodiments is considered in conjunction with the following drawings, in which:

[0023] FIG. 1 shows a schematic diagram of an example wireless network, in accordance with embodiments of the present disclosure;

[0024] FIG. 2 shows a schematic diagram of an example wireless device, in accordance with embodiments of the present disclosure;

[0025] FIG. 3 shows a schematic diagram of an example network node, in accordance with embodiments of the present disclosure;

[0026] FIG. 4 is an example signal flow diagram illustrating measurement reporting and machine learning (ML) model management, in accordance with embodiments of the present disclosure;

[0027] FIG. 5 is an example signal flow diagram illustrating measurement reporting and machine learning (ML) model management, in accordance with embodiments of the present disclosure;

[0028] FIG. 6 is an example structure for UE measurement sample storage in a circular buffer in the UE, in accordance with embodiments of the present disclosure;

[0029] FIG. 7 is an example block flow diagram illustrating a process for managing UE measurement sample storage in a circular buffer in the UE, in accordance with embodiments of the present disclosure;

[0030] FIG. 8 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein;

[0031] FIG. 9 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein;

[0032] FIG. 10 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0033] FIG. 11 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0034] FIG. 12 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0035] FIG. 13 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0036] FIG. 14 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0037] FIG. 15 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0038] FIG. 16 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.

[0039] FIG. 17 is a flowchart of operations according to certain example embodiments implemented, for example, by an apparatus as described herein.Abbreviations

[0040] The following meanings for the abbreviations used herein apply:

[0041] Al Artificial Intelligence

[0042] AMF Access and Mobility Management Function

[0043] CHF Charging Function

[0044] CM Connection Management

[0045] CN Core Network

[0046] DNN Data Network Name

[0047] ES Energy Saving

[0048] FM Fault Management

[0049] FQDN Fully Qualified Domain Name

[0050] ML Machine Learning

[0051] MnS Management Service

[0052] NAS Non-Access Stratum

[0053] NF Network Function

[0054] NSC Network Slice Consumer

[0055] NSP Network Slice Provider

[0056] NWDAF Network Data Analytics Function

[0057] PCF Policy Control Function

[0058] PDU Packet Data Unit

[0059] PGW Packet Data Network GateWay

[0060] PM Performance Management

[0061] QoS Quality of Service

[0062] SMF Session Management Function

[0063] SMS Short Message Service

[0064] SMSF SMS Function

[0065] SUPI Subscription Permanent Identifier

[0066] UDM Unified Data Management

[0067] UE User EquipmentDetailed Description

[0068] The examples and embodiments set forth below represent information to enable those skilled in the art to practice the subject disclosure. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the description and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the description.

[0069] In the following description, numerous specific details are set forth. However, it is understood that embodiments may be practiced without these specific details. In other instances, well-known circuits, structures, and techniques have not been shown in detail in order not to obscure the understanding of the description. Those of ordinary skill in the art, with the included description, will be able to implement appropriate functionality without undue experimentation.

[0070] The following embodiments are examples. 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.

[0071] 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).

[0072] As used herein, "plurality" means two or more. As used herein, a "set" of items may include one or more of such items. As used herein, whether in the subject disclosure or the claims, the terms "comprising", "including", "carrying", "having", "containing", "involving", and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases "consisting of and "consisting essentially of, respectively, are closed or semi-closed transitional phrases with respect to claims. Use of ordinal terms such as "first", "second", "third", etc., in the claims or the subject disclosure to modify an element does not by itself connote any priority, precedence, or order of one element over another or the temporal order in which acts of a method are performed, but are used merely as labels to distinguish one element having a certain name from another element having a same name (but for use of the ordinal term) to distinguish the elements. As usedherein, "and / or" and "at least one of' means that the listed items are alternatives, but the alternatives also include any combination of the listed items.

[0073] 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).

[0074] 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 NRNB (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 non-terrestrial network (NTN) or non-ground network device such as a satellite network device, a low earth orbit (LEO) satellite and a geosynchronous earth orbit (GEO) satellite, or an aircraft network device.

[0075] Moreover, in connection of split radio access network (RAN), the network device may refer to a centralised 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 some embodiments, the DUs may comprise e.g. a radiolink 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) and an internet protocol (IP) layers. 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.

[0076] 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.

[0077] 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.

[0078] Before explaining the examples according to the subject disclosure in detail, certain general principles of a wireless communication system are briefly explained with reference to FIGs. 1 to 3 to assist in understanding the technology underlying the described examples.

[0079] FIG. 1 illustrates an example of a wireless network 100 that may be used for wireless communications. Wireless network 100 includes wireless devices, such as UEs 110 (e.g., 110A-110B), and network nodes 120 / 130, such as radio access nodes 120 (e.g., 120A- 120B, which may be network nodes like eNBs, gNBs, etc.), connected to one or more furthernetwork nodes 130 over an interconnecting network 125. The network 100 may use any suitable deployment scenarios. UEs 110 within coverage area 115 may each be capable of communicating directly with radio access nodes 120 over a wireless or air interface. In some embodiments, UEs 110 may also be capable of communicating with each other via D2D communication.

[0080] As an example, UE 110A may communicate with radio access node 120A over a wireless or air interface. That is, UE 110A may transmit wireless signals to and / or receive wireless signals from radio access node 120A. The wireless signals may contain voice traffic, data traffic, control signals, and / or any other suitable information.

[0081] As used herein, the term "user equipment" (UE) (e.g., UE 110) has the full breadth of its ordinary meaning and may refer to any type of wireless device which may communicate with a network node (e.g., network node 120) and / or with another UE (e.g., different to UE 110) in a cellular or mobile or wireless communication system. Examples of UE are target device, D2D UE, machine type UE or UE capable of machine-to-machine (M2M) communication, personal digital assistant, tablet, mobile terminal, smart phone, laptop embedded equipped (LEE), laptop mounted equipment (LME), USB dongles, ProSe UE, vehicle-to-vehicle (V2V) UE, V2X UE, MTC UE, eMTC UE, FeMTC UE, UE Cat 0, UE Cat Ml, narrow band loT (NB-IoT) UE, UE Cat NB1, etc. Example embodiments of a UE are described in more detail below with respect to FIG. 2.

[0082] In some embodiments, an area of wireless signal coverage 115 associated with a radio access node (e.g., 120) may be referred to as a cell. However, particularly with respect to the 5thgeneration (5G) / New Radio (NR) mobile communication concepts, beams, such as the herein described multicast radio beams (MRBs) may be used within cells for communication.

[0083] With respect to a beam-based mobile communication system, the radio access node / base station (e.g., network node 120) may transmit a beamformed signal to the UE 110 in one or more transmit directions (transmission beam, Tx beam). The UE 110 may receive the beamformed signal from the network node 120 in one or more receive directions (reception beam, Rx beam). The UE 110 may also transmit a beamformed signal to the network node 120 in one or more directions and the network node 120 may receive the beamformed signal from the UE 110 in one or more directions. The network node 120 and the UE 110 may determine the best receive and transmit directions, e.g., best in the sense ofthese directions leading to the highest link quality or fulfilling other quality conditions in the most suitable manner, for each of the base station / UE pairs.

[0084] The interconnecting network 125 may refer to any interconnecting system capable of transmitting audio, video, signals, data, messages, etc., or any combination of the preceding. The interconnecting network 125 may include all or a portion of a public switched telephone network (PSTN), a public or private data network, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a local, regional, or global communication or computer network such as the Internet, a wireline or wireless network, an enterprise intranet, or any other suitable communication link, including combinations thereof.

[0085] In some embodiments, the network node 130 may be a core network node, managing the establishment of communication sessions and other various other functionalities for UEs 110. Examples of network node 130 may include mobile switching center (MSC), MME, serving gateway (SGW), packet data network gateway (PGW), operation and maintenance (O&M), operations support system (OSS), SON, positioning node (e.g., Enhanced Serving Mobile Location Center, E-SMLC), location server node, Minimization of Drive Tests (MDT) node, etc. UEs 110 may exchange certain signals with the network node 130 using the non-access stratum (NAS) layer. In non-access stratum signaling, signals between UEs 110 and the network node 130 may be transparently passed through the radio access network. In some embodiments, radio access nodes 120 may interface with one or more network nodes 130 over an internode interface.

[0086] As used herein, the term "network node" has the full breadth of its ordinary meaning and may correspond to any type of radio access node (e.g., network node 120) or any network node, which may communicate with a UE and / or with another network node in a cellular or mobile or wireless communication system. Examples of network nodes are NodeB, MeNB, SeNB, a network node may belonging to MCG or SCG, base station (BS), multi-standard radio (MSR) radio access node such as MSR BS, eNodeB, network controller, radio network controller (RNC), base station controller (BSC), relay, donor node controlling relay, base transceiver station (BTS), access point (AP), transmission point, transmission node, RRU, RRH, node in distributed antenna system (DAS), core network node (e.g., MSC, MME, etc.), O&M, OSS, Self-organizing Network (SON), positioning node (e.g., E-SMLC), MDT, test equipment, etc. Example embodiments of a network node are described in more detail below with respect to FIG. 3.

[0087] In some embodiments, network node 120 may be a distributed radio access node. The components of the radio access node 120, and their associated functions, may be separated into two main units (or sub-radio network nodes) which may be referred to as the central unit (CU) and the distributed unit (DU). Different distributed radio network node architectures are possible. For instance, in some architectures, a DU may be connected to a CU via dedicated wired or wireless link (e.g., an optical fiber cable) whereas in other architectures, a DU may be connected a CU via a transport network. Also, how the various functions of the network node 120 are separated between the CU(s) and DU(s) may vary depending on the chosen architecture.

[0088] In some embodiments, network nodes 120 may communicate with each other over terrestrial or other connections. The communication between the network nodes 120 may, e.g., in a 5G / NR communication system may be achieved by using an Xn interface connecting the network nodes 120.

