Artificial intelligence (AI) / machine learning (ML) assisted beam management procedures

Abstract

Embodiments are related to a fifth generation (5G) or sixth generation (6G) wireless communications system. A method for a user equipment (UE) of a wireless system, comprises accessing information for a first set of resources and a second set of resources associated with beamforming operations of an antenna array, obtaining measurements for the second set of resources, accessing beam information to configure a set of beams associated with the first set of resources, the beam information generated by a machine learning (ML) model using the measurements for the second set of resources, and configuring the antenna array to generate the set of beams using the beam information. Other embodiments are described and claimed.

Description

ARTIFICIAL INTELLIGENCE (AI)ZMACHINE LEARNING (ML) ASSISTED BEAM MANAGEMENT PROCEDURESCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 575,223 filed April 5. 2024, entitled “ARTIFICIAL INTELLIGENCE (AI) / MACHINE LEARNING (ML) ASSISTED BEAM MANAGEMENT PROCEDURES”, and U.S. Provisional Application No. 63 / 554,762 filed February 16, 2024, entitled “ARTIFICIAL INTELLIGENCE (AI)ZMACHINE LEARNING (ML) ASSISTED BEAM MANAGEMENT PROCEDURES”, the contents of both which are herein incorporated by reference in their entirety.BACKGROUND

[0002] In a wireless system, a beam is a highly focused directional transmission pattern created through advanced beamforming techniques by an antenna array located at a base station. Instead of transmitting broad signals to cover large areas uniformly, the base station carefully adjusts the amplitude and phase of signals across multiple antenna elements to form one or more beams aimed at specific users or clusters of users. This directional steering significantly enhances received signal quality, mitigates interference, and improves overall spectral efficiency. By periodically transmitting reference signals, such as Synchronization Signal Blocks (SSBs) in fifth generation (5G) new radio (NR) systems, and using beam management procedures defined by Third Generation Partnership Project (3GPP), the network can dynamically track the best-performing beam for each user device. As users move or as radio conditions shift due to factors like traffic load or environmental changes, beams can be reoriented or reshaped to maintain optimal connectivity. These techniques, often combined with massive Multiple-Input. Multiple-Output (MIMO) antenna arrays, enable multiple beams to serve different users simultaneously, resulting in greater capacity, more stable connectivity, and enhanced user experiences, especially at the cell edges or in challenging propagation scenarios.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0003] To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

[0004] FIG. 1 illustrates a block diagram of a wireless system in accordance with one embodiment.

[0005] FIG. 2 illustrates a radio access network (RAN) in accordance with one embodiment.

[0006] FIG. 3 illustrates a functional framework in accordance with one embodiment.

[0007] FIG. 4 illustrates a message flow in accordance with one embodiment.

[0008] FIG. 5 illustrates a message flow in accordance with one embodiment.

[0009] FIG. 6 illustrates a logic diagram in accordance with one embodiment.

[0010] FIG. 7 illustrates a message flow in accordance with one embodiment.

[0011] FIG. 8 illustrates an apparatus in accordance with one embodiment.

[0012] FIG. 9 illustrates an apparatus in accordance with one embodiment.

[0013] FIG. 10 illustrates a logic flow in accordance with one embodiment.

[0014] FIG. 11 illustrates a logic flow in accordance with one embodiment.

[0015] FIG. 12 illustrates a core network (CN) in accordance with one embodiment.

[0016] FIG. 13 illustrates an operating environment of a CN in accordance with one embodiment.

[0017] FIG. 14 illustrates an operating environment of a CN in accordance with one embodiment.

[0018] FIG. 15 illustrates a wireless network in accordance with one embodiment.

[0019] FIG. 16 illustrates an apparatus in accordance with one embodiment.

[0020] FIG. 17 illustrates a computer readable medium in accordance with one embodiment.DETAILED DESCRIPTION

[0021] The present disclosure generally relates to wireless technology, and more specifically to beamforming techniques, procedures and architecture suitable for a wireless network. Some embodiments are particularly directed to a 3GPP Fifth Generation (5G) network, a 3GPP 5G New Radio (NR) network, and / or a 3GPP Sixth Generation (6G) network, as defined by a radio access network 1 (RANI) and / or radio access network 2 (RAN2) working groups. Although some embodiments are described in the context of a 5G network, a 5G NR network, or a 6G network, embodiments may apply to other ty pes of wireless networks as well. Embodiments are not limited in this context.

[0022] Embodiments herein may be related to one or more 3GPP standards, technical reports (TR), or work items (WI). Non-limiting examples include 5G NR Release 19 (Rel- 19); 3GPP TS 38.214 titled “NR; Physical layer procedures for data,” Version 18.4.0 (2024- 09); 3GPP TS 38.306 titled “NR; User Equipment (UE) radio access capabilities.” Version 18.3.0 (2024-09); 3GPP TS 38.331 titled “NR; Radio Resource Control (RRC); Protocol specification,” Version 18.3.0 (2024-09); and / or 3GPP TR 38.843 titled “Study on Artificial Intelligence (AI)ZMachine Learning (ML) for NR air interface,” Version 18.0.0 (2024-01), including any progeny, revisions, or variants. Specifically, some embodiments are related to RP-234039 titled “Artificial Intelligence (AI)ZMachine Learning (ML) for NR Air Interface,” 3GPP TSG RAN Meeting#102, Edinburgh, Scotland, December 2023 (Rel-19 Work Item (WI) on AI / ML).

[0023] In legacy 5G NR systems, layer 1 (LI preference signal received power (RSRP) / signal to interference plus noise ratio (SINR) measurements can be performed based on configured periodic channel state information (CSI)-reference signal (RS) or synchronization signal block (SSB) resources and the related RSRP / SINR can be reported to a base station (e.g., a gNB) in periodic, semi-persistent, or aperiodic CSI reports. However, for systems with a large number of beams in a spatial domain, or for the case when the UE location changes due to mobility thereby leading to changing beams in a temporal domain, tracking the optimal beam becomes a prohibitively complex task with respect to the number of LI measurements / reports as well as the reference signal overhead that should be configured to the UE.

[0024] In various embodiments, AI / ML based techniques are implemented wherein the measurement / reporting and RS overhead for beam management, in both the spatial domain and the temporal domain, can be optimized. Embodiments relate to different aspects of measurement reporting that can enable AI / ML based beam management including signaling and configuration for AI / ML model life cycle management (LCM). Some embodiments implement techniques designed to reduce the overhead and latency of LI beam measurement and reporting. The disclosed embodiments and examples may apply to one or both use-cases of beam prediction: (1) spatial-domain DL beam prediction for a first set of beams (referred to as “Set A”) based on measurement results of a second set of beams (referred to as “Set B”); and (2) temporal-domain DL beam prediction for the Set A of beams based on the historic measurement results of the Set B of beams. It is worthy to note that references to a “network” (or “NW”) and an access node (e.g., a base station, eNB, gNB, etc.) are made interchangeably without necessarily implying a particular distinctionbetween these in the context of the disclosed embodiments and examples. Embodiments are not limited in this context.

[0025] Embodiments may include or relate to technical solutions for accomplishing one or more of the following objectives: (1) beam management that includes downlink (DL) transmit (Tx) beam prediction for both a user equipment (UE)-sided model and a network (NW)-sided model, encompassing RANI and / or RAN2; (2) spatial-domain DL Tx beam prediction for a first set of beams (e.g., Set A) based on measurement results of a second set of beams (e.g.. Set B), also referred to as beam management case 1 (“BM-Casel”): (3) temporal DL Tx beam prediction for Set A of beams based on historic measurement results of Set B of beams, also referred to as beam management case 2 (“BM-Case2”); (4) specify necessary signaling mechanisms to facilitate life cycle management (LCM) operations specific to the beam management use cases, if any; and (5) enabling methods to ensure consistency between training and inference regarding NW-side additional conditions (if identified) for inference at the UE.

[0026] Embodiments implement a common framework design to support both BM-Casel and BM-Case2. Embodiments describe different sub-use-cases for AI / ML aided beam management framework. Further, the common framework design and sub-use-cases remain consistent with the conclusions from 3GPP TR 38.843 titled “Study on Artificial Intelligence (AI)ZMachine Learning (ML) for NR Air Interface’', version 2.0.1 (2023-12), and initial assessments of possible specification impact to enable these sub-use-cases for 5G NR Rel-19.

[0027] The present disclosure will now be described with reference to the attached drawing figures, wherein like reference numerals are used to refer to like elements throughout, and wherein the illustrated structures and devices are not necessarily drawn to scale. As utilized herein, terms “component,” “system,” “interface,” and the like are intended to refer to a computer-related entity, hardware, software (e.g., in execution), and / or firmware. For example, a component can be a processor (e.g., a microprocessor, a controller, or other processing device), a process running on a processor, a controller, an object, an executable, a program, a storage device, a computer, a tablet PC and / or a user equipment (e.g., mobile phone, etc.) with a processing device. By way of illustration, an application running on a server and the server can also be a component. One or more components can reside within a process, and a component can be localized on one computer and / or distributed between two or more computers. A set of elements or a set of other components can be described herein, in which the term “set” can be interpreted as “one or more.”

[0028] Further, these components can execute from various computer readable storage media having various data structures stored thereon such as with a module, for example. The components can communicate via local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network, such as. the Internet, a local area network, a wide area network, or similar network with other systems via the signal).

[0029] As another example, a component can be an apparatus with specific functionality provided by mechanical parts operated by electric or electronic circuitry', in which the electric or electronic circuitry can be operated by a software application or a firmware application executed by one or more processors. The one or more processors can be internal or external to the apparatus and can execute at least a part of the software or firmware application. As yet another example, a component can be an apparatus that provides specific functionality' through electronic components without mechanical parts; the electronic components can include one or more processors therein to execute software and / or firmware that confer(s), at least in part, the functionality of the electronic components.

[0030] Use of the word exemplary is intended to present concepts in a concrete fashion. As used in this application, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or”. That is, unless specified otherwise, or clear from context, “X employs A or B” is intended to mean any of the natural inclusive permutations. That is, if X employs A; X employs B; or X employs both A and B, then “X employs A or B” is satisfied under any of the foregoing instances. In addition, the articles “a” and “an” as used in this application and the appended claims should generally be construed to mean “one or more” unless specified otherwise or clear from context to be directed to a singular form. Furthermore, to the extent that the terms “including”, “includes”, “having”, “has”, “with”, or variants thereof are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term “comprising.” Additionally, in situations wherein one or more numbered items are discussed (e.g., a “first X”, a “second X”, etc.), in general the one or more numbered items may be distinct or they may be the same, although in some situations the context may indicate that they are distinct or that they are the same.

[0031] As used herein, the term “circuitry” may refer to, be part of, or include an Application Specific Integrated Circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or group), or associated memory (shared, dedicated, or group) operably coupled to the circuitry' that execute one or more software or firmware programs, a combinational logic circuit, or other suitable hardware components that provide the describedfunctionality. In some embodiments, the circuitry may be implemented in, or functions associated with the circuitry may be implemented by, one or more software or firmware modules. In some embodiments, circuitry may include logic, at least partially operable in hardware.

[0032] FIG. 1 illustrates an example of a wireless communication system 100a. For purposes of convenience and without limitation, the example wireless communication system 100a is described in the context of the long-term evolution (LTE) and fifth generation (5G) new radio (NR) (5G NR) cellular networks communication standards as defined by one or more 3GPP technical specifications (TSs) and / or technical reports (TRs). However, other types of wireless standards are possible.

[0033] The wireless communication system 100a includes UE 102a and UE 102b (collectively referred to as the "UEs 102"). In this example, the UEs 102 are illustrated as smartphones (e.g., handheld touchscreen mobile computing devices connectable to one or more cellular networks). In other examples, any of the UEs 102 can include other mobile or non-mobile computing devices, such as consumer electronics devices, cellular phones, smartphones, feature phones, tablet computers, wearable computer devices, personal digital assistants (PDAs), pagers, wireless handsets, desktop computers, laptop computers, in- vehicle infotainment (IVI), in-car entertainment (ICE) devices, an Instrument Cluster (IC), head-up display (HUD) devices, onboard diagnostic (OBD) devices, dashtop mobile equipment (DME), mobile data terminals (MDTs), Electronic Engine Management System (EEMS), electronic / engine control units (ECUs), electronic / engine control modules (ECMs), embedded systems, microcontrollers, control modules, engine management systems (EMS), networked or "smart" appliances, machine-type communications (MTC) devices, machine- to-machine (M2M) devices, Internet of Things (loT) devices, or combinations of them, among others.

[0034] In some implementations, any of the UEs 102 may be loT UEs, which can include a network access layer designed for low-power loT applications utilizing short-lived UE connections. An loT UE can utilize technologies such as M2M or MTC for exchanging data with an MTC server or device using, for example, a public land mobile network (PLMN), proximity services (ProSe), device-to-device (D2D) communication, sensor networks, loT networks, or combinations of them, among others. The M2M or MTC exchange of data may be a machine-initiated exchange of data. An loT network describes interconnecting loT UEs, which can include uniquely identifiable embedded computing devices (within the Internet infrastructure), with short-lived connections. The loT UEs may execute backgroundapplications (e.g., keep-alive messages or status updates) to facilitate the connections of the loT network.

[0035] The UEs 102 are configured to connect (e.g.. communicatively couple) with a radio access network (RAN) 112. In some implementations, the RAN 112 may be a next generation RAN (NG RAN), an evolved UMTS terrestrial radio access network (E- UTRAN), or a legacy RAN, such as a UMTS terrestrial radio access network (UTRAN) or a GSM EDGE radio access network (GERAN). As used herein, the term "NG RAN" may refer to a RAN 112 that operates in a 5G NR wireless communication system 100a, and the term "E-UTRAN" may refer to a RAN 112 that operates in an LTE or 4G wireless communication system 100a.

[0036] To connect to the RAN 112. the UEs 102 utilize connections (or channels) 118 and 120, respectively, each of which can include a physical communications interface or layer, as described below. In this example, the connections 118 and 120 are illustrated as an air interface to enable communicative coupling, and can be consistent with cellular communications protocols, such as a global system for mobile communications (GSM) protocol, a code-division multiple access (CDMA) network protocol, a push-to-talk (PTT) protocol, a PTT over cellular (POC) protocol, a universal mobile telecommunications system (UMTS) protocol, a 3GPP LTE protocol, a 5G NR protocol, or combinations of them, among other communication protocols.

[0037] The UE 102b is shown to be configured to access an access point (AP) 104 (also referred to as "WLAN node 104," "WLAN 104," "WLAN Termination 104," "WT 104" or the like) using a connection 122. The connection 122 can include a local wireless connection, such as a connection consistent with any IEEE 802.11 protocol, in which the AP 104 would include a wireless fidelity' (Wi-Fi) router. In this example, the AP 104 is shown to be connected to the Internet without connecting to the core network of the wireless system, as described in further detail below.

[0038] The RAN 112 can include one or more nodes such as RAN nodes 106a and 106b (collectively referred to as "RAN nodes 106" or "RAN node 106") that enable the connections 118 and 120. As used herein, the terms "access node," "access point," or the like may describe equipment that provides the radio baseband functions for data or voice connectivity, or both, between a network and one or more users. These access nodes can be referred to as base stations (BS), gNodeBs, gNBs, eNodeBs, eNBs, NodeBs, RAN nodes, rode side units (RSUs), transmission reception points (TRxPs or TRPs), and the link, and can include ground stations (e.g.. terrestrial access points) or satellite stations providingcoverage within a geographic area (e.g., a cell), among others. As used herein, the term "NG RAN node" may refer to a RAN node 106 that operates in an 5G NR wireless communication system 100a (for example, a gNB), and the term "E-UTRAN node" may refer to a RAN node 106 that operates in an LTE or 4G wireless communication system 100a (e.g., an eNB). In some implementations, the RAN nodes 106 may be implemented as one or more of a dedicated physical device such as a macrocell base station, or a low' power (LP) base station for providing femtocells, picocells or other like cells having smaller coverage areas, smaller user capacity, or higher bandwidth compared to macrocells.

[0039] In some implementations, some or all of the RAN nodes 106 may be implemented as one or more software entities running on server computers as part of a virtual netwmrk, which may be referred to as a cloud RAN (CRAN) or a virtual baseband unit pool (vBBUP). The CRAN or vBBUP may implement a RAN function split, such as a packet data convergence protocol (PDCP) split in which radio resource control (RRC) and PDCP layers are operated by the CRAN / vBBUP and other layer two (e.g., data link layer) protocol entities are operated by individual RAN nodes 106; a medium access control (MAC) / physical layer (PHY) split in which RRC, PDCP. MAC, and radio link control (RLC) layers are operated by the CRAN / vBBUP and the PHY layer is operated by individual RAN nodes 106; or a "lower PHY" split in which RRC, PDCP, RLC, and MAC layers and upper portions of the PHY layer are operated by the CRAN / vBBUP and low'er portions of the PHY layer are operated by individual RAN nodes 106. This virtualized framework allows the freed-up processor cores of the RAN nodes 106 to perform, for example, other virtualized applications. In some implementations, an individual RAN node 106 may represent individual gNB distributed units (DUs) that are connected to a gNB central unit (CU) using individual Fl interfaces (not shown in FIG. 1). In some implementations, the gNB-DUs can include one or more remote radio heads or RFEMs, and the gNB-CU may be operated by a server that is located in the RAN 112 (not shown) or by a server pool in a similar manner as the CRAN / vBBUP. Additionally or alternatively, one or more of the RAN nodes 106 may be next generation eNBs (ng-eNBs), including RAN nodes that provide E-UTRA user plane and control plane protocol terminations toward the UEs 102, and are connected to a 5G core network (e.g., CN 114) using a next generation interface.

[0040] In vehicle-to-everything (V2X) scenarios, one or more of the RAN nodes 106 may be or act as RSUs. The term "Road Side Unit" or "RSU" refers to any transportation infrastructure entity used for V2X communications. A RSU may be implemented in or by a suitable RAN node or a stationary (or relatively stationary) UE, where a RSU implementedin or by a UE may be referred to as a "UE-type RSU." a RSU implemented in or by an eNB may be referred to as an "eNB-type RSU," a RSU implemented in or by a gNB may be referred to as a "gNB-type RSU," and the like. In some implementations, an RSU is a computing device coupled with radio frequency circuitry located on a roadside that provides connectivity support to passing vehicle UEs 102 (vUEs 102). The RSU may also include internal data storage circuitry to store intersection map geometry, traffic statistics, media, as well as applications or other software to sense and control ongoing vehicular and pedestrian traffic. The RSU may operate on the 5.9 GHz Direct Short Range Communications (DSRC) band to provide very low latency communications required for high speed events, such as crash avoidance, traffic warnings, and the like. Additionally or alternatively, the RSU may operate on the cellular V2X band to provide the aforementioned low latency communications, as well as other cellular communications services. Additionally or alternatively, the RSU may operate as a Wi-Fi hotspot (2.4 GHz band) or provide connectivity to one or more cellular networks to provide uplink and downlink communications, or both. The computing device(s) and some or all of the radio-frequency circuitry of the RSU may be packaged in a weatherproof enclosure suitable for outdoor installation, and can include a network interface controller to provide a wired connection (e.g., Ethernet) to a traffic signal controller or a backhaul network, or both.

[0041] Any of the RAN nodes 106 can terminate the air interface protocol and can be the first point of contact for the UEs 102. In some implementations, any of the RAN nodes 106 can fulfill various logical functions for the RAN 112 including, but not limited to, radio network controller (RNC) functions such as radio bearer management, uplink and downlink dynamic radio resource management and data packet scheduling, and mobility management.

[0042] In some implementations, the UEs 102 can be configured to communicate using orthogonal frequency division multiplexing (OFDM) communication signals with each other or with any of the RAN nodes 106 over a multicarrier communication channel in accordance with various communication techniques, such as, but not limited to, OFDMA communication techniques (e.g., for downlink communications) or SC-FDMA communication techniques (e.g., for uplink communications), although the scope of the techniques described here not limited in this respect. The OFDM signals can comprise a plurality of orthogonal subcarriers.

[0043] The RAN nodes 106 can transmit to the UEs 102 over various channels. Various examples of downlink communication channels include Physical Broadcast Channel (PBCH), Physical Downlink Control Channel (PDCCH), and Physical Downlink Shared Channel (PDSCH). Other types of downlink channels are possible. The UEs 102 cantransmit to the RAN nodes 106 over various channels. Various examples of uplink communication channels include Physical Uplink Shared Channel (PUSCH), Physical Uplink Control Channel (PUCCH), and Physical Random Access Channel (PRACH). Other types of uplink channels are possible.

