Core network management for artificial intelligence at user equipment

Abstract

Certain aspects of the present disclosure provide techniques for core network management of artificial intelligence (AI) deployed at a user equipment (UE). An example method for wireless communication by an apparatus includes obtaining an indication to train a machine learning (ML) model for deployment at a UE; transmitting a request for training data associated with the one or more UEs to a first network entity; obtaining training data from the first network entity; training an ML model based at least in part on the training data; transmitting an indication of training completion of the ML model to a second network entity; and transmitting the trained ML model and an indication to the first network entity to transmit the trained ML model to the UE.

Description

[0001] introduction Technical Field

[0002] Various aspects of this disclosure relate to wireless communication, and more specifically to techniques for managing artificial intelligence used in user equipment.

[0003] Related technical descriptions Wireless communication systems are widely deployed to provide a variety of telecommunications services, such as telephone, video, data, messaging, broadcasting, or other similar services. These wireless communication systems may employ multiple access technologies that enable communication with multiple users by sharing available wireless communication system resources.

[0004] Despite significant technological advancements in wireless communication systems over the years, challenges remain. For example, complex and dynamic environments can still attenuate or block signals between wireless transmitters and receivers. Therefore, there is a continuous desire to improve the technical performance of wireless communication systems, including, for example: improving communication speed and data carrying capacity; improving the efficiency of shared communication media; reducing the power consumed by transmitters and receivers during communication; improving the reliability of wireless communication; avoiding redundant transmission and / or reception and related processing; improving the coverage area of ​​wireless communication; increasing the number and types of devices that can access the wireless communication system; increasing the ability of different types of devices to communicate with each other; and increasing the number and types of available wireless communication media. Therefore, there is a need for further improvements to wireless communication systems to overcome the aforementioned technical challenges and other obstacles. Summary of the Invention

[0005] One aspect provides a method for wireless communication by a device. The method includes obtaining an instruction to train a machine learning (ML) model for deployment at a user equipment (UE); transmitting a request to a first network entity for training data associated with one or more UEs; obtaining the training data from the first network entity; training the ML model at least in part based on the training data; transmitting an instruction to a second network entity regarding the completion of training of the ML model; and transmitting the trained ML model to the first network entity and an instruction to transmit the trained ML model to the UE.

[0006] Another aspect provides a method for wireless communication by a device. The method includes: obtaining from a first network entity a first request for training data associated with one or more UEs; transmitting a second request for the training data to the one or more UEs; obtaining the training data from the one or more UEs; transmitting the training data to the first network entity; obtaining from the first network entity a machine learning (ML) model trained at least in part based on the training data and an instruction to transmit the ML model to at least one UE; and transmitting the ML model to the at least one UE.

[0007] Another aspect provides a method for wireless communication by a device. The method includes: obtaining a request for training data associated with the device from a first network entity; transmitting the training data to the first network entity; obtaining an ML model trained at least partially based on the training data from the first network entity; and communicating with a second network entity at least partially based on the ML model.

[0008] Other aspects provide: one or more means operable to, configured to, or otherwise adapted to perform any portion of any method described herein (e.g., such that performance can be implemented by only one means or in a distributed manner across multiple means); one or more non-transitory computer-readable media comprising instructions that, when executed by one or more processors of the one or more means, cause the one or more means to perform any portion of any method described herein (e.g., such that instructions can be included in only one computer-readable medium or in a distributed manner across multiple computer-readable media, such that instructions can be executed by only one processor or by multiple processors in a distributed manner, such that the one or more means can perform any portion of any method described herein). Each device in the apparatus may include one or more processors, and / or enable execution to be performed by only one device or in a distributed manner across multiple devices; one or more computer program products embodied on one or more computer-readable storage media including code for performing any part of any method described herein (e.g., enabling the code to be stored in only one computer-readable medium or in a distributed manner across computer-readable media); and / or one or more devices including one or more components for performing any part of any method described herein (e.g., enabling execution to be performed by only one device or by multiple devices in a distributed manner). By way of example, an apparatus may include a processing system, a device having a processing system, or a processing system cooperating via one or more networks.

[0009] For illustrative purposes, the following description and figures illustrate certain features. Attached Figure Description

[0010] The accompanying drawings depict certain features of the various aspects described herein and should not be considered as limiting the scope of this disclosure.

[0011] Figure 1 An example wireless communication network is depicted.

[0012] Figure 2 An example decomposed base station architecture is described.

[0013] Figure 3 Various aspects of the example base station and example user equipment (UE) are described.

[0014] Figure 4A , Figure 4B , Figure 4C and Figure 4D Various example aspects of data structures used in wireless communication networks are described.

[0015] Figure 5 An example of a service-based architecture for the core network is described.

[0016] Figure 6 An example artificial intelligence (AI) architecture that can be used for AI-enhanced wireless communication is illustrated.

[0017] Figure 7 An example AI architecture of a first wireless device communicating with a second wireless device is illustrated.

[0018] Figure 8 This is an example block diagram of an artificial neural network.

[0019] Figure 9 An example AI management framework for managing AI deployed at the UE is described.

[0020] Figure 10 The process flow for communication between network entities and UEs in the system is described.

[0021] Figure 11 A method for wireless communication is described.

[0022] Figure 12 Another method for wireless communication is described.

[0023] Figure 13 Another method for wireless communication is described.

[0024] Figure 14 Various aspects of the example communication device are described.

[0025] Figure 15 Various aspects of the example communication device are described.

[0026] Figure 16 Various aspects of the example communication device are described. Detailed Implementation

[0027] This disclosure provides apparatus, methods, processing systems, and computer-readable media for core network management of artificial intelligence (AI) deployed at user equipment (UE).

[0028] In some cases, wireless communication systems (e.g., 5G New Radio (NR) systems and / or any future wireless communication systems) can employ AI to perform any of a variety of wireless communication operations, such as Channel State Feedback (CSF) estimation or encoding / decoding, beam management, device localization, etc. In some cases, an AI-based CSF encoder (e.g., a neural network (NN)-based encoder) can be deployed at the UE to generate compressed Channel State Information (CSI), such as a compressed pre-decoder matrix, and the UE can transmit the compressed CSI to network entities. An ML-based CSF decoder (e.g., an NN-based decoder) can be deployed at the network entity to decompress the CSI and use it for channel scheduling and / or configuration of communication links with the UE and / or other UEs. In some cases, ML models can allow the UE to perform beam management tasks, such as beam selection, beam fault detection, and / or beam fault recovery. For example, for beam selection, an ML model can allow the UE to predict channel conditions associated with one or more narrow beams based on measurements associated with one or more wide beams. In some cases, ML models can allow the UE and / or network entities to predict the UE's location under non-line-of-sight conditions for use in communication between the UE and the network entity.

[0029] Technical challenges of AI-based wireless communication include, for example, techniques for designing lifecycle management (LCM) of ML models for UE deployment, which can be implemented without substantial changes to existing wireless communication systems to avoid costly and time-consuming hardware and / or software upgrades. Implementing AI-based wireless communication at the UE involves various LCM tasks, including, for example, data collection (for model training and / or model monitoring), model training, model management (e.g., model monitoring, evaluation, activation, or deactivation), model inference, and model storage. More specifically, there is a need to cost-effectively and efficiently integrate the data collection processes for UE model training and the transfer of ML models to specific UEs into certain wireless communication systems.

[0030] The aspects described in this paper overcome the aforementioned technical problems by providing core network (CN) management for ML models deployed at or on the UE. For example, the CN's data analytics framework can be enhanced to facilitate UE model training and / or model deployment, as described in this paper. Figure 9Further described. In the data analytics framework, the model training function can perform model training on ML models deployed on the UE, and the data collection function (e.g., the Data Collection Application Function (DCAF)) can communicate with clients at one or more UEs to collect training data for model training at the model training function. The data collection function can provide the UE and / or the model training function with access to one or more ML models for deployment to the UE and / or for model training. In some cases, the application server (e.g., a model server operated by the UE vendor) can perform various ML management tasks between the application server and the CN and / or between the application server and the UE, as described herein. Figure 9 Further details are provided. For example, the application server can configure training data sampling and reporting at the UE client.

[0031] The core network management techniques for AI-based UE deployments described herein offer a variety of beneficial effects and / or advantages. These techniques provide an efficient, dynamic, and / or cost-effective solution for managing ML models for UE deployments. For example, enhancing the CN's data analytics framework to support the management of ML models for UE deployments allows network operators to implement AI-based wireless communication without significant changes to the CN and / or radio access network. Therefore, a data analytics framework for managing ML models for UE deployments avoids costly and time-consuming hardware and / or software upgrades to wireless communication systems.

[0032] The data analytics framework for managing ML models used in UE deployments facilitates the allocation of LCM tasks to various entities, such as network entities and other entities. Some entities may include those that design, manufacture, sell, and / or supply communication equipment and / or communication hardware (e.g., modems, system-on-a-chip (SoC), and / or system-in-package (SiP)). For example, since some UE designers may prefer to manage and / or monitor certain LCM tasks (e.g., model performance monitoring, model training, model deployment, etc.), the CN's data analytics framework can facilitate the exposure of certain application functions (e.g., model training and / or data collection functions) to a third-party managed application server, which can be managed by the entity. In other words, the CN's data analytics framework can provide dynamic allocation of LCM tasks to network operators and / or other entities (such as UE designers, manufacturers, suppliers, and / or vendors).