[0089] Example wireless communication systems are architectures standardized by the 3rd Generation Partnership Project (3GPP). A latest 3GPP based development is often referred to as the long-term evolution (LTE) of the Universal Mobile Telecommunications System (UMTS) radio-access technology (RAT). The various development stages of the 3GPP specifications are referred to as releases. More recent developments of the LTE are often referred to as LTE Advanced (LTE-A). The LTE (LTE-A) employs a radio mobile architecture known as the Evolved Universal Terrestrial Radio Access Network (E-UTRAN) and a core network known as the Evolved Packet Core (EPC). Base stations of such systems are known as evolved or enhanced Node Bs (eNBs) and provide E-UTRAN features such as user plane Packet Data Convergence / Radio Link Control / Medium Access Control / Physical layer protocol (PDCP / RLC / MAC / PHY) and control plane Radio Resource Control (RRC) protocol terminations towards the communication devices. Other RAT examples comprise those provided by base stations of systems that are based on technologies such as WLAN and / or Worldwide Interoperability for Microwave Access (WiMax). A base station may provide coverage for an entire cell or similar radio service area. Core network elements include Mobility Management Entity (MME), Serving Gateway (S-GW) and Packet Gateway (P-GW).

[0090] An example of a suitable communications system is the 5G or NR concept. Network architecture in NR may be similar to that of LTE-A. Base stations of NR systems may be known as next generation Node Bs (gNBs). Changes to the network architecture maydepend on the need to support various radio technologies and finer Quality of Service (QoS) support, and some on-demand requirements for QoS levels to support Quality of Experience (QoE) of user point of view. Also network aware services and applications, and service and application aware networks may bring changes to the architecture. Those are related to Information Centric Network (ICN) and User-Centric Content Delivery Network (UC-CDN) approaches. NR may use multiple input-multiple output (MIMO) antennas, many more base stations or nodes than the LTE (a so-called small cell concept), including macro sites operating in co-operation with smaller stations and perhaps also employing a variety of radio technologies for better coverage and enhanced data rates.

[0091] Future networks may utilize network functions virtualization (NFV) which is a network architecture concept that proposes virtualizing network node functions into "building blocks" or entities that may be operationally connected or linked together to provide services. A virtualized network function (VNF) may comprise one or more virtual machines running computer program codes using standard or general type servers instead of customized hardware. Cloud computing or data storage may also be utilized. In radio communications this may mean node operations to be carried out, at least partly, in a server, host or node operationally coupled to a remote radio head. It is also possible that node operations will be distributed among a plurality of servers, nodes, or hosts. It should also be understood that the distribution of labor between core network operations and base station operations may differ from that of the LTE or even be non-existent.

[0092] An example 5G core network (CN) comprises functional entities. The CN is connected to a UE via the radio access network (RAN). An UPF (User Plane Function) whose role is called PSA (PDU Session Anchor) may be responsible for forwarding frames back and forth between the DN (data network) and the tunnels established over the 5G towards the UEs exchanging traffic with the data network (DN). The UPF is controlled by an SMF (Session Management Function) that receives policies from a PCF (Policy Control Function). The CN may also include an AMF (Access & Mobility Function).

[0093] Generally, all concepts disclosed herein may be applicable to different communication networks, comprising but not limited to LTE, LTE-A, 5G, 5G advanced, 6G, and other future or already implemented networks.

[0094] FIG. 2 is a schematic diagram of an apparatus for the UE 110. In an embodiment, the apparatus may comprise the UE 110, in yet another embodiment the apparatus is comprised in the UE 110, and in another embodiment the apparatus is the UE 110. Theapparatus may comprise a wireless device. The apparatus may comprise at least one processor 220 and at least memory 230 storing computer program instructions that, when executed by the at least one processor 220, cause the apparatus to carry out the embodiments of the UE 110 described herein. UE 110 includes a transceiver 210, processor 220, memory 230, and a network interface 240. In some embodiments, the transceiver 210 facilitates transmitting wireless signals to and receiving wireless signals from radio access node 120 (e.g., via transmitted s) (Tx), receiver(s) (Rx) and antenna(s)). The processor 220 executes instructions to provide some or all of the functionalities described herein as being provided by UE 110, and the memory 230 stores the instructions executed by the processor 220. In some embodiments, the processor 220 and the memory 230 form processing circuitry.

[0095] The processor 220 may include any suitable combination of hardware to execute instructions and manipulate data to perform some or all of the described functions of UE 110 described herein. In some embodiments, the processor 220 may include, for example, one or more computers, one or more central processing units (CPUs), one or more microprocessors, one or more application specific integrated circuits (ASICs), one or more field programmable gate arrays (FPGAs) and / or other logic.

[0096] The memory 230 is generally operable to store instructions, such as a computer program, software, an application including one or more of logic, rules, algorithms, code, tables, etc. and / or other instructions capable of being executed by a processor 220. Examples of memory 230 include computer memory (for example, Random Access Memory (RAM) or Read Only Memory (ROM)), mass storage media (for example, a hard disk), removable storage media (for example, a Compact Disk (CD) or a Digital Video Disk (DVD)), and / or or any other volatile or non-volatile, non- transitory computer-readable and / or computerexecutable memory devices that store information, data, and / or instructions that may be used by the processor 220 of UE 110. For example, the memory 230 includes computer program code causing the processor 220 to perform processing according to the methods described herein.

[0097] The network interface 240 is communicatively coupled to the processor 220 and may refer to any suitable device operable to receive input for UE 110, send output from UE 110, perform suitable processing of the input or output or both, communicate to other devices, or any combination thereof. The network interface 240 may include appropriate hardware (e.g., port, modem, network interface card, etc.) and software, including protocol conversion and data processing capabilities, to communicate through a network.

[0098] Other embodiments of UE 110 may include additional components beyond those shown in FIG. 2 that may be responsible for providing certain aspects of the wireless device’ s functionalities, including any of the functionalities described herein and / or any additional functionalities (including any functionality necessary to support the mechanisms according to the subject disclosure). As an example, UE 110 may include input devices and circuits, output devices, and one or more synchronization units or circuits, which may be part of the processor 220. Input devices include mechanisms for entry of data into UE 110. For example, input devices may include input mechanisms, such as a microphone, input elements, a display, etc. Output devices may include mechanisms for outputting data in audio, video and / or hard copy format. For example, output devices may include a speaker, a display, etc.

[0099] In some embodiments, the wireless device UE 110 may comprise a series of modules configured to implement the functionalities of the wireless device described herein. Moreover, in some embodiments, the UE 110 may also comprise means for the functionalities described herein. A non-transitory computer readable medium with computer executable instructions stored thereon executed by the processor 220 of the UE 110 to perform the functionalities as described herein may also be provided.

[0100] It will be appreciated that the various modules may be implemented as combination of hardware and software, for instance, the processor, memory, and transceiver(s) of UE 110 shown in FIG. 2. Some embodiments may also include additional modules to support additional and / or optional functionalities.

[0101] FIG. 3 is a schematic diagram of an example of an apparatus for a radio access node 120 or network node 130. The apparatus may comprise at least one processor 220 and at least memory 230 storing computer program instructions that, when executed by the at least one processor 220, cause the apparatus to carry out the embodiments of the core network node 130 or radio access node 120 described herein. The example a radio access node (e.g., 120) or a core network node (e.g., 130) may include one or more of a transceiver 310, processor 320, memory 330, and network interface 340. In some embodiments, the transceiver 310 facilitates transmitting wireless signals to and receiving wireless signals from wireless devices, such as UE 110 (e.g., via transmitter(s) (Tx), receiver(s) (Rx), and antenna(s)). The processor 320 executes instructions to provide some or all of the functionalities described herein as being provided by a radio access node (e.g., 120) or a core network node (e.g., 130), the memory 330 stores the instructions executed by the processor 320. In some embodiments, the processor 320 and the memory 330 form processingcircuitry. The network interface 340 may communicate signals to backend network components, such as a gateway, switch, router, Internet, Public Switched Telephone Network (PSTN), core network nodes or radio network controllers, etc.

[0102] The processor 320 may include any suitable combination of hardware to execute instructions and manipulate data to perform some or all of the described functions of a radio access node (e.g., 120) or a core network node (e.g., 130), such as those described herein. In some embodiments, the processor 320 may include, for example, one or more computers, one or more central processing units (CPUs), one or more microprocessors, one or more application specific integrated circuits (ASICs), one or more field programmable gate arrays (FPGAs) and / or other logic.

[0103] The memory 330 is generally operable to store instructions, such as a computer program, software, an application including one or more of logic, rules, algorithms, code, tables, etc. and / or other instructions capable of being executed by a processor 320. Examples of memory 330 include computer memory (for example, Random Access Memory (RAM) or Read Only Memory (ROM)), mass storage media (for example, a hard disk), removable storage media (for example, a Compact Disk (CD) or a Digital Video Disk (DVD)), and / or or any other volatile or non-volatile, non- transitory computer-readable and / or computerexecutable memory devices that store information. For example, the memory 330 includes computer program code causing the processor 320 to perform processing according to the methods described herein.

[0104] In some embodiments, the network interface 340 is communicatively coupled to the processor 320 and may refer to any suitable device operable to receive input for a radio access node (e.g., 120) or a core network node (e.g., 130), send output from a radio access node (e.g., 120) or a core network node (e.g., 130), perform suitable processing of the input or output or both, communicate to other devices, or any combination of the preceding. The network interface 340 may include appropriate hardware (e.g., port, modem, network interface card, etc.) and software, including protocol conversion and data processing capabilities, to communicate through a network.

[0105] Other embodiments of a radio access node (e.g., 120) or a core network node (e.g., 130) may include additional components beyond those shown in FIG. 3 that may be responsible for providing certain aspects of the node’s functionalities, including any of the functionalities described herein and / or any additional functionalities (including any functionality necessary to support the solutions described herein). The various differenttypes of radio access nodes or core network nodes may include components having the same physical hardware but configured (e.g., via programming) to support different radio access technologies, or may represent partly or entirely different physical components.

[0106] Processors, interfaces, and memory similar to those described with respect to FIG. 3 may be included in other nodes (such as UE 110, network node 120, etc.). Other nodes may optionally include or not include a wireless interface (such as the transceiver described in FIG. 3).