[0044] In some implementations, a downlink resource grid can be used for downlink transmissions from any of the RAN nodes 106 to the UEs 102, while uplink transmissions can utilize similar techniques. The grid can be a time-frequency grid, called a resource grid or time-frequency resource grid, which is the physical resource in the downlink in each slot. Such a time-frequency plane representation is a common practice for OFDM systems, which makes it intuitive for radio resource allocation. Each column and each row of the resource grid corresponds to one OFDM symbol and one OFDM subcarrier, respectively . The duration of the resource grid in the time domain corresponds to one slot in a radio frame. The smallest time-frequency unit in a resource grid is denoted as a resource element. Each resource grid comprises a number of resource blocks, which describe the mapping of certain physical channels to resource elements. Each resource block comprises a collection of resource elements; in the frequency domain, this may represent the smallest quantity of resources that currently can be allocated. There are several different physical downlink channels that are conveyed using such resource blocks.

[0045] The PDSCH carries user data and higher-layer signaling to the UEs 102. The PDCCH carries information about the transport format and resource allocations related to the PDSCH channel, among other things. It may also inform the UEs 102 about the transport format, resource allocation, and hybrid automatic repeat request (HARQ) information related to the uplink shared channel. Downlink scheduling (e.g., assigning control and shared channel resource blocks to the UE 102b within a cell) may be performed at any of the RAN nodes 106 based on channel quality information fed back from any of the UEs 102. The downlink resource assignment information may be sent on the PDCCH used for (e.g., assigned to) each of the UEs 102.

[0046] The PDCCH uses control channel elements (CCEs) to convey the control information. Before being mapped to resource elements, the PDCCH complex-valued symbols may first be organized into quadruplets, which may then be permuted using a subblock interleaver for rate matching. In some implementations, each PDCCH may be transmitted using one or more of these CCEs, in which each CCE may correspond to nine sets of four physical resource elements collectively referred to as resource element groups (REGs). Four Quadrature Phase Shift Keying (QPSK) symbols may be mapped to eachREG. The PDCCH can be transmitted using one or more CCEs, depending on the size of thedownlink control information (DC1) and the channel condition. In LTE, there can be four or more different PDCCH formats defined with different numbers of CCEs (e.g., aggregation level, L=l, 2, 4, or 8).100471 Some implementations may use concepts for resource allocation for control channel information that are an extension of the above-described concepts. For example, some implementations may utilize an enhanced PDCCH (EPDCCH) that uses PDSCH resources for control information transmission. The EPDCCH may be transmitted using one or more enhanced CCEs (ECCEs). Similar to above, each ECCE may correspond to nine sets of four physical resource elements collectively referred to as an enhanced REG (EREG). An ECCE may have other numbers of EREGs.

[0048] The RAN nodes 106 are configured to communicate with one another using an interface 132. In examples, such as where the wireless communication system 100a is an LTE system (e.g., when the core network (CN) 114 is an evolved packet core (EPC) network), the interface 132 may be an X2 interface 132. The X2 interface may be defined between two or more RAN nodes 106 (e.g., two or more eNBs and the like) that connect to the CN 114, or between two eNBs connecting to CN 114, or both. In some implementations, the X2 interface can include an X2 user plane interface (X2-U) and an X2 control plane interface (X2-C). The X2-U may provide flow control mechanisms for user data packets transferred over the X2 interface, and may be used to communicate information about the delivery of user data between eNBs. For example, the X2-U may provide specific sequence number information for user data transferred from a master eNB to a secondary eNB; information about successful in sequence delivery of PDCP protocol data units (PDUs) to a UE 102 from a secondary eNB for user data; information of PDCP PDUs that were not delivered to a UE 102; information about a current minimum desired buffer size at the secondary eNB for transmitting to the UE user data, among other information. The X2-C may provide intra-LTE access mobility functionality, including context transfers from source to target eNBs or user plane transport control; load management functionality; intercell interference coordination functionality, among other functionality.

[0049] In some implementations, such as where the wireless communication system 100a is a 5G NR system (e.g., when the CN 114 is a 5G core network), the interface 132 may be an Xn interface 132. The Xn interface may be defined between two or more RAN nodes 106 (e.g., two or more gNBs and the like) that connect to the 5G CN 114, between a RAN node 106 (e.g., a gNB) connecting to the 5G CN 114 and an eNB, or between two eNBs connecting to the 5G CN 114, or combinations of them. In some implementations, the Xn interface can include an Xn user plane (Xn-U) interface and an Xn control plane (Xn-C)interface. The Xn-U may provide non-guaranteed delivery of user plane PDUs and support / provide data forwarding and flow control functionality. The Xn-C may provide management and error handling functionality, functionality to manage the Xn-C interface; mobility support for UE 102 in a connected mode (e.g., CM-CONNECTED) including functionality to manage the UE mobility for connected mode between one or more RAN nodes 106, among other functionality. The mobility support can include context transfer from an old (source) serving RAN node 106 to new (target) serving RAN node 106, and control of user plane tunnels between old (source) serving RAN node 106 to new (target) serving RAN node 106. A protocol stack of the Xn-U can include a transport network layer built on Internet Protocol (IP) transport layer, and a GPRS tunneling protocol for user plane (GTP-U) layer on top of a user datagram protocol (UDP) or IP layer(s), or both, to carry user plane PDUs. The Xn-C protocol stack can include an application layer signaling protocol (referred to as Xn Application Protocol (Xn-AP or XnAP)) and a transport network layer (TNL) that is built on a stream control transmission protocol (SCTP). The SCTP may be on top of an TP layer, and may provide the guaranteed delivery of application layer messages. In the transport IP layer, point-to-point transmission is used to deliver the signaling PDUs. In other implementations, the Xn-U protocol stack or the Xn-C protocol stack, or both, may be same or similar to the user plane and / or control plane protocol stack(s) shown and described herein.

[0050] The RAN 112 is shown to be communicatively coupled to a CN 114 (referred to as a "CN 114"). The CN 114 includes multiple network elements and / or network functions (NFs), such as network element 108a and network element 108b (collectively referred to as the "network elements 108"), which are configured to offer various data and telecommunications services to customers / subscribers (e.g., users of UEs 102) who are connected to the CN 114 using the RAN 112. The components of the CN 114 may be implemented in one physical node or separate physical nodes and can include components to read and execute instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium). In some implementations, network functions virtualization (NFV) may be used to virtualize some or all of the network node functions described here using executable instructions stored in one or more computer- readable storage mediums, as described in further detail below. A logical instantiation of the CN 114 may be referred to as a network slice, and a logical instantiation of a portion of the CN 114 may be referred to as a network sub-slice. NFV architectures and infrastructures may be used to virtualize one or more network functions, alternatively performed by proprietary hardware, onto physical resources comprising a combination of industry-standard server hardware, storage hardware, or switches. In other words, NFV systems can be used to execute virtual or reconfigurable implementations of one or more network components or functions, or both.[00511 In some implementations, the CN 114 may be a 5G core network (referred to as "5GC CN 114" or "5G CN 114"), and the RAN 112 may be connected with the CN 114 using a next generation interface 124. In some implementations, the next generation interface 124 may be split into two parts, a next generation user plane (NG-U) interface 128, which carries traffic data between the RAN nodes 106 and a user plane function (UPF), and the SI control plane (NG-C) interface 126, which is a signaling interface between the RAN nodes 106 and access and mobility management functions (AMFs). Examples where the CN 114 is a 5G core network are discussed in more detail with regard to later figures.

[0052] In some implementations, the CN 114 may be an evolved packet core (EPC) (referred to as "EPC CN 114" or the like), and the RAN 112 may be connected with the CN 114 using an SI interface 124. In some implementations, the SI interface 124 may be split into two parts, an SI user plane (Sl-U) interface 128, which carries traffic data between the RAN nodes 106 and the serving gateway (S-GW), and the SI -MME interface 126, which is a signaling interface between the RAN nodes 106 and mobility management entities (MMEs).

[0053] The CN 114 may include MME, SGW, SGSN, HSS, PGW, PCRF, and / or other NFs coupled with one another over various interfaces (or “reference points”) (not shown). The CN 114 may be a 5GC including an AUSF, AMF, SMF, UPF, NSSF, NEF, NRF, PCF, UDM, AF, and / or other NFs coupled with one another over various service-based interfaces and / or reference points. The 5GC may enable edge computing by selecting operator / 3rd party services to be geographically close to a point that the UE 102 is attached to the network. This may reduce latency and load on the network. In edge computing implementations, the 5GC may select a UPF close to the UE 102 and execute traffic steering from the UPF to a data network (DN) 134 via an N6 interface. This may be based on the UE subscription data, UE location, and information provided by the AF, which allows the AF to influence UPF (re)selection and traffic routing.

[0054] The DN 134 may represent various network operator services, Internet access, or third party services that may be provided by one or more servers including, for example, the application server 110. The application server 110 may be an element offering applications that use IP bearer resources with the core network (e.g., UMTS packet services (PS) domain, LTE PS data services, among others). The application server 110 can also beconfigured to support one or more communication services (e.g., VoIP sessions, PTT sessions, group communication sessions, social networking services, among others) for the UEs 102 using the CN 114. The application server 110 can use an IP communications interface 130 to communicate with one or more network element 108a or network element 108b.

[0055] The DN 134 may be an operator external public, a private PDN, or an intra-operator packet data network, for example, for provision of IMS services. In this embodiment, the application server 110 can be coupled to an IMS via an S-CSCF or the I-CSCF. In some implementations, the DN 134 may represent one or more local area DNs (LADNs), which are DNs (or DN names (DNNs)) that is / are accessible by a UE 102 in one or more specific areas. Outside of these specific areas, the UE 102 is not able to access the LADN / DN.

[0056] Additionally or alternatively, the DN 134 may be an Edge DN 134, which is a (local) Data Network that supports the architecture for enabling edge applications. In these embodiments, the application server 110 may represent the physical hardware systems / devices providing app server functionality and / or the application software resident in the cloud or at an edge compute node that performs server function(s). In some embodiments, the application server 110 provides an edge hosting environment that provides support required for Edge Application Server's execution.

[0057] In some embodiments, the 5GS can use one or more edge compute nodes to provide an interface and offload processing of wireless communication traffic. In these embodiments, the edge compute nodes may be included in, or co-located with one or more RAN 112. For example, the edge compute nodes can provide a connection between the RAN 112 and UPF in the 5GC. The edge compute nodes can use one or more NFV instances instantiated on virtualization infrastructure within the edge compute nodes to process wireless connections to and from the RAN 112 and a UPF.

[0058] FIG. 2 illustrates a wireless communication system 200. The wireless communication system 200 is a sub-system of the wireless communication system 100a illustrated in FIG. 1. The wireless communication system 200 depicts a UE 202 connected to an access node 216 of a wireless communication system 100a over a connection 212. An example of an access node 216 is a gNB 204. The UE 202 and connection 212 are similar to the UE 102 and the connections 118, 120 described with reference to FIG. 1. The gNB 204 is similar to the RAN node 106, and represents an implementation of the RAN node 106 as a gNB with a dual-architecture.[0059| As depicted in FIG. 2, the gNB 204 is divided into two physical entities referred to a centralized or central unit (CU) and a distributed unit (DU). The gNB 204 may comprise a gNB-CU 214 and one or more gNB-DU 210. The gNB-CU 214 is further divided into a gNB-CU control plane (gNB-CU-CP) 206 and a gNB-CU user plane (gNB-CU-UP) 208. The gNB-CU-CP 206 and the gNB-CU-UP 208 communicate over an El interface. The gNB-CU-CP 206 communicates with one or more gNB-DU 210 over an Fl-C interface.The gNB-CU-UP 208 communicates with the one or more gNB-DU 210 over an Fl-U interface.

[0060] The gNB-CU-CP 206 and the gNB-CU-UP 208 provides support for higher layers of a protocol stack such as Service Data Adaptation Protocol (SDAP), Packet Data Convergence Protocol (PDCP) and RRC. The gNB-DU 210 provides support for lower layers of the protocol stack such as Radio Link Control (RLC), MAC layer, and PHY layer. In some implementations, there is a single gNB-CU 214 for each gNB 204 that controls multiple gNB-DU 210. For example, the gNB 204 may have hundreds of gNB-DU 210 connected to a single gNB-CU 214. Each gNB-DU 210 is able to support one or more cells, where one gNB 204 can potentially control hundreds of cells in a 5G NR system.

[0061] As previously discussed, in a 5G NR system, such as defined by 3GPP TS 38.473 and / or TS 28.423, the UE 202 can enter different RRC states, such as an idle state and a connected state. The UE 202 can also enter an inactive state where the UE 202 is registered with the network but not actively transmitting data. A resume procedure can prepare the UE 202 for subsequent data transmission by causing the UE 202 to switch from an inactive state to a connected state. In 5G NR, the RRC states for a 5G NR enabled UE can include RRC IDLE, RRC INACTIVE, and RRC CONNECTED states. When not transmitting data in an RRC_CONNECTED state, the UE 202 can switch to an RRC INACTIVE state but remain registered with the network.

[0062] FIG. 3 illustrates a functional framework 300. The functional framework 300 is an example of an AI / ML framework or architecture for a wireless network. The functional framework 300 covers a general functional architecture addressing both model-ID-based life cycle management (LCM) and functionality-based LCM. Therefore, some of the functions or data flows, information flows, or instruction flows (e.g., indicated by arrows) shown in FIG. 3 might not always be relevant for a given LCM approach. As an illustrative example, consider a scenario where the network performs functionality-based LCM and where models are not identified in the network, while the UE concurrently performs model-level management (e.g., model selection, model switching, model activation, model deactivation,etc.). In this case, the “Model Training” or “Model Storage” functions with their respective procedures, may be regarded as less relevant from a network perspective.

[0063] As seen in FIG. 3, the functional framework 300 comprises various types of specialized circuitry, such as a data collection circuitry 304, a model training circuitry 306, a model management circuitry 308, a model inference circuitry 310, and a model storage circuitry 312. The functional framework 300 may use circuitry, such as processing circuitry, to execute a set of functions for a network entity, such as a UE, a base station, a NF of a core network, an 0AM, and other network devices or network nodes. The functions may be implemented as circuits, such as an application specific integrated circuit (ASIC) or field programmable gate array (FPGA). The functions may also be implemented as program instructions stored in memory that when executed by processing circuitry causes the processing circuitry to perform the functions. Embodiments are not limited in this context.

[0064] The data collection circuitry 304 performs a function that provides input data to the model training circuitry 306, model management circuitry 308, and model inference circuitry 310 functions. The data collection circuitry 304 collects and forwards different types of data, such as training data 314 needed as input for the model training circuitry 306, monitoring data 316 needed as input for the model management circuitry 308 to manage a trained ML model 320 or AI / ML functionalities of the trained ML model 320, and inference data 318 needed as input for the model inference circuitry 310.

[0065] The model training circuitry 306 executes a function that performs AI / ML model training, validation, and testing for one or more ML models 302, which may also generate model performance metrics which can be used as part of the model testing procedure. The model training circuitry 306 is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) based on training data 314 delivered by the data collection circuitry 304.

[0066] The model training circuitry 306 trains an ML model 302 and it outputs a trained ML model 320. The model training circuitry 306 delivers trained, validated, and tested trained ML model 320 to the model storage circuitry 312. Additionally, or alternatively, the model training circuitry 306 delivers an updated version of a trained ML model 320 to the model storage circuitry 312.

[0067] The model management circuitry 308 executes a function that oversees the operation (e.g., selection / (de)activation / switching / fallback) and monitoring (e.g., performance) of trained ML models 320 or AI / ML functionalities. The model management circuitry 308 is also responsible for making decisions to ensure the proper inferenceoperation based on data received from the data collection circuitry 304 and the model inference circuitry 310.

[0068] The model management circuitry 308 may generate and forward management instructions 328 to the model inference circuitry 310. A management instruction 328 comprises information needed as input to manage the model inference circuitry 310. The information may include selection, activation, deactivation, or switching of trained ML models 320 or AI / ML-based functionalities, fallback to non-AI / ML operation (e.g., not relying on inference process), and other management operations.

[0069] The model management circuitry 308 may also generate and forward a model request 324 to the model storage circuitry 312. The model request 324 comprises information such as a model transfer request, a model delivery request, and other management instructions for the model storage circuitry 312.

[0070] The model management circuitry 308 may also generate and forward feedback 330 to the model training circuitry 306. The feedback 330 comprises information needed as input for the model training circuitry 306, such as performance feedback, retraining requests, updating requests, reinforcement learning, and so forth.

[0071] The model inference circuitry 310 executes a function that provides outputs from the process of applying trained ML models 320 or AI / ML functionalities, using the inference data 318 provided by the data collection circuitry 304 as an input. The model inference circuitry 310 is also responsible for data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) based on the inference data 318 delivered by the data collection circuitry 304. The model inference circuitry 310 receives the inference data 318 as input, analyzes the inference data 318, and it generates an inference output 326. The inference output 326 may be delivered to various circuitry executing functions for any number of downstream tasks for a 5G or 6G wireless system. The model inference circuitry 310 may also deliver the inference output 326 to the model management circuitry 308 to monitor the performance of the trained ML models 320 or AI / ML functionalities.

[0072] The model storage circuitry 312 executes a function responsible for storing trained ML models 320 that can be used to perform an inference function by the model inference circuitry 310. It is worthy to note that the model storage circuitry 312 in FIG. 3 is only intended as a reference point (if any) when applicable for protocol terminations, model transfer / delivery, and related processes. It should be stressed that its purpose does not encompass restricting the actual storage locations of models. Therefore, the specification impact of all data / information / instruction flows (i.e., the arrows in FIG. 3) to / from thisfunction may vary for a given implementation. The model storage circuitry 312 may be used to deliver a trained ML model 320 to the model inference circuitry 310.

[0073] The functional framework 300 may be implemented for various use cases for a 5G or 6G wireless network. Non-limiting examples of use cases include: (1) beam management that includes downlink (DL) transmit (Tx) beam prediction for both a user equipment (UE)- sided model and a network (NW)-sided model, encompassing RANI and / or RAN2; (2) spatial-domain DL Tx beam prediction for a first set of beams (e.g., Set A) based on measurement results of a second set of beams (e.g., Set B), also referred to as beam management case 1 (“BM-Casel”); (3) temporal DL Tx beam prediction for Set A of beams based on historic measurement results of Set B of beams, also referred to as beam management case 2 (“BM-Case2”); (4) specify necessary signaling mechanisms to facilitate life cycle management (LCM) operations specific to the beam management use cases, if any; and (5) enabling methods to ensure consistency between training and inference regarding NW-side additional conditions (if identified) for inference at the UE. A particular use case for the functional framework 300 may be determined for a given implementation. Elowever, the functional framework 300 provides a robust AI / ML architecture for a variety of sub use cases and is diverse enough to support various requirements on a variety of gNB- UE collaboration levels.

[0074] FIG. 4 illustrates a message flow 400. The message flow 400 is an example of a message flow 400 for beam management between various devices of the wireless communication system 100a, such as an access node 216 like gNB 204 and a UE 202. Embodiments are not limited to this example.

[0075] As previously described, AI / ML based techniques are implemented wherein the measurement / reporting and RS overhead for beam management, in both the spatial domain and the temporal domain, can be optimized. Embodiments herein relate to different aspects of measurement reporting that can enable AI / ML based beam management including signaling and configuration for AI / ML model life cycle management (LCM). Some embodiments implement techniques designed to reduce the overhead and latency of LI beam measurement and reporting. Unless mentioned otherwise, the disclosed embodiments and examples may apply to both use-cases of beam prediction: (1) spatial-domain DL beam prediction for a Set A 406 of beams based on measurement results of a Set B 408 of beams; and (2) temporal-domain DL beam prediction for the Set A 406 of beams based on the historic measurement results of the Set B 408 of beams. It is worthy to note that references to a “network” (or “NW”) and an access node 216 such as a base station (or “gNB”) are made interchangeably without necessarily implying a particular distinction between these inthe context of the disclosed embodiments and examples. Embodiments are not limited in this context.

[0076] The beam management use case for AI / ML can be broadly divided into two subcases, namely spatial domain beam management and temporal domain beam management. Regardless of the specific use case, a common AI / ML model training and deployment framework can be considered as follows for a wireless communication system as shown in FIG. 4, where a machine learning model with offline training / validation is considered. The model is offline in the sense that dataset for training and testing are pre-generated, and the model is pre-trained in a non-real time manner.

[0077] In Rel-19, RANI focuses on offline model training and testing and different model life cycle management procedures related to offline model training to train a trained ML model 320 to assist in beam management operations. Two use-cases for beam management using AI / ML includes BM-Casel and BM-Case2.

[0078] Spatial Domain Beam Management (BM-Casel) predicts the optimal beam at UE 202 or gNB 204 without an exhaustive search with an aim to reduce measurement and reporting latency. The sub use cases include DL Tx (or UL Rx) beam prediction at the gNB 204, DL Rx (or UL Tx) beam prediction at UE 202, and joint DL Tx (or UL Rx) and UL Rx (or DL Rx) beam pair prediction. Embodiments are not limited to these examples..