[0033] The core network management techniques for AI for UE deployment described in this paper can provide improved performance for managing ML models used in UE deployment, such as reduced latency, increased throughput, and / or increased capacity. Since the core network is designed to provide edge computing resources (e.g., edge servers) for wireless communication, model training and / or data collection functions can perform LCM tasks (e.g., training data collection, model training, and / or model deployment) that meet the performance expectations and / or specifications required for wireless communication.

[0034] In some cases, the data analytics framework used to manage ML models deployed for UEs can be executed on dedicated AI / ML computing devices, such as cloud servers with one or more neural network processors, one or more graphics processing units, or any suitable AI / ML accelerator. Dedicated AI / ML computing devices can have a more efficient ability to perform AI / ML computations compared to general-purpose processors such as microprocessors (e.g., lower latency, increased throughput, etc.).

[0035] Introduction to wireless communication networks The techniques and methods described herein can be used in a variety of wireless communication networks. Although aspects herein may be described using terms commonly associated with 3G, 4G, 5G, 6G, and / or other generations of wireless technologies, aspects of this disclosure are equally applicable to other communication systems and standards not explicitly mentioned herein.

[0036] Figure 1 An example of a wireless communication network 100 in which the aspects described herein can be implemented is depicted.

[0037] Generally, wireless communication network 100 includes various network entities (alternatively, network elements or network nodes). Network entities are typically communication devices and / or communication functions performed by communication devices (e.g., user equipment (UE), base station (BS), components of a BS, servers, etc.). Because such communication devices are part of wireless communication network 100 and facilitate wireless communication, they may be referred to as wireless communication devices. For example, various functions of the network and various devices associated with and interacting with the network may be considered network entities. Furthermore, wireless communication network 100 includes terrestrial and non-terrestrial aspects (also referred to herein as non-terrestrial network entities). Terrestrial aspects include terrestrial network entities such as terrestrial network entities (e.g., BS 102), and non-terrestrial aspects include satellite 140 and airborne platforms. These non-terrestrial aspects may include onboard network entities (e.g., one or more BSs) capable of communicating with other network elements (e.g., terrestrial BSs) and UEs.

[0038] In the depicted example, wireless communication network 100 includes BS 102, UE 104 and one or more core networks (such as Evolved Packet Core (EPC) 160 and 5G Core (5GC) network 190) that interoperate to provide communication services over various communication links, including wired and wireless links.

[0039] Figure 1 Various example UEs 104 are described, which may more generally include: cellular phones, smartphones, Session Initiation Protocol (SIP) phones, laptops, personal digital assistants (PDAs), satellite radios, global positioning systems, multimedia devices, video devices, digital audio players, cameras, game consoles, tablets, smart devices, wearable devices, vehicles, electricity meters, air pumps, large or small kitchen appliances, healthcare devices, implants, sensors / actuators, displays, Internet of Things (IoT) devices, always-on (AON) devices, edge processing devices, data centers, or other similar devices. UE 104 may also be more generally referred to as mobile devices, wireless devices, stations, mobile stations, subscriber stations, mobile subscriber stations, mobile units, subscriber units, wireless units, remote units, remote devices, access terminals, mobile terminals, wireless terminals, remote terminals, mobile phones, and others.

[0040] BS 102 communicates wirelessly with UE 104 via communication link 120 (e.g., transmitting or receiving signals to or from UE 104). Communication link 120 between BS 102 and UE 104 may include uplink (UL) (also known as reverse link) transmission from UE 104 to BS 102 and / or downlink (DL) (also known as forward link) transmission from BS 102 to UE 104. In various aspects, communication link 120 may utilize multiple-input multiple-output (MIMO) antenna techniques, including spatial multiplexing, beamforming, and / or transmit diversity.

[0041] BS 102 may typically include: NodeB, enhanced NodeB (eNB), next-generation enhanced NodeB (ng-eNB), next-generation NodeB (gNB or gNodeB), access point, transceiver base station, radio base station, radio transceiver, transceiver functionality, transmit / receive point, and / or others. Each of BS 102 provides communication coverage for a corresponding coverage area 110, which may sometimes be referred to as a cell, and in some cases may overlap (e.g., a small cell 102' may have a coverage area 110' that overlaps with the coverage area 110 of a macro cell). For example, BS may provide communication coverage for macro cells (covering a relatively large geographic area), pico cells (covering a relatively small geographic area, such as a stadium), femtocells (covering a relatively small geographic area (e.g., a home)), and / or other types of cells.

[0042] Generally, a cell can refer to a portion, partition, or segment of wireless communication coverage served by a network entity within a wireless communication network. A cell can have geographical characteristics (such as a geographical coverage area) and radio frequency characteristics (such as time and / or frequency resources dedicated to the cell). For example, multiple cells using different frequency resources (e.g., bandwidth portions) and / or different time resources can cover a specific geographical coverage area. As another example, a single cell can cover a specific geographical coverage area. In some contexts (e.g., carrier aggregation scenarios and / or multi-connectivity scenarios), the terms "cell" or "serving cell" can refer to or correspond to a specific carrier frequency (e.g., component carrier) used for wireless communication, and "cell group" can refer to or correspond to multiple carriers used for wireless communication. As an example, in a carrier aggregation scenario, a UE can communicate on multiple component carriers corresponding to multiple (serving) cells in the same cell group, and in a multi-connectivity (e.g., dual-connectivity) scenario, a UE can communicate on multiple component carriers corresponding to multiple cell groups.

[0043] While BS 102 is described as a single communication device in various aspects, it can be implemented in a variety of configurations. For example, to cite a few examples, one or more components of the base station can be decomposed, including a central unit (CU), one or more distributed units (DU), one or more radio units (RU), a near-real-time (near-RT) RAN intelligent controller (RIC), or a non-real-time (non-RT) RIC. In another example, various aspects of the base station can be virtualized. More generally, a base station (e.g., BS 102) can include components located at a single physical location or components located at various physical locations. In examples where the base station includes components located at various physical locations, the various components can each perform functions, such that the various components collectively achieve functionality similar to a base station located at a single physical location. In some aspects, a base station including components located at various physical locations can be referred to as a decomposed radio access network architecture (such as an open RAN (O-RAN) or virtualized RAN (VRAN) architecture). Figure 2 An example decomposed base station architecture is depicted and described.

[0044] Different BSs 102 within the wireless communication network 100 can also be configured to support different radio access technologies (such as 3G, 4G, and / or 5G). For example, a BS 102 configured for 4G LTE (collectively referred to as Evolved Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access Network (E-UTRAN)) can interface with EPC 160 via a first backhaul link 132 (e.g., S1 interface). A BS 102 configured for 5G (e.g., 5G NR or Next Generation RAN (NG-RAN)) can interface with 5GC 190 via a second backhaul link 184. BSs 102 can communicate directly or indirectly with each other (e.g., via EPC 160 or 5GC 190) via a third backhaul link 134 (e.g., X2 interface), which can be wired or wireless.

[0045] Wireless communication network 100 can subdivide the electromagnetic spectrum into various categories, bands, channels, or other characteristics. In some aspects, subdivision is provided based on wavelength and frequency, where frequency may also be referred to as carrier, subcarrier, channel, tone, or subband. For example, 3GPP currently defines frequency range 1 (FR1) as including 410MHz to 7125MHz, which is often (interchangeably) referred to as “sub-6GHz”. Similarly, 3GPP currently defines frequency range 2 (FR2) as including 24,250MHz to 71,000MHz, which is sometimes (interchangeably) referred to as “millimeter wave” (“mmW” or “mmWave”). In some cases, FR2 can be further defined according to subranges (such as a first subrange FR2-1 including 24,250MHz to 52,600MHz and a second subrange FR2-2 including 52,600MHz to 71,000MHz). Base stations configured to communicate using mmWave / near mmWave radio bands (e.g., mmWave base stations such as BS 180) can utilize beamforming (e.g., 182) with UEs (e.g., 104) to improve path loss and range.

[0046] The communication link 120 between BS 102 and, for example, UE 104 can be via one or more carriers, which may have different bandwidths (e.g., 5MHz, 10MHz, 15MHz, 20MHz, 100MHz, 400MHz and / or other MHz) and may be aggregated in various ways. The carriers may be adjacent to each other or may not be adjacent to each other. The allocation of carriers may be asymmetric with respect to DL and UL (e.g., more or fewer carriers may be allocated to DL compared to UL).

[0047] Compared to lower-frequency communication, communication using higher frequency bands may have higher path loss and shorter range. Therefore, some base stations (e.g., Figure 1 The beamforming 182 of the BS 180 (180) with the UE 104 can be used to improve path loss and range. For example, the BS 180 and UE 104 may each include multiple antennas, such as antenna elements, antenna panels, and / or antenna arrays, to facilitate beamforming. In some cases, the BS 180 may transmit beamformed signals to the UE 104 in one or more transmit directions 182''. The UE 104 may receive beamformed signals from the BS 180 in one or more receive directions 182''. The UE 104 may also transmit beamformed signals to the BS 180 in one or more transmit directions 182''. The BS 180 may also receive beamformed signals from the UE 104 in one or more receive directions 182''. The BS 180 and UE 104 may then perform beamforming training to determine the optimal receive and transmit directions for each of the BS 180 and UE 104. It is worth noting that the transmit and receive directions of the BS 180 may or may not be the same. Similarly, the transmission and reception directions of UE 104 may or may not be the same.