[0107] In some embodiments, the radio access node 120 or the core network node 130 may comprise a series of modules configured to implement the functionalities of the radio access node 120 or the core network node 130 described herein. Moreover, in some embodiments, the radio access node 120 or the core network node 130 may also comprise means for the functionalities described herein. A non-transitory computer readable medium with computer executable instructions stored thereon executed by the processor 320 of the network node 120 / 130 to perform the functionalities as described herein may also be provided.

[0108] It will be appreciated that the various modules may be implemented as combination of hardware and software, for instance, the processor, memory, and transceiver(s) of the radio access node 120 or the core network node 130 shown in FIG. 3. Some embodiments may also include additional modules to support additional and / or optional functionalities.

[0109] Before referring to FIGs. 4 to 7 and describing principles according to the disclosure, summarizing information and aspects related to the subject disclosure will be provided. It should be noted that all concepts described herein, although described, e.g., for one communication direction, e.g., for downlink communication, are applicable for the other direction as well, e.g., in the uplink (UL) communication. Moreover, concepts described for one entity, e.g., a UE 110, are applicable to another entity, e.g., a base station or network node 120, when considering for example another communication direction or another network setting as will be apparent to the skilled person.

[0110] Today's wireless communication systems require UEs 110 to perform radio measurements. Radio measurement data obtained by performing radio measurements relate, e.g., to channel properties of a communication link and may be information that can be measured directly, e.g. a signal strength, but also comprises information calculated based on directly measured information, such as channel quality information and the like. Radiomeasurement data is, thus, collected on layer 1 (LI, physical layer, lower layer) of the respective network protocol. For example, in LTE and / or 5G, radio measurements may - beyond others - relate to channel state information (CSI). The reports of CSI (CSI-reports) to the network indicate how good or bad a measured channel is, for example based on a particular reference signal, and, thus, how well the network is perceived by the UE 110 at a specific time. A CSLreport may comprise one or more of the following major components, also denoted as report quantities (which are, e.g., also listed and explained in 3GPP TS 38.214 vl8.1.0): channel quality information (CQI), precoding matrix indicator (PMI), CSL RS resource indicator, synchronization signal (SS) / physical broadcast channel (PBCH) resource block indicator (SSBRI), layer indicator (LI), rank indicator (RI) and / or Ll- reference signal received power (RSRP).[OHl] Radio measurement data obtained by UEs 110 by performing radio measurements on CSLRS be filtered and used to aid the radio access node 120 in making mobility decisions, e.g., a decision to hand a UE 110 from one radio access node 120 to another radio access node. These radio measurements are made on a plurality of radio access nodes which could be considered as candidate cells for handover. Furthermore, the radio measurements can be associated with the cell which transmitted the reference signals.

[0112] Radio measurement data can also be obtained by UEs 110 by performing radio measurements on positioning reference signals (PRS). Measurements on PRS can be made on plurality of radio access nodes and / or transmission-reception points (TRPs). Radio measurements on PRS can further be used to calculate a UE’s 110 position, e.g., a latitude, longitude, and altitude. Furthermore, the radio measurements can be associated with the cell which transmitted the PRS. For example, in LTE and / or 5G, these radio measurements on PRS and / or a UE’s 110 position, e.g., its latitude, longitude, and altitude, can be reported to the network over the LTE Positioning Protocol (LPP) through an LPP Provide Location Information message, which is further explained in 3GPP TS 37.355 vl8.41.0 and 3GPP TS 38.305 vl8.3.0.

[0113] There exist various reasons why a network seeks to collect radio measurement data from UEs 110. Most generally, radio measurements are typically used for supporting the functioning of a mobile network. For example, CSI (e.g., as configured with information element (IE) CSI-ReportConfig as described in 3GPP TS 38.331 vl8.1.0, clause 6.3.2) is currently measured in real-time (i.e., immediately after receiving a channel / signal) and immediately (after suitable processing time) reported to the network or, more specifically,to a network node 120 to support selection of appropriate network nodes, channels, frequencies, beams, and the like. In other words, immediate radio measurement reports, such as the currently defined CSI-reports and RRC measurement reports, are sent at a next available opportunity to use physical channel resources for reporting (e.g., configured PUSCH / PUCCH opportunity) because the radio measurement data may be time-critical for the operation of the network. However, radio measurements may also facilitate other purposes that may be less time-critical.

[0114] For example, recent developments in mobile communication consider the use of AI / ML models for beam management as well as for other applications. Generally speaking, there exist benefits of augmenting the air-interface with features enabling improved support of AI / ML based algorithms or, in other words, AI / ML models. AI / ML functionality over radio interface may be associated with life cycle management (LCM) of an AI / ML model, e.g., model training, model deployment, model inference, model monitoring, model updating, and the like.

[0115] In particular, in order to maintain high performance in the network while the UEs 110 are operating with AI / ML models, the network requires feedback from the UEs 110 such that it can be determined whether poor performance related to the operation of an AI / ML model is due to the model itself or due to channel characteristics such as channel fading, multi-path, signal strength, e.g., reference signal received power (RSRP), signal quality, e.g., reference signal received quality (RSRQ), or signal-to-noise and interference ratio (SINR). However, models trained for a particular scenario should generally perform well, and by the time that poor performance is detected, the information which could have been useful, such as historical measurement information and model output information generated from that historical measurement information, would be lost.

[0116] Additionally, in order to train, update, or finetune these AI / ML models, physical layer data from UEs 110 may be used as training data. Physical layer data comprises LI radio measurements, which include calculations based on LI measurement data, such as channel state information (e.g., CQI, PMI, CRI, SSBRI, LI, and / or RI) and beam predictions. Currently, these reports can only be provided by the UE 110 in radio resource control (RRC) connected state on a physical uplink control channel (PUCCH) or a physical uplink shared channel (PUSCH) in immediately after real-time measurements, i.e., very frequently if data without holes in time is required. Hence, if the training of AI / ML models (or any otherfunctionality at the network side) requires large amounts of physical layer data from the UEs 110, the air interface will be overloaded.

[0117] To overcome the large impact to air interface and signaling overhead due to these frequent report transfers, UEs 110 are tapped to collect field measurements to enable the network to use UEs 110 to collect mobile network data, and thus, reduce the need for frequent performance monitoring-related reports and to reduce the need for traditional drive tests, which are associated with high costs, time commitments, and the requirement of providing access to areas with test vehicles.

[0118] Referring now to FIG. 4, a signal flow diagram of an example signaling process is illustrated between a UE 410 and a network function 420 for Network-Initiated AI / ML management. The network function 420 configures the UE 410 for radio measurement and reporting and / or for measuring / reporting AI / ML model performance metrics. The UE 410 can provide Performance or Assistance information (e.g., radio measurements, AI / ML performance metrics, etc.) to the network element 420. The network function 420 can manage evaluation of AI / ML model performance and training / retraining of the AI / ML model. Thereafter, the network function 420 can provide management instructions (e.g., an updated AI / ML model-functionality, updated RRC configuration information, and / or the like) to the UE 410.

[0119] Referring now to FIG. 5, a signal flow diagram is provided that illustrates an example signaling process between a UE 510, a network function 520, and a trace control element (TCE) 530 in a mobile network.

[0120] Although illustrated in FIG. 5 as performing radio measurements for a particular network node (e.g., 520), the UE 510 is not, per se, limited or restricted with regard to how many network nodes, RANs, eNBs, gNBs, and / or the like, upon which or for which the UE 510 can be configured to perform measurements such as radio measurements or AI / ML model performance measurements.

[0121] In some embodiments, in Step 1 (Inference Configuration), the network node 520 or the UE 510 can configure the UE 510 to perform inference based on a measurement and reporting configuration such as a CSI-ReportConfig, an RRC- MeasConfig, or an LPP RequestLocationlnformation, or any other type of measurement reporting configuration.

[0122] In some embodiments, in Step 2 (Monitoring Configuration), the network node 520 or the UE 510 can configure the UE 510 to store and report monitoring dataassociated with the inference configuration or configurations provided in Step 1. In some embodiments, the Monitoring Configuration can comprise one or more of the following elements:- Buffer depth - number of samples to store, where a sample could be one or more measurements and one or more inference outputs associated with the measurements. Alternatively, buffer depth may be configured in terms of the size of the buffer in bits or in some other unit.- Monitoring report triggering conditions - events upon which the UE should transmit its monitoring buffer, which could include one or more performance degradation thresholds, a measurement window with a periodicity, or a rolling timer which periodically triggers a report. Some examples of performance degradation thresholds follow.1) Inference output accuracy < 80% - for a beam prediction model, this could mean that the best beam was chosen correctly less than 80% of the time, and for a positioning prediction model, this could mean that there is positioning accuracy calculation which indicates a positioning accuracy of less than 80%.2) For beam prediction, one metric could be the percentage of time that the best beam was within the top N beams, e.g., the percentage of time that the best beam was within the top 4 reported beams.- Associated measurement and reporting configuration - an ID pointing to one of the measurement and reporting configurations exemplified in Step 1.- Monitoring configuration ID - an ID to which a measurement and reporting configuration points to associate itself with the monitoring configuration.

[0123] In some embodiments, in Steps 3 - 4, the UE 510 can perform / make measurements, and input the measurements into an AI / ML model to generate an inference output from the AI / ML model. The measurement(s) and inference output(s) are stored locally in the monitoring buffer in the UE 510. In some embodiments, alongside the measurement(s) and inference output(s) stored in the monitoring buffer at the UE 510 can be stored one or more of the following information, which the UE may optionally be configured to report to the network:- System frame number- Slot number- OFDM symbol number- Timestamp, which could be a GNSS-based timestamp

[0124] In some embodiments, in Step 5, the UE 510 reports the inference output to the network node 520.

[0125] In some embodiments, in Step 6, the UE 510 can detect that AI / ML model performance has degraded past a configured threshold.

[0126] In some embodiments, in Step 7, the UE 510 can transmit a buffered monitoring report to the network node 520.

[0127] In some embodiments, in Step 8, the network node 520 can detect that performance has degraded past a configured (e.g., by 0AM) or pre-configured (e.g., implementation-based) threshold.