[0079] In some embodiments, the access node 216 trains the ML model 302 and / or executes the trained ML model 320 to support beamforming operations. For example, for DL Tx (or UL Rx) beam prediction at the gNB 204, the gNB 204 may use an antenna array 402. In some embodiments, the antenna array 402 may be implemented as a two- dimensional (2D) planar array at gNB 204 with multiple analog Tx beams and a fixed Rx beam or optimal Rx beam selection at the UE 202. The antenna array 402 may send multiple beams with a set of resources 404 such as a CSI-RS for CSI acquisition. For example, the resources 404 may comprise multiple sets of resources, such as a Set A 406 and a Set B 408. Only a subset of the resources 404, such as Set B 408, is measured by UE 202 in LI using the measurement unit 410. The measurement unit 410 obtains a set of measurements 412 for the subset of gNB beams. The Set A 406 beams are used for generating beams 418 by the antenna array 402 to communicate data with the antenna array 402 of the gNB 204.

[0080] In some embodiments, the UE 202 sends a measurement report 414 with measurements 412 to the gNB 204. Based on these measurements 412 (or a function of these measurements 412), the trained ML model 320 executed by the access node 216 predicts the top K beam indices explicitly or implicitly from these subsets of measurements 412 leadingto latency reduction for optimal beam selection, especially for cases with large number of analog beams at gNB 204. The access node 216 sends beam information 416 generated by the trained ML model 320, wherein the beam information 416 comprises configuration information to configure beams 418 for the antenna array 402 of the UE 202. The antenna array 402 of the gNB 204 and the antenna array 402 of the UE 202 may exchange beams 418 in accordance with the configuration information.

[0081] In some embodiments, the UE 202 trains the ML model 302 and / or executes the trained ML model 320 to support beamforming operations. In this case, the UE 202 sends the measurements 412 directly to the local trained ML model 320 to predict the top K beam indices from the measurements 412. In some instances, the UE 202 may still send the measurement report 414 with measurements 412 to the gNB 204 for other purposes, such as testing the quality of the trained ML model 320 deployed to the UE 202, offloading inferencing operations from the UE 202 when the trained ML model 320 deployed for the UE 202 degrades in performance or becomes too large for proper execution on the limited resources (e.g., compute, memory, power, etc.) of the UE 202, re-training the trained ML model 320 for re-deployment to the UE 202, and other use cases.

[0082] The beam information 416 may comprise configuration information for the antenna array 402 to form a set of beams 418. In some embodiments, such as for 5G NR, the beam information 416 may include Transmission Configuration Indicator (TCI) states. TCI states are standardized mechanisms that facilitate beam-based operation. They provide a way for the network (e.g., gNB 204) to instruct the UE 202 about which spatial transmission parameters to use when decoding certain downlink channels. A TCI state effectively “points” the UE 202 toward a particular beam or set of reference signals from which it can derive the channel characteristics needed for proper demodulation and beamforming. Each TCI state is associated with one or more reference signals — most commonly Channel State Information-Reference Signals (CSI-RS) or Synchronization Signal Blocks (SSBs). These reference signals define the spatial transmission characteristics (e.g., a specific beam direction) that the UE should assume for the corresponding physical channel. The TCI state identifies which reference signals the UE should treat as quasi co-located with the data or control channel. Quasi-co-location (QCL) means the UE 202 can assume that certain channel properties (like delay spread, Doppler spread, and spatial characteristics) of the reference signal also apply to the data or control transmission. Through QCL relationships, the UE 202 knows how to form its receive beam — aligning its antenna array weights to “look” in the direction indicated by the reference signals of that TCI state. During initial setup or reconfiguration, the gNB 204 provides a set of TCI states to the UE 202 throughRadio Resource Control (RRC) signaling. Each TCI state includes identifiers of the reference signals and QCL parameters. For each downlink transmission (e g., PDSCH for data or PDCCH for control), the gNB 204 can dynamically indicate which TCI state the UE 202 should use. This is done via Downlink Control Information (DCI) signaling. Instead of reconfiguring beams every time, the network quickly "points" to a pre-defined TCI state, allowing rapid beam switching. Since multiple TCI states can be preconfigured, the network can switch the UE 202 between different beams without reconfiguring the underlying parameters from scratch. This flexibility is crucial for managing mobility scenarios, interference conditions, or dynamic changes in the radio environment. The gNB 204 selects another TCI state (pointing to another beam) for subsequent transmissions, enabling agile and efficient beam management. Different TCI states can be used for the Physical Downlink Control Channel (PDCCH) and Physical Downlink Shared Channel (PDSCH). For example, the UE might use one TCI state, pointing to a certain CSI-RS resource, for decoding control information, while a different TCI state associated with another reference signal beam could be used to decode the data channel. This allows optimal beam selection independently for control and data. Consequently, a TCI state is a “pointer” to a beam and the associated channel parameters the UE 202 needs for proper demodulation. By preconfiguring multiple TCI states and choosing among them dynamically, the gNB 204 gains a powerful tool for beam-based transmissions in 5G NR, enabling efficient and responsive beamforming operations.

[0083] For DL Rx (UL Tx) beam prediction at the UE 202, the UE 202 may also use an antenna array 402. In some embodiments, the antenna array 402 may be implemented as a 2D planar array at UE 202 with multiple analog Rx beams and fixed or optimal Tx beam at the gNB 204. The trained ML model 320 is provided with LI measurements 412 (or functions thereof) for a subset of the UE-side beams and the top K beam indices can be predicted explicitly or implicitly from a subset of measurements 412 thereby reducing the latency of beam acquisition or tracking at the UE 202.

[0084] For joint DL Tx (UL Rx) and UL Rx (DL Rx) beam pair prediction, measurements are taken for pairs of gNB 204 and UE 202 analog beams and the trained ML model 320 uses these LI measurements 412 (or a function thereof), to predict the top-K beam pairs similar to joint P2 / P3 process thereby providing major latency gains in 2-sided beam selection.

[0085] Temporal Domain Beam Management (BM-Case2) predicts the best beam at gNB 204 or UE 202 for future time instants (prediction window) given LI measurement observations of beams from a window of preceding time instants (observation window).Time series data can be collected from UEs moving across an environment and long shortterm memory (LSTM) based models can be used to predict the best beam. An example of a prediction window and an observation window are described in more detail with reference to FIG. 6.

[0086] AI / ML model implementation in most cases may be dependent on implementation and specific models will vary. However, the functionalities to enable AI / ML models, regardless of implementation details to be integrated into 5G-NR networks will remain the same. Current networks offer a myriad of beam measurement and reporting functionalities which may be further augmented to enable AI / ML use cases. Some examples for AI / ML model integration into 5G and the related specification impact are provided below.

[0087] Where the set of all DL Tx beams in the system from which the AI / ML model is expected to predict the best beam is known as Set A, the input to the AI / ML model is known as a Set B which has a cardinality typically much lesser than Set A and may or may not be a subset of Set A.

[0088] For spatial domain beam prediction (BM-Case-1), spatial domain DL Tx beam prediction can be applied towards initial beam acquisition using SSB beams or beam tracking using narrower CSI-RS beams in conjunction with current beam reporting framework for faster and more accurate beam selection. The AI / ML model may reside either at UE 202 or gNB 204 and the major specification impact will be based on LI beam measurement / reporting and reference signal transmissions.

[0089] As an example of this use-case, the AI / ML model may reside at the network to enable UE 202 and ML-aided gNB 204 beam tracking using DL measurements. The process can function as follows: (1) gNB 204 triggers CSI-RS for CSI based on periodic / aperiodic beam report from UE 202 if Tx beam drops below threshold; (2) gNB 204 transmits M CSI- RS based on a Set B of beams where M is less than a cardinality of Set A of total number of CSI-RS beams; (3) UE 202 measures Ll-RSRP / SINR and reports to gNB 204; and (4) gNB 204 can use M measurement or a function of the measurements as input to the gNB-side AI / ML model to predict best or top-K Tx (or Rx) beam.

[0090] In this case, the UE 202 does not need to know which CSI-RS beams are transmitted. For example, gNB 204 can sample the spatial domain based on its own implementation. The scheme relies on the LI beam report (e.g., measurement report 414) from the UE 202 for best beam prediction at the gNB 204. Potential gain is from reduced CSI-RS transmissions for measurement. A similar procedure can also be used for a UE-side model where the UE 202 uses the measurements 412 from CSI-RS beams as input to itsAI / ML model to predict the best or top-K DL Tx beams and reports this to the network. For the UE-side model, a potential additional gain may come from smaller beam reports depending on the value of K for top-K beams. The value of K is generally expected to be smaller than M which is the dimensionality of the AI / ML model input.

[0091] The spatial domain DL Tx beam prediction problem can be further sub-divided into two sub-problems. For BM-Casel, when Set A 406 is a subset of Set B 408, the input to the AI / ML model (e.g., Set B) is formed of beams which are a part of Set A. This is typically for the case when narrow analog beam measurements from CSI-RS transmissions are used to predict the best narrow beams as shown in FIG. 4. The AI / ML model is trained to map a subset of measurements 412 to all the measurements for beams in Set A or to the index of the best beam in Set A. In this case, CSI-RS is expected to be used for both measurement and data collection. For this use case, specification impact can be expected from CSI-RS transmissions for Set B measurements and consequent overhead reduction from smaller number of required measurements.

[0092] For BM-Casela, when Set B is not an element of Set B. the input to the AI / ML model (e.g., Set B) is based on a different analog beamforming assumption than the Set A of beams from which the AI / ML model is expected to predict the best beam. Typically, the measurement for Set B is assumed to be on wide SSB beams and the AI / ML takes these as an input to predict the best narrow CSI-RS beam which can be used for data transmission. This method can potentially reduce latency of initial access procedures by enabling the UE 202 to measure on fewer SSBs and directly providing the best CSI-RS beam for further communication. For this case, specification impact can be expected from SSB transmissions related to Set B, such as configuration of Set B and simplification of the initial access procedure and consequent latency reduction due to smaller number of overall measurements needed to predict the best narrow CSI-RS beam.

[0093] As previously described, AI / ML based techniques are considered wherein the measurement / reporting and RS overhead for beam management, in both the spatial domain and the temporal domain, can be optimized. Various embodiments herein relate to different aspects of measurement reporting that can enable AI / ML based beam management including signaling and configuration for AI / ML model life cycle management (LCM). Some embodiments implement techniques designed to reduce the overhead and latency of LI beam measurement and reporting. Unless mentioned otherwise, the disclosed embodiments and examples may apply to both use-cases of beam prediction: (1) spatial-domain DL beam prediction for a Set A of beams based on measurement results of a Set B of beams; and (2) temporal -domain DL beam prediction for the Set A of beams based on the historicmeasurement results of the Set B of beams. Examples for these two use-cases of beam prediction are described in detail in the following section.

[0094] In one embodiment, for beam management assisted by AI / ML models at NW side or UE side, a UE 202 may be configured with a Set A of beams (set of all beams which can be used for DL / UL) and a Set B of beams on which the UE 202 is expected to perform measurements. In one example, the UE 202 is configured with a single Set A and a single Set B where the beams in Set B correspond to joint / DL TCI states which are included in Set A. In another case, the UE 202 can be configured with a single Set A and single Set B wherein the beams in Set B are corresponding to joint / DL TCI states which may not all be included in Set A. In one example the configuration can be based on RRC signaling. In yet another example, the UE 202 may be configured with multiple variants of Set A and Set B using RRC signaling and may be activated with one Set A and Set B as a pair or respectively from the list of configurations based on triggering by MAC-CE and / or DCI signaling.

[0095] For example, for UE-sided model at least for BM Case-1, for the inference results report, two resource sets can be configured for Set A and Set B separately in the CSI report configuration for the report. In some cases, only resource set for Set B is configured. The UE 202 performs measurement on the resource set for Set B for inference, and UE 202 is not expected to measure resource set for Set A for inference. The beam information in the inference report refers to the resource set for Set A. For both BM-Casel and BM-Case2. for UE-sided model for inference, when Set A and Set B are configured within CSI report configuration, two CSI-ResourceConfigld s are configured for Set A and Set B separately.

[0096] In one embodiment, a configuration of Set A, Set B or Set A / B pair may be associated with a specific AI / ML model which expects inputs from the configured Set B and outputs associated with the configured Set A. In this case, configuring multiple Set A, Set B and / or Set A / B pairs and choosing any one of these configurations using RRC / MAC- CE / DCI may also implicitly allow AI / ML model selection or switching at the gNB 204 or UE 202. For UE side models, a UE 202 may be additionally configured with an association of one or more models with one or more configured sets A / B when more than one model is assumed to be used. In this case, models may be identified individually or may be grouped to form model-groups that may be subsequently referenced when associating to one or more configurations for sets A / B. Such dependencies on model configurations can also be interpreted as application of additional conditions at the network side for a UE-side model based on configuration related to the model provided by the network to the UE 202.

[0097] In one embodiment, for a NW side model, a UE 202 may be configured only with a Set B of beams without configuration of a Set A of beams. In another example, a UE 202 may be configured with a Set A of beams and not be explicitly configured with a Set B of beams but may be configured with LI measurement resources which implicitly map to Set B of beams for the network side model.

[0098] In one embodiment, for a UE-sided model, a UE 202 may be configured with a Set A of beams but not explicitly with a Set B of beams. The Set B of beams may be chosen by the UE 202 for measurement and the corresponding reference signal transmissions can be requested by the UE 202 from the gNB 204. Alternatively, the gNB 204 can configure a set of measurement resources which are associated to reference signals which are either explicitly configured as part of Set A or have TCI states which are included in Set A configuration.

[0099] In one embodiment, for models trained at the UE side, a UE 202 may expect that the Set B used during model training will be the same Set B used during model inference. This may be configured for all configured / activated beams in Set B using RRC / DCI / MAC- CE signaling wherein, the signaling may be a flag which, if set to 1 (true) enables the UE 202 to assume that the configured Set B was the same as the one used during training of the corresponding or associated model. In another example, the configuration could be based on an index of the configured and activated Set B wherein a UE 202 may be further provided with configurations to identify the index of the Set B which was used for training the model being used for inference. In one example, the UE 202 may expect the Set B index of training and inference to be identical.

[0100] For example, where the “flag’?-based indication described here can be mapped to the provision of an Associated ID below. For UE sided model in beam management, support associated ID, it is assumed the associated ID at least can be configured within CSI framework. The UE 202 may assume the similar properties of a DL Tx beam or beam set / list associated with the same associated ID.

[0101] In one embodiment, for a network side model, a UE 202 may expect the network to configure the same LI measurement resources, which were used during model training, for the purpose of model inference.

[0102] In one embodiment, a UE 202 may use different Set B configurations for the same AI / ML model and in this case, being configured with different Set B may not imply model switching / selection. Additionally, in this case, the assumption of Set B may not be the same across training and inferencing phases. The UE 202 may assume different Set B for trainingand inferencing. For a UE-side model, it may be up to the UE 202 to switch / select AI / ML models based on Set A / B / A+B configuration and the UE 202 may or may not notify the network of model switching / selection decisions.

[0103] In one embodiment, when a UE 202 is configured with a Set B of measurement beams, the UE 202 may be expected to measure on all the configured beams within Set B in one shot and update the measurements for beams in Set B for use at UE side model or reports measurements using Ll-reporting for use at NW side model. In another embodiment, a UE 202 may be expected to measure K instances of the measurement RS in Set B and then update Set B measurements using a function of all the K LI measurements. In one example of the embodiment, the value of K may be provided to the UE 202 by its serving gNB 204 using RRC / MAC-CE signaling. In another example, a UE may be configured instead with K aperiodic measurement resource corresponding to the beams in Set B and the UE may be expected to measure all the K aperiodic instances of the measurement RS and then update measurements for beams in Set B.

[0104] In one embodiment, a UE 202 may be configured with periodic or semi-persistent measurement resources for beams corresponding to configured / activated Set B and the UE 202 may be expected to update the measurements for beams in Set B after every measurement.

[0105] In one embodiment, for a UE side model activated for use during inferencing, a UE 202 may need to report the model input type to the gNB 204, such as whether the model expects LI measurements and / or beam indexes with or without assistance information. In one example, the model input type may be signaled implicitly based on the identity of the model being used or based on the configuration / activation of Set A / B / A+B based on prior offline coordination between UE 202 and network sides. In one alternative, the configuration of Set A / B / A+B may be associated with a model input and / or output type and this may also be used to identify the input and / or output of the model being used.

[0106] In one embodiment, for both network or UE side models, the model output or a function of the model output thereof may correspond to top-K identified beams where the value of K may be configured to the UE by RRC / MAC-CE or DCI configuration. In one example, multiple values of K can be configured and one of the candidate values can be selected by MAC-CE / DCI. In one example, the value of K may also be associated with the identity of the AI / ML model. In another example, the value of K may be configured to the UE 202 along with the Set A configuration explicitly or implicitly, such as a fraction of the number of beams configured in Set A for the implicit indication. The value of K may beindependent of the model output type, such as a beam index-based output or Ll-RSRP / SINR based output. This can be mapped to both model inferencing and model performance monitoring related methods. The purpose, such as for inferencing or monitoring, may be transparent to the UE 202.

[0107] In one embodiment, for UE side models, a UE 202 may be expected to report a quality metric associated with the reported predicted information regarding the DL Tx beams. In an example, the quality metric represents a measure of the confidence of the model outputs or the confidence of one or both of: the reported top-K beams and their corresponding Ll-RSRP values that may be derived from the model outputs. In this case, the LI reporting can additionally include the quantity associated with the confidence of the reported beams. In one example, the reported quantity can be an output from the AI / ML model and corresponding to the top-K or all of the beams in the activated Set A.

[0108] In one embodiment, representations of the Channel Impulse Response (CIR) obtained based on the measurements on CSI-RS resources may be used as inputs to the AI / ML model at the UE 202 or network (gNB 204) sides, wherein such representations may involve quantization of one or more of: the channel timing response ( relative timing of the detected paths for a time domain representation of the channel), amplitude or power for each of the detected paths for a time domain representation of the channel), and channel phase response (phase information corresponding to each of the detected paths for a time domain representation of the channel). The maximum number of detected paths to be used as model input may be specified or configured to a UE 202 via higher layers with the interpretation that if a UE 202 reports fewer than the maximum number of detected paths (for the case of AI / ML model being at the network side), it is understood that only the reported paths were detected.

[0109] In an example of the embodiment, the channel timing response (corresponding to the timing of the detected paths for the estimated CIR) may be defined based on the relative time of the detected first arrival path (FAP) with respect to the DL Rx timing for the UE 202 that may correspond to the start of the DL subframe in which the associated CSI-RS resource was received and timing of additional paths relative to the timing of the FAP. In a further example, this definition may be applied only for the case when a UE 202 is configured to report CIR related measurements to the gNB 204, e.g., for network-sided AI / ML models.

[0110] In another example of the embodiment, the power of the detected paths of the CIR may be reported using Ll-RSRP -like or SINR or SNR metrics where the Ll-RSRP or SINRor SNR is measured on the resource elements (REs) corresponding to CS1-RS resource for an indicated beam in Set B. In a further example, this definition may be applied only for the case when a UE 202 is configured to report CIR related measurements to the gNB 204, e.g., for network-sided AI / ML models.

[0111] In another example of the embodiment, the phase for each of the detected paths of the CIR may be defined such that the phase for the FAP is normalized to 0 degrees or 0 radians, and those for the subsequent paths are relative to the phase of the FAP. In a further example, this definition may be applied only for the case when a UE 202 is configured to report CIR related measurements to the gNB 204, e g., for network-sided AI / ML models.

[0112] In one embodiment, the configuration of measurement resource for a UE 202 can be such that the UE 202 can be expected to make multiple one-shot measurements on a set of configured resources where the measurements are within a time-window (observation window 602) configured to the UE via RRC or MAC-CE or DCI signaling.

[0113] In one embodiment, the time window for multiple measurements can be configured to the UE 202 as part of the Set B configurations. In one option, a UE 202 may be configured with periodic measurement resources and the start and end for the measurement window on which the UE 202 measures to update the measurements for Set B. In another example, a UE 202 may be configured with periodic measurement resources and the start and length for the measurement window on which the UE measures to update the measurements for Set B. In yet another example, a UE 202 may be provided with periodic measurement windows via indication of a start of the window, duration (or length) of the window, and a periodicity (latter using a combination of one or more of: numbers of subframes, slots, and symbols). The start of the window may be indicated with respect to System Frame Number (SFN) #0 or with respect to the end of the PDCCH carrying the DCI format if a DCI format is used to activate the measurement window. In such a case, the UE 202 may be expected to perform measurements using CSI-RS resources corresponding to the beams in Set B that occur within instances of the measurement window. In another example, a UE 202 may be provided via higher layers (RRC signaling) with configuration of multiple measurement windows that may or may not overlap in time. Further, for this example, UE 202 may be provided with an association between one or more measurement window and one or more models in case multiple candidate models may be considered at the UE side.