[0048] The wireless communication network 100 also includes a Wi-Fi AP 150 that communicates with a Wi-Fi station (STA) 152 via a communication link 154 in, for example, unlicensed spectrum in 2.4 GHz and / or 5 GHz.

[0049] Some UEs 104 may use device-to-device (D2D) communication links 158 to communicate with each other. The D2D communication link 158 may use one or more sidelink channels, such as physical sidelink broadcast channel (PSBCH), physical sidelink discovery channel (PSDCH), physical sidelink shared channel (PSSCH), physical sidelink control channel (PSCCH), and / or physical sidelink feedback channel (PSFCH).

[0050] EPC 160 may include various functional components, including: such as the Mobility Management Entity (MME) 162 in the illustrated example, other MMEs 164, Serving Gateway 166, Multimedia Broadcast Multicast Service (MBMS) Gateway 168, Broadcast Multicast Service Center (BM-SC) 170, and / or Packet Data Network (PDN) Gateway 172. MME 162 can communicate with Home Subscriber Server (HSS) 174. MME 162 is the control node that handles signaling between UE 104 and EPC 160. Generally, MME 162 provides bearer and connectivity management.

[0051] Generally, user Internet Protocol (IP) packets are transmitted through Serving Gateway 166, which is itself connected to PDN Gateway 172. PDN Gateway 172 provides UE IP address allocation and other functions. PDN Gateway 172 and BM-SC 170 are connected to IP service 176, which may include, for example, the Internet, intranet, IP Multimedia Subsystem (IMS), packet switching (PS) streaming service, and / or other IP services.

[0052] The BM-SC 170 provides functions for MBMS user service dispatch and delivery. The BM-SC 170 can serve as an entry point for content provider MBMS transmissions, can be used to authorize and initiate MBMS bearer services within a Public Land Mobile Network (PLMN), and / or can be used to schedule MBMS transmissions. The MBMS gateway 168 can be used to distribute MBMS services to BS 102 belonging to a Broadcast-Specific Service Single Frequency Network (MBSFN) area, and / or can be responsible for session management (start / stop) and collecting eMBMS-related billing information.

[0053] 5GC 190 may include various functional components, including: Access and Mobility Management Function (AMF) 192, other AMFs 193, Session Management Function (SMF) 194, and User Plane Function (UPF) 195. AMF 192 can communicate with Unified Data Management (UDM) 196.

[0054] AMF 192 is the control node that handles signaling between UE 104 and 5GC 190. AMF 192 provides services such as Quality of Service (QoS) flow and session management.

[0055] Internet Protocol (IP) packets are transmitted via UPF 195, which is connected to IP service 197 and provides the UE with IP address allocation and other functions for 5GC 190. IP service 197 may include, for example, the Internet, intranet, IMS, PS streaming service and / or other IP services.

[0056] In various aspects, to give a few examples, network entities or network nodes can be implemented as aggregated base stations, decomposed base stations, components of base stations, integrated access and backhaul (IAB) nodes, relay nodes, and sidelink nodes.

[0057] Figure 2An example decomposed base station 200 architecture is depicted. The decomposed base station 200 architecture may include one or more central units (CUs) 210, which may communicate directly with the core network 220 via a backhaul link, or indirectly with the core network 220 through one or more decomposed base station units, such as a near real-time (near-RT) RAN Intelligent Controller (RIC) 225 via an E2 link, or a non-real-time (non-RT) RIC 215 associated with a Service Management and Orchestration (SMO) framework 205, or both. CUs 210 may communicate with one or more distributed units (DUs) 230 via corresponding midhaul links (such as F1 interfaces). DUs 230 may communicate with one or more radio units (RUs) 240 via corresponding fronthaul links. RUs 240 may communicate with a corresponding UE 104 via one or more radio frequency (RF) access links. In some specific implementations, UE 104 may be served simultaneously by multiple RUs 240.

[0058] Each unit in a cell (e.g., CU 210, DU 230, RU 240, and near-RT RIC 225, non-RT RIC 215, and SMO frame 205) may include or be coupled to one or more interfaces configured to receive or transmit signals, data, or information (collectively, signals) via a wired or wireless transmission medium. Each of the cells, or an associated processor or controller that provides instructions to the cell's communication interface, may be configured to communicate with one or more other cells via the transmission medium. For example, these cells may include a wired interface configured to receive signals or transmit signals to one or more other cells via a wired transmission medium. Additionally or alternatively, a cell may include a wireless interface that may include a receiver, transmitter, or transceiver (such as a radio frequency (RF) transceiver) configured to receive signals on a wireless transmission medium or transmit signals to one or more other cells, or both.

[0059] In some aspects, CU 210 can host one or more higher-level control functions. Such control functions may include Radio Resource Control (RRC), Packet Data Convergence Protocol (PDCP), Serving Data Adaptation Protocol (SDAP), etc. Each control function can be implemented using an interface configured to signal to other control functions hosted by CU 210. CU 210 can be configured to handle user plane functions (e.g., Central Unit-User Plane (CU-UP)), control plane functions (e.g., Central Unit-Control Plane (CU-CP)), or combinations thereof. In some implementations, CU 210 can be logically split into one or more CU-UP units and one or more CU-CP units. When implemented in an O-RAN configuration, CU-UP units can communicate bidirectionally with CU-CP units via an interface such as an E1 interface. CU 210 can be implemented to communicate with DU 230 for network control and signaling, as needed.

[0060] DU 230 may correspond to a logical unit comprising one or more base station functions for controlling the operation of one or more RU 240s. In some aspects, DU 230 may host one or more of the Radio Link Control (RLC) layer, the Media Access Control (MAC) layer, and one or more high physical (PHY) layers (such as modules for forward error correction (FEC) encoding and decoding, scrambling, modulation and demodulation, etc.), depending at least in part on the functional breakdown, such as those defined by the 3rd Generation Partnership Project (3GPP). In some aspects, DU 230 may also host one or more low PHY layers. Each layer (or module) may be implemented using an interface configured to communicate signaling with other layers (and modules) hosted by DU 230 or with control functions hosted by CU 210.

[0061] Lower-layer functionality can be implemented by one or more RU 240s. In some deployments, the RU240 controlled by the DU 230 may correspond to a logical node that hosts RF processing functions or low-PHY layer functions (such as performing Fast Fourier Transform (FFT), Inverse FFT (iFFT), digital beamforming, Physical Random Access Channel (PRACH) extraction and filtering, or both, at least in part based on functional decomposition (such as lower-layer functional decomposition). In such architectures, the RU 240 may be implemented to handle over-the-air (OTA) communications with one or more UE 104s. In some specific implementations, the real-time and non-real-time aspects of control plane and user plane communications with the RU 240 may be controlled by the corresponding DU 230. In some scenarios, this configuration enables the implementation of the DU 230 and CU 210 in cloud-based RAN architectures (such as vRAN architectures).

[0062] SMO framework 205 can be configured to support RAN deployment and provisioning of both non-virtualized and virtualized network elements. For non-virtualized network elements, SMO framework 205 can be configured to support the deployment of dedicated physical resources for RAN coverage requirements, which can be managed via operation and maintenance interfaces such as the O1 interface. For virtualized network elements, SMO framework 205 can be configured to interact with cloud computing platforms such as Open Cloud (O-Cloud) 290 to perform network element lifecycle management (such as instantiating virtualized network elements) via cloud computing platform interfaces such as the O2 interface. Such virtualized network elements may include, but are not limited to, CU 210, DU 230, RU 240, and near-RT RIC 225. In some specific implementations, SMO framework 205 may communicate with the hardware aspects of the 4G RAN (such as Open eNB (O-eNB) 211) via the O1 interface. Additionally, in some implementations, the SMO framework 205 may communicate directly with one or more DU 230s and / or one or more RU 240s via the O1 interface. The SMO framework 205 may also include a non-RT RIC 215 configured to support the functionality of the SMO framework 205.

[0063] The non-RT RIC 215 can be configured to include logical functions that enable non-real-time control and optimization of RAN elements and resources, including AI / ML workflows for model training and updates, or policy-based guidance for applications / features in the near-RT RIC 225. The non-RT RIC 215 can be coupled to or communicate with the near-RT RIC 225 (e.g., via an A1 interface). The near-RT RIC 225 can be configured to include logical functions that enable near real-time control and optimization of RAN elements and resources via data collection and actions through an interface (e.g., via an E2 interface) connecting one or more CU 210s, one or more DU 230s, or both, and O-eNBs to the near-RT RIC 225.

[0064] In some implementations, to generate AI / ML models to be deployed in the near-RT RIC 225, the non-RT RIC 215 may receive parameters or external enrichment information from an external server. This information can be utilized by the near-RT RIC 225 and may be received from non-network data sources or network functions at the SMO framework 205 or the non-RT RIC 215. In some examples, the non-RT RIC 215 or the near-RT RIC 225 may be configured to tune RAN behavior or performance. For example, the non-RT RIC 215 may monitor long-term trends and patterns in performance and employ AI / ML models to perform corrective actions via the SMO framework 205 (such as reconfiguration via O1) or by creating RAN management policies (such as A1 policies).