[0128] In some embodiments, in Step 9, the network node 520 can request a buffered monitoring report from the UE 510.

[0129] In some embodiments, in Step 10, the UE 510 can retrieve buffered monitoring data from the internal buffer at the UE 510. In Step 11, the UE 510 then transmits a buffered monitoring report comprising some or all of the buffered monitoring data to the network node 520.

[0130] In some embodiments, in Step 12, if configured to do so, the network node 520 may transmit the buffered monitoring data to a data collection entity, e.g., TCE 530 or an NWDAF in the mobile network.

[0131] In some embodiments, a performance monitoring report may comprise one or more of the following element(s):- Measurements used to produce inference output- Inference output, for example if the network is not configured to store historical inference outputs and zero or more of the following elements, which the UE may optionally be configured to report, some of which could become mandatory if range selection and specific sample selection are supported:- System frame number- Slot number- Symbol number- Timestamp and zero or more of the following elements, which the UE may optionally be configured to report, to help identify an underperforming model or to forward buffered monitoring data to a server which trains UE-side AI / ML models:- AI / ML model ID- AI / ML model dataset ID- Associated ID - a standardized identifier to map to data collection configurations for training UE-side AI / ML models

[0132] In some embodiments, the monitoring report could carry up to the number of measurement and inference instances supported by the UE’s buffer and could be sent in parts in case the monitoring report is bigger than a maximum message / container size allowed by the control plane or user plane.

[0133] Referring now to FIG. 6, The following table visualizes the sample storage in the buffer. As an example, identifier A is an “Associated Id”, mapping the collected data to a particular configuration, whereas identifiers B and C, and others, could be a model-specific identifier, or a cell identifier in case measurements are made and inference is performed on cells other than the serving cell.

[0134] A potential problem arises when UEs 510 in RRC connected mode perform radio measurements and store these measurements locally at the UE 510 because of finite buffer space for storing the measurements and the periodic need to clear the buffer. However, after the UE 510 deletes / clears the local buffer of UE performed radio measurements, the network may request a UE measurements report containing recent relevant radio measurements, which the UE 510 no longer has because it deleted / cleared the local buffer of UE-performed radio measurements.

[0135] One possible approach for solving this problem is to use a circular buffer at the UE 510, such as a circular buffer having a variable depth. The UE 510 can therefore store radio measurements locally in the variable-depth circular buffer and deletion / clearing of the variable-depth circular buffer is carried out in tranches, meaning that only a portion of the UE-performed radio measurements stored in the variable-depth circular buffer are cleared / deleted at each instance of clearing / deletion; namely, it is the oldest UE-performed radio measurements stored in the circular buffer that are cleared at each instance of deletion / clearing.

[0136] In some embodiments, UE measurement information, which is input to a ML model in the UE 510, could be in a proprietary form specific to the ML model’ s development. As such, what is stored in the circular buffer could be in a form useful to the mobile network, e.g., which could be used by the mobile network in certain embodiments to calculate ML model / inference performance metrics or to input into its own algorithm to compare to the UE’s ML model output information provided by the UE 510.

[0137] In some embodiments, the inference output of a ML model in the UE 510 could be useful for finetuning or updating the ML model. Further, in some embodiments, the inference output of a ML model that has been processed into a form useful to the mobile network could be useful for making monitoring decisions. Each of these forms can be considered to be outputs of the ML model.

[0138] In some embodiments, the inference outputs from the ML model are stored directly in the circular buffer at the UE 510. In other embodiments, the inference outputs from the ML model are postprocessed by the UE 510 before being stored in the circular buffer. In some embodiments, the postprocessing of inference outputs from the ML model and / or the inference outputs themselves can be considered ‘inference output information.’ In some embodiments, when inference outputs from the ML model are postprocessed for form the inference output information, the inference output information can correspond to, be associated with, be paired with, be connected to, be correlated with, be mapped to, or otherwise be connected to one or more inference outputs from the ML model. In some embodiments, the inference output information can comprise one or more of: predicted channel state information, predicted beam index information, predicted beam layer 1 reference signal received power (Ll-RSRP) information, predicted UE latitude information, predicted UE longitude information, predicted line of sight information, or predicted nonline of sight information.

[0139] In some embodiments, the UE 510 can be configured to store UE measurement information alongside the inference output information in the circular buffer. For example, the UE 510 can be configured to store, in the circular buffer, UE measurement information comprising one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR values, one or more Ll-RSRP values, one or more filtered Layer 3 -RSRP (L3- RSRP) values, one or more Positioning Reference Signal-RSRP (PRS-RSRP) values, one or more PRS-Reference Signal Received Path Power (PRS-RSRPP) values, one or more powerdelay profiles, one or more power delay values, one or more channel impulse responses, one or more frequency error measurements, or one or more timing error measurements.

[0140] In some embodiments, the UE 510 can be configured to receive one or more indications or configuration information from the

[0141] Referring now to FIG. 7, illustrated is a process for storing radio measurements locally at the UE 510 using a circular buffer approach. A circular buffer is a data storage structure that enables the continuous collection of data by overwriting old data. Take for example a circular buffer with four entries. The first entry in the buffer has an index=0 and the last entry in the buffer has an index=3. To calculate the next buffer index, inext, calculate (icurrent + 1) % (sbuffer + 1), where iCUrrent is the current buffer index, Sbuffer is the size or depth of the buffer, i.e., how many samples it can store, and % is the modulus operator, which simply returns the remainder of a division operation.

[0142] The example in Table 1 below illustrates the concept further using the same buffer depth of 4.Table 1

[0143] When the buffer index reaches the end, i.e., icurrent = 3, the next index, i.e., inext, resets to the start of the buffer, z = 0.

[0144] Performance monitoring is an important function to enable AI / ML functionalities, enabled by AI / ML models, to ensure that their use does not negatively impact NW performance. It is currently understood that AI / ML models do not generalize well to radio access network (RAN) level use cases such as channel state information (CSI) prediction and compression, beam management (BM), and positioning. As a result, performance monitoring is used to determine when an AI / ML functionality should be switched from or deactivated, falling back to a non-AI / ML method. The outputs of theperformance monitoring process can also be used as input to model finetuning, wherein previously unseen scenarios are used to update to the model to increase its applicability.

[0145] In some embodiments, radio measurements can be made of select cells and / or reference signals (RS), but not others. The radio measurements can be fed into an AI / ML model that generates inferred radio measurements about other cells or RS. Along with the inferred radio measurements, the AI / ML model can output various information or data about the performance of the AI / ML model itself. In other embodiments, the UE 110 or networkside device such as network node 120 / 130 can review the output from the AI / ML in view of the radio measurements input to the AI / ML and information about the AI / ML, such as an AI / ML indicator / identifier and / or an indication of at least one dataset used to train the AI / ML. Based on this review, the UE 110 or network node 120 / 130 can determine the performance of the AI / ML model itself. The information about the AI / ML model can comprise, e.g., an indication of the AI / ML model ID, an indication of one or more indexes supported by the AI / ML model, an indication of an AI / ML model type, and / or other information that can be used to identify the radio measurements in the UE measurement report or ML model monitoring report as belonging to a specific AI / ML model. In some embodiments, the performance of the AI / ML model itself can be determined, generally, based on confidence metrics, uncertainty metrics, metrics that indicate whether model outputs should be trusted, and / or the like.

[0146] In some embodiments, the radio measurements input to the AI / ML model can be collected / measured during a particular (e.g., predetermined) time period. In some embodiments, radio measurements can be collected / measured before, during, and / or after the UE 110 receives a request for the UE to monitor the AI / ML model performance.

[0147] In some embodiments, the approach described herein enables the UE 110 to be configured to store historical measurement and inference data from an AI / ML functionality or model for future reporting.

[0148] It is assumed that UE-based and NW -based performance monitoring exist and that the output of the performance monitoring procedure(s) can be used as triggers to retrieve a historical measurement and inference data log from a UE. For UE-based performance monitoring, the UE could be configured with one or more performance thresholds which could detect instantaneous poor performance degradation, or performance degradation over a period of time. NW-based performance monitoring could be considered proprietary, but itcould result in a request for monitoring data. In both cases, the UE would transmit a monitoring data report to the NW.

[0149] To avoid burdensome memory requirements for data logging, the UE 110 periodically overwrites its performance monitoring data buffer, as in a circular buffer, which could have a variable depth. Performance monitoring reports could include the entire buffer, a range of samples from the buffer, or specific samples from the buffer.

[0150] The circular buffer mechanism limits the storage burden on the UE, and the delayed reporting mechanism optimizes the use of air resources by withholding measurements during times of good performance, while allowing for their transmission when part of a log containing samples of poor performance.

[0151] Because the data used for monitoring could also help improve existing models through finetuning, the network node 120 could forward the collected data to a server for training models, optionally tagged with information that could be used to identify the model.

[0152] In some embodiments, the method may rely on RAN signaling in downlink (DL), e.g., in an RRC message, to facilitate the configuration of the UE 110, which enables radio measurement logging in RRC connected state and may disable transmitting radio measurement results that would otherwise be reported in real-time. In some embodiments, the method introduces logged radio measurement report configuration that enables UE to log measurements, e.g., for CSI and beam prediction, also for UEs 110 in RRC connected state. The RAN (e.g., network node 120) may, in some embodiments, determine that logged data collection is required and send a logged radio measurement report configuration, so that the UE’s physical layer data may - at least in part in some embodiments - be generated by radio measurements, but not be transmitted immediately after generation (e.g., as legacy immediate radio measurement reports), but rather once one or more logged reporting condition(s) for logged radio measurement reports are met.

[0153] This means that the UE procedures applied after configuration of logging for radio measurements imply that even though the physical layer of the UE 110 may generate the measurements (or measurement data), the generated measurements are first processed internally in the UE 110. The UE 110 may also store the radio measurements in the UEs memory until a predefined logged radio measurement reporting condition is detected.