[0114] In an embodiment, a UE 202 may be provided with multiple configurations for observation window 602 and / or measurement prediction window 604 and eachconfiguration may be associated with a UE-sided model for model-level identification and / or to an AI / ML-based functionality. Accordingly, the configuration of measurement prediction window 604 may affect selection of the model in case of multiple models or activation of a model, or could be interpreted as applicability condition for a model to ensure consistency between training and inference phases.

[0115] In another embodiment, a UE 202 may be provided with one or multiple configurations for prediction windows 604 during which the beam prediction (spatial and / or temporal) based on UE-sided ML models may apply, and each configuration may be associated with a UE-sided model for model-level identification and / or to an AI / ML-based functionality. Accordingly, the configuration of prediction window 604 may affect selection of the model in case of multiple models or activation of a model, or could be interpreted as applicability condition for a model to ensure consistency between training and inference phases. For example, for UE-sided model for BM-Case2, for inference results report, information may be provided to support configuration of UE with N future time instances for inference by NW when applicable, how to determine reference time for the time instances, and duration values of the N time instances that can be predicted.

[0116] In one embodiment, a UE 202 may be expected to report a time index for each measurement. In one example, the time index may be an absolute time index, e.g., a time stamp, or, alternatively, it can be an index which indicates the time instances with respect to a configured measurement window. When provided with multiple measurement windows, UE 202 may be also expected to report the measurement window configuration index corresponding to which the measurements were performed.

[0117] Implementing AI / ML models for beamforming procedures offer certain advantages for life cycle management (LCM). In one embodiment, a UE 202 may be expected to determine one or more key performance indicators (KPIs) corresponding to AI / ML model performance for a UE side model. A KPI is a measurement or calculated metric that may be used to quantify the performance of an AI / ML model. In this case, the UE may be configured with multiple KPI types by RRC configuration. The RRC configuration can be based on UE capability reporting, where the UE 202 may report supported KPI types which may or may not be linked to specific AI / ML models. UE 202 can be configured by the NW to calculate a specific KPI type by MAC-CE or DCI signaling to select from one of multiple configured KPI types. Alternatively, a UE 202 may be configured via higher layers with one or more KPI ty pes that the UE 202 may be expected to determine. In a further example of the embodiment, a UE 202 may be configured to report the determined one or more KPIs tothe serving cell. Such information may be used by the network for LCM of the models at the UE-side.

[0118] In one embodiment, the KPIs reported by a UE 202 may also include a time index information related to the Al / ML model outputs and / or inputs which were used to evaluate the respective KPIs. In an example, the time index information corresponds to a time stamps corresponding to the model input and / or model output. For time information associated with model input, it could correspond to the start of the measurement window (or start of the first measurement window if multiple measurement windows are used) that was used for determining inputs to the ML model. For time information associated with model output, it could correspond to start of the prediction window 604, such as for temporal beam prediction or could also be applicable for spatial beam prediction for Set A beams.

[0119] In one embodiment, the configuration and selection of KPI types may be associated with selecting and activating an AI / ML model. In one example, a KPI may be associated to one AI / ML model by UE 202 implementation or higher layer signaling and activating a KPI may implicitly signal a switching or selection of AI / ML model. In one embodiment, the UE 202 can be configured to periodic, aperiodic or semi-persistent reporting of KPI values corresponding to an AI / ML Model inference using Uplink Control Information (UCI) on PUCCH or PUSCH or using MAC-CE over PUSCH or using RRC signaling over PUSCH. Examples for the RRC signaling option include: (i) using RRC containers / procedures similar to ULInformationTransfer procedure defined in 3GPP TS 38.331, and (ii) UE Assistance Information (UAI) reporting procedure defined in 3GPP TS 38.331.

[0120] In one embodiment, the KPI configured to a UE 202 may also include a reference or threshold value for the KPI which the UE 202 may be expected to use to report quality of the measured / calculated KPI for reporting. In one example, the UE 202 may report a KPI only if the value of the KPI differs from the configured threshold by a predetermined or configured amount. In this case, the UE 202 may trigger an event based on the difference of the KPI and the threshold and provide aperiodic event-triggered report to the network only when the event is triggered.

[0121] In one embodiment, based on the output of the UE-side AI / ML model, a UE 202 may request the NW to transmit reference signals using the top-K beams where the UE 202 can then perform LI measurements to determine the actual Ll-RSRP / SINR and report actual LI measurement or a KPI based on the measurement to the network for performance monitoring. In another example, if the model output is an LI measurement quantity7or is analogous to an LI measurement quantity, the UE 202 may choose to report the modeloutput or a KPI based on the model output instead of performing further measurements. In another example, the UE 202 may report the difference between the model output and the actual measurements of the corresponding beams.

[0122] In one embodiment, a UE 202 may be configured with a time window for measurement based on model output or collection of actual model outputs for model performance monitoring. In one example, the time window may be independent of the observation window 602 and prediction time prediction window 604 configured to the UE 202. In another example, the time window is identical to the observation window 602 and / or prediction window 604.

[0123] In one embodiment, for a UE side AI / ML model, a UE 202 may be configured with reference signal (RS) resources which are specific to model monitoring and different from model inferencing. In one example of this embodiment, the UE 202 may be configured with periodic CSI-RS resources corresponding to TCI states included in Set B of beams and the UE 202 may be expected to perform model monitoring based on LI measurements on these RSs and report the model inference based on the input from these measurements. In another example, the UE 202 may calculate a KPI based on the model inference and report the KPI. In yet another example, the UE 202 may compare the model output or any function of the output, e.g., a configured or pre-determined KPI, to a threshold value and send a report to the NW only if the comparison triggers a pre-determined event. In this case, the reporting can be aperiodic and the resources for aperiodic reporting can be either from existing UL resources or the UE 202 may request UL resources for the event-triggered aperiodic reporting. The reporting can be over UCI using PUSCH or PUCCH, or over MAC-CE using PUSCH, or using RRC signaling over PUSCH. In another example, the UE 202 may expect to be triggered by the gNB using MAC-CE and / or DCI signaling or configured by higher layers (RRC signaling) to perform model monitoring on pre-configured periodic or aperiodic or semi-persistent reference signal resources and report the model output or a calculated KPI to the gNB 204.

[0124] In one embodiment, a UE 202 may determine, based on current model inferencing, that model performance monitoring and / or reporting may be required and request reference signal transmissions from the gNB 204 corresponding to TCI states in a single or one of multiple configured sets B. In response, the gNB 204 may activate or trigger reference signal transmissions in an aperiodic or semi-persistent manner. In one example, these aperiodic or semi-persistent reference signal configurations may be provided apriori to the UE 202 based on RRC and / or MAC-CE configurations and may be further activated by MAC-CE and / or DCI signaling respectively.

[0125] In one embodiment, for a network side AI / ML model, a UE 202 may be configured with one or multiple Set B of beams wherein, the UE 202 may be configured to report LI measurements (RSRP / SINR) for performance monitoring for each activated Set B of beams to the NW in a periodic or aperiodic manner. The UE 202 may be pre-configured with periodic UL resources for periodic reporting or may be provided dynamically with UL resources for aperiodic or semi-persistent reporting based on MAC-CE or DCI signaling.

[0126] In one embodiment, a UE 202 may monitor LI measurements on one or more activated Set B of beams and determine if the LI measurement quantity corresponding to one or more beams in a Set B is degraded beyond a threshold, and if the condition is satisfied, trigger an event for aperiodic LI reporting on the UL to report either the LI measurements or a KPI calculated from the LI measurements to the gNB 204. In one example, the threshold can be associated with each Set B and may be configured to the UE. In another example, the threshold can be a fixed in the specifications value for each LI measurement quantity. In another example, the UE event trigger may be based on PUCCH transmission and based on this event trigger the gNB 204 may be expected to assign UL resources for reporting. In another example, the UE event trigger and the associated LI measurement report can be sent to the gNB 204 using a MAC-CE over PUSCH.

[0127] In one embodiment, for a UE-sided AI / ML model, for model monitoring at the UE- side, the UE 202 may request reference signal transmissions from the gNB 204 corresponding to beams in Set A but not in Set B. in order to determine the accuracy of the model outputs, such as model accuracy KPI-based model monitoring. In another alternative, the resources 404 requested by the UE 202 for model monitoring can also be the same as the resources 404 corresponding to the Set B of beams which are measured and used for model input. The UE 202 can compare the measured RSRP on the requested resources 404 to the output from the AI / ML model for the corresponding reference signals in Set A. If the model outputs differ from the measurements by a threshold, which can be configured to the UE 202 or can be a fixed threshold, the UE 202 may be required to report to the network that the current model is sub-optimal, such as indicate AI / ML model failure. Alternatively, when the outputs of the AI / ML model degrade beyond a threshold, the UE 202 may be configured to switch to another AI / ML model or to non-AI / ML fallback mechanisms. In yet another example, if the UE 202 reports model failure, the network may be expected to provide UE 202 with an alternate model via model transfer / delivery, or deliver a new or updated training dataset to the UE for retuning or retraining the current model. Here, a UE 202 may correspond to a UE-side Over-the-top (OTT) server that may receive the model or dataset from a network entity, e.g., network-side server.

[0128] In one embodiment, for a UE-sided AI / ML model, for model monitoring at the NW-side, the UE 202 may be configured with a set of measurement resources which the UE 202 may perform LI measurements and also perform model inference to get a predicted LI measurement for beams corresponding to the configured measurement resources. The UE 202 may be configured to report both the LI measurements as well as the predicted LI measurement from the model with respective identification, or the UE 202 may be configured to report the difference between the predicted LI measurement and the actual value of the LI measurement for the configured measurement resources. Further, thresholds or conditions may be specified or configured to a UE 202 to report the absolute or differential metrics for a subset of the measurement resources. In an example, a threshold on the difference between predicted LI measurement and actual LI measurement may be specified or configured such that UE 202 reports the metrics only when the difference is larger than the threshold. The reporting may be based on periodic reporting instances whereby UE 202 reports the configured absolute or differential metrics using periodic or semi -statically configured resources. Alternatively, the reporting could be event-driven and triggered by specified or configured conditions for such reporting. The network can use the reported values from the UE to perform model monitoring based on the model accuracy- related KPI. The network can subsequently configure the UE to deactivate the current model, deliver a new AI / ML model via model transfer / delivery, activate or switch to an already- pre-configured model, or deliver a new dataset to the UE 202 to re-train or retune the current model.

[0129] In one embodiment, the requirements for AI / ML model failure at the UE-side with UE-side model monitoring can be specified or configured to UE 202, such as the UE 202 is configured with a threshold and number of measurements that must be made. The UE 202 will then need to make the configured number of measurements and compare the measured metric (RSRP or Ll-RSRP) to the model output and if all or a certain percentage, higher than a configured or specified threshold percentage, of the model outputs are degraded beyond the specified or configured threshold, the UE may declare model failure. The model failure declaration can be based on UCI transmission by the UE on preconfigured PUCCH resources similar to SR. For more latency-tolerant cases, the model failure declaration could utilize MAC-CE signaling as another option. A response to the model failure indication maybe expected from the network wherein the network can indicate the UE 202 to switch to non-AI / ML fallback i.e., deactivate the current AI / ML model, or switch to a different AI / ML model which is already pre-configured, or deliver a new AI / ML model through model transfer or delivery-, or deliver a new training dataset to retrain or retune the currentmodel. In another example, the Al / ML model failure can be determined by the network based on UE reporting when model monitoring is on the UE side.

[0130] With respect to data collection for the model, the data collection circuitry 304 may collect data for training, or inference, or model monitoring purposes. In one embodiment, for a network side AI / ML model, when data collection circuitry 304 is triggered by the network for a configured Set A of beams, a UE 202 may be expected to report its Rx beam assumption corresponding to the LI measurements on Set A. In one example, the Rx beam assumption could be reported implicitly by the UE 202, such that the UE 202 has used the same Rx beam for all measurements in Set A. In another example, the UE 202 may report that it has used its best Rx beam for each reported LI measurement. In one option, the UE 202 may also be configured by the gNB 204 to report LI measurements using a specific Rx beam assumption. This configuration may either be based on RRC signaling or may be provided along with the trigger for data collection on Set A. In one option, the Rx beam assumption may also be configured along with the configuration of a Set A. In case more than one Set A is configured, the UE 202 can expect to be configured with an Rx beam assumption for each configured Set A for data collection.

[0131] In one embodiment, the Rx beam assumption may also be associated with an AI / ML model type and different AI / ML models may have different Rx beam assumptions. In one example, the Rx beam assumption configured to a Set A for data collection may also be assumed by the UE 202 for any configured / activated Set B which is associated with the same Set A and this assumption may be used by the UE 202 for reporting LI measurements on any Set B of beams associated with the Set A.

[0132] In one embodiment, for a UE-side AI / ML model, the UE 202 may trigger data collection from network in the form of resources 404 transmissions on which the UE 202 can perform LI measurements. In this case, the UE 202 may be preconfigured with one or multiple Set A configurations via RRC signaling wherein the Set A configurations may also include reference signal configurations or measurement resources for the UE to perform LI measurement. For data collection the UE 202 may trigger any of the preconfigured Set A through UCI over PUCCH or PUSCH and the gNB 204 can be expected to transmit resources 404 in an aperiodic or semi-persistent or periodic manner.

[0133] In one embodiment, the configuration for data collection can be differentiated based on training, inference, or model monitoring. In case of training data collection, the UE 202 may be configured with containers using higher layer signaling , such as RRC orMAC-CE to report the collected measurements of a Set A of beams (if configured) or a set of measurement resources.

[0134] With respect to model inference operations, in one embodiment of a UE-side AI / ML model, when the model output or the top-K beams based on the model output is a beam which is not part of the Set B of beams, a UE 202 can indicate in the UL LI beam report that the reported beam is part of Set A of beams but not part of a Set B of beams, such as the beams in the LI report have not been measured by the UE. In another embodiment, if the model output is Ll-RSRP or Ll-SINR, and the top-K measurements correspond to beams in Set A but not in Set B, the LI beam report sent by the UE 202 on the basis of this model inference may include information to signal to the gNB 204 that the reported LI measurement on the reported beams is an output from the UE-side model and not an actual LI measurement. Alternatively, if the UE side model output or a function of the model output for top-K beams contains beams in Set A but not in Set B, the UE 202 may signal to the gNB 204 to transmit resources 404 corresponding to the TCI states for the beams from the model output which are in Set A but are not in Set B such that the UE can measure LI RSRP / SINR on the transmitted reference signals and then send the LI beam report which contains actual LI measurements made on all reported beams.

[0135] In various embodiments, a trained ML model 320 may be selected, activated, deactivated, switched, or otherwise managed for the UE 202 and / or gNB 204. In one embodiment, a UE 202 may report its capability for simultaneous monitoring of performance of UE-sided AI / ML model and performance of one or more non-AI / ML-based beam prediction methods. Based on reported UE capabilities, the serving gNB 204 may configure the UE 202 to compare performance of one or more UE-sided AI / ML model(s) to that of one or more non-AI / ML-based beam prediction methods. Results from such comparisons could be used by the UE 202 to perform model LCM, including one or more of: model update, model switching, or fallback to non-AI / ML methods. The decisions for the different model LCM aspects may be based on configured or specified thresholds for one or more KPIs for which the UE 202 may be configured to determine. Alternatively, such results may be reported by the UE 202 to the serving gNB 204 for model LCM performed by the network.

[0136] In another embodiment, a UE 202 may be configured to monitor performance of UE-sided AI / ML model and performance of one or more non-AI / ML-based beam prediction methods wherein one or more of: the prediction window and the measurement window- for the AI / ML model-based beam prediction and non-AI / ML method-based beam prediction may not be the same. In this case, the prediction and / or measurement window7(asapplicable) for non-AI / ML-based beam prediction may be separately configured to the HE 202 by the network. Alternatively, in this case, the prediction window 604 and / or measurement window (as applicable) for non-AI / ML-based beam prediction may be defined relative to the corresponding windows for AI / ML model-based beam prediction.

[0137] It is worthy to note that other examples of model activation / selection / switching and evaluation of additional network-side conditions for UE-sided models have been discussed previously, such as in the context of configurations related to model application for beam management use-cases.

[0138] For beam indication and application, based on the output of a NW-side AI / ML or UE-side AI / ML model inference, a UE 202 may expect to be indicated with a joint / DL / UL TCI state which is not from the set of activated TCI states. The beam indication may be done using a new DCI format which indicates TCI states which are configured to be part of the Set A of beams or from the list of all configured TCI states wherein the bit- width of the beam indication field is increased from 3 bits to n bits wheren~TCI where Ara is the total number of TCI states in Set A or the total number of configured TCI states. In an alternative, the beam indication can be performed using DCI format 1_1 / 1_2 wherein the bit width of the TCI indication field is determined based on higher layer configuration for AI / ML based beam indication, such as the bit width is increased to n bits from the current 3 bits.

[0139] FIG. 5 illustrates a message flow 500 suitable for implementing embodiments as described herein. Specifically, the message flow 500 is an example of a message flow for BM-Case2. Embodiments are not limited to this example.

[0140] For BM-Case2, a temporal domain beam prediction technique is used to predict the gNB 204 and / or UE 202 beams used for future transmission and reception. In general, the beam prediction process includes two phases, the observation phase 502 and the prediction phase 504. The beam prediction can be implemented at the BS-side, or the UE-side, or both sides.

[0141] In observation phase 502, measurements are made, e.g., Ll-RSRP is collected. The gNB 204 performs a full or partial CSA / SSB beam sweep 506. The UE 202 performs measurements 412 using the measurement unit 410, and sends a measurement report 414 to the gNB 204. The measurement report 414 may include the measurements 412, such as RSRP feedback 508. The beam sweep 506 and the RSRP feedback 508 are repeated for a defined set of time intervals.

[0142] In prediction phase 504. the measurements are fed into the beam prediction model 512, such as trained ML model 320, that generates a set of predicted beams 514. The gNB 204 and / or UE 202 uses the predicted beams 514 to transmit and receive data in the prediction phase 504.

[0143] FIG. 6 illustrates a logic diagram 600. The logic diagram 600 provides an example of the trained ML model 320 using measurement values 606 to generate predicted values 608 used for beam indices 610 of the antenna array 402 of the gNB 204. Embodiments are not limited to this example.

[0144] An AI / ML based implementation of the temporal domain beam prediction function allows for measurements over an observation window 602 to be fed into a trained ML model 320 which will then predict the measurements for a prediction window 604 and these predictions can be used to infer the best beam indices 610 or set of beams to be used. For the temporal domain beam prediction problem, in addition to the LI measurement and reporting updates as discussed previously, it may be necessary to enable configuration of an observation window 602 and a prediction window 604. Depending on where the trained ML model 320 resides, one or both may need to be configured to the UE 202. Furthermore, depending on model implementation and if the trained ML model 320 is transferred from one node to another, it may also be necessary7to configure model selection at the inference node based on different prediction and / or observation window lengths.

[0145] Finally, if the trained ML model 320 resides at the UE-side, in order to perform measurements 412 during the observation phase 502, the UE 202 may need to trigger aperiodic reference signal transmissions from the gNB 204.

[0146] FIG. 7 illustrates a message flow 700. The message flow 700 is an example of a procedure for configuring and activating AI / ML models for BM related use cases, such as BM-Casel and BM-Case 2, for example. Embodiments are not limited to this example.

[0147] RAN2 on a set of terminology7with respect to BM related use cases. Supported functionalities refer to functionals that a UE 202 can indicate by using UE capability7information (e g., via RRC / LPP signaling). Applicable functionalities refer to functionalities that a UE 202 is ready to apply for inference operations. Activated functionalities refer to functionalities already enabled for performing inference operations.

[0148] As depicted in FIG. 7, the message flow 700 illustrates a signaling procedure on applicable functionality reporting for beam management using a UE-sided model. The network, via the gNB 204, sends a message 702 such as a UECapability Enquiry. The UECapabilityEnqiry message initiates the procedure to a UE 202 reporting its AI / MLsupported functionalities. The UE 202 receives the message 702, and it sends a message 704 such as a UECapablity Information message to the network, containing supported functionalities at the UE side. In message 704, the UE 202 reports its UE-capability information / parameters. such as Rel-19 AI / ML-specific FGs (including components and corresponding value ranges). These AI / ML-specific UE capability information / parameters will depend on how FGs are defined including the granularity.

[0149] The gNB 204 receives the message 704, and it sends a message 706 such as an RRCReconfiguration message. The RRCReconfiguration may include configuration information provided from NW to UE: (1) UE is allowed to do UAI reporting via OtherConflg; (2) network may provide NW-side additional condition using RRC signaling; and (3) inference configuration of supported functionalities. Between receiving the message 706 and sending back a message 708, the UE 202 decides the applicable functionalities based on NW-side additional conditions (if provided), UE-side additional conditions (internally known by UE 202) and model availability in device. Other configuration may be considered by UE 202 (e.g. inference configuration). The applicable functionality is decided by the UE 202 if NW-side additional condition is not provided.