[0065] Figure 3 Various aspects of examples BS 102 and UE 104 are described.

[0066] Generally, BS 102 includes various processors (e.g., 318, 320, 330, 338, and 340), antennas 334a to 334t (collectively referred to as 334), transceivers 332a to 332t (collectively referred to as 332) including modulators and demodulators, and other aspects that enable the wireless transmission of data (e.g., data source 312) and the wireless reception of data (e.g., data sink 314). For example, BS 102 can transmit and receive data between BS 102 and UE 104. BS 102 includes a controller / processor 340 that can be configured to implement the various functions described herein related to wireless communication.

[0067] Generally, UE 104 includes various processors (e.g., 358, 364, 366, 370, and 380), antennas 352a to 352r (collectively referred to as 352), transceivers 354a to 354r (collectively referred to as 354) including modulators and demodulators, and other aspects that enable the wireless transmission of data (e.g., retrieval from data source 362) and the wireless reception of data (e.g., provision to data sink 360). UE 104 includes a controller / processor 380 that can be configured to implement the various functions described herein related to wireless communication.

[0068] Regarding example downlink transmission, BS 102 includes a transmission processor 320 that can receive data from data source 312 and control information from controller / processor 340. This control information may be for a Physical Broadcast Channel (PBCH), Physical Control Format Indicator Channel (PCFICH), Physical Hybrid Automatic Repeat Request (HARQ) Indicator Channel (PHICH), Physical Downlink Control Channel (PDCCH), Group Common PDCCH (GC PDCCH), and / or others. In some examples, this data may be for a Physical Downlink Shared Channel (PDSCH).

[0069] The transmitter processor 320 can process (e.g., encode and symbol map) data and control information to obtain data symbols and control symbols, respectively. The transmitter processor 320 can also generate reference symbols (such as those for the primary synchronization signal (PSS), secondary synchronization signal (SSS), PBCH demodulation reference signal (DMRS), and channel state information reference signal (CSI-RS)).

[0070] The transmit (TX) multiple-input multiple-output (MIMO) processor 330 can perform spatial processing (e.g., pre-decoding) on ​​data symbols, control symbols, and / or reference symbols where applicable, and can provide the output symbol stream to the modulators (MODs) in transceivers 332a to 332t. Each modulator in transceivers 332a to 332t can process its corresponding output symbol stream to obtain an output sample stream. Each modulator can further process (e.g., convert to analog, amplify, filter, and up-convert) the output sample stream to obtain a downlink signal. The downlink signal from the modulators in transceivers 332a to 332t can be transmitted via antennas 334a to 334t, respectively.

[0071] To receive downlink transmissions, UE 104 includes antennas 352a to 352r that receive downlink signals from BS 102 and provide the received signals to demodulators (DEMODs) in transceivers 354a to 354r, respectively. Each demodulator in transceivers 354a to 354r can adjust (e.g., filter, amplify, down-convert, and digitize) the corresponding received signal to obtain an input sample. Each demodulator can further process the input sample to obtain the received symbols.

[0072] The RX MIMO detector 356 acquires received symbols from all demodulators in transceivers 354a to 354r, performs MIMO detection on the received symbols where applicable, and provides the detected symbols. The receive processor 358 processes (e.g., demodulates, deinterleaves, and decodes) the detected symbols, provides the decoded data for UE 104 to data sink 360, and provides the decoded control information to controller / processor 380.

[0073] Regarding the example uplink transmission, UE 104 further includes a transmission processor 364 that receives and processes data from data source 362 (e.g., for PUSCH) and control information from controller / processor 380 (e.g., for Physical Uplink Control Channel (PUCCH)). Transmission processor 364 can also generate reference symbols for reference signals (e.g., for Sounding Reference Signal (SRS)). Symbols from transmission processor 364 may be pre-decoded by TX MIMO processor 366, where applicable, further processed by modulators in transceivers 354a to 354r (e.g., for SC-FDM), and transmitted to BS 102.

[0074] At BS 102, uplink signals from UE 104 can be received by antennas 334a to 334t, processed by demodulators in transceivers 332a to 332t, detected where applicable by RX MIMO detector 336, and further processed by receiver processor 338 to obtain decoded data and control information transmitted by UE 104. Receiver processor 338 can provide the decoded data to data sink 314 and the decoded control information to controller / processor 340.

[0075] Memory 342 and memory 382 can store data and program code for BS 102 and UE 104, respectively.

[0076] Scheduler 344 can schedule UE to transmit data on the downlink and / or uplink.

[0077] In various respects, BS 102 can be described as transmitting and receiving various types of data associated with the methods described herein. In these contexts, “transmitting” can refer to various mechanisms that output data, such as from data source 312, scheduler 344, memory 342, transmit processor 320, controller / processor 340, TX MIMO processor 330, transceivers 332a to 332t, antennas 334a to 334t, and / or other aspects described herein. Similarly, “receiving” can refer to various mechanisms that acquire data, such as from antennas 334a to 334t, transceivers 332a to 332t, RX MIMO detector 336, controller / processor 340, receive processor 338, scheduler 344, memory 342, and / or other aspects described herein.

[0078] In various respects, UE 104 can also be described as transmitting and receiving various types of data associated with the methods described herein. In these contexts, “transmitting” can refer to various mechanisms that output data, such as from data source 362, memory 382, ​​transmit processor 364, controller / processor 380, TX MIMO processor 366, transceivers 354a to 354t, antennas 352a to 352t, and / or other aspects described herein. Similarly, “receiving” can refer to various mechanisms that acquire data, such as from antennas 352a to 352t, transceivers 354a to 354t, RX MIMO detector 356, controller / processor 380, receive processor 358, memory 382, ​​and / or other aspects described herein.

[0079] In some respects, the processor can be configured to perform various operations (such as those associated with the methods described herein) and to send (output) data to or receive data from another interface configured to send or receive data, respectively.

[0080] In various aspects, artificial intelligence (AI) processors 318 and 370 may perform AI processing for BS 102 and / or UE 104, respectively. AI processor 318 may include AI accelerator hardware or circuitry, such as one or more neural processing units (NPUs), one or more neural network processors, one or more tensor processors, one or more deep learning processors, etc. AI processor 370 may similarly include AI accelerator hardware or circuitry. As an example, AI processor 370 may perform AI-based beam management, AI-based channel state feedback (CSF), AI-based antenna tuning, and / or AI-based positioning (e.g., Global Navigation Satellite System (GNSS) positioning). In some cases, AI processor 318 may use hardware-accelerated AI inference and / or AI training to process feedback (e.g., CSF) from UE 104. AI processor 318 may, for example, use hardware-accelerated AI inference associated with the CSF to decode compressed CSF from UE 104. In some cases, AI processor 318 may perform certain RAN-based functions, including, for example, network planning, network performance management, energy-efficient network operation, etc.

[0081] Figure 4A , Figure 4B , Figure 4C and Figure 4D Describes the use of wireless communication networks (such as Figure 1 All aspects of the data structure of the wireless communication network 100.

[0082] Specifically, Figure 4A Figure 400 is an example of the first subframe within a 5G (e.g., 5G NR) frame structure. Figure 4B Figure 430 illustrates an example of a DL channel within a 5G subframe. Figure 4C Figure 450 illustrates an example of the second subframe within a 5G frame structure, and Figure 4D Figure 480 illustrates an example of a UL channel within a 5G subframe.

[0083] Wireless communication systems can utilize Orthogonal Frequency Division Multiplexing (OFDM) with a cyclic prefix (CP) on both the uplink and downlink. Such systems can also support half-duplex operation using Time Division Duplex (TDD). OFDM and Single-Carrier Frequency Division Multiplexing (SC-FDM) will (e.g., as...) Figure 4B and Figure 4D The system bandwidth (as depicted in the text) is divided into multiple orthogonal subcarriers. Each subcarrier can be modulated with data. Modulation symbols can be transmitted in the frequency domain using OFDM and / or in the time domain using SC-FDM.

[0084] Wireless communication frame structures can be frequency division duplex (FDD), where for a specific set of subcarriers, subframes within that set are dedicated to either deep (DL) or ultra-low (UL). Wireless communication frame structures can also be time division duplex (TDD), where for a specific set of subcarriers, subframes within that set are dedicated to both DL and UL.

[0085] exist Figure 4A and Figure 4C In this example, the wireless communication frame structure is TDD, where D stands for DL, U for UL, and X is flexibly used between DL and UL. The UE can configure the time slot format using the received Time Slot Format Indicator (SFI) (dynamically via DL Control Information (DCI) or semi-statically / statically via Radio Resource Control (RRC) signaling). In the depicted example, a 10ms frame is divided into 10 equal-sized 1ms subframes. Each subframe may include one or more time slots. In some examples, each time slot may include 12 or 14 symbols, depending on the Cyclic Prefix (CP) type (e.g., 12 symbols per time slot for extended CP, or 14 symbols per time slot for regular CP). Subframes may also include micro-time slots, which typically have fewer symbols than the entire time slot. Other wireless communication technologies may have different frame structures and / or different channels.