[0154] In some embodiments, lower layer data (i.e. the data obtained by performing the radio measurement including postprocessing / calculation of parameters) can be provided internally to an upper layer of the UE 110 (e.g., layer 3, RRC) for storing (logging) the data.The generation and transmission of the logged radio measurement report may then happen in the upper layer only after transmission criteria for logged radio measurement reports is satisfied. Alternatively, the radio measurements may be stored in the lower layer (e.g. LI, L2) and logged radio measurement reports are provided to the upper layer upon a reporting event is met or when the upper layer requests the logged radio measurement report. The UE 110 may also, in some embodiments, specifically preprocess the radio measurements and / or the logged radio measurement reports according to a specific purpose as configured by the network node 120 (e.g., internal AI / ML operations on data, data formatting, etc.).

[0155] In some example circumstances, the network node 120 (e.g., gNB) or the UE 110 may detect probable AI / ML model performance degradation (or otherwise determine it may have occurred). Such a degradation of AI / ML-enabled functionality at the UE 110 may reflect a problem with the AI / ML model training, the dataset(s) used for performing AI / ML model training, or the like. In some embodiments, in an instance in which the UE 110 detects such AI / ML model performance degradation, the UE 110 may prepare an AI / ML model performance report to be sent to the network node 120 / 130. In other embodiments, in an instance in which the UE 110 detects such AI / ML model performance degradation, the UE 110 may simply provide an indication of the detected possible AI / ML model performance degradation to the network node 120 / 130, thereby allowing the network to determine whether the network node 120 will request comprehensive radio measurement data and AI / ML model outputs / inferred measurements and / or information about the AI / ML model from the UE 110. In some embodiments, the network node 120 may then request comprehensive (all or most) radio measurement data and the like from the UE 110. In other embodiments, the network node 120 may not request any radio measurement data or may only require a limited set of radio measurement data. The network node 120 triggers logged data collection from the UE 110 to get the sought metrics.

[0156] Alternatively, in other example circumstances, the network node 120 (e.g., gNB) may perform AI / ML-enabled functionality training or retraining. As a result, the network node 120 may request comprehensive radio measurement data from the UEs 110. In some embodiments, the network node 120 may then not request any radio measurement data in a real-time or may only require a limited set of radio measurement data in real-time. The network node 120 triggers logged data collection from the UE 110 to get the sought metrics for AI / ML training purposes.

[0157] For example, if the AI / ML-enabled functionality is for beam management, the network node 120 may trigger CSI-reports for data collection from the UE to get CSI data and beam measurements. CSI measurement results do not have to be delivered in real-time (immediately) but can be combined with logging triggers and reporting triggers to obtain logged CSI-reports. For this purpose, the network node 120 may send an RRC configuration with CSI-report configuration towards the UE 110.

[0158] 3GPP TR 38.843 studied AI / ML for the air interface, including CSI prediction, CSI compression, beam prediction, and positioning use cases. Therein, the following definition is provided for model monitoring: A procedure that monitors the inference performance of the AI / ML model, and the following definition is provided for model update: process of updating the model parameters and / or model structure of a model, which could take monitoring data as input. The following information from the TR describe monitoring and its purpose. “Monitoring Data: Data needed as input for the Management of AI / ML models or AI / ML functionalities.” - for example, monitoring outputs could be used to make a decision to activate or deactivate an AI / ML model or functionality.

[0159] The following metrics / methods for AI / ML model monitoring in lifecycle management per use case are considered:- Monitoring based on inference accuracy, including metrics related to intermediate KPIs

[0160] For beam management:- For monitoring at the gNB side, and if needed, calculated performance metrics or data required for performance metric calculation, can at least be generated by the gNB. - For this, measurements from the UE could be required.- Type 1 performance monitoring:1) Configuration / Signalling from gNB to UE for measurement and / or reporting1. UE may have different operations a. Option 1 (NW-side performance monitoring): UE sends reporting to NW (e.g., for the calculation of performance metric at NW)- UE reporting of beam measurement s) based on a set of beams indicated by gNB.

[0161] For positioning:- If monitoring based on model output: e.g., estimated UE location corresponding to model output for direct AI / ML positioning, estimated intermediate parameter(s) corresponding to model output for AI / ML assisted positioning, ground-truth label corresponding to model inference output for both direct and AI / ML assisted positioning,- Signalling from monitoring entity to request measurement(s) (if needed).

[0162] In some embodiments, the network node 120 can configure the UE 110 for measurement and reporting by providing a buffered monitoring configuration, which may be provided as a field in a measurement reporting configuration such as CSI-ReportConfig, ReportConfigNR, or LPP RequestLocationlnformation.

[0163] In some embodiments, the buffered monitoring configuration can include an indication of Buffer Depth. Buffer Depth could be provided in terms of number of samples, where a sample could be defined as, for one AI / ML model output instance, AI / ML model inputs, AI / ML model outputs and any of the optional elements related to timestamping the samples.

[0164] In some embodiments, the buffered monitoring configuration can include an indication of Trigger Configuration (Trigger Config). Several trigger configuration options are as follows:- Performance Based - If not otherwise triggered, the UE would transmit its monitoring buffer to the NW once a performance degradation threshold has been reached. The performance monitoring threshold could, for example, be defined over a period of time in which the AI / ML functionality performed under a configured performance threshold.- Sample Based - If not otherwise triggered, the UE would transmit its monitoring buffer to the NW after the configured number samples had been observed, regardless of the buffer size.- Timer Based - If not otherwise triggered, the UE would periodically transmit its monitoring buffer to the NW.

[0165] In some embodiments, the buffered monitoring configuration can include Model Information. This information / field may enable the UE 110 to provide some proprietary information identifying the model or dataset being used to perform inference.

[0166] In some embodiments, if multiple buffered monitoring configurations are possible, the following field may be included in BufferedMonitoringConfig.

[0167] In some embodiments, to manage multiple BufferedMonitoringConfigs, the AddMod and Remove mechanism, such as defined in 3GPP TS 38.331, could be used, wherein AddMod is used to add or modify a BufferedMonitoringConfig, identified by a BufferedMonitoringConfigld, where a modification is performed by providing a fully new or partially modified (through the use of OPTIONAL fields) configuration. Remove is used to remove a BufferedMonitoringConfig, identified by a BufferedMonitoringConfigld.

[0168] The buffered monitoring report could be requested through a variety of different mechanisms, such as:- MAC-CE1) A single bit could trigger the transmission of the full buffer if only one monitoring report is configured.2) Multiple bits could trigger the transmission of the full buffer if multiple monitoring reports are configured. The multiple bits would identify a monitoring report configuration by a monitoring report configuration ID.- RRC Message - A UE-InformationRequest, or other RRC message could contain a buffer request structure with the following fields.1) Monitoring Report Config ID, if multiple monitoring reports are configured2) Sample Selection1. Full buffer2. Range of samples from the buffer - e.g., starting and ending at system frame 512 and 768 or starting at a time provided by a timestamp and a number of milliseconds thereafter3. Selection of samples from the buffer - e.g., a list of system frame number and / or slot number and / or OFDM symbol number identifying the inference output transmission time.

[0169] In some embodiments, an RRC Message, such as IdM for motion Response, or another RRC message could contain a buffered monitoring report structure with the fields described in the Embodiment for Buffered Monitoring Report Contents, on a per-sample basis, i.e., each sample could include lists of multiple input measurements and inference outputs.

[0170] In some embodiments, an LPP Message, such as LPP ProvideLocationlnformation or another LPP message could contain a buffered monitoring report structure as described in the RRC example.

[0171] In some embodiments, a Layer 1 (LI) Message (physical layer message) can be used that, e.g., contains a buffered monitoring report structure such as that illustrated the RRC example in FIG. 6.

[0172] In some embodiments, a Trace Message, such as TraceRecord or another trace message could contain a buffered monitoring report structure as described in the RRC example illustrated in FIG. 6.

[0173] If a NW node is configured by another entity such as an Operations,Administration, and Management (0AM) entity to collect performance monitoring data from UEs, the NW node could be configured to forward performance monitoring data to adata collection entity such as a Trace Collection Entity (TCE) or a Network Data Analytics Function (NWDAF).

[0174] The configuration toward the NW could contain filtering information such as:- Model ID- Dataset ID- Associated ID- UE hardware model, which could be represented by a Type Allocation Code (TAC)- Time of day information- Cell information, e.g., cell ID, use case- Measurement types

[0175] In some embodiments, the TCE 430 may transfer the collected data to a server for training AI / ML models for the use of finetuning, for example to address performance issues found in scenarios in which the AI / ML model does not perform as well as desired, for example due to insufficient training in similar scenarios.

[0176] In some embodiments, the radio measurement report configuration may indicate one or more logged radio measurement conditions. Logged radio measurement conditions may relate to any conditions indicating what to log, when to log, and how to perform radio measurements. The logged radio measurement report configuration may be received in a layer 3 RRC message.

[0177] Further, an indication from the network to activate logged radio measurements may take several forms and may, e.g., be the logged radio measurement report configuration and / or a separate logged radio measurement report configuration activation command. Hence, the UE 110 may start logging directly after receiving the logged radio measurement report configuration or wait until it receives a respective logged radio measurement report configuration activation command.

[0178] The one or more logged radio measurement conditions may comprise at least one of a logged radio measurement trigger defining when to log radio measurements, a report defining a content of the logged radio measurements, and a logged radio measurement termination condition. The logged radio measurement trigger may comprise a periodic, an aperiodic, or a semipersi stent logging trigger indicating when to log or when to start logging of radio measurements. Hence, the logged radio measurement trigger indicates which radiomeasurements to obtain and when the radio measurements are captured, processed, and stored.

[0179] A report quantity, i.e., content of the radio measurements, e.g., measurement values and / or calculated metrics, may be one or more of CQI, PMI, CRI, SSBRI, LI, RI, or other values / metrics obtained based on radio measurements. The logging termination condition may define when to stop logging measurements and may comprise at least one of a maximum payload size for the logged data report, a maximum memory size to be used for the logged radio measurements, or maximum time for logging the measurements.