[0150] The UE 202 sends a message 708 for applicable functionality reporting to the gNB 204 in the following scenarios: (1) upon being configured to provide applicable functionality and upon change of applicable functionality via UAI; (2) as response to NW- side additional condition requesting applicable functionality reporting in message 706 or other network configurations (e.g. inference configuration).

[0151] The gNB 204 sends a message 710 such as an RRCReconfiguration message to the UE 202. The network configures inference configuration to UE after applicable functionality reporting, if inference configuration based on supported functionality is not provided in message 706, such as providing inference configuration in message 710. If inference configuration based on supported functionality is provided in message 706, it is up to network implementation whether to provide an updated configuration or not.

[0152] Once configured, the UE 202 and the gNB 204 may exchange model management messages for model management 712 of the AI / ML model, such as activation of a model, deactivation of a model, inference configuration, performance monitoring information, and so forth.

[0153] RAN2 also agreed the applicable functionality may be activated by receiving its inference configuration when it is provided in message 710, optionally including the initial activation state, the initial state of applicable functionality if inference configuration ofsupported functionality is provided in message 706. additional L1 / L2 signaling for activation / deactivation, if multiple applicable functionalities can be activated at the same time, and what is the granularity of functionality. NW-side additional condition is assumed as associated ID in RAN2.

[0154] RANI further outlined the following options for applicability7for inference for UE- side model, referred to as options 1, 2, and 3.

[0155] In option 1. in message 706. the following configurations are provided from NW to UE: (1) UE is allowed to do UAI reporting via OtherConfig; (2) and (3) NW configures one or more CSI-ReportConfig for inference configuration, where the associated ID may be configured in CSI framework as working assumption applied. Some lEs in the CSI report configuration can be removed or modified. CSI report configuration for UE-side model inference might not be activated immediately upon receiving message 706. In message 708, UE 202 reports applicability of the above CSI-ReportConfig^ one or more of the above CSI- ReportConfig to be reported, activation (including when / how) of inference report after obtaining the applicability from UE 202 in message 708, and whether message 710 is optional.

[0156] In option 2, in message 706, following configurations are provided from NW to UE: (1) UE is allowed to do UAI reporting via OtherConfig; (2) NW configures one set or multiple sets of inference related parameters, where the set of inference related parameters is not configured by CSI-ReportConfig, and the set of inference related parameters includes Set A related information, Set B related information, report content related information, and for BM-Case 2, time instances related information for measurements, time instances related information for prediction; and (3) the associated ID(s) may be configured, wherein the associated ID(s) may be part of one set of the inference related parameters, or independently from the one set of the inference related parameters. In message 708, UE 202 reports applicability of the above one or multiple sets of inference related parameters, where the associated ID information may be associated. In message 710, NW configures configuration(s) for CSI report for inference.

[0157] In option 3, for message 706, the following configurations are provided from NW to UE: (1) UE is allowed to do UAI reporting via OtherConfig: and (2) the associated ID(s) may be provided to UE, e.g., a new RRC parameter. In message 708, UE reports by UAI the applicable one or multiple sets of inference related parameters may be included. The set of inference related parameters may include Set A related information, Set B related information, report content related information, and for BM-Case2, time instances relatedinformation for measurements, time instances related information for prediction, where not applicable may also be replied by UE and if the inference related parameters are not supported for reporting, only the applicability’s or not is reported in message 708. The associated ID(s) may be included as part of the inference related parameters, or independently from the set of the inference related parameters. In message 710, NW configures configuration(s) for CSI report for inference. There is no impact of configuring CSI report configuration for non- Al beam management in RRC Reconfiguration.

[0158] Other options are available as well. For example, in message 706, the following configurations are provided from NW to UE: (1) UE is allowed to do UAI reporting via OtherConfig: (2) the applicability report is based on options (A) and / or (B). The container design is configurable. In option (A) one or more of CSI-ReportConfig for inference configuration, wherein the associated ID may be configured in CSI framework as working assumption applied. CSI report configuration for UE-side model inference should not be activated immediately upon receiving message 706. In option (B), one set or multiple sets of inference related parameters for applicability report only (not for inference). The container design is configurable. The set of inference related parameters selected from the IES in / or the IES referred by CSI-ReportConfig as a starting point, e.g., the associated ID (this doesn’t imply the associated ID is mandatory). Set A related information, Set B related information, report content related information, and for BM-Case2, time instances related information for measurements, and time instances related information for prediction. In message 708, UE 202 reports applicability for all the above (A) one or more CSI- ReportConfig and / or (B) set(s) of inference related parameters. Other information along with the applicability may be configured as well. If option (A) is configured in message 706, applicable aperiodic CSI Report and semi -persistent CSI report can be activated / triggered by NW after the applicability reported, and applicable periodic CSI Report is considered as activated only if the applicability of the corresponding CSI- ReportConfig is reported in RRCReconfigurationComplete. In message 710, the NW can optionally configure CSI-ReportConfig for inference configuration in RR(' Reconfiguration. where the associated ID may be configured in CSI framework as working assumption applied. The message 710 may be optional if UE has already been configured with CSI- ReportConfig in message 706.

[0159] FIG. 8 illustrates an apparatus 800 suitable for implementation as a UE 202 in the wireless communications wireless communication system 100a. As previously discussed, the UE 202 may take measurements and actions based on one or more measurement criteriaas defined by 3 GPP standards or non-3GPP standards. Embodiments are not limited in this context.

[0160] As depicted in FIG. 8, the apparatus 800 may comprise a processor circuitry 804, a memory 806 with a measurement manager 810. an interface 818, a data storage device 822, and radio-frequency (RF) 820. The interface 818 may store machine-readable instructions (e.g., program code) for software applications that when executed causes the software applications to perform certain defined functions. Non-limiting examples of the applications may include an encoder / decoder such as codec 808 and a measurement generator 802. The codec 808 encodes and decodes messages. The measurement generator 802 performs measurements for the UE 202 using one or more sensors. The apparatus 800 may optionally include a set of platform components (not shown) suitable for a UE 202, such as input / output devices, memory’ controllers, different memory types, network interfaces, hardware ports, and so forth.

[0161] The apparatus 800 for the UE 202 may receive and decode information form one or more messages from an access node 216 via the RF circuitry 820. The access node 216 may comprise part of a RAN node implemented as, for example, a NodeB, an eNB, or a gNB 204 of the wireless communications wireless communication system 100a. The UE 202 may also receive, decode and measure a set of resources 404 (e.g., reference signals) from the access node 216.

[0162] The measurement manager 810 receives the decoded information and performs measurements for the resources 404 using a measurement object 812. The measurement generator 802 retrieves a measurement object 812, a set of one or more measurement criteria 814, and it begins to generate measurement values 816 according to the set of measurement criteria 814.

[0163] A measurement object 812 refers to a specific entity or parameter that is being measured or monitored by the UE 202. It represents a target for measurement or evaluation within the radio frequency (RF) environment. The measurement object 812 is used in the context of various measurement procedures and functions within the 5G or 6G network. Examples for the measurement object 812 includes measurements for: (1) signal quality, signal strength, or other relevant parameters of serving cells and neighboring cells to assist in handover decisions and interference management; (2) reference signals received from the gNB 204 or cells in the vicinity to estimate signal quality, timing, and other characteristics for purposes like beamforming, channel estimation, or synchronization; (3) radio resource blocks (RBs) to determine quality or interference level of specific RBs to evaluate thesuitability for data transmission; (4) channel quality indicators (CQls) for channel conditions to provide feedback to the gNB 204, which aids in link adaptation and scheduling decisions; (5) interference levels caused by neighboring cells or sources to assess the impact on the communication link; (6) parameters related to movement, speed, velocity, or pathloss to assist handover, location-based services, or mobility management; and (7) resources 404 used for beamforming operations of an antenna array 402 for the UE 202 and / or the gNB 204. These are just a few examples of a measurement object 812 in 5G or 6G. Measurement object 812 is used for performance monitoring, network optimization, and providing relevant information for efficient and reliable communication in the network.

[0164] For example, the measurement object 812 may include the reference signals communicated between the UE 202 and the gNB 204. In this case, the measurement object 812 may comprise, for example, reference signals for Ll-RSRP / SINR measurement such as CSI-RS or SSB resources, SS-RSRP measurement, reference signals for SS-RSRQ measurement, BFD reference signals, RLM reference signals, SDT reference signals, or any other signals suitable for measurement or relaxed measurement in the wireless communications wireless communication system 100a.

[0165] The apparatus 800 for the UE 202 may include the interface 818. The interface 818 may be arranged to send or receive, to or from a data storage device 822, resource information for resources 404, measurement values 816 for measurement report 414. beam information 416 received from the access node 216, and other information associated with a 5G NR or 6G system. The data storage device 822 may be located external to the UE 202 (off-device) or the data storage device 822 may be located internal to the UE 202 (on- device). When the data storage device 822 is implemented on-device, the data storage device 822 may comprise volatile or non-volatile memory, as described in more detail with reference to FIG. 16.

[0166] The UE measurement information may comprise one or more measurements 412 or measurement values 816 as measured by the measurement units 410, the measurement generator 802 and / or measurement criteria 814 for the UE 202. The UE 202 may be provisioned with the measurement criteria 814 by an original equipment manufacturer (OEM) or as received from the access node 216 via RRM, RRC, or other control signaling.

[0167] The UE 202 may also comprise a model manager 824. The model manager 824 manages the trained ML model 320 when instantiated for the UE 202. For example, the model manager 824 may activate the trained ML model 320, deactivate the trained ML model 320. update the trained ML model 320, replace the trained ML model 320 withanother Al / ML model, measure performance of the trained ML model 320, and perform other management operations.

[0168] FIG. 9 illustrates an apparatus 900 suitable for implementation as an access node 216 in the wireless communication system 100a. The access node 216 is an example of the gNB 204. As previously discussed, the access node 216 may receive UE capability information, associated with the trained ML model 320, from the UE 202. The access node 216 may send UE configuration information, such as model information for the trained ML model 320 or beam information 416. to the UE 202. The access node 216 may use the UE capability information to generate the model information or the beam information 416 in accordance with one or more 3GPP standards or non-3GPP standards. Embodiments are not limited in this context.

[0169] As depicted in FIG. 9, the apparatus 900 may comprise a processor circuitry 902, a memory 904, an interface 906, a data storage device 908, and RF circuitry 910. The memory 904 may store machine-readable instructions (e.g., program code) for software applications that when executed causes the software applications to perform certain defined functions. Non-limiting examples of the applications may include an encoder / decoder such as codec 912, the measurement manager 914, the model manager 920, and / or the trained ML model 320. The codec 912 encodes and decodes messages. The apparatus 900 may optionally include a set of platform components (not shown) suitable for an access node 216, such as input / output devices, memory controllers, different memory types, network interfaces, hardware ports, and so forth.

[0170] The apparatus 900 for the access node 216 may receive one or more messages from one or more UEs 202 via the RF circuitry 910. The access node 216 may comprise part of a RAN node implemented as, for example, a NodeB, an eNB, or a gNB 204 of the wireless communication system 100a.

[0171] The codec 912 receives and decodes encoded messages from the UE 202. The codec 912 encodes and sends encoded messages to the UE 202.

[0172] The access node 216 includes an interface 906 to send or receive, to or from a data storage device 908 for a wireless communication system 100a. The access node 216 also includes processor circuitry 902 communicatively coupled to the memory 904, the processor circuitry 902 to execute a codec 912 to decode a message from a UE 202 with UE capability information, such as information associated with the trained ML model 320, and measurements 412 and / or measurement values 816 in a measurement report 414. The UE capability information may include, for example, supported functionalities for an AI / MLmodel such as trained ML model 320, applicable functionalities associated with the trained ML model 320, and / or activated functionalities associated with the trained ML model 320. Supported functionalities refer to functionalities that UE 202 can indicate by using UE capability' information (via RRC / LPP signaling). Applicable functionalities refer to functionalities that the UE is ready to apply for inference. Activated functionalities refer to functionalities already enabled for performing inference

[0173] The access node 216 may also comprise a model manager 920. The model manager 920 manages the trained ML model 320 when instantiated for the access node 216. For example, the model manager 920 may activate the trained ML model 320, deactivate the trained ML model 320, update the trained ML model 320, replace the trained ML model 320 with another AI / ML model, measure performance of the trained ML model 320, and perform other management operations.

[0174] Operations for the disclosed embodiments may be further described with reference to the following figures. Some of the figures may include a logic flow. Although such figures presented herein may include a particular logic flow, it can be appreciated that the logic flow merely provides an example of how the general functionality as described herein can be implemented. Further, a given logic flow does not necessarily have to be executed in the order presented unless otherwise indicated. Moreover, not all acts illustrated in a logic flow may be required in some embodiments. In addition, the given logic flow may be implemented by a hardware element, a software element executed by a processor, or any combination thereof. The embodiments are not limited in this context.

[0175] FIG. 10 illustrates an embodiment of a logic flow 1000. The logic flow 1000 may be representative of some or all of the operations executed by one or more embodiments described herein. For example, the logic flow 1000 my include some or all of the operations performed by devices or entities within the wireless communication system 100a, the wireless communication system 200, the functional framework 300, the message flow 400, the message flow 500, the logic diagram 600, the message flow 700, the apparatus 800, the apparatus 900, or any UE operable therein. More particularly, the logic flow 1000 illustrates operations for the UE 202 associated with an AI / ML model such as the ML model 302 and / or trained ML model 320. Embodiments are not limited in this context.

[0176] In block 1002, logic flow 1000 comprises accessing information for a first set of resources and a second set of resources associated with beamforming operations of an antenna array. In block 1004, the logic flow 1000 comprises obtaining measurements for the second set of resources. In block 1006, the logic flow 1000 comprises accessing beaminformation to configure a set of beams associated with the first set of resources, the beam information generated by a machine learning (ML) model using the measurements for the second set of resources. In block 1008, the logic flow 1000 comprises configuring the antenna array to generate the set of beams using the beam information. In block 1010, the logic flow 1000 comprises generating the set of beams using the antenna array to send information over a wireless medium.

[0177] By way of example, a UE 202 of a wireless communication system 100a, comprises an interface 818 and processor circuitry 804 to access information for a first set of resources such as Set A 406 and a second set of resources such as Set B 408 associated with beamforming operations of an antenna array 402. The processor circuitry 804 obtains measurements 412 for the second set of resources Set B 408. The processor circuitry 804 accesses beam information 416 to configure a set of beams 418 associated with the first set of resources Set A 406. The beam information 416 is generated by an ML model 302, such as trained ML model 320, using the measurements 412 for the second set of resources Set B 408. The processor circuitry 804 configures the antenna array 402 to generate the set of beams 418 using the beam information 416. The antenna array 402 generates the set of beams 418 to send information over a wireless medium using radio-frequencies.

[0178] In some embodiments, for example, the first set of resources Set A 406 and the second set of resources Set B 408 comprise a joint set of transmission configuration indicator (TCI) states. In some embodiments, for example, the first set of resources Set A 406 includes a subset of TCI states of the second set of resources Set B 408. In some embodiments, for example, the second set of resources Set B 408 includes a subset of TCI states of the first set of resources Set A 406.

[0179] In some embodiments, for example, the processor circuitry 804 generates the beam information 416 by the trained ML model 320 at the UE 202. In some embodiments, for example, the access node 216 generates the beam information 416 by the trained ML model 320 at the access node 216, and it sends the beam information 416 to the UE 202.

[0180] In some embodiments, for example, when the trained ML model 320 is implemented by the access node 216, the processor circuitry 804 decodes a first message from the access node 216, the first message including the information for the first set of resources Set A 406 or the second set of resources Set B 408. In some embodiments, for example, the first message includes Set A 406, Set B 408, or a combination of Set A 406 and Set B 408. This information may be sent in a single message, or in multiple messages, as RRC messages. Alternatively, this information may be stored in the memory 806 of theUE 202. The processor circuitry 804 retrieves this information from the memory 806 when configuring the antenna array 402 for performing beamforming operations.

[0181] In some embodiments, for example, the processor circuitry 804 encodes a second message for the access node 216. the second message including a measurement report 414 with the measurements 412 for the second set of resources Set B 408.

[0182] In some embodiments, for example, the processor circuitry 804 encodes a fourth message for the access node 216. the fourth message includes a quality metric associated with the beam information 416 generated by the ML model. For example, the UE 202 may send a confidence value as the quality metric, where the confidence value represents a confidence level in the output of the trained ML model 320, such as the beam information 416.

[0183] In some embodiments, for example, the beam information 416 corresponds to a topic identified beams 418 where a value for K is configured for the trained ML model 320 using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI).

[0184] In some embodiments, for example, the processor circuitry 804 decodes a third message from the access node 216. the third message includes the beam information 416 generated by the trained ML model 320 using the measurements 412 for the second set of resources Set B 408.

[0185] In some embodiments, for example, the processor circuitry 804 decodes a third message from the access node 216, the third message includes the beam information 416 generated by the trained ML model 320 using historical measurements 412 for the second set of resources Set B 408. In this case, the access node 216 may collect measurements 412 from multiple UEs 202 within a given cell over time, train the trained ML model 320 using the historical measurements 412, and deploy the trained ML model 320 for using new measurements 412 from the UE 202 to predict or generate the beam information 416.

[0186] FIG. 11 illustrates an embodiment of a logic flow 1100. The logic flow 1100 may be representative of some or all of the operations executed by one or more embodiments described herein. For example, the logic flow 1100 my include some or all of the operations performed by devices or entities within the wireless communication system 100a, the wireless communication system 200. the functional framework 300, the message flow 400, the message flow 500, the logic diagram 600, the message flow 700. the apparatus 800, the apparatus 900, or any access node 216 operable therein, such as gNB 204. More particularly, the logic flow 1100 illustrates operations for the access node 216 associatedwith an Al / ML model such as the ML model 302 and / or trained ML model 320. Embodiments are not limited in this context.

[0187] In block 1102, logic flow 1100 comprises accessing information for a first set of resources and a second set of resources associated with beamforming operations of an antenna array. In block 1104, the logic flow 1100 comprises obtaining measurements for the second set of resources. In block 1106, the logic flow 1100 comprises generating beam information to configure a set of beams associated with the first set of resources by a machine learning (ML) model using the measurements for the second set of resources. In block 1108, the logic flow 1100 comprises encoding the beam information in a message for a user equipment (UE) to configure the antenna array to generate the set of beams using the beam information. In block 1110, the logic flow 1100 comprises generating the set of beams using an antenna array to send information over a wireless medium.

[0188] By way of example, an access node 216 of a wireless communication system 100a, comprises an interface 906 and processor circuitry' 902 arranged to access information for a first set of resources Set A 406 and a second set of resources Set B 408 associated with beamforming operations of an antenna array 402. The processor circuitry 902 obtains measurements 412 for the second set of resources Set B 408. The processor circuitry 902 generates beam information 416 to configure a set of beams 418 associated with the first set of resources Set A 406 by a trained ML model 320 using the measurements 412 for the second set of resources Set B 408. The processor circuitry 902 encodes the beam information 416 in a message for a UE 202 to configure the antenna array 402 to generate the set of beams 418 using the beam information 416.

[0189] In some embodiments, for example, the first set of resources Set A 406 and the second set of resources Set B 408 comprise ajoint set of transmission configuration indicator (TCI) states. In some embodiments, for example, the first set of resources Set A 406 includes a subset of TCI states of the second set of resources Set B 408. In some embodiments, for example, the second set of resources Set B 408 includes a subset of TCI states of the first set of resources Set A 406.

[0190] In some embodiments, for example, when the trained ML model 320 is implemented by the access node 216, the processor circuitry 902 encodes a first message for the UE 202, the first message including the information for the first set of resources Set A 406 or the second set of resources Set B 408. In some embodiments, for example, the first message includes Set A 406, Set B 408, or a combination of Set A 406 and Set B 408. This information may be sent in a single message, or in multiple messages, as RRC messages.Alternatively, this information may be stored in the memory 806 of the UE 202, in which case the access node 21 does not need to encode and send the first message.

[0191] In some embodiments, for example, the processor circuitry 902 decodes a second message from the UE 202, the second message including a measurement report 414 with the measurements 412 for the second set of resources Set B 408.

[0192] In some embodiments, for example, the processor circuitry 902 encodes a third message for the UE 202, the third message including the beam information 416 generated by the trained ML model 320 using the measurements 412 for the second set of resources Set B 408.

[0193] In some embodiments, for example, the processor circuitry 902 encodes a third message for the UE 202, the third message includes the beam information 416 generated by the ML model using historical measurements 412 for the second set of resources Set B 408. In this case, the access node 216 may collect measurements 412 from multiple UEs 202 within a given cell over time, train the trained ML model 320 using the historical measurements 412, and deploy the trained ML model 320 for using new measurements 412 from the UE 202 to predict or generate the beam information 416.