[0086] In some respects, the number of time slots within a subframe (e.g., the time slot duration within a subframe) is based on a parameter set that defines the frequency-domain subcarrier spacing and symbol duration, as further described herein. In some respects, given a parameter set μ, each subframe has 2 μ The number of time slots is 1. Therefore, parameter sets (µ) 0 through 6 allow for 1, 2, 4, 8, 16, 32, and 64 time slots per subframe, respectively. In some cases, extended CP (e.g., 12 symbols per time slot) can be used with specific parameter sets; for example, parameter set 2 allows for 4 time slots per subframe. Subcarrier spacing and symbol length / duration are functions of the parameter set. The subcarrier spacing can be equal to... kHz, where μ is the parameter set from 0 to 6. As an example, the parameter set... Corresponding to a subcarrier spacing of 15 kHz, and the parameter set This corresponds to a subcarrier spacing of 960 kHz. Symbol length / duration is negatively correlated with subcarrier spacing. Figure 4A , Figure 4B , Figure 4C and Figure 4D It provides a slot format with 14 symbols per slot (e.g., regular CP) and a parameter set with 4 slots per subframe. Example. In this case, the slot duration is 0.25 ms, the subcarrier spacing is 60 kHz, and the symbol duration is approximately 16.67 μs.

[0087] like Figure 4A , Figure 4B , Figure 4C and Figure 4D As depicted, the resource grid can be used to represent the frame structure. Each time slot includes a resource block (RB) (also known as a physical RB (PRB)) extending for, for example, 12 consecutive subcarriers. The resource grid is divided into multiple resource elements (REs). The number of bits carried by each RE depends on the modulation scheme, including, for example, quadrature phase shift keying (QPSK) or quadrature amplitude modulation (QAM).

[0088] like Figure 4A As illustrated in the example, some REs in the RE carry information for the UE (e.g., Figure 1 and Figure 3 The reference (pilot) signal (RS) for the UE (104) may include a demodulation RS (DMRS) and / or a channel state information reference signal (CSI-RS) for channel estimation at the UE. The RS may also include a beam measurement RS (BRS), a beam refinement RS (BRRS), and / or a phase tracking RS (PT-RS).

[0089] Figure 4B Examples of various DL channels within a subframe of a frame are illustrated. The Physical Downlink Control Channel (PDCCH) carries the DCI within one or more Control Channel Elements (CCEs), each CCE comprising, for example, nine RE groups (REGs), each REG comprising, for example, four consecutive REs in an OFDM symbol.

[0090] The Primary Synchronization Signal (PSS) can be located within symbol 2 of a specific subframe of the frame. The PSS is generated by the UE (e.g., Figure 1 and Figure 3 104) is used to determine subframe / symbol timing and physical layer identifier.

[0091] The secondary synchronization signal (SSS) can be located in symbol 4 of a specific subframe of the frame. The SSS is used by the UE to determine the physical layer cell identifier group number and radio frame timing.

[0092] Based on the Physical Layer Identifier and Physical Layer Cell Identifier Group Number, the UE can determine the Physical Cell Identifier (PCI). Based on the PCI, the UE can determine the location of the aforementioned DMRS. The Physical Broadcast Channel (PBCH), carrying the Master Information Block (MIB), can be logically grouped with the PSS and SSS to form a Synchronization Signal (SS) / PBCH block. The MIB provides the System Frame Number (SFN) and the number of Restricted Frames (RBs) in the system bandwidth. The Physical Downlink Shared Channel (PDSCH) carries user data, broadcast system information (such as System Information Blocks (SIBs)) not transmitted via the PBCH, and / or paging messages.

[0093] like Figure 4C As illustrated, some REs in the REs carry DMRS for channel estimation at the base station (indicated as R for a particular configuration, but other DMRS configurations are possible). The UE can transmit DMRS for PUCCH and DMRS for PUSCH. PUSCH DMRS can be transmitted, for example, in the first or second symbol before the PUSCH. PUCCH DMRS can be transmitted in different configurations depending on whether a short or long PUCCH is being transmitted and depending on the specific PUCCH format used. UE104 can transmit a Sounding Reference Signal (SRS). SRS can be transmitted, for example, in the last symbol of a subframe. SRS can have a comb structure, and the UE can transmit SRS on one of the comb teeth. SRS can be used by the base station for channel quality estimation to enable frequency-dependent scheduling of the UL.

[0094] Figure 4D Examples of various UL channels within a subframe of a frame are illustrated. The PUCCH can be located as indicated in one configuration. The PUCCH carries uplink control information (UCI), such as scheduling requests, channel quality indicators (CQI), pre-decoding matrix indicators (PMI), rank indicators (RI), and HARQ ACK / NACK feedback. The PUSCH carries data and may additionally be used to carry buffer status reports (BSR), power clearance reports (PHR), and / or UCI.

[0095] Example Core Network Architecture As described herein, a wireless communication system may include a core network (CN) (e.g., 5GC network 190) that enables connectivity to data networks (e.g., the Internet, intranets, private data networks, etc.). In some cases, the CN may enable connectivity to application servers, such as those hosting video streaming services, social media services, virtual reality and / or augmented reality services, gaming services, etc.

[0096] Figure 5 An example service-based architecture for CN 500 is illustrated. In this example, CN 500 communicates with RAN 502 and UE 504, for example, as described in this article. Figure 1 As described, CN 500 can facilitate communication between UE 504 and data network 506, which may include, for example, the Internet and / or an intranet.

[0097] CN 500 includes Access and Mobility Management Function (AMF) 508, Session Management Function (SMF) 510, User Plane Function (UPF) 512, one or more Application Functions (AF) 514, Network Repository Function (NRF) 516, and Network Open Function (NEF) 518.

[0098] In a service-based architecture, the network functions of CN 500 communicate with each other via a common bus 520. The common bus 520 is used to communicate control plane services between network functions. Control plane services may include control signaling, such as mobility management and / or session management signaling, while user plane services may include application data between the UE and the application server. In this example, the network functions include AMF 508, SMF 510, AF 514, NRF 516, and NEF 518. It should be noted that CN 500 may include other network functions besides or instead of these examples.

[0099] A service-based architecture enables a cloud-based CN. For example, any of the various functions of CN 500 can be or include logical functions hosted on or on computing devices such as network entities, computers, servers, virtual servers, etc. One or more functions can be virtualized to allow virtual network entities to operate using a shared computing platform (e.g., a cloud computing platform). For example, a network entity can be configured to host, execute, and / or support any of the various functions of CN 500. In some aspects, any of the various functions of CN 500 can correspond to a network entity hosting a given function and / or a shared computing platform. It should be noted that other architectures for CNs, such as reference point architectures, roaming architectures, etc., can be used in addition to or in place of service-based architectures.

[0100] For example, AMF 508 can perform registration management, connection management, reachability management, mobility management, access authentication, and access authorization for UE 504. SMF 510 can perform Protocol Data Unit (PDU) session management, such as allocating and managing UE Internet Protocol (IP) addresses. UPF 512 routes and forwards user plane services between RAN 502 and data network 506.

[0101] AF 514 is a control plane function that interacts with other functions (such as AMF 508, SMF 510, and UPF 512) to provide support for one or more specific services. For example, AF 514 may include control plane functions for managing video streaming services, social media services, and / or video game services. In some cases, AF 514 may be co-located at network entities (such as base stations, CUs, DUs, and / or RUs) to facilitate reduced latency and / or reduced transmission bandwidth between network entities and services.

[0102] NRF 516 can be used as a repository that allows network functions to register their services and then allows other network functions to discover those services and corresponding network functions. For example, for network function registration, upon initial activation and / or reconfiguration, a network function (e.g., a specific AF 514) can register the services managed at the network function with NRF 516, and NRF 516 can store network function profiles for later discovery by other network functions. For network function discovery, a network function can request information associated with a specific network function from NRF 516, and NRF 516 can provide the requested information to the network function.

[0103] NEF 518 can support the secure opening of capabilities and events associated with CN 500 and / or UE 504 to external network entities (not shown) and enables the secure provisioning of information from external network entities to CN 500. For example, network function capabilities and events can be securely developed by NEF 518 to support, for example, third-party services (e.g., analytics and monitoring of streaming services), application functions, edge computing, etc.

[0104] It should be noted that any of the network entities in CN 500 (e.g., AMF 508, SMF 510, UPF 512, AF 514, NRF 516 and / or NEF 518) may perform functions other than or in lieu of those functions described for the respective entity.

[0105] In some aspects, CN 500 may include a data analytics framework with Network Data Analysis Function (NWDAF) 522 and Data Collection Application Function (DCAF) 524. NWDAF 522 may provide analysis to network functions in CN 500 and / or Network Controller 550. The analysis may include, for example, performance statistics and / or predictions associated with the operation of CN 500 and / or RAN 502. For example, NWDAF 522 may predict UE mobility, e.g., as a prediction of the routes the UE will take through network coverage areas and the corresponding network entities that can serve communication with the UE along such routes. As another example, NWDAF 522 may provide network slice monitoring, which may include monitoring of network performance and quality of service on one or more network slices and / or one or more users / subscribers. Network slice monitoring may allow communication service providers to meet the terms of service license agreements, e.g., agreements that specify certain QoS levels for subscribers or users. In some respects, the NWDAF 522 can perform ML model training on ML models (e.g., ML models for network analysis) deployed at or in CN 500 and / or RAN 502.