[0180] The logged radio measurement report configuration may further indicate one or more logged radio measurement reporting conditions. The UE 110 may, in response to determining that at least one (e.g., one, multiple, or all as being defined or specified before) of the one or more logged radio measurement reporting conditions is met, transmit a logged radio measurement report comprising the logged radio measurement data to the network node. In other words, in some embodiments, it may be defined that the logged radio measurement report is only transmitted if all of the defined one or more logged radio measurement reporting conditions are met, whereas in other embodiments, it may be defined that only some of the logged radio measurement reporting conditions have to be met for reporting. The one or more logged radio measurement reporting conditions may define an opportunity when the logged radio measurement report is to be transmitted and an uplink resource for transmission of the logged radio measurement report. The opportunity may comprise at least one of a periodical condition and an event-based condition. The uplink resource may, e.g., be PUCCH, PUSCH or PDCCH, which may be a periodic resource, an aperiodic resource, or a semipersi stent resource. This means that the logged radio measurement reporting conditions defined whether the logged radio measurement reports are transmitted periodically (i.e., at a specific periodicity) or event-triggered (e.g., threshold of maximum logged data has been reached, network node 120 requests logged radio measurement report, or the like).

[0181] As said, the logged radio measurement report configuration may be received as an RRC message, e.g., in an RRC IE, such as the CSI-ReportConfig IE as described above. The UE 110 may then provide the logged radio measurement report configuration (or at least part of it) from layer 3 to layer 1 because the radio measurements are performed in layer 1. The logged radio measurement report configuration (or a respective part of it) may then be stored in layer 1 and read when required for performing radio measurements.

[0182] The logged radio measurement report may be at least in part transmitted in a layer 3 RRC message and / or at least in part transmitted as layer 1 PHY uplink control information (e.g., via PUSCH, PUCCH, and the like). In some examples, the logged radio measurement report, which may be significantly larger than immediate radio measurement reports, may be sent in two distinct messages, one message comprising a part of the content on layer 1 and one message comprising another part of the content on layer 3. Where to transmit the logged radio measurement report may also be dynamic or configurable such that up to a specified size (e.g., less or equal to 1 kBit) of the report, the logged radio measurement report may be transmitted on layer 1 (e.g., as uplink control information on PUSCH / PUCCH), and if the report has a size larger than the specified size (e.g., larger 1 kBit onwards), the logged radio measurement report may be transmitted on layer 3.

[0183] If the logged radio measurement report is (at least in part) transmitted via a layer 3 RRC message, the UE 110 may obtain radio measurements according to the logged radio measurement report configuration (i.e., according to the logged radio measurement conditions), and provide at least a part of radio measurements from layer 1 to layer 3 for logging (i.e., storing) over time.

[0184] Radio measurements may relate to measured channel properties, e.g., raw data and / or calculated metrics thereof, additionally with further included data and possibly in a format similar to a legacy immediate radio measurement report. In other embodiments, the UE 110 may not provide (at least a part of) the radio measurements from the lower to the upper layer but may generate (at least a part of) the final logged radio measurement report in layer 1 and provide the generated (at least a part of the) logged radio measurement report from layer 1 to layer 3 for transmission. The UE 110 may provide the logged radio measurement report and / or the radio measurements from the lower layer to the upper layer upon logged radio measurement condition(s) apply (i.e., for each measurement sample in time), upon a logging termination condition is met, or upon one or more logged radio measurement reporting conditions are met. The upper layer may also request logged radio measurement data and / or logged radio measurement report if it determines that one or more logged radio measurement reporting condition are met. As is apparent, if radio measurements are stored in part in layer 1 and in part in layer 3 in these embodiments, in which the logged radio measurement report is transmitted via layer 3 RRC message, the radio measurements stored in the different layers may be mutually exclusive to avoid double storing and reporting.

[0185] In some embodiments, the UE 110 may also receive, from the network node, at some point in time a logged radio measurement report deactivation command, which deactivates transmission of logged radio measurement reports with respect to all or part of respective configured report quantities. This enables the network node 120 to also stop data collection and - possibly - start / continue legacy immediate radio measurement reporting.

[0186] If the UE 110 was already configured for immediate radio measurement reporting (e.g., as described above for providing real-time CSI-reports), the indication to activate logged radio measurements (e.g., the logged radio measurement configuration or a separate MAC CE) further may be an immediate radio measurement report deactivation command indicating a deactivation of immediate radio measurements. Alternatively, the deactivation of immediate radio measurements may be indicated by a MAC CE indicating an immediate radio measurement report deactivation command. Yet alternatively, the immediate radio measurement reports may not be deactivated or only part of the content may be omitted. This may be considered in the logged radio measurement report configuration, in an immediate radio measurement report configuration, in another radio measurement report configuration (e.g., which comprises both, the logged radio measurement report configuration and the immediate radio measurement report configuration), or yet a separate message. If the UE 110 receives such an immediate radio measurement report deactivation command, the UE 110 may transition from immediate radio measurements to logged radio measurements.

[0187] The UE 110 may receive the immediate radio measurement report configuration, which may indicate one or more immediate reporting conditions in the same or a separate message as the logged radio measurement report configuration. The UE 110 may then directly or upon receiving an immediate radio measurement report activation command (for activating the immediate radio measurement report configuration), perform radio measurements and, in response to determining that at least one (e.g., one, multiple, or all as being defined or specified before) of the one or more immediate reporting conditions is met, transmit an immediate radio measurement report comprising the latest radio measurements to the network node.

[0188] In some embodiments, the logged radio measurements are configured to be performed more frequently over time, reported less frequently and / or reported after a greater period in time than the immediate radio measurements. For example, the logged radio measurement reports may be configured with a lower periodicity and / or a larger offset inslots for reporting over physical uplink control channel, a larger timing offset for reporting over physical uplink shared channel, and / or a longer layer 3 timer for reporting over dedicated control channel than the immediate radio measurement reports.

[0189] In some further embodiments, the logged radio measurement reports may include at least one additional report quantity compared to the immediate radio measurement reports. This flexibility to configured reporting times and report quantities differently for logged and immediate radio measurement reporting is achieved by providing the UE 110 with two distinct configurations, in some embodiments comprised in one RRC message, e.g., in a channel state information report configuration information element, CSI-ReportConfig IE (as described above), which allow activation and deactivation of one or the other flexibly, e.g., by sending respective activation and deactivation commands. If the UE 110 is configured to log and immediately report radio measurements during the same time period, the logged radio measurement reports transmission of immediate radio measurement reports, which are usually smaller than the logged radio measurement reports) may be prioritized over transmission of logged radio measurement reports. Hence, if both transmissions are scheduled for the same transmission time, the immediate radio measurement report is prioritized due to its time sensitivity. The transmission of the logged radio measurement report may be postponed, e.g., to the next possible transmission slot or according to configured conditions.

[0190] A content of the logged radio measurements may in some embodiments comprise at least one of channel state measurements and beam measurements. As is apparent, the content may, in some embodiments, be defined by the logged radio measurement conditions, namely, as report quantities. The channel state measurements may comprise one or more of channel quality information, a precoding matrix indicator, a channel state information resource indicator, a synchronization signal block, a layer indicator, a rank indicator, and a level 1 reference signal received power. This said, the channel state measurements may relate to CSI report quantities as described above. The beam measurements may comprise beam predictions, e.g., determined by the UE 110 by AI / ML algorithms.

[0191] The network node 120 may be configured to determine that data collection of logged radio measurement is to be initiated (at some point in the future) with the UE 110. Hence, the network node 120 may determine that it will at some point in the future (e.g., now or later) plan to obtain logged radio measurements from the UE 110, e.g., because thenetwork node 120 needs training data for AI / ML model training or the like. In some embodiments, the network node 120 may also generally configure at least some UEs for logged radio measurement reporting.

[0192] In some embodiments, the network node 120 then transmits logged radio measurement report configuration to the UE 110. In some embodiments, this may already indicate to the UE 110 to perform logged radio measurements but the network node 120 may also transmit a further logged radio measurement report activation command to the UE 110. The logged radio measurement report configuration may be received in a layer 3 radio resource control message.

[0193] The logged radio measurement report configuration may further indicate one or more logged radio measurement reporting conditions (e.g. those described with respect to FIG. 4) and the network node 120 may further receive a logged radio measurement report comprising logged radio measurements from the user equipment according to the one or more logged radio measurement reporting conditions. The logged radio measurement report may be at least in part received in a layer 3 radio resource control message and / or the logged radio measurement report may also be at least in part transmitted as layer 1 physical layer uplink control information. The network node 120 may also transmit, to the UE 110, a logged radio measurement report deactivation command for deactivating transmission of logged measurement reports with respect to all or part of respective configured report quantities.

[0194] The indication to activate logged radio measurements may further be an immediate radio measurement report deactivation command indicating a deactivation of immediate radio measurements. Alternatively, a deactivation of immediate radio measurements may be indicated by a MAC CE indicating an immediate radio measurement report deactivation command. In embodiments, the network node 120 may transmit, to the UE 110, an immediate radio measurement report configuration indicating one or more immediate reporting conditions for performing radio measurements and correspondingly receives, from the UE 110, an immediate radio measurement report comprising the latest radio measurements according to the immediate radio measurement report configuration.

[0195] In some embodiments, the network node 120 may transmit, to the UE 110, an immediate radio measurement report configuration activation command for activating the immediate radio measurement report configuration for starting performing radio measurements and transmitting immediate radio measurement reports. If the network node 120 has configured the UE 110 with immediate radio measurement reporting, the loggedradio measurements as configured with the logged radio measurement report configuration may be performed more frequently in time but reported less frequently in time than the immediate radio measurements. In other words, the frequency (e.g., periodicity in symbols, slots, etc.) at which radio measurements for logging are obtained can be higher than the frequency at which immediate radio measurements are obtained. At the same time, the transmission rate, i.e., how often reports are sent to the network node 120, can be lower for the logged radio measurements because the radio measurements are logged and not reported at any reporting opportunity as the immediate radio measurement reports.

[0196] The immediate measurement report configuration and the logged radio measurement report configuration may be received within one message. The logged radio measurement configuration and / or the immediate measurement report configuration may be included in a channel state information report configuration information element, CSI- ReportConfig IE as described above.