[0194] In some embodiments, for example, the processor circuitry 902 decodes a fourth message from the UE 202, the fourth message includes a quality metric associated with the beam information 416 generated by the trained ML model 320.

[0195] In some embodiments, for example, the beam information 416 corresponds to a topic identified beams where a value for K is configured for the trained ML model 320 using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI).

[0196] FIG. 12 illustrates a network architecture 1200. FIG. 12 illustrates block diagrams of NF components (NFs) and interfaces in connection with embodiments / aspects described herein. In the 5G network architecture of FIG. 12, a next generation (NG) radio access network (RAN) (NG-RAN) comprises a functional split feature that splits a gNodeB (gNB) (also referred to as an '‘NG RAN,” '‘NG RAN node,” or the like) into a gNB-Centralized Unit (CU) (gNB-CU) that implements the upper layer of gNB function and gNB-Distributed Unit (DU) (gNB-DU) that implements the lower layer gNB function. The 5G core NFs and gNB-CU can be implemented as Virtualized Network Functions (VNFs), and the gNB-CU and / or gNB-DU can be implemented as Physical Network Function(s) (PNF(s). An Operator can create a virtualized 5G networks by using the European Telecommunications Standards Institute (ETSI) network functions virtualization (NFV) lifecycle management function toinstantiate a Network Service (NS) in the cloud that includes various VNFs (e.g.. 5G core NFs, gNB-CU), PNFs (e.g. gNB-DU), and VNF Forwarding Graph(s) (VNFFG(s)).

[0197] FIG. 12 illustrates an architecture of a network architecture 1200 including a second CN 1234 in accordance with various embodiments. The network architecture 1200 is similar to the wireless communication system 100a, and may illustrate equipment, devices and network elements similar to those described with reference to the wireless communication system 100a. As depicted in FIG. 12, the network architecture 1200 includes a user equipment (UE) 1224, a RAN 1226 or access node (AN); and a DN 1230, which can all be the same or similar to similarly named elements as discussed herein. The DN 1230 is the same or similar to the DN 134, and it can implement, for example, operator services, Internet access or 3rd party services, as discussed further below. The CN 114 maybe implemented as a 5GC or 5GS, and it can include an Authentication Server Function (AUSF) 1216; an AMF 1218; an SMF 1220; a NEF 1204; a PCF 1208; an NRF 1206; a Unified Data Management (UDM) 1210; an application function (AF) 1212; a SCP 1222; a user plane function (UPF) 1228; Network Slice-Specific Authentication and Authorization Function (NSSAAF) 1214; and an NSSF 1202, each with respective components for processing corresponding 5GC network functions (NFs).

[0198] The UPF 1228 can act as an anchor point for intra- RAT and inter- RAT mobility-, an external protocol data unit (PDU) session point of interconnect to DN 203, and a branching point to support multi-homed PDU session. The UPF 1228 can also perform packet routing and forwarding, perform packet inspection, enforce the user plane part of policy rules, lawfully intercept packets (UP collection), perform traffic usage reporting, perform QoS handling for a user plane (e.g., packet filtering, gating, uplink (UL)Zdownlink (DL) rate enforcement), perform Uplink Traffic verification (e.g., Service Data Flow (SDF) to Quality of Service (QoS) flow mapping), transport level packet marking in the uplink and downlink, and perform downlink packet buffering and downlink data notification triggering. UPF 1228 can include an uplink classifier to support routing traffic flows to a data network. The DN 1230 can represent various network operator services, Internet access, or third party services. DN 1230 can include, or be similar to. application server XQ30 discussed previously. The UPF 1228 can interact with the SMF 1220 via an N4 reference point between the SMF 1220 and the UPF 1228.

[0199] The AUSF 1216 can store data for authentication of UE 202 and handle authentication-related functionality . The AUSF 1216 can facilitate a common authentication framework for various access types. The AUSF 1216 can communicate with the AMF 1218 via an N12 reference point between the AMF 1218 and the AUSF 1216; and cancommunicate with the UDM 1210 via an N13 reference point between the UDM 1210 and the AUSF 1216. Additionally, the AUSF 1216 can exhibit an Nausf service-based interface.

[0200] The AMF 1218 can be responsible for registration management (e.g., for registering UE 202, etc.), connection management, reachability management, mobility management, and lawful interception of AMF -related events, and access authentication and authorization. The AMF 1218 can be a termination point for an Ni l reference point between the AMF 1218 and the SMF 1220. The AMF 1218 can provide transport for SM messages between the UE 202 and the SMF 1220. and act as a transparent proxy for routing SM messages. AMF 1218 can also provide transport for SMS messages between UE 202 and a Short Message Service (SMS) function (SMSF) (not shown by FIG. 12). AMF 1218 can act as Security7Anchor Function (SEAF), which can include interaction with the AUSF 1216 and the UE 202, receipt of an intermediate key that was established as a result of the UE 202 authentication process. Where Universal Subscriber Identity Module (USIM) based authentication is used, the AMF 1218 can retrieve the security material from the AUSF 1216. AMF 1218 can also include a Security Context Management (SCM) function, which receives a key from the SEAF that it uses to derive access-network specific keys. Furthermore, AMF 1218 can be a termination point of a RAN CP interface or RAN connection point interface, which can include or be an N2 reference point between the RAN 1226 and the AMF 1218; and the AMF 1218 can be a termination point of Non Access Stratum (NAS) layer (Nl) signaling, and perform NAS ciphering and integrity protection.

[0201] AMF 1218 can also support NAS signaling with a UE 202 over an N3 Interworking Function (IWF) interface. The N3 IWF can be used to provide access to untrusted entities. N3IWF can be a termination point for the N2 interface between the RAN 1226 and the AMF 1218 for the control plane, and can be a termination point for the N3 reference point between the RAN 1226 and the UPF for the user plane. As such, the AMF 1218 can handle N2 signaling from the SMF 1220 and the AMF 1218 for PDU sessions and QoS, encapsulate / de-encapsulate packets for IPSec and N3 tunneling, mark N3 user-plane packets in the uplink, and enforce QoS corresponding to N3 packet marking considering QoS requirements associated with such marking received over N2. N31WF can also relay uplink and downlink control-plane NAS signaling between the UE 202 and AMF 1218 via an Nl reference point between the UE 202 and the AMF 1218, and relay uplink and downlink user-plane packets between the UE 202 and UPF 1228. The N3IWF also provides mechanisms for IPsec tunnel establishment with the UE 202. The AMF 1218 can exhibit an Namf service-based interface, and can be a termination point for an N14 reference pointbetween two AMFs and an N17 reference point between the AMF 1218 and a 5G- Equipment Identity Register (EIR) (not shown by FIG. 12).

[0202] The UE 202 can need to register with the AMF 1218 in order to receive network services. Registration Management (RM) is used to register or deregister the UE 202 with the network (e.g., AMF 1218), and establish a UE context in the network (e.g., AMF 1218). The UE 202 can operate in an RM-REGISTERED state or an RM-DEREGISTERED state. In the RM-DEREGISTERED state, the UE 202 is not registered with the network, and the UE context in AMF 1218 holds no valid location or routing information for the UE 202 so the UE 202 is not reachable by the AMF 1218. In the RM-REGISTERED state, the UE 202 is registered with the network, and the UE context in AMF 1218 can hold a valid location or routing information for the UE 202 so the UE 202 is reachable by the AMF 1218. In the RM-REGISTERED state, the UE 202 can perform mobility’ Registration Update procedures, perform periodic Registration Update procedures triggered by expiration of the periodic update timer (e.g., to notify the network that the UE 202 is still active), and perform a Registration Update procedure to update UE capability information or to re-negotiate protocol parameters with the network, among others.

[0203] The AMF 1218 can store one or more RM contexts for the UE 202, where each RM context is associated with a specific access to the network. The RM context can be a data structure, database object, etc. that indicates or stores, inter alia, a registration state per access type and the periodic update timer. The AMF 1218 can also store a 5GC MM context that can be the same or similar to the (E)MM context discussed previously. In various embodiments, the AMF 1218 can store a CE mode B Restriction parameter of the UE 202 in an associated MM context or RM context. The AMF 1218 can also derive the value, when needed, from the UE's usage setting parameter already stored in the UE context (and / or MM / RM context).

[0204] Connection Management (CM) can be used to establish and release a signaling connection between the UE 202 and the AMF 1218 over the N1 interface. The signaling connection is used to enable NAS signaling exchange between the UE 202 and the CN 1234, and comprises both the signaling connection between the UE and the Access Netw ork (AN) (e.g., Radio Resource Control (RRC) connection or UE-N3IWF connection for non-3GPP access) and the N2 connection for the UE 202 between the AN (e.g., RAN 1226) and the AMF 1218. The UE 202 can operate in one of two CM states, CM-1DLE mode or CM- CONNECTED mode. When the UE 202 is operating in the CM-IDLE state / mode, the UE 202 can have no NAS signaling connection established with the AMF 221 over the N1 interface, and there can be RAN 1226 signaling connection (e g., N2 and / or N3connections) for the UE 202. When the UE 202 is operating in the CM-CONNECTED state / mode, the UE 202 can have an established NAS signaling connection with the AMF 1218 over the N1 interface, and there can be a RAN 1226 signaling connection (e.g., N2 and / or N3 connections) for the UE 202. Establishment of an N2 connection between the RAN 1226 and the AMF 1218 can cause the UE 202 to transition from CM-IDLE mode to CM-CONNECTED mode, and the UE 202 can transition from the CM-CONNECTED mode to the CM-IDLE mode when N2 signaling between the RAN 1226 and the AMF 1218 is released.

[0205] The SMF 1220 can be responsible for SM (e g., session establishment, modify and release, including tunnel maintain between UPF and AN node); UE IP address allocation and management (including optional authorization); selection and control of UP function; configuring traffic steering at UPF to route traffic to proper destination; termination of interfaces toward policy control functions; controlling part of policy enforcement and QoS; lawful intercept (for SM events and interface to LI system); termination of SM parts of NAS messages; downlink data notification; initiating AN specific SM information, sent via AMF over N2 to AN; and determining SSC mode of a session. SM can refer to management of a PDU session, and a PDU session or “session” can refer to a PDU connectivity service that provides or enables the exchange of PDUs between a UE 202 and a DN 1230 identified by a Data Network Name (DNN). PDU sessions can be established upon UE 202 request, modified upon UE 202 and 5GC CN 1234 request, and released upon UE 202 and 5GC CN 1234 request using NAS SM signaling exchanged over the N1 reference point between the UE 202 and the SMF 1220. Upon request from an application server, the 5GC CN 1234 can trigger a specific application in the UE 202. In response to receipt of the trigger message, the UE 202 can pass the trigger message (or relevant parts / information of the trigger message) to one or more identified applications in the UE 202. The identified application(s) in the UE 202 can establish a PDU session to a specific DNN. The SMF 1220 can check whether the UE 202 requests are compliant with user subscription information associated with the UE 202. In this regard, the SMF 1220 can retrieve and / or request to receive update notifications on SMF 1220 level subscription data from the UDM 1210.

[0206] The SMF 1220 can include the following roaming functionality: handling local enforcement to apply QoS SLAs (VPLMN); charging data collection and charging interface (VPLMN); lawful intercept (in VPLMN for SM events and interface to LI system); and support for interaction with external DN 1230 for transport of signaling for PDU session authorization / authentication by external DN 1230. An N16 reference point between two SMFs 1220 can be included in the network architecture 1200, which can be between anotherSMF 1220 in a visited network and the SMF 1220 in the home network in roaming scenarios. Additionally, the SMF 1220 can exhibit the Nsmf service-based interface.

[0207] The NEF 1204 can provide means for securely exposing the services and capabilities provided by 3GPP network functions for third party, internal exposure / re- exposure. Application Functions (e.g., AF 1212), edge computing or fog computing systems, etc. In such embodiments, the NEF 1204 can authenticate, authorize, and / or throttle the AFs 1212. NEF 1204 can also translate information exchanged with the AF 1212 and information exchanged with internal network functions. For example, the NEF 1204 can translate between an AF-Service-Identifier and an internal 5GC information. NEF 1204 can also receive information from other network functions (NFs) based on exposed capabilities of other network functions. This information can be stored at the NEF 1204 as structured data, or at a data storage NF using standardized interfaces. The stored information can then be re-exposed by the NEF 1204 to other NFs and AFs, and / or used for other purposes such as analytics. Additionally, the NEF 1204 can exhibit an Nnef service-based interface.

[0208] The NRF 1206 can support service discovery functions, receive NF discovery requests from NF instances, and provide the information of the discovered NF instances to the NF instances. NRF 1206 also maintains information of available NF instances and their supported services. As used herein, the terms “instantiate,’' “instantiation,” and the like can refer to the creation of an instance, and an “instance” can refer to a concrete occurrence of an object, which can occur, for example, during execution of program code. Additionally, the NRF 1206 can exhibit the Nnrf service-based interface.

[0209] The PCF 1208 can provide policy rules to control plane function(s) to enforce them, and can also support unified policy framework to govern network behavior. The PCF 1208 can also implement a front end (FE) to access subscription information relevant for policy decisions in a Uniform Data Repository (UDR) or user datagram protocol of the UDM 1210. The PCF 1208 can communicate with the AMF 1218 via an N15 reference point between the PCF 1208 and the AMF 1218, which can include a PCF 1208 in a visited network and the AMF 1218 in case of roaming scenarios. The PCF 1208 can communicate with the application function AF 1212 via an N5 reference point between the PCF 1208 and the AF 1212; and with the SMF 1220 via an N7 reference point between the PCF 1208 and the SMF 1220. The network architecture 1200 and / or CN 1234 can also include an N24 reference point between the PCF 1208 (in the home network) and a PCF 1208 in a visited network. Additionally, the PCF 1208 can exhibit an Npcf service-based interface.

[0210] The UDM 1210 can handle subscription-related information to support the network entities' handling of communication sessions, and can store subscription data of UE 202. For example, subscription data can be communicated between the UDM 1210 and the AMF 1218 via an N8 reference point between the UDM 1210 and the AMF 1218. The UDM 1210 can include two parts, an application FE and a Uniform Data Repository (UDR) (the FE and UDR are not shown by FIG. 12). The UDR can store subscription data and policy data for the UDM 1210 and the PCF 1208, and / or structured data for exposure and application data (including PFDs for application detection, application request information for multiple UEs UE 202) for the NEF 1204. The Nudr service-based interface can be exhibited by the UDR to allow the UDM 1210, PCF 1208, and NEF 1204 to access a particular set of the stored data, as well as to read, update (e.g., add, modify), delete, and subscribe to notification of relevant data changes in the UDR. The UDM 1210 can include a UDM-FE, which is in charge of processing credentials, location management, subscription management and so on. Several different front ends can serve the same user in different transactions. The UDM-FE accesses subscription information stored in the UDR and performs authentication credential processing, user identification handling, access authorization, registration / mobility management, and subscription management. The UDR can interact with the SMF 1220 via an N10 reference point between the UDM 1210 and the SMF 1220. UDM 1210 can also support SMS management, wherein an SMS-FE implements the similar application logic as discussed previously. Additionally, the UDM 1210 can exhibit the Nudm service-based interface.

[0211] The AF 1212 can provide application influence on traffic routing, provide access to the NCE, and interact with the policy framework for policy control. The NCE can be a mechanism that allows the 5GC CN 1234 and AF 1212 to provide information to each other via NEF 1204, which can be used for edge computing implementations. In such implementations, the network operator and third party services can be hosted close to the UE 202 access point of attachment to achieve an efficient service delivery through the reduced end-to-end latency and load on the transport netw ork. For edge computing implementations, the 5GC can select a UPF 1228 close to the UE 202 and execute traffic steering from the UPF 1228 to DN 1230 via the N6 interface. This can be based on the UE subscription data, UE location, and information provided by the AF 1212. In this way, the AF 1212 can influence UPF (re)selection and traffic routing. Based on operator deployment, when AF 1212 is considered to be a trusted entity, the netw ork operator can permit AF 1212 to interact directly with relevant NFs. Additionally, the AF 1212 can exhibit a Naf servicebased interface.

[0212] The NSSF 1202 can select a set of network slice instances serving the UE 202. The NSSF 1202 can also determine allowed NSSAI and the mapping to the subscribed single Network Slice Selection Assistance Information (S-NSSAIs), if needed. The NSSF 1202 can also determine the AMF 1218 set to be used to serve the UE 202, or a list of candidate AMF 1218 based on a suitable configuration and possibly by querying the NRF 1206. The selection of a set of network slice instances for the UE 202 can be triggered by the AMF 1218 with which the UE 202 is registered by interacting with the NSSF 1202, which can lead to a change of AMF 1218. The NSSF 1202 can interact with the AMF 1218 via an N22 reference point between AMF 1218 and NSSF 1202; and can communicate with another NSSF 1202 in a visited network via an N31 reference point (not shown by FIG. 12). Additionally, the NSSF 1202 can exhibit an Nnssf service-based interface.

[0213] The CN 1234 can include an SMSF, which can be responsible for SMS subscription checking and verification, and relaying SM messages to / from the UE 202 to / from other entities, such as an SMS-GMSC / IWMSC / SMS-router. The SMS can also interact with AMF 1218 and UDM 1210 for a notification procedure that the UE 202 is available for SMS transfer (e.g., set a UE not reachable flag, and notifying UDM 1210 when UE 202 is available for SMS).

[0214] The CN 1234 can also include other elements that are not shown by FIG. 12, such as a Data Storage system / architecture, a 5G-EIR, a SEPP, and the like. The Data Storage system can include a SDSF, an UDSF. and / or the like. Any NF can store and retrieve unstructured data into / from the UDSF (e.g., UE contexts), via N18 reference point between any NF and the UDSF (not shown by FIG. 12. Individual NFs can share a UDSF for storing their respective unstructured data or individual NFs can each have their own UDSF located at or near the individual NFs. Additionally, the UDSF can exhibit an Nudsf service-based interface (not shown by FIG. 12. The 5G-EIR can be an NF that checks the status of PEI for determining whether particular equipment / entities are blacklisted from the network; and the SEPP can be a non-transparent proxy that performs topology hiding, message filtering, and policing on inter-PLMN control plane interfaces.

[0215] Additionally, there can be many more reference points and / or service-based interfaces between the NF services in the NFs; however, these interfaces and reference points have been omitted from FIG. 12 for clarity. In one example, the CN 1234 can include an Nx interface, which is an inter-CN interface between the Mobility Management Entity (MME) and the AMF 1218 in order to enable interworking between CN 1234 and other CN. Other example interfaces / reference points can include an N5g-Equipment Identity Register (EIR) service-based interface exhibited by a 5G-EIR, an N27 reference point between theNetwork Repository Function (NRF) in the visited network and the NRF in the home network; and an N31 reference point between the NSSF in the visited network and the NSSF in the home network. Further, any of the above functions, entities, etc. can include or be comprised by a component as referred to herein.

[0216] The SCP 1222 (or individual instances of the SCP 1222) supports indirect communication (see e.g., 3GPP TS 23.501 section 7.1.1); delegated discovery (see e.g., 3GPP TS 23.501 section 7.1.1); message forwarding and routing to destination NF / NF service(s), communication security (e.g., authorization of the NF Service Consumer to access the NF Service Producer API) (see e.g., 3GPP TS 33.501), load balancing, monitoring, overload control, etc.; and discovery and selection functionality for UDM(s), AUSF(s), UDR(s), PCF(s) with access to subscription data stored in the UDR based on UE's SUPI, SUCI or GPSI (see e g., 3GPP TS 23.501 section 6.3). Load balancing, monitoring, overload control functionality provided by the SCP may be implementation specific. The SCP 1222 may be deployed in a distributed manner. More than one SCP 1222 can be present in the communication path between various NF Services. The SCP 1222, although not an NF instance, can also be deployed distributed, redundant, and scalable.

[0217] The DN 1230 may represent various network operator services, Internet access, or third party7services that may be provided by one or more servers including, for example, application server 110. In some implementations, the DN 1230 may be, or include, one or more edge compute nodes. Additionally or alternatively, the DN 1230 may be an Edge DN 1230, which is a (local) Data Network that supports the architecture for enabling edge applications. In these embodiments, the application server 110 may represent the physical hardware systems / devices providing app server functionality and / or the application software resident in the cloud or at an edge compute node that performs server function(s). In some embodiments, the application server 110 provides an edge hosting environment that provides support required for Edge Application Server's execution.

[0218] FIG. 13 illustrates a distributed NAS architecture 1300. As an alternative to the centralized NAS architecture 400, it may be possible to have a “distributed NAS” architecture where a UE would engage in direct dialogue with each individual network function in a core network, without going through a single point of contact comparable to the AMF in 5GC.