[0106] Network controller 550 can configure and manage CN 500 and / or RAN 502. The network controller enables the operation, management, and maintenance of CN 500 and / or RAN 502. For example, network controller 550 can collect performance data, events, and / or alarms reported by CN 500 and / or RAN 502 and provide visualization of such data. Network controller 550 can provide a platform for provisioning and / or maintaining CN 500 and / or RAN 502. In some cases, network controller 550 may be or include, for example, a near-RT RIC 225, a non-RT RIC 215, and / or SMO framework 205 in an open RAN or cloud-based RAN architecture. In some cases, network controller 550 may be or include an Operations, Administration, and Maintenance (OAM) host or server.

[0107] The NWDAF 522 can interact with the DCAF 524 to collect data from UE applications running at the UE 504 (e.g., streaming service applications, game service applications, social media service applications, etc.) as input for analytics generation and / or ML model training at the NWDAF 522. Data collection requests from the NWDAF can trigger the DCAF 524 to collect data from the UE application. The UE application running at the UE 504 can establish a connection to the DCAF 524 via a PDU session through the user (or data) plane, and the DCAF 524 communicates with and collects data from the UE application.

[0108] As discussed herein, a reference to the RAN performing certain operations may refer to one or more network entities (e.g., base stations, non-terrestrial networks, and / or one or more decomposed entities thereof) performing the operations. As discussed herein, a reference to a CN performing certain operations may refer to one or more physical and / or logical network entities (e.g., network functions and / or application functions) performing the operations.

[0109] Example of artificial intelligence for wireless communication Some aspects described in this paper can be implemented, at least in part, using some form of artificial intelligence (AI), such as the process of using machine learning (ML) models to infer or predict output data based on input data. Example ML models may include mathematical representations of one or more relationships between various objects (e.g., within a wireless network environment or system) to provide outputs representing one or more predictions or inferences relating to one or more of these objects. Once the ML model has been trained, it can be deployed to process data that is all or part of the training data and provide outputs representing one or more predictions or inferences based on the input data.

[0110] Machine learning (ML) can often be characterized by a learning type that generates a specific type of learning model that performs a particular type of task. For example, different types of machine learning include supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning.

[0111] Supervised learning algorithms typically model the relationships and dependencies between input features (e.g., feature vectors) and one or more target outputs. Supervised learning uses labeled training data, which consists of data including one or more inputs and a desired output. Supervised learning can be used to train models to perform tasks such as classification (where the goal is to predict discrete values) or regression (where the goal is to predict continuous values). Some example supervised learning algorithms include nearest neighbor, Naive Bayes, decision trees, linear regression, support vector machines (SVM), and artificial neural networks (ANN).

[0112] Unsupervised learning algorithms process unlabeled input data and train models that take the input and transform it into output to solve real-world problems. Examples of unsupervised learning tasks are clustering (where the model's output can be cluster labels), dimensionality reduction (where the model's output is an output feature vector with fewer features than the input feature vector), and outlier detection (where the model's output is a value indicating how the input differs from typical examples in the dataset). An example unsupervised learning algorithm is k-means.

[0113] Semi-supervised learning algorithms process datasets containing both labeled and unlabeled examples, where the number of unlabeled examples is typically much greater than the number of labeled examples. However, the goal of semi-supervised learning is to achieve the goals of supervised learning. Typically, a semi-supervised model includes a model trained to generate pseudo-labels for the unlabeled data, which is then combined with the labeled data to train a second classifier that utilizes a larger amount of overall training data to improve task performance.

[0114] Reinforcement learning algorithms use observations gathered by an agent from its interactions with the environment to take actions that maximize reward or minimize risk. Reinforcement learning is a continuous and iterative process in which the agent learns from its experience with the environment until it has explored, for example, the full range of possible states. An example type of reinforcement learning algorithm is adversarial networks. Reinforcement learning can be particularly beneficial when used to improve or attempt to optimize the behavior of models deployed in dynamically changing environments, such as wireless communication networks.

[0115] ML models can be deployed in one or more devices (e.g., network entities such as base stations and / or user equipment) to support various wired and / or wireless communication aspects of a communication system. For example, ML models can be trained to identify patterns and relationships in data corresponding to networks, devices, air interfaces, etc. ML models can improve operations associated with one or more aspects, such as transceiver circuitry control, frequency synchronization, timing synchronization, channel state estimation, channel equalization, channel state feedback, modulation, demodulation, device location, transceiver tuning, beamforming, signal decoding / decoding, network routing, load balancing, and power saving (to name just a few). AI-enhanced transceiver circuitry control may include, for example, filter tuning, transmit power control, gain control (including automatic gain control), phase control, power management, etc.

[0116] The aspects described herein can be used to describe technical solutions for performing certain tasks and various technical problems by applying specific types of ML models (such as ANNs). However, it should be understood that other types of ML models may be used in addition to or in place of ANNs. Therefore, unless explicitly stated otherwise, the topic of ML models is not necessarily intended to be limited to ANN solutions. Furthermore, it should be understood that, unless otherwise specifically stated, terms such as “AI model,” “ML model,” “AI / ML model,” and “trained ML model” are intended to be used interchangeably.

[0117] Figure 6 An example AI architecture 600 is illustrated and can be used for AI-enhanced wireless communication. As shown in the figure, architecture 600 includes multiple logical entities, such as a model training host 602, a model inference host 604, a data source 606, and an agent 608. The AI ​​architecture can be used in any of the various use cases for wireless communication, such as those listed above.

[0118] The model inference host 604 in architecture 600 is configured to run an ML model based on inference data 612 provided by data source 606. The model inference host 604 may produce an output 614 (e.g., a prediction or inference, such as discrete or continuous values) based on the inference data 612 and then provide it as input to agent 608.

[0119] Agent 608 can be a component or entity of a wireless communication system, including, for example, a radio access network (RAN), a wireless local area network, a device-to-device (D2D) communication system, etc. For example, Agent 608 can be a user equipment (UE), a base station, or any of its decomposed network entities (including centralized units (CUs), distributed units (DUs), and / or radio units (RUs)), an access point, a radio station, a RAN intelligent controller (RIC) in a cloud-based RAN, etc., as described herein. Figures 1 to 3 and Figure 5 As described. In addition, the type of agent 608 may also depend on the type of task performed by model inference host 604, the type of inference data 612 provided to model inference host 604, and / or the type of output 614 generated by model inference host 604.

[0120] For example, if the output 614 from the model inference host 604 is associated with beam management, then the agent 608 may be or include a UE, DU, or RU. As another example, if the output 614 from the model inference host 604 is associated with transmit / receive scheduling, then the agent 608 may be a CU or DU.

[0121] After agent 608 receives output 614 from model inference host 604, agent 608 may determine whether to take action based on the output. For example, if agent 608 is a DU or RU, and the output from model inference host 604 is associated with beam management, agent 608 may determine whether to change or modify the transmit and / or receive beams based on output 614. If agent 608 determines to take action based on output 614, agent 608 may instruct the action to at least one action subject 610. For example, if agent 608 determines to change or modify the transmit and / or receive beams used for communication between agent 608 and the subject (e.g., UE) of action 610, agent 608 may transmit a beam switching instruction to the subject (e.g., UE) of action 610. As another example, agent 608 may be a UE, and output 614 from model inference host 604 may be one or more predicted channel characteristics of one or more beams. For example, model inference host 604 may predict the channel characteristics of a beam set based on measurements of another beam set. Based on the predicted channel characteristics, agent 608 (such as UE) may transmit a request to action subject 610 (such as BS) to switch to a different beam for communication. In some cases, agent 608 and action subject 610 are the same entity.

[0122] Data source 606 can be configured to collect data that can be used as training data 616 for training an ML model or as inference data 612 for feeding ML model inference operations. Specifically, data source 606 can collect data from any of various entities (e.g., UE and / or BS), which may include the subject of operation 610, and provide the collected data to model training host 602 for ML model training. For example, after action subject 610 (e.g., UE) receives beam configuration from agent 608, action subject 610 can provide performance feedback associated with the beam configuration to data source 606, where the performance feedback can be used by model training host 602 to monitor and / or evaluate ML model performance, such as whether the output 614 provided to agent 608 is accurate. In some examples, if the output 614 provided to agent 608 is inaccurate (or its accuracy is below an accuracy threshold), model training host 602 can determine, for example, to modify or retrain the ML model used by model inference host 604 via ML model deployment / update.

[0123] In some respects, the model training host 602 may be deployed at the same or a different entity as the entity that deploys the model inference host 604, or together with that entity. For example, to offload model training processing that might affect the performance of the model inference host 604, the model training host 602 may be deployed at a model server, as further described herein. Furthermore, in some cases, training and / or inference may be distributed across devices in a decentralized or federated manner.

[0124] In some other respects, ML models are deployed at or on the UE to perform any of the various ML operations associated with wireless communication between the UE and network entities (e.g., BS). More specifically, the model infers the host (such as...) Figure 6 The model inference host (604) can be deployed at or on the UE to predict, infer, encode and / or decode information associated with wireless communication between the UE and network entities, such as encoding CSF and / or predicting CSI, beam management operations and / or device location.