[0197] The logged radio measurements may comprise at least one of channel state measurements and beam measurements. The channel state measurements may further comprise one or more of channel quality information, a precoding matrix indicator, a channel state information resource indicator, a synchronization signal block, a layer indicator, a rank indicator, and a level 1 reference signal received power. The beam measurements may comprise beam predictions.

[0198] In an 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 method steps as disclosed herein may include software and / or hardware components (e.g., 110, 120, 220, 230, 320, 330) of the apparatus. For example, at least one processor and at least one memory storing thereon computer program codes can comprise 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 areseparate means for performing A, B and C, or partially or fully overlapping means for performing A, B, C.

[0199] Referring now to FIG. 8, a process flow diagram of a method 600 is illustrated. The method 600 can comprise: using a machine learning (ML) model in a user equipment (UE) to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network, at 602. The method 600 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements, at 604. The method 600 can further comprise: in response to a UE measurement reporting event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information, at 606.

[0200] Referring now to FIG. 9, a process flow diagram of a method 700 is illustrated. The method 700 can comprise: receiving, at a user equipment (UE), from a mobile network, an indication configuring a circular buffer in the UE to have a specified buffer depth, at 702. The method 700 can further comprise: using a machine learning (ML) model in the UE to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network, at 704. The method 700 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements, at 706. The method 700 can further comprise: in response to a UE measurement reporting event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information, at 708.

[0201] Referring now to FIG. 10, a process flow diagram of a method 800 is illustrated. The method 800 can comprise: receiving, at a user equipment (UE), from a mobile network, an indication configuring the UE to maintain, with UE measurement information and inference output information in a circular buffer in the UE, timing-related information identifying a respective time at which respective of the inference output information was generated, at 802. The method 800 can further comprise: using a ML model in the UE to generate inference output(s) based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in the mobile network, theinference output information being associated with the inference output(s), at 804. The method 800 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information, UE measurement information associated with the plurality of UE measurements, and timing-related information associated with the inference output information, at 806. The method 800 can further comprise: in response to a UE measurement reporting event, transmitting a UE measurement report comprising at least a portion of the inference output information, the UE measurement information corresponding to the at least a portion of the inference output information, and the timing-related information, at 808.

[0202] Referring now to FIG. 11, a process flow diagram of a method 900 is illustrated. The method 900 can comprise: receiving, at a user equipment (UE), from a mobile network, an indication configuring the UE to maintain, with UE measurement information and inference output information in a circular buffer in the UE, one or more identifiers associated with one or more machine learning (ML) models used in the UE to perform inference, at 902. The method 900 can further comprise: using a ML model in the UE to generate inference output(s) based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in the mobile network, the inference output information being associated with the inference output(s), at 904. The method 900 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information, UE measurement information associated with the plurality of UE measurements, and one or more identifiers associated with the ML model used to generate the inference output(s), at 906. The method 900 can further comprise: in response to a UE measurement reporting event, transmitting a UE measurement report comprising at least a portion of the inference output information, the UE measurement information corresponding to the at least a portion of the inference output information, and the one or more identifiers associated with the ML model used to generate the inference output(s), at 908.

[0203] Referring now to FIG. 12, a process flow diagram of a method 1000 is illustrated. The method 1000 can, optionally, comprise: providing, to an element or function in the mobile network, a ML functionality performance report comprising the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information, at 1002. The method 1000 can comprise or further comprise: receiving, from a user equipment (UE), a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements ofone or more cells in the mobile network, the inference output information being inferred by a machine learning (ML) model in the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network, at 1004.

[0204] Referring now to FIG. 13, a process flow diagram of a method 1100 is illustrated. The method 1100 can, optionally, comprise: providing, to the UE, an indication configuring the UE to perform a lifecycle management (LCM) action with regard to using an ML model of the ML functionality in the UE, at 1102. The method 1100 can comprise or further comprise: receiving, at a base station, from a user equipment (UE), a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements of one or more cells in the mobile network, the inference output information being inferred by a machine learning (ML) model in the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network, at 1104.

[0205] Referring now to FIG. 14, a process flow diagram of a method 1200 is illustrated. The method 1200 can comprise: receiving, at a base station, from a user equipment (UE), a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements of one or more cells in the mobile network, the inference output information being inferred by a machine learning (ML) model in the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network, at 1202. The method 1200 can, optionally, further comprise: storing, at the base station, the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information, at 1204.

[0206] Referring now to FIG. 15, a process flow diagram of a method 1300 is illustrated. The method 1300 can, optionally, comprise: providing, to a user equipment (UE), an indication configuring the UE to perform inference, using a machine learning (ML) model of a ML functionality in the UE, at 1302. The method 1300 can comprise or further comprise: receiving, at a base station, from the UE, a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements of one or more cells in the mobile network, the inference output information being inferred by the ML model of the ML functionality in the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network, at 1304.

[0207] Referring now to FIG. 16, a process flow diagram of a method 1400 is illustrated. The method 1400 can comprise: using a machine learning (ML) model in a user equipment (UE) to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network, at 1402. The method 1400 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements, at 1404. The method 1400 can, optionally, further comprise: determining the ML model performance of the ML model, at 1406. The method 1400 can, optionally, further comprise: determining whether the ML model performance of the ML model is below the ML model performance threshold, which constitutes a UE performance monitoring event, at 1408. The method 1400 can further comprise: in response to a UE performance monitoring event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information, at 1410.

[0208] Referring now to FIG. 17, a process flow diagram of a method 1500 is illustrated. The method 1500 can comprise: using a machine learning (ML) model in a user equipment (UE) to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network, at 1502. The method 1500 can further comprise: maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements, at 1504. The method 1500 can, optionally, further comprise: receiving, from the mobile network, a UE measurement reporting request, which represents a UE performance monitoring event, at 1506. Alternatively, the method 1500 can, optionally, further comprise: determining an occurrence of a measurement reporting period which represents a UE performance monitoring event, the measurement reporting period being indicated to the UE by the mobile network during prior signaling, at 1508. The method 1500 can further comprise: in response to a UE performance monitoring event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information, at 1510.

[0209] The herein described procedures may be applied per model or per functionality level (identified by an identifier) or across models or functionalities of a given entity, e.g.,as a UE feature. It should be understood that the apparatuses described herein may comprise or be coupled to other units or modules etc., such as radio parts or radio heads, used in or for transmission and / or reception. Although the apparatuses have been described as one entity, different modules and memory may be implemented in one or more physical or logical entities.

[0210] It is noted that although embodiments have been described in relation to LTE and 5G NR, similar principles may be applied in relation to other networks and communication systems where enforcing fast connection re-establishment is required. Therefore, although certain embodiments were described above by way of example with reference to certain example architectures for wireless networks, technologies and standards, embodiments may be applied to any other suitable forms of communication systems than those illustrated and described herein.

[0211] It is also noted herein that although the above describes example embodiments, there are several variations and modifications which may be made to the disclosed solution without departing from the scope of the subject disclosure.

[0212] In general, the various example embodiments may be implemented in hardware or special purpose circuits, software, logic or any combination thereof. Some aspects of the subject disclosure may be implemented in hardware, whereas other aspects may be implemented in firmware or software which may be executed by a controller, microprocessor or other computing device, although the subject disclosure is not limited thereto. Although various aspects of the subject disclosure may be illustrated and described as block diagrams, flow charts, or using some other pictorial representation, it is well understood that these blocks, apparatus, systems, techniques or methods described herein may be implemented in, as non-limiting examples, hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof.

[0213] Example embodiments of the subject disclosure may be implemented by computer software executable by a data processor of the mobile device, such as in the processor entity, or by hardware, or by a combination of software and hardware. Computer software or program, also called program product, including software routines, applets and / or macros, may be stored in any apparatus-readable data storage medium and they comprise program instructions to perform particular tasks. A computer program product may comprise one or more computer- executable components which, when the program is run,are configured to carry out embodiments. The one or more computer-executable components may be at least one software code or portions of it.

[0214] Further in this regard it should be noted that any blocks of the logic flow as in the figures may represent program processes, or interconnected logic circuits, blocks and functions, or a combination of program processes and logic circuits, blocks and functions. The software may be stored on such physical media as memory chips, or memory blocks implemented within the processor, magnetic media such as hard disk or floppy disks, and optical media such as for example DVD and the data variants thereof, CD. The physical media is a non-transitory media.

[0215] The memory may be of any type suitable to the local technical environment and may be implemented using any suitable data storage technology, such as semiconductorbased memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory. The data processors may be of any type suitable to the local technical environment, and may comprise one or more of general- purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), FPGA, gate level circuits and processors based on multi-core processor architecture, as non-limiting examples.

[0216] Example embodiments of the subject disclosure may be practiced in various components such as integrated circuit modules. The design of integrated circuits is by and large a highly automated process. Complex and powerful software tools are available for converting a logic level design into a semiconductor circuit design ready to be etched and formed on a semiconductor substrate.

[0217] Moreover, in accordance with the foregoing description, the embodiments described herein reflect possible embodiments of the herein presented solution, which are further combinable between embodiments and / or between sets of embodiments. These embodiments do not define the entire scope of the disclosure of this application; nor do these embodiments define or limit the scope of the invention.

[0218] The above-noted aspects and features may be implemented in systems, apparatuses, methods, articles and non-transitory computer-readable media depending on the desired configuration. The subject disclosure may be implemented in and used with a number of different types of devices, including but not limited to cellular phones, tablet computers, wearable computing devices, portable media players, and any of various other computing devices.

[0219] Even though the invention has been described above with reference to an example according to the accompanying drawings, it is clear that the invention 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.

[0220] With regard to the illustrated signal flow diagrams and flowcharts depicting methods provided in the drawings and described above, it will be understood that each block or signal and combination of blocks and signals may be implemented by various means, such as hardware, firmware, processor, circuitry, and / or other communication devices associated with execution of software including one or more computer program instructions.