[0219] The distributed NAS architecture 1300 illustrates a network architecture that is similar to the network architecture 1200. As depicted in FIG. 13, the UE 1322 may communicate with the CN 1304 via a CU-CP 1302, which is similar to the gNB-CU-CP 206of the gNB-CU 214 of the gNB dual-architecture implementation as described with reference to FIG. 2. To enable the distributed NAS architecture 1300, a Network Function (NF) identifier (ID) (NF-ID) parameter may need to be included in RRC signaling, so that the CU-CP 1302 is able to route the NAS message to the relevant NF in the CN 1304. The NF-ID parameter may be encoded as a NF type (e.g.L‘SMSFT?, “SMF”,L‘PCF’ etc.), or it may be a unique NF identifier. For instance, the CU-CP 1302 routes the NAS message over the N2 reference point via the message bus 1324 to a network function in the CN 1304 based on the NF-ID parameter. In this manner, the UE 1322 may communicate NAS messages directly with the NSSAAF 1314, the AUSF 1316, the AMF 1318. or the SMF 1320 over secure connections (SCs) 1326, 1328, 1330 and 1332, respectively, without having to traverse or go through the AMF 1318, as in the centralized NAS architecture 400.

[0220] The distributed NAS architecture 1300 provides a more efficient way for the UE 1322 to directly communicate with the various network functions of the CN 1304. However, the distributed NAS architecture 1300 may need improved security techniques and measures to allow secure connections and / or security associations to protect the direct connects. In the distributed NAS architecture 1300, the UE 1322 may need to establish a direct secure connection with each NF with which the UE engages in NAS signaling.

[0221] LEVERAGING TRUSTED SECURITY INFORMATION

[0222] Embodiments herein relate to how direct secure connections between the UE 1322 and each NF can be efficiently established. Embodiments herein may include the assumption that, upon initial registration, the UE establishes Non-Access Stratum (NAS) master security association. The NAS master security association information is stored in the UE 1322 as well as a SEAF (not shown) external to the AMF 1318. When the UE 1322 needs to contact a new network function, or vice versa, the UE 1322 and the new NF may establish a security association for each secure connection (e.g., SC 1326, 1328, 1330, and 1332) by bootstrapping security parameters from the NAS master security association.

[0223] FIG. 14 illustrates a distributed NAS architecture 1400 similar to the distributed NAS architecture 1300. The distributed NAS architecture 1400 may include a CN 1410 with various network functions, such as NSSF 1412, NEF 1414, NRF 1416. PCF 1418, UDM 1420, AF 1422. SEAF 1408. AUSF 1424. AMF 1426, SMSF 1428, and LMF 1430. A UE 1432 may communicate with each network function via the gNB-DU 1402, the gNB- CU-CP 1406 and the gNB-CU-CP 1406 of a RAN. The UE 1432 may establish direct connections to a given network function, such as secure connections 1438, 1440, 1442,1444. and 1446. The secure connections do not traverse a single entity, such as the AMF 1426, when transporting NAS messages.

[0224] When the UE 1432 performs Initial Registration with the CN 1410, the UE 1432 is authenticated with the HPLMN based on long-term credentials stored in the UE 1432 and the UDM 1420. As part of the authentication procedure, a master security association is established between the UE 1432 and the serving network. The information for the NAS master security association is stored in the UE 1432 as well as in the SEAF 1408. When the UE 1432 needs to contact a new NF. or vice versa, the UE 1432 and the new NF function establish a security association by bootstrapping security parameters from the NAS master security association. It is worthy to note that the AMF 1426 is depicted as a standalone function distinct from the SEAF 1408 because a next-generation mobile system does not necessarily need to have the AMF 1426. If the next generation mobile system does have an AMF 1426, the SEAF 1408 can be collocated with the AMF 1426, as depicted in FIG. 4.

[0225] The individual secure connections between the UE 1432 and the various NF of the CN 1410 may be established using the NAS master security information stored in the UE 1432 and the SEAF 1408, as depicted in FIGS. 8A, 8B.

[0226] FIG. 15 schematically illustrates a wireless network 1500 in accordance with various embodiments. The wireless network 1500 may include a UE 1502 in wireless communication with an AN 1524. The UE 1502 and AN 1524 may be similar to, and substantially interchangeable with, like-named components described elsewhere herein.

[0227] The UE 1502 may be communicatively coupled with the AN 1524 via connection 1546. The connection 1546 is illustrated as an air interface to enable communicative coupling, and can be consistent with cellular communications protocols such as an LTE protocol, a 5G NR, or a 6G protocol operating at mmWave or sub-6GHz frequencies.

[0228] The UE 1502 may include a host platform 1504 coupled with a modem platform 1508. The host platform 1504 may include application processing circuitry 1506, which may be coupled with protocol processing circuitry71510 of the modem platform 1508. The application processing circuitry 1506 may run various applications for the UE 1502 that source / sink application data. The application processing circuitry 1506 may further implement one or more layer operations to transmit / receive application data to / from a data network. These layer operations may include transport (for example UDP) and Internet (for example, IP) operations

[0229] The protocol processing circuitry 1510 may implement one or more of layer operations to facilitate transmission or reception of data over the connection 1546. The layeroperations implemented by the protocol processing circuitry 1510 may include, for example, MAC, RLC, PDCP, RRC and NAS operations.

[0230] The modem platform 1508 may further include digital baseband circuitry’ 1534 that may implement one or more layer operations that are “below” layer operations performed by the protocol processing circuitry 1510 in a network protocol stack. These operations may include, for example, PHY operations including one or more of HARQ-ACK functions, scrambling / des crambling, encoding / decoding, layer mapping / de-mapping, modulation symbol mapping, received symbol / bit metric determination, multi-antenna port precoding / decoding, which may include one or more of space-time, space-frequency or spatial coding, reference signal generation / detection, preamble sequence generation and / or decoding, synchronization sequence generation / detection, control channel signal blind decoding, and other related functions.

[0231] The modem platform 1508 may further include transmit circuitry71514, receive circuitry 1516, RF circuitry 1518, and RF front end RFFE 1520, which may include or connect to one or more antenna panels 1522. Briefly, the transmit circuitry 1514 may include a digital -to-analog converter, mixer, intermediate frequency (IF) components, etc.; the receive circuitry71516 may include an analog-to-digital converter, mixer, IF components, etc.; the RF circuitry71518 may include a low-noise amplifier, a power amplifier, power tracking components, etc.; RFFE 1520 may include filters (for example, surface / bulk acoustic wave filters), switches, antenna tuners, beamforming components (for example, phase-array antenna components), etc. The selection and arrangement of the components of the transmit circuitry 1514, receive circuitry71516, RF circuitry71518, RFFE 1520, and antenna panels 1522 (referred generically as “transmit / receive components”) may be specific to details of a specific implementation such as, for example, whether communication is TDM or FDM, in mmWave or sub-6 gHz frequencies, etc. In some embodiments, the transmit / receive components may be arranged in multiple parallel transmit / receive chains, may be disposed in the same or different chips / modules, etc.

[0232] In some embodiments, the protocol processing circuitry 1510 may include one or more instances of control circuitry (not shown) to provide control functions for the transmit / receive components.

[0233] A UE reception may be established by and via the antenna panels 1522, RFFE 1520, RF circuitry71518, receive circuitry 1516, digital baseband circuitry 1512, and protocol protocol processing circuitry71510. In some embodiments, the antenna panels 1522may receive a transmission from the AN 1524 by receive-beamforming signals received by a plurality of antennas / antenna elements of the one or more antenna panels 1522.

[0234] A UE transmission may be established by and via the protocol processing circuitry 1510. digital baseband circuitry 1512, transmit circuitry 1514. RF circuitry 1518. RFFE 1520, and antenna panels 1522. In some embodiments, the transmit components of the UE 1502 may apply a spatial filter to the data to be transmitted to form a transmit beam emitted by the antenna elements of the antenna panels 1522.

[0235] Similar to the UE 1502, the AN 1524 may include a host platform 1526 coupled with a modem platform 1530. The host platform 1526 may include application processing circuitry 1528 coupled with protocol processing circuitry 1532 of the modem platform 1530. The modem platform 1530 may further include digital baseband circuitry 1534, transmit circuitry 1536, receive circuitry' 1538, RF circuitry 1540, RFFE circuitry 1542, and antenna panels 1544. The components of the AN 1524 may be similar to and substantially interchangeable with like-named components of the UE 1502. In addition to performing data transmission / reception as described above, the components of the host platform 1504 may perform various logical functions that include, for example, RNC functions such as radio bearer management, uplink and downlink dynamic radio resource management, and data packet scheduling.

[0236] FIG. 16 is a block diagram illustrating an apparatus 1600 with various components, according to some example embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein. Specifically, FIG. 16 shows a diagrammatic representation of hardware resources 1630 including one or more processors (or processor cores) 1610, one or more memory devices 1622, and one or more communication resources 1626, each of which may be communicatively coupled via a bus 1620 or other interface circuitry'. For embodiments where node virtualization (e.g., NFV) is utilized, a hypervisor 1602 may be executed to provide an execution environment for one or more network slices / sub-slices to utilize the hardware resources 1630.

[0237] The processors 1610 may include, for example, a processor 1612 and a processor 1614. The processors 1610 may be, for example, a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a DSP such as a baseband processor, an ASIC, an FPGA, a radio-frequency integrated circuit (RFIC), another processor (including those discussed herein), or any suitable combination thereof.

[0238] The memory devices 1622 (or storage devices) may include main memory, disk storage, or any suitable combination thereof. The memory devices 1 22 may include, but are not limited to, any type of volatile, non-volatile, or semi-volatile memory such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), Flash memory, solid-state storage, etc.

[0239] The communication resources 1626 may include interconnection or network interface controllers, components, or other suitable devices to communicate with one or more peripheral devices 1604 or one or more databases 1606 or other network elements via a network 2008. For example, the communication resources 1626 may include wired communication components (e.g., for coupling via USB, Ethernet, etc.), cellular communication components, NFC components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components, and other communication components.

[0240] The instructions 1616, instructions 1618, instructions 1624, instructions 1628, and / or instructions 1632 may comprise software, a program, an application, an applet, an app, or other executable code for causing at least any of the processors 2010 to perform any one or more of the methodologies discussed herein. The instructions 1616, instructions 1618, instructions 1624, instructions 1628, and / or instructions 1632 may reside, completely or partially, within at least one of the processors 1610 (e.g., within the processor’s cache memory), the memory devices 1622. or any suitable combination thereof. Furthermore, any portion of the instructions 1616, instructions 1618, instructions 1624, instructions 1628, and / or instructions 1632 may be transferred to the hardware resources 1630 from any combination of the peripheral devices 1604 or the databases 1606. Accordingly, the memory of processors 1610, the memory devices 1622, the peripheral devices 1604, and the databases 1606 are examples of computer-readable and machine-readable media.

[0241] For one or more embodiments, at least one of the components set forth in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and / or methods as set forth in the example section below. For example, the baseband circuitry as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below. For another example, circuitry associated with a UE, base station, network element, etc. as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below in the example section.

[0242] FIG. 17 illustrates a computer readable media 1702. The computer readable media 1702 may store one or more computer executable instructions 1704 to implemented one or more embodiments as described herein. Various aspects or features described herein can be implemented as a method, apparatus, or article of manufacture using standard programming and / or engineering techniques. The term ‘"article of manufacture7’ as used herein is intended to encompass a computer program accessible from any computer-readable device, carrier, or media. For example, computer-readable computer readable media 1702 can include but are not limited to magnetic storage devices (e.g., hard disk, floppy disk, magnetic strips, etc.), optical disks (e.g.. compact disk (CD), digital versatile disk (DVD), etc ), smart cards, and flash memory devices (e.g., EPROM, card, stick, key drive, etc.). Additionally, various storage media described herein can represent one or more devices and / or other machine- readable media for storing information. The term “machine-readable medium” can include, without being limited to, wireless channels and various other media capable of storing, containing, and / or carrying instruction(s) and / or data. Additionally, a computer program product can include a computer readable medium having one or more instructions or codes operable to cause a computer to perform functions described herein.

[0243] Communications media embody computer executable instructions 1704 or computer-readable instructions, data structures, program modules or other structured or unstructured data in a data signal such as a modulated data signal, e.g., a carrier wave or other transport mechanism, and includes any information delivery or transport media. The term “modulated data signal” or signals refers to a signal that has one or more of its characteristics set or changed in such a manner as to encode information in one or more signals. By way of example, and not limitation, communication media include wired media, such as a wired network or direct-wired connection, and wireless media such as acoustic, RF. infrared and other wireless media.

[0244] An exemplary storage medium can be coupled to processor, such that processor can read information from, and write information to, the storage medium. In the alternative, storage medium can be integral to processor. Further, in some aspects, processor and storage medium can reside in an ASIC. Additionally, ASIC can reside in a user terminal. In the alternative, processor and storage medium can reside as discrete components in a user terminal. Additionally, in some aspects, the processes and / or actions of a method or algorithm can reside as one or any combination or set of codes and / or instructions on a machine-readable medium and / or computer readable medium, which can be incorporated into a computer program product.

[0245] While the disclosed subject matter has been described in connection with various embodiments and corresponding Figures, where applicable, it is to be understood that other similar embodiments can be used or modifications and additions can be made to the described embodiments for performing the same, similar, alternative, or substitute function of the disclosed subject matter without deviating therefrom. Therefore, the disclosed subject matter should not be limited to any single embodiment described herein, but rather should be construed in breadth and scope in accordance with the appended examples below.

[0246] In particular regard to the various functions performed by the above described components (assemblies, devices, circuits, systems, etc.), the terms (including a reference to a “means”) used to describe such components are intended to correspond, unless otherwise indicated, to any component or structure which performs the specified function of the described component (e.g., that is functionally equivalent), even though not structurally equivalent to the disclosed structure which performs the function in the herein illustrated exemplary implementations of the disclosure. In addition, while a particular feature can have been disclosed with respect to only one of several implementations, such feature can be combined with one or more other features of the other implementations as can be desired and advantageous for any given or particular application.

[0247] The various elements of the devices as previously described with reference to the figures include various hardware elements, software elements, or a combination of both. Examples of hardware elements include devices, logic devices, components, processors, microprocessors, circuits, processors, circuit elements (e.g., transistors, resistors, capacitors, inductors, and so forth), integrated circuits, application specific integrated circuits (ASIC), programmable logic devices (PLD), digital signal processors (DSP), field programmable gate array (FPGA), memory units, logic gates, registers, semiconductor device, chips, microchips, chip sets, and so forth. Examples of software elements include software components, programs, applications, computer programs, application programs, system programs, software development programs, machine programs, operating system software, middleware, firmware, software modules, routines, subroutines, functions, methods, procedures, software interfaces, application program interfaces (API), instruction sets, computing code, computer code, code segments, computer code segments, words, values, symbols, or any combination thereof. However, determining whether an embodiment is implemented using hardware elements and / or software elements varies in accordance with any number of factors, such as desired computational rate, power levels, heat tolerances, processing cycle budget, input data rates, output data rates, memory resources, data bus speeds and other design or performance constraints, as desired for a given implementation.

[0248] One or more aspects of at least one embodiment are implemented by representative instructions stored on a machine-readable medium which represents various logic within the processor, which when read by a machine causes the machine to fabricate logic to perform the techniques described herein. Such representations, known as “intellectual property (IP) cores" are stored on a tangible, machine readable medium and supplied to various customers or manufacturing facilities to load into the fabrication machines that make the logic or processor. Some embodiments are implemented, for example, using a machine-readable medium or article which may store an instruction or a set of instructions that, when executed by a machine, causes the machine to perform a method and / or operations in accordance with the embodiments. Such a machine includes, for example, any suitable processing platform, computing platform, computing device, processing device, computing system, processing system, processing devices, computer, processor, or the like, and is implemented using any suitable combination of hardware and / or software. The machine-readable medium or article includes, for example, any suitable type of memory unit, memory device, memory article, memory medium, storage device, storage article, storage medium and / or storage unit, for example, memory, removable or non-removable media, erasable or non-erasable media, writeable or re-writeable media, digital or analog media, hard disk, floppy disk, Compact Disk Read Only Memory (CD-ROM), Compact Disk Recordable (CD-R). Compact Disk Rewriteable (CD-RW), optical disk, magnetic media, magneto-optical media, removable memory cards or disks, various types of Digital Versatile Disk (DVD), a tape, a cassette, or the like. The instructions include any suitable ty pe of code, such as source code, compiled code, interpreted code, executable code, static code, dynamic code, encrypted code, and the like, implemented using any suitable high-level, low-level, object-oriented, visual, compiled and / or interpreted programming language.

[0249] As utilized herein, terms “component,” “system,” “interface,” and the like are intended to refer to a computer-related entity, hardware, software (e.g., in execution), and / or firmware. For example, a component is a processor (e g., a microprocessor, a controller, or other processing device), a process running on a processor, a controller, an object, an executable, a program, a storage device, a computer, a tablet PC and / or a user equipment (e.g., mobile phone, etc.) with a processing device. By way of illustration, an application running on a server and the server is also a component. One or more components reside within a process, and a component is localized on one computer and / or distributed between two or more computers. A set of elements or a set of other components are described herein, in which the term “set” can be interpreted as “one or more.”

[0250] Further, these components execute from various computer readable storage media having various data structures stored thereon such as with a module, for example. The components communicate via local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network, such as. the Internet, a local area network, a wide area network, or similar network with other systems via the signal).

[0251] As another example, a component is an apparatus with specific functionality provided by mechanical parts operated by electric or electronic circuitry', in which the electric or electronic circuitry is operated by a software application or a firmware application executed by one or more processors. The one or more processors are internal or external to the apparatus and execute at least a part of the software or firmware application. As yet another example, a component is an apparatus that provides specific functionality- through electronic components without mechanical parts; the electronic components include one or more processors therein to execute software and / or firmware that confer(s), at least in part, the functionality of the electronic components.

[0252] Use of the word exemplary is intended to present concepts in a concrete fashion. As used in this application, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or”. That is, unless specified otherwise, or clear from context, “X employs A or B” is intended to mean any of the natural inclusive permutations. That is, if X employs A; X employs B; or X employs both A and B, then “X employs A or B” is satisfied under any of the foregoing instances. In addition, the articles “a” and “an” as used in this application and the appended examples should generally be construed to mean “one or more” unless specified otherwise or clear from context to be directed to a singular form. Furthermore, to the extent that the terms “including”, “includes”, “having”, “has”, “with”, or variants thereof are used in either the detailed description or the examples, such terms are intended to be inclusive in a manner similar to the term “comprising.” Additionally, in situations wherein one or more numbered items are discussed (e.g., a “first X”, a “second X”, etc.), in general the one or more numbered items may be distinct or they may be the same, although in some situations the context may indicate that they are distinct or that they are the same.

[0253] As used herein, the term “circuitry” may refer to, be part of, or include a circuit, an integrated circuit (IC), a monolithic IC, a discrete circuit, a hybrid integrated circuit (H1C), an Application Specific Integrated Circuit (ASIC), an electronic circuit, a logic circuit, a microcircuit, a hybrid circuit, a microchip, a chip, a chiplet, a chipset, a multi-chip module (MCM), a semiconductor die, a system on a chip (SoC), a processor (shared, dedicated, orgroup), a processor circuit, a processing circuit, or associated memory (shared, dedicated, or group) operably coupled to the circuitry that execute one or more software or firmware programs, a combinational logic circuit, or other suitable hardware components that provide the described functionality. In some embodiments, the circuitry' is implemented in, or functions associated with the circuitry are implemented by, one or more software or firmware modules. In some embodiments, circuitry’ includes logic, at least partially operable in hardware. It is noted that hardware, firmware and / or software elements may be collectively or individually referred to herein as “logic"’ or “circuit.”

[0254] Some embodiments are described using the expression “one embodiment” or “an embodiment” along with their derivatives. These terms mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. The appearances of the phrase “in one embodiment"’ in various places in the specification are not necessarily all referring to the same embodiment. Moreover, unless otherwise noted the features described above are recognized to be usable together in any combination. Thus, any features discussed separately can be employed in combination with each other unless it is noted that the features are incompatible with each other.

[0255] Some embodiments are presented in terms of program procedures executed on a computer or network of computers. A procedure is here, and generally, conceived to be a self-consistent sequence of operations leading to a desired result. These operations are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical, magnetic or optical signals capable of being stored, transferred, combined, compared, and otherwise manipulated. It proves convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like. It should be noted, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to those quantities.

[0256] Further, the manipulations performed are often referred to in terms, such as adding or comparing, which are commonly associated with mental operations performed by a human operator. No such capability' of a human operator is necessary7, or desirable in most cases, in any of the operations described herein, which form part of one or more embodiments. Rather, the operations are machine operations. Useful machines for performing operations of various embodiments include general purpose digital computers or similar devices.