[0125] Figure 7 An example AI architecture for a first wireless device 702 communicating with a second wireless device 704 is illustrated. The first wireless device 702 can be as described in this document. Figures 1 to 3 The UE described herein. Similarly, the second wireless device can be as described herein. Figures 1 to 3 The base station described. Note that the AI ​​architecture of the first wireless device 702 can be applied to the second wireless device 704.

[0126] The first wireless device 702 may be or may include a chip, a system-on-a-chip (SoC), a system-in-package (SiP), a chipset, a package, or a device, which includes one or more processors, processing blocks, or processing elements (collectively, “processor 710”) and one or more memory blocks or elements (collectively, “memory 720”).

[0127] For example, in transmit mode, processor 710 can transform information (e.g., packets or data blocks) into modulation symbols. As digital baseband signals (e.g., digital in-phase (I) and / or quadrature (Q) baseband signals representing corresponding symbols), processor 710 can output the modulation symbols to transceiver 740. Processor 710 can be coupled to transceiver 740 for transmitting and / or receiving signals via one or more antennas 746. In this example, transceiver 740 includes radio frequency (RF) circuitry 742, which can be coupled to antenna 746 via interface 744. For example, interface 744 can include switches, duplexers, double-signalers, multiplexers, etc. RF circuitry 742 can, for example, use a digital-to-analog converter to convert digital signals into analog baseband signals. RF circuitry 742 can include any of various circuits, including, for example, baseband filters, mixers, frequency synthesizers, power amplifiers, and / or low-noise amplifiers. In some cases, RF circuitry 742 can up-convert baseband signals to one or more carrier frequencies for transmission. Antenna 746 can transmit RF signals, which can be received at the second wireless device 704.

[0128] In receive mode, RF signals received via antenna 746 (e.g., from a second wireless device 704) can be amplified and converted to a baseband frequency (e.g., down-conversion). The received baseband signal can be filtered and converted into digital I or Q signals for digital signal processing. Modem 710 can receive the digital I or Q signals and further process the digital signals, for example, by demodulating them.

[0129] One or more ML models 730 may be stored in memory 720 and accessible by processor 710. In some cases, different ML models 730 with different characteristics may be stored in memory 720, and a particular ML model 730 may be selected based on its characteristics and / or application, as well as the characteristics and / or conditions of the first wireless device 702 (e.g., power state, mobility state, battery reserve, temperature, etc.). For example, ML models 730 may have different inference data and output pairings (e.g., different types of inference data produce different types of outputs), and predictions (e.g., Figure 6The output of 614 is associated with different accuracy levels (e.g., 80%, 90%, or 95% accuracy), different time delays associated with generating predictions (e.g., processing time less than 10ms, 100ms, or 1 second), different ML model sizes (e.g., file size), different coefficients or weights, etc.

[0130] Processor 710 can use ML model 730 based on input data (e.g., Figure 6 The inferred data 612) is used to generate output data (e.g., Figure 6 (614), for example, as in this article for Figure 6 The inference host 604 is described above. The ML model 730 can be used to perform any of a variety of AI augmentation tasks, such as those described above.

[0131] For example, the ML model 730 can take measurements of a reference signal (e.g., corresponding to a wide beam) as input to predict channel characteristics associated with different reference signals (e.g., corresponding to a narrow beam within a wide beam, another wide beam, a narrow beam outside a wide beam, etc.). Input data may include measurements of one or more reference signals or pilot signals, such as Channel Quality Indicator (CQI), Signal-to-Noise Ratio (SNR), Signal-to-Interference Plus Noise Ratio (SINR), Signal-to-Noise Plus Distortion Ratio (SNDR), Received Signal Strength Indicator (RSSI), Reference Signal Received Power (RSRP), Reference Signal Received Quality (RSRQ), and / or Block Error Rate (BLER). Output data may include one or more predicted measurements (or characteristics) of one or more reference signals or pilot signals, which may differ from the reference signals or pilot signals associated with the input data. In some respects, one or more reference signals or pilot signals for which one or more measurements are predicted may be considered "virtual resources" because they are not actually transmitted, but the measurements are predicted as if they were transmitted. In some respects, one or more reference signals or pilot signals for which one or more measurements are predicted may actually be transmitted, but not actually measured by the first wireless device 702. In some respects, the ML model 730 may be used to predict certain operations of the RF circuit 742, such as filter tuning, envelope tracking, automatic gain control, mixing, frequency synthesis, digital predistortion, etc. It should be noted that other input and / or output data may be used in addition to or instead of the examples described herein.

[0132] In some respects, model server 750 can perform any of a variety of ML model lifecycle management (LCM) tasks for first wireless device 702 and / or second wireless device 704. Model server 750 can operate as model training host 602 and use training data to update ML model 730. In some cases, model server 750 can operate as data source 606 to collect and host training data, inference data, and / or performance feedback associated with ML model 730. In some respects, model server 750 can host various types and / or versions of ML model 730 for download by first wireless device 702 and / or second wireless device 704.

[0133] In some cases, model server 750 can monitor and evaluate the performance of ML model 730 to trigger one or more LCM tasks. For example, model server 750 can determine whether to activate or deactivate the use of a specific ML model at first wireless device 702 and / or second wireless device 704, and model server 750 can provide such instructions to the respective first wireless device 702 and / or second wireless device 704. In some cases, model server 750 can determine whether to switch to a different ML model 750 used at first wireless device 702 and / or second wireless device 704, and model server 750 can provide such instructions to the respective first wireless device 702 and / or second wireless device 704. In other examples, model server 750 can also act as a central server for decentralized machine learning tasks such as federated learning.

[0134] Figure 8 This is an exemplary block diagram of an example Artificial Neural Network (ANN) 800.

[0135] ANN 800 can receive input data 806, which may include one or more bit data 802, preprocessed data (optionally) output from preprocessor 804, or some combination thereof. Here, data 802 may include training data, validation data, application-related data, etc., for example, depending on the development and / or deployment phase of ANN 800. In some other implementations, preprocessor 804 may be included within ANN 800. Preprocessor 804 may, for example, process all or part of data 802, which may result in some data in data 802 being changed, replaced, deleted, etc. In some implementations, preprocessor 804 may add additional data to data 802.

[0136] ANN 800 includes at least one first layer 808 of artificial neurons 810 to process input data 806 and provide the resulting first layer output data to at least a portion of at least one second layer 814 via edge 812. The second layer 814 processes the data received via edge 812 and provides second layer output data to at least a portion of at least one third layer 818 via edge 816. The third layer 818 processes the data received via edge 816 and provides third layer output data to at least a portion of a final layer 822 comprising one or more neurons via edge 820 to provide output data 824. All or part of the output data 824 may be further processed in some way by (optionally) a post-processor 826. Thus, in some examples, ANN 800 may provide output data 828 based on output data 824, post-processed data output from post-processor 826, or some combination thereof. In some other embodiments, post-processor 826 may be included within ANN 800. Postprocessor 826 may process all or part of the output data 824, which may result in output data 828 being at least partially different from output data 824, for example, due to data being altered, replaced, deleted, etc. In some implementations, postprocessor 826 may be configured to add additional data to output data 824. In this example, the second layer 814 and the third layer 818 represent intermediate or hidden layers that can be arranged in a hierarchical or other similar structure. Although not explicitly shown, one or more additional intermediate layers may exist between the second layer 814 and the third layer 818.

[0137] The structure and training of the artificial neurons 810 in each layer can be customized according to the specific requirements of the application. Within a given layer of an ANN, some or all of the neurons can be configured to process the information provided to that layer and output corresponding transformation information from that layer. For example, the transformation information from the layer can represent a weighted sum of input information associated with or based on a nonlinear activation function or other activation functions used to “activate” the artificial neurons in the next layer. Artificial neurons in such layers can be activated by or in response to weights and biases that can be adjusted during the training process. The weights of various artificial neurons can act as parameters controlling the strength of connections between layers or artificial neurons, while the biases can act as parameters controlling the direction of connections between layers or artificial neurons. Activation functions can select or determine whether an artificial neuron sends its output to the next layer in response to the data received by the artificial neuron. Different activation functions can be used to model different types of nonlinear relationships. By introducing nonlinearity into the ML model, activation functions allow the ML model to “learn” complex patterns and relationships in the input data (e.g., Figure 6(606 in the example). Some non-exhaustive example activation functions include linear functions, binary step functions, sigmoid, tanh, ReLU and its variants, exponential linear unit (ELU), Swish, Softmax, etc.

[0138] Design tools (such as computer applications and programs) can be used to select the appropriate structure of the ANN 800, multiple layers, and multiple artificial neurons in each layer, as well as to select activation functions, loss functions, training procedures, etc. Once the initial model is designed, it can be trained using training data. Training data may include one or more datasets within which the ANN 800 can detect, identify, label, or discern patterns. Training data can represent various types of information, including written, visual, audio, environmental context, operational attributes, etc. During training, the parameters of the artificial neurons 810 can be changed, such as to minimize or otherwise reduce the loss function or cost function. The training process can be repeated multiple times to fine-tune the ANN 800 in each iteration.

[0139] Various ANN model architectures are available for consideration. For example, in a feedforward ANN architecture, each artificial neuron 810 in a layer receives information from the previous layer and similarly generates information for the next layer. In a convolutional ANN architecture, some layers can be organized as filters that extract features from data (e.g., training data and / or input data). In a recursive ANN architecture, some layers may have connections that allow data to be processed across time, such as for processing information with temporal structure, such as time series data prediction.