[0221] For example, one or more of the procedures described above may be embodied by computer program instructions. In this regard, the computer program instructions which embody the procedures described above may be stored by a memory of the apparatus employing an example embodiment and executed by processing circuitry, such as a processor. As will be appreciated, any such computer program instructions, such as instructions, may be loaded onto a computer or other programmable apparatus (for example, hardware) to produce a machine, such that the resulting computer or other programmable apparatus implements the functions specified in the flowchart blocks. These computer program instructions may also be stored in a computer-readable memory that may direct a computer or other programmable apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture the execution of which implements the function specified in the flowchart blocks. The computer program instructions may also be loaded onto a computer or other programmable apparatus to cause a series of operations to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide operations for implementing the functions specified in the flowchart blocks.

[0222] Accordingly, blocks of the flowcharts support combinations of means for performing the specified functions and combinations of operations for performing the specified functions. It will also be understood that one or more blocks of the flowcharts, andcombinations of blocks in the flowcharts, can be implemented by special purpose hardwarebased computer systems which perform the specified functions, or combinations of special purpose hardware and computer instructions.

[0223] Many modifications and other embodiments of the inventions set forth herein will come to mind to one skilled in the art to which these inventions pertain having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is to be understood that the inventions are not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Moreover, although the foregoing descriptions and the associated drawings describe example embodiments in the context of certain example combinations of elements and / or functions, it should be appreciated that different combinations of elements and / or functions may be provided by alternative embodiments without departing from the scope of the appended claims. In this regard, for example, different combinations of elements and / or functions than those explicitly described above are also contemplated as may be set forth in some of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

Claims

Claims1. A user equipment (UE) comprising: at least one processor; and at least one memory storing instructions thereon that, when executed by the at least one processor, cause the UE to perform at least: using a machine learning (ML) model to generate inference output information based at least upon a plurality of UE measurements, the plurality of UE measurements being associated with one or more cells in a mobile network; maintaining, at the UE, in a circular buffer, the inference output information and UE measurement information associated with the plurality of UE measurements; and in response to receiving, at the UE, a UE measurement reporting request event, transmitting a UE measurement report comprising at least a portion of the inference output information and the UE measurement information corresponding to the at least a portion of the inference output information.

2. The UE of claim 1, wherein the inference output information corresponds to one or more outputs of the ML model.

3. The UE of any prior claim, wherein the inference output information comprises one or more of: predicted channel state information (CSI), predicted beam index (BI) information, predicted beam Layer 1 -Reference Signal Received Power (Ll-RSRP) information, predicted UE latitude information, predicted UE longitude information, predicted line of sight (LOS) information, or predicted non-line of sight (NLOS) information.

4. The UE of any prior claim, UE measurement information comprises one or more of: one or more RSRP values, one or more Reference Signal Received Quality (RSRQ) values, one or more Signal-to-Interference-plus-Noise Ratio (SINR) values,one or more Ll-RSRP values, one or more filtered Layer 3 (L3)-RSRP values, one or more Positioning Reference Signal (PRS)-RSRP values, one or more PRS-Reference Signal Received Path Power (RSRPP) values, one or more power delay profiles (PDP), one or more power delay (PD) values, one or more channel impulse responses (CIR), one or more frequency error measurements, or one or more timing error measurements.

5. The UE of any prior claim, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the UE to perform: receiving, from the mobile network, an indication configuring the circular buffer in the UE to have a specified buffer depth.

6. The UE of any prior claim, wherein the UE measurement report comprises only a portion of contents of the circular buffer.

7. The UE of any one of claims 1-5, wherein the UE measurement report comprises an entire contents of the circular buffer.

8. The UE of any prior claim, wherein the UE measurement reporting request event is based on at least one of the following occurring: the UE receiving, from the mobile network, a UE measurement reporting request.

9. The UE of claim 8, wherein the UE measurement reporting request comprises one or more of: an indication to report the entire contents of the circular buffer; a time-range of samples to report from the contents of the circular buffer; or a selection of specific samples to report from the contents of the circular buffer;10. The UE of any prior claim, wherein the UE measurement report further comprises one or more identifiers associated with the ML model, and wherein the instructions storedon the at least one memory, when executed by the at least one processor, further cause the UE to perform: maintaining, in the circular buffer at the UE, with the inference output information, the one or more identifiers identifying the ML model used to generate the inference output information, wherein the ML model identifiers correspond to at least the inference output information in the UE measurement report.

11. The UE of claim 10, wherein the one or more identifiers associated with the ML model comprise at least one of: a model identifier (ID) associated with the ML model, an associated ID corresponding to a network configuration used during data collection, an identifier of a data collection configuration used to train the ML model, an identifier of a dataset used to train the ML model, a UE hardware ID, or a UE vendor ID.

12. The UE of any prior claim, wherein the UE measurement report further comprises one or more timing-related information associated with the inference output information, and wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the UE to perform: maintaining, in the circular buffer at the UE, the one or more timing-related information identifying a respective time at which respective of the inference output information was generated, wherein the timing-related information corresponds to at least the inference output information in the UE measurement report.

13. The UE of claim 12, wherein the one or more timing-related information comprise at least one of: a system frame number (SFN), a slot number, an orthogonal frequency division multiplexing symbol number,a global navigation satellite system timestamp, a UE timestamp, or another timestamp.

14. A base station configured for communication in a mobile network, the base station comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the base station to perform at least: receiving, from a user equipment (UE), in response to a base station- initiated measurement reporting event, a UE measurement report comprising at least UE measurement information and inference output information, the UE measurement information being associated with one or more UE measurements of one or more cells in the mobile network, and the inference output information being inferred by one or more machine learning (ML) models in a ML functionality of the UE based at least upon the one or more UE measurements of the one or more cells in the mobile network.

15. The base station of claim 14, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to an element or function in the mobile network, an ML functionality performance report comprising the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information.

16. The base station of any one of claims 14-15, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: storing, at the base station, the UE measurement information and the inference output information or information associated with the UE measurement information and the inference output information.

17. The base station of any one of claims 14-16, wherein the instructions stored on theat least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the UE to perform a lifecycle management (LCM) action with regard to using the one or more ML models in the LE.

18. The base station of any one of claims 14-17, wherein the inference output information in the LE measurement report is associated with one or more inference outputs from an ML model in the ML functionality in the UE.

19. The base station of claim 18, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the LE to perform inference, using the ML model of the ML model functionality in the LE.

20. The base station of any one of claims 18-19, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: determining, based at least upon the LE measurement report, an ML model performance of the ML model used to generate the one or more inference outputs associated with the inference output information; and determining whether the ML model performance of the ML model is below a ML model performance threshold.

21. The base station of any one of claims 18-20, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the LE to maintain the LE measurement information and the inference output information in a circular buffer maintained at the LE.

22. The base station of any one of claims 18-21, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the basestation to perform: providing, to the UE, an indication configuring the UE to generate and transmit the UE measurement report comprising the UE measurement information and the inference output information.

23. The base station of any one of claims 18-22, wherein the UE measurement report further comprises one or more identifiers associated with the ML model.

24. The base station of any one of claims 18-23, wherein the indication that the UE is to perform inference using the ML model comprises or is provided in one or more of: a CSI Report Configuration message; a Radio Resource Control (RRC) Report Configuration message; a New Radio (NR) Measurement Report Configuration message; or an LTE Positioning Protocol (LPP) Request Location Information message.

25. The base station of any one of claims 14-24, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the circular buffer in the UE.

26. The base station of claim 25, wherein the inference output information comprises one or more of: predicted CSI, predicted BI information, predicted beam Ll-RSRP information, predicted UE latitude information, predicted UE longitude information, predicted LOS information, or predicted NLOS information.

27. The base station of any one of claims 25-26, wherein the indication configuring the circular buffer in the UE further configures the UE to store the UE measurementinformation in the buffer in the form of one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR Ratio values, one or more LI -RSRP values, one or more filtered L3-RSRP values, one or more PRS-RSRP values, one or more PRS-RSRPP values, one or more PDP, one or more PD values, one or more CIRs, one or more frequency error measurements, or one or more timing error measurements.

28. The base station of any one of claims 26-27, wherein the indication configuring the circular buffer in the UE further configures the circular buffer in the UE to have a specified buffer depth.

29. The base station of one of claims 14-28, wherein the base station-initiated measurement reporting event comprises one or more of: the base station providing, to the LE, a LE measurement reporting request.

30. The base station of claim 29, wherein the UE measurement reporting request comprises one or more of: an indication to report all contents of the circular buffer, a time-range of samples to report from the contents of the circular buffer, or a selection of specific samples to report from the contents of the circular buffer.

31. The base station of any one of claims 14-30, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the UE to include one or more ML model identifiers in the LE measurement report.

32. The base station of claim 31, wherein the UE measurement report comprises the one or more ML model identifiers associated with the one or more ML models used by the UE to generate the inference output information in the UE measurement report.

33. The base station of claim 32, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: determining, based upon the one or more ML model identifiers in the UE measurement report, the one or more ML models used by the UE to generate the inference output information.

34. The base station of any one of claims 29-33, wherein the one or more identifiers comprise at least one of: one or more model identifiers (IDs) associated with the one or more ML models, one or more associated IDs corresponding to one or more network configurations used during data collection, one or more identifiers of one or more data collection configurations used to train the one or more ML models, one or more identifiers of one or more datasets used to train the one or more ML models, a UE hardware ID, or a UE vendor ID.

35. The base station of any one of claims 14-34, wherein the instructions stored on the at least one memory, when executed by the at least one processor, further cause the base station to perform: providing, to the UE, an indication configuring the UE to include timing-related information in the UE measurement report, the timing-related information being associated with the ML model used to generate the inference output information.

36. The base station of claim 35, wherein the timing-related information comprise at least one of:an SFN, a slot number, an OFDM symbol number, a GNSS timestamp, a UE timestamp, or another timestamp.

37. The base station of any one of claims 14-36, wherein the UE measurement information comprises one or more of: one or more RSRP values, one or more RSRQ values, one or more SINR values, one or more LI -RSRP values, one or more filtered L3-RSRP values, one or more PRS-RSRP values, one or more PRS-RSRPP values, one or more PDPs, one or more PD values, one or more CIRs, one or more frequency error measurements, or one or more timing error measurements.

Images