[0257] Some embodiments are described using the expression "coupled" and "connected" along with their derivatives. These terms are not necessarily intended as synonyms for each other. For example, some embodiments are described using the terms “connected” and / or “coupled” to indicate that two or more elements are in direct physical or electrical contact with each other. The term "coupled.” however, also means that two or more elements are not in direct contact with each other, but yet still co-operate or interact with each other.

[0258] Various embodiments also relate to apparatus or systems for performing these operations. This apparatus is specially constructed for the required purpose or it comprises a general purpose computer as selectively activated or reconfigured by a computer program stored in the computer. The procedures presented herein are not inherently related to a particular computer or other apparatus. Various general purpose machines are used with programs written in accordance with the teachings herein, or it proves convenient to construct more specialized apparatus to perform the required method steps. The required structure for a variety of these machines are apparent from the description given.

[0259] It is emphasized that the Abstract of the Disclosure is provided to allow a reader to quickly ascertain the nature of the technical disclosure. It is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the examples. In addition, the following examples are hereby incorporated into the Detailed Description, with each example standing on its own as a separate embodiment. In the appended examples, the terms "including" and "in which" are used as the plain-English equivalents of the respective terms "comprising" and "wherein," respectively. Moreover, the terms "first," "second," "third," and so forth, are used merely as labels, and are not intended to impose numerical requirements on their objects.

[0260] The techniques described herein may be implemented with privacy safeguards to protect user privacy. Furthermore, the techniques described herein may be implemented with user privacy safeguards to prevent unauthorized access to personal data and confidential data. The training of the Al models described herein is executed to benefit all users fairly, without causing or amplifying unfair bias.

[0261] According to some embodiments, the techniques for the models described herein do not make inferences or predictions about individuals unless requested to do so through an input. According to some embodiments, the models described herein do not learn from and are not trained on user data without user authorization. In instances where user data is permitted and authorized for use in Al features and tools, it is done in compliance with a user’s visibility settings, privacy choices, user agreement and descriptions, and theapplicable law. According to the techniques described herein, users may have full control over the visibility of their content and who sees their content, as is controlled via the visibility settings. According to the techniques described herein, users may have full control over the level of their personal data that is shared and distributed between different Al platforms that provide different functionalities. According to the techniques described herein, users may choose to share personal data with different platforms to provide services that are more tailored to the users. In instances where the users choose not to share personal data with the platforms, the choices made by the users will not have any impact on their ability to use the services that they had access to prior to making their choice.

[0262] According to the techniques described herein, users may have full control over the level of access to their personal data that is shared with other parties. According to the techniques described herein, personal data provided by users may be processed to determine prompts when using a generative Al feature at the request of the user, but not to train generative Al models. In some embodiments, users may provide feedback while using the techniques described herein, which may be used to improve or modify the platform and products. In some embodiments, any personal data associated with a user, such as personal information provided by the user to the platform, may be deleted from storage upon user request. In some embodiments, personal information associated with a user may be permanently deleted from storage when a user deletes their account from the platform.

[0263] According to the techniques described herein, personal data may be removed from any training dataset that is used to train Al models. The techniques described herein may utilize tools for anonymizing member and customer data. For example, user’s personal data may be redacted and minimized in training datasets for training Al models through delexicalisation tools and other privacy enhancing tools for safeguarding user data. The techniques described herein may minimize use of any personal data in training Al models, including removing and replacing personal data. According to the techniques described herein, notices may be communicated to users to inform how their data is being used and users are provided controls to opt-out from their data being used for training Al models.

[0264] According to some embodiments, tools are used with the techniques described herein to identify and mitigate risks associated with Al in all products and Al systems. In some embodiments, notices may be provided to users when Al tools are being used to provide features.

[0265] A FIRST SET OF EXAMPLES

[0266] Example 1. A system and method of wireless communication for a fifth generation (5G) or beyond or new radio (NR) system: wherein for beam management assisted by AI / ML models at NW side or UE side, a UE is configured with one or more “Set A of beams’' (set of all beams which can be used for DL / UL) and one or more “Set B of beams'’ on which the UE is expected to perform measurements.

[0267] Example 2. The system and method of example 1, wherein a configuration of Set A, Set B or Set A / B pair is associated with a specific AI / ML model which expects inputs from the configured Set B and outputs associated with the configured Set A.

[0268] Example 3. The system and method of example 1, wherein, for a NW side model, a UE is configured only with a Set B of beams without configuration of a Set A of beams.

[0269] Example 4. The system and method of example 1, wherein, for a NW side model, a UE is configured with a Set A of beams and not explicitly configured with a Set B of beams but is configured with LI measurement resources which implicitly map to Set B of beams.

[0270] Example 5. The system and method of example 1, wherein, for models trained at the UE side, the UE expects that the Set B used during model training will be the same Set B used during model inference.

[0271] Example 6. The system and method of example 1, wherein, for a network side model, a UE expects the network to configure the same LI measurement resources, which were used during model training, for the purpose of model inference.

[0272] Example 7. The system and method of example 1, wherein a UE may use different Set B configurations for the same AI / ML model and in this case, being configured with different Set B does not imply model switching / selection.

[0273] Example 8. The system and method of example 1, wherein when a UE is configured with a Set B of measurement beams, the UE is expected to measure on all the configured beams within Set B in one shot and update the measurements for beams in Set B for use at UE side model or reports measurements using LI -reporting for use at NW side model.

[0274] Example 9. The system and method of example 1, wherein a UE is expected to measure K instances of the measurement RS in Set B and then update Set B measurements using a function of all the K LI measurements.

[0275] Example 10. The system and method of example 9, wherein the value of K is provided to the UE by its serving gNB using RRC or MAC-CE signalling.

[0276] Example 11. The system and method of example 1, wherein, for UE side models, a UE is expected to report a quality metric associated with the reported predicted information regarding the DL Tx beams.

[0277] Example 12. The system and method of example 11. wherein the quality metric represents a measure of the confidence of the model outputs or the confidence of one or both of: the reported top-K beams and their corresponding Ll-RSRP values that may be derived from the model outputs.

[0278] Example 13. The system and method of example 1, wherein representations of the Channel Impulse Response (CIR) obtained based on the measurements on CSI-RS resources are used as inputs to the AI / ML model at the UE or network (gNB) sides, wherein such representations may involve quantization of one or more of: the channel timing response ( relative timing of the detected paths for a time domain representation of the channel), amplitude or power for each of the detected paths for a time domain representation of the channel), and channel phase response (phase information corresponding to each of the detected paths for a time domain representation of the channel).

[0279] Example 14. The system and method of example 1, wherein the UE is provided with multiple configurations for observation / measurement windows and each configuration is associated with a UE-sided model for model-level identification and / or to an AI / ML-based functionality.

[0280] Example 15. The system and method of example 1, wherein the UE is provided with one or multiple configurations for prediction windows during which the beam prediction (spatial and / or temporal) based on UE-sided ML models may apply, and each configuration is associated with a UE-sided model for model-level identification and / or to an AI / ML-based functionality7.

[0281] Example 16. The system and method of example 1, wherein the UE is expected to report a time index for each measurement.

[0282] Example 17. The system and method of example 16. wherein the time index is an absolute time index, e.g., a time stamp.

[0283] Example 18. The system and method of example 16, wherein the time index is an index which indicates the time instances with respect to a configured measurement window.

[0284] Example 19. The system and method of example 1, wherein the UE is expected to determine a KPI corresponding to AI / ML model performance for a UE side model.

[0285] Example 20. The system and method of example 19, wherein the KPIs reported by the UE also include a time index information related to the AI / ML model outputs and / or inputs which were used to evaluate the respective KPIs.

[0286] Example 21. The system and method of example 19, wherein the configuration and selection of KPI ty pes are associated with selecting and activating an AI / ML model.

[0287] Example 22. The system and method of example 19, wherein the KPI configured to the UE includes a reference or threshold value for the KPI which the UE is expected to use to report quality of the measured / calculated KPI for reporting.

[0288] Example 23. The system and method of example 1. wherein based on the output of the UE-side AI / ML model, the UE may request the NW to transmit reference signals using the top-K beams where the UE can then perform LI measurements to determine the actual Ll-RSRP / SINR and report actual LI measurement or a KPI based on the measurement to the network for performance monitoring.

[0289] Example 24. The system and method of example 1, wherein the UE is configured with a time window for measurement based on model output or collection of actual model outputs for model performance monitoring.

[0290] Example 25. The system and method of example 1, wherein for a UE side AI / ML model, the UE is configured with reference signal (RS) resources which are specific to model monitoring and different from model inferencing.

[0291] Example 26. The system and method of example 1, wherein, for a network side AI / ML model, the UE is configured with one or multiple Set B of beams wherein, the UE is configured to report LI measurements (RSRP / SINR) for performance monitoring for each activated Set B of beams to the NW in a periodic or aperiodic manner.

[0292] Example 27. The system and method of example 1, wherein, for a network side AI / ML model, when data collection is triggered by the network for a configured Set A of beams, the UE is expected to report its Rx beam assumption corresponding to the LI measurements on Set A.

[0293] Example 28. The system and method of example 1, wherein the Rx beam assumption is associated with an AI / ML model type and different AI / ML models may have different Rx beam assumptions.

[0294] Example 29. The system and method of example 1, wherein, for UE-side AI / ML model, for model inference if the model output or the top-K beams based on the model output is a beam which is not part of the Set B of beams, the UE can indicate in the UL LI beam report that the reported beam is part of Set A of beams but not part of a Set B of beams.

[0295] Example 30. The system and method of example 1 , wherein, if the model output is Ll-RSRP or Ll-SINR, and the top-K measurements correspond to beams in Set A but not in Set B, the LI beam report sent by the UE on the basis of this model inference includes information to signal to the gNB that the reported LI measurement on the reported beams is an output from the UE-side model and not an actual LI measurement.

[0296] Example 31. The system and method of example 1, wherein the UE may report its capability for simultaneous monitoring of performance of UE-sided AI / ML model and performance of one or more non-AI / ML-based beam prediction methods.

[0297] Example 32. The system and method of example 31, wherein, based on reported UE capability, the serving gNB configures the UE to compare performance of one or more UE- sided AI / ML model(s) to that of one or more non-AI / ML-based beam prediction methods.

[0298] Example 33. The system and method of example 1, wherein the UE may be configured to monitor performance of UE-sided AI / ML model and performance of one or more non-AI / ML-based beam prediction methods wherein one or more of: the prediction window and the measurement window for the AI / ML model-based beam prediction and non- AI / ML method-based beam prediction are not the same.

[0299] Example 34. The system and method of example 1, wherein, for a UE-sided model, the UE may be configured with measurement resources and a threshold, wherein if all or a subset of measurements on the configured resources are different from the predict measurement quantity from the AII / ML model corresponding to the same resources by a threshold amount, the UE may declare model failure at the UE side.

[0300] Example 35. The system and method of example 34 wherein the UE may report model failure using PUCCH or MAC-CE and the network may respond by either deactivating the current model, switching the model to another preconfigured model, delivering a new model via model transfer or delivery or delivering a dataset for retraining or retuning the current model.

[0301] Example 36. The system and method of example 1, wherein based on the inference from an AI / ML model, a UE might be indicated with a TCI state which is not part of the set of activated TCI states, using a new DCI format or using an existing DCI format with higher layer configuration to determine the bit width of the existing TCI state indication field.

[0302] A SECOND SET OF EXAMPLES

[0303] Example 1. A method for a user equipment (UE) of a wireless system, comprising: accessing information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; obtaining measurements for the second set of resources; generating beam information to configure a set of beams associated with the first set of resources by a machine learning (ML) model using the measurements for the second set of resources; and encoding a message comprising the beam information.

[0304] Example 2. The method of example 1, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

[0305] Example 3. The method of example 1, wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

[0306] Example 4. The method of example 1, comprising decoding a first message from a base station, the first message comprising the information for the first set of resources or the second set of resources.

[0307] Example 5. The method of example 1, comprising encoding a second message for the base station, the second message comprising a measurement report with the measurements for the second set of resources.

[0308] Example 6. The method of example 1, comprising encoding a third message for the base station, the third message comprising a quality' metric associated with the beam information generated by the ML model.

[0309] Example 7. The method of example 1, comprising encoding a fourth message for the base station, the fourth message comprising a performance metric for the ML model or a function, based on a periodic reporting configuration or when the performance metric is above or below a defined threshold value.

[0310] Example 8. The method of example 1, comprising generating the beam information by the ML model using historical measurements for the second set of resources.

[0311] Example 9. The method of any of examples 1 to 8, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE). or downlink control information (DCI); DL layer 1 -reference signal received power (LI -RSRP) values corresponding to the top-K identified beams; or DL layer 1-signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

[0312] Example 10. The method of example 9, comprising encoding the message with the beam information comprising the DL LI -RSRP values or the DL Ll-SINR values and an indication whether the DL LI -RSRP values or the DL Ll-SINR values are generated by the ML model or are measured values.

[0313] Example 11. An apparatus for a user equipment (HE) of a wireless system, comprising: an interface to communicate resource information; and circuitry operably coupled to the interface, the circuitry to: access information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; obtain measurements for the second set of resources; generate beam information to configure a set of beams associated with the first set of resources by a machine learning (ML) model using the measurements for the second set of resources; and encode a message comprising the beam information.

[0314] Example 12. The apparatus of example 11, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

[0315] Example 13. The apparatus of example 11. wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

[0316] Example 14. The apparatus of example 11. the circuitry to decode a first message from a base station, the first message comprising the information for the first set of resources or the second set of resources.

[0317] Example 15. The apparatus of example 11. the circuitry to encode a second message for the base station, the second message comprising a measurement report with the measurements for the second set of resources.

[0318] Example 16. The apparatus of example 11. the circuitry to encode a third message for the base station, the third message comprising a quality metric associated with the beam information generated by the ML model.

[0319] Example 17. The apparatus of example 11. the circuitry to encode a fourth message for the base station, the fourth message comprising a performance metric for the ML model or a function, based on a periodic reporting configuration or when the performance metric is above or below a defined threshold value.

[0320] Example 18. The apparatus of example 1 1, the circuitry to generate the beam information by the ML model using historical measurements for the second set of resources.

[0321] Example 19. The apparatus of any of examples 11 to 18, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control(RRC) message, a medium access control (MAC) control element (CE). or downlink control information (DCI);DL layer 1 -reference signal received power (Ll-RSRP) values corresponding to the top-K identified beams; or DL layer 1 -signal to interference plus noise ratio (L1-S1NR) values corresponding to the top-K identified beams.

[0322] Example 20. The apparatus of any of examples 11 to 18, the circuitry’ to: obtain training data for ML model development, where the training data is generated based on measurements performed by the UE and data obtained from a network-side server or a UE- side server; train the ML model using the measurements; and deploy the trained ML model for inference at the UE.

[0323] Example 21. An apparatus for a base station of a wireless system, comprising: an interface to communicate resource information; and circuitry operably coupled to the interface, the circuitry to: determine information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; decode beam information from a user equipment (UE), the beam information generated by a machine learning (ML) model using a set of measurements for the second set of resources; and generate a DL Tx beam using the first set of resources and the beam information by the antenna array.

[0324] Example 22. The apparatus of example 20, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

[0325] Example 23. The apparatus of example 20. wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

[0326] Example 24. The apparatus of any of examples 20 to 23, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI);DL layer 1 -reference signal received power (Ll-RSRP) values corresponding to the top-K identified beams; or DL layer 1-signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

[0327] Example 25. The apparatus of any of examples 20 to 23, the circuitry to encode a message for the UE, the message comprising configuration information to configure the UE to report measurements based on the first set of resources or the second set of resources using higher layer signaling.

[0328] Example 26. A method for a base station of a wireless system, comprising: determining information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; decoding beam information from a user equipment (UE), the beam information generated by a machine learning (ML) model using a set of measurements for the second set of resources; and generating a DL Tx beam using the first set of resources and the beam information by the antenna array.

[0329] Example 27. The method of example 26, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

[0330] Example 28. The method of example 26, wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

[0331] Example 29. The method of any of examples 26 to 28, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI);DL layer 1 -reference signal received power (LI -RSRP) values corresponding to the top-K identified beams; or DL layer 1 -signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

[0332] Example 30. The method of any of examples 26 to 28, comprising encoding a message for the UE, the message comprising configuration information to configure the UE to report measurements based on the first set of resources or the second set of resources using higher layer signaling.

[0333] Example 31. An apparatus comprising means to perform a method in any preceding example 1 to 30.

[0334] Example 32. Machine-readable storage including machine-readable instructions, when executed, to implement a method or realize an apparatus in any preceding example 1 to 31.

Claims

CLAIMSWhat is claimed is:

1. A method for a user equipment (UE) of a wireless system, comprising: accessing information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; obtaining measurements for the second set of resources; generating beam information to configure a set of beams associated with the first set of resources by a machine learning (ML) model using the measurements for the second set of resources; and encoding a message comprising the beam information.

2. The method of claim 1, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

3. The method of claim 1, wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

4. The method of claim 1, comprising decoding a first message from a base station, the first message comprising the information for the first set of resources or the second set of resources.

5. The method of claim 4, comprising encoding a second message for the base station, the second message comprising a measurement report with the measurements for the second set of resources.

6. The method of claim 5. comprising encoding a third message for the base station, the third message comprising a quality metric associated with the beam information generated by the ML model.

7. The method of claim 6. comprising encoding a fourth message for the base station, the fourth message comprising a performance metric for the ML model or a function, based on a periodic reporting configuration or when the performance metric is above or below a defined threshold value.

8. The method of claim 1. comprising generating the beam information by the ML model using historical measurements for the second set of resources.

9. The method of any of claims 1 to 8, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI);DL layer 1 -reference signal received power (Ll-RSRP) values corresponding to the top-K identified beams; orDL layer 1 -signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

10. The method of claim 9, comprising encoding the message with the beam information comprising the DL Ll-RSRP values or the DL Ll-SINR values and an indication whether the DL Ll-RSRP values or the DL Ll-SINR values are generated by the ML model or are measured values.

11. An apparatus for a user equipment (UE) of a wireless system, comprising: an interface to communicate resource information; and circuitry operably coupled to the interface, the circuitry' to: access information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; obtain measurements for the second set of resources; generate beam information to configure a set of beams associated with the first set of resources by a machine learning (ML) model using the measurements for the second set of resources; and encode a message comprising the beam information.

12. The apparatus of claim 11, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

13. The apparatus of claim 11, wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

14. The apparatus of claim 11, the circuitry to decode a first message from a base station, the first message comprising the information for the first set of resources or the second set of resources.

15. The apparatus of claim 14, the circuitry to encode a second message for the base station, the second message comprising a measurement report with the measurements for the second set of resources.

16. The apparatus of claim 15, the circuitry to encode a third message for the base station, the third message comprising a quality metric associated with the beam information generated by the ML model.

17. The apparatus of claim 16, the circuitry to encode a fourth message for the base station, the fourth message comprising a performance metric for the ML model or a function, based on a periodic reporting configuration or when the performance metric is above or below a defined threshold value.

18. The apparatus of claim 11, the circuitry to generate the beam information by the ML model using historical measurements for the second set of resources.

19. The apparatus of any of claims 11 to 18. wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE), or downlink control information (DCI);DL layer 1 -reference signal received power (Ll-RSRP) values corresponding to the top-K identified beams; orDL layer 1 -signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

20. The apparatus of any of claims 11 to 18. the circuitry to: obtain training data for ML model development, where the training data is generated based on measurements performed by the UE and data obtained from a network-side server or a UE-side server; train the ML model using the measurements; and deploy the trained ML model for inference at the UE.

21. An apparatus for a base station of a wireless system, comprising: an interface to communicate resource information; andcircuitry operably coupled to the interface, the circuitry to: determine information for a first set of resources and a second set of resources associated with downlink (DL) transmit (Tx) beamforming operations of an antenna array; decode beam information from a user equipment (UE), the beam information generated by a machine learning (ML) model using a set of measurements for the second set of resources; and generate a DL Tx beam using the first set of resources and the beam information by the antenna array.

22. The apparatus of claim 20, wherein the first set of resources or the second set of resources comprise channel state information-reference signal (CSI-RS) resources or synchronization signal block (SSB) resources with corresponding transmission configuration indicator (TCI) states.

23. The apparatus of claim 20, wherein the second set of resources comprises a subset of transmission configuration indicator (TCI) states of the first set of resources.

24. The apparatus of any of claims 20 to 23, wherein the beam information corresponds to: a top-K identified beams where a value for K greater than or equal to 1 is configured for the ML model or a function using a radio resource control (RRC) message, a medium access control (MAC) control element (CE). or downlink control information (DCI);DL layer 1 -reference signal received power (Ll-RSRP) values corresponding to the top-K identified beams; orDL layer 1 -signal to interference plus noise ratio (Ll-SINR) values corresponding to the top-K identified beams.

25. The apparatus of any of claims 20 to 23, the circuitry to encode a message for the UE, the message comprising configuration information to configure the UE to report measurements based on the first set of resources or the second set of resources using higher layer signaling.

Images