[0140] In the autoencoder ANN architecture, compact representations of data can be processed, and models can be trained to predict or potentially reconstruct the original data from a reduced set of features. The autoencoder ANN architecture can be used for tasks related to dimensionality reduction and data compression.

[0141] Generative adversarial network (GAN) architectures can include generator ANNs and discriminator ANNs trained to compete with each other. GANs are ANN architectures that can be used for tasks related to generating synthetic data or improving the performance of other models.

[0142] Transformer ANN structures utilize attention mechanisms that enable models to process input sequences in a parallel and efficient manner. Attention mechanisms allow the model to focus on different parts of the input sequence at different times. These attention mechanisms can be implemented using a series of layers called attention layers to compute, compute, determine, or select a weighted sum of input features based on the similarity between different elements of the input sequence. A transformer ANN structure may include a series of feedforward ANN layers that learn a non-linear relationship between the input and output sequences. The output of the transformer ANN structure can be obtained by applying a linear transformation to the output of the final attention layer. Transformer ANN structures are particularly useful for tasks involving sequence modeling or other similar processing.

[0143] Another example type of ANN structure is a model with one or more invertible layers. This type of model can be inverted or "unfolded" to reveal the input data used to generate the output of the layers.

[0144] Other examples of ANN model architectures include fully connected neural networks (FCNN) and long short-term memory (LSTM) networks.

[0145] ANN 800 or other ML models can be implemented in various types of processing circuits, as well as their memory and applicable instructions, for example, as described in this paper. Figure 6 and Figure 7 As described. For example, the model can be implemented using general-purpose hardware circuitry such as one or more central processing units (CPUs) and one or more graphics processing units (GPUs). One or more ML accelerators, such as tensor processing units (TPUs), embedded neural processing units (eNPUs), or other dedicated processors, and / or field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), etc., can also be used to develop ANN models. Various programming tools can be used to develop ANN models.

[0146] There are deployable ML models (such as...) Figure 8 Various model training techniques and processes used at some point before or after ANN 800.

[0147] As part of the model development process, information suitable for training data formats may be collected or otherwise created for training the ML model accordingly. For example, training data may be collected or otherwise created regarding information associated with received / transmitted signal strength, interference, and resource usage data, as well as any other relevant data that may be used to train the model to solve one or more problems or challenges in a communication system. In some instances, all or part of the training data may originate from one or more user equipment (UEs), one or more network entities, or one or more other devices in a wireless communication system. In some cases, all or part of the training data may be aggregated from multiple sources, such as one or more UEs, one or more network entities, the Internet, etc. For example, wireless network architectures (such as self-organizing networks (SONs) or mobile-driven test (MDT) networks) may be adapted to support data collection for ML model applications. In another example, training data may be generated or collected online, offline, or both online and offline by UEs, network entities, or other devices, and all or part of such training data may be transmitted or shared (real-time or near real-time), such as through store-and-forward functions. Offline training may refer to, for example, creating and using a static training dataset in a batch manner, while online training may refer to collecting and using training data in real-time or near real-time. For example, online or offline training can be used to train and / or fine-tune the ML model at a network device (e.g., a UE). For offline training, data collection and training can occur offline at the network side (e.g., at a base station or other network entity) or at the UE side. For online training, the training of the ML model at the UE side can be performed locally at the UE or by a server device (e.g., a server configured, controlled, and / or hosted by an entity that designs, manufactures, sells, and / or supplies UEs and / or UE hardware) based on data provided from the UE to the server device in real-time or near real-time.

[0148] In some cases, all or part of the training data may be shared within the wireless communication system, or even shared (or obtained from outside) outside the wireless communication system.

[0149] Once the ML model has been trained on the training data, its performance can be evaluated. In some scenarios, evaluation / validation tests can be conducted using a validation dataset, which may include data not present in the training data, to compare the model's performance against a baseline or other benchmark information. If the model's performance is deemed unsatisfactory, fine-tuning the model may be beneficial, for example, by changing its architecture, retraining it on the data, or using different optimization techniques. Once the model's performance is deemed satisfactory, it can be deployed accordingly. In some cases, the model may be updated in some way, such as by changing or replacing all or part of the model, or by undergoing further training, to name just a few examples.

[0150] As ANN (such as Figure 8 As part of the training process of an ANN (Artificial Neural Network 800), parameters affecting the function of artificial neurons and layers can be adjusted. For example, backpropagation can be used to train an ANN by iteratively adjusting the weights and / or biases of certain artificial neurons associated with the error between the model's predicted output and the expected output, which may be known or otherwise considered acceptable. Backpropagation may include forward passes, loss functions, backward passes, and parameter updates that can be performed during training iterations. The process can be repeated a certain number of times for each set of training data until the weights of the artificial neurons / layers are sufficiently adjusted.

[0151] Backpropagation, associated with the loss function, measures how well a model can predict the desired output for a given input. Optimization algorithms can be used during training to adjust weights and / or biases to reduce or minimize the loss function, thus improving model performance. Various optimization algorithms exist that can be used in conjunction with backpropagation or other training techniques. Some initial examples include gradient descent-based and stochastic gradient descent-based optimization algorithms. Stochastic gradient descent (or ascent) can be used to adjust weights / biases to minimize or otherwise reduce the loss function. Mini-batch gradient descent, a variant of gradient descent, involves updating weights / biases using mini-batch training data instead of the entire dataset. Momentum techniques can accelerate the optimization process by adding momentum terms to update or otherwise influence certain weights / biases.

[0152] Adaptive learning rate techniques adjust the learning rate of an optimization algorithm that is associated with one or more characteristics of the training data. Batch normalization techniques can be used to normalize the input to a model in order to stabilize the training process and potentially improve the model's performance.

[0153] The "drop-out" technique can be used to randomly discard some of the artificial neurons from the model during the training process, for example, to reduce overfitting and potentially improve the model's generalization.

[0154] The “early stopping” technique can be used to stop an ongoing training process early, such as when the performance of a model using a validation dataset begins to degrade.

[0155] Another example technique includes data augmentation, which generates additional training data by applying transformations to all or part of the training information.

[0156] Transfer learning techniques can be used, which involve using a pre-trained model as a starting point for training a new model. This can be useful when training data is limited or when there are multiple tasks that are related to each other.

[0157] Multi-task learning techniques can be used, which involve training a model to perform multiple tasks simultaneously to potentially improve the model's performance on one or more tasks. In some cases, hyperparameters can be input and applied during the training process.

[0158] Another example technique that can be useful for ML models is some form of "pruning." Pruning techniques, which can be performed during the training process or after the model has been trained, involve removing unnecessary (e.g., because they have no effect on the output), less necessary (e.g., because their effect on the output is negligible), or potentially redundant features from the model. In some cases, pruning techniques can reduce the complexity of the model or improve its efficiency without compromising its expected performance.

[0159] Pruning techniques can be particularly useful in the context of wireless communication, where available resources, such as power and bandwidth, may be limited. Some example pruning techniques include weight pruning, neuron pruning, layer pruning, structural pruning, and dynamic pruning. Pruning techniques can, for example, reduce the amount of data corresponding to a model that may need to be transmitted or stored.

[0160] Weight pruning techniques may involve removing weights from a model. Neuron pruning techniques may involve removing neurons from a model. Layer pruning techniques may involve removing layers from a model. Structural pruning techniques may involve removing connections between neurons in a model. Dynamic pruning techniques may involve adjusting the pruning strategy of a model to adapt to one or more characteristics of the data or environment. For example, in some wireless communication devices, dynamic pruning techniques may be more aggressive in pruning models used in low-power or low-bandwidth environments and less aggressive in pruning models used in high-power or high-bandwidth environments. In some aspects, pruning techniques may also be applied to training data, for example, to remove outliers. In some specific implementations, preprocessing techniques on all or part of the training dataset can improve model performance or facilitate faster model convergence. For example, training data may be preprocessed to modify or remove unnecessary, irrelevant, incorrect, or other identifiable data. Such preprocessing of training data may, for example, lead to a reduction in potential overfitting or otherwise improve the performance of the trained model.

[0161] One or more of the example training techniques presented above can be used as part of the training process. As shown above, some example training processes that can be used to train ML models include supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning techniques.

[0162] Decentralized, distributed, or shared learning, such as federated learning, enables training on data distributed across multiple devices or organizations without centralizing the data or training. Federated learning can be particularly useful in data-sensitive or privacy-constrained situations, or when centralizing data is impractical, inefficient, or costly. For example, in the context of wireless communications, federated learning can be used to improve performance by allowing ML models to be trained on data collected from a wide range of devices and environments. For instance, ML models can be trained on data collected from a large number of wireless devices in a network, such as distributed wireless communication nodes, smartphones, or Internet of Things (IoT) devices, to improve network performance and efficiency. Using federated learning, a user equipment (UE) or other device can receive a full or partial copy of the model and perform local training on this copy using locally available training data. Such a device can then provide updates about the locally trained model (e.g., trainable parameter gradients) to one or more other devices, such as network entities or servers, where updates from other similar devices (such as other UEs) can be aggregated and used to provide updates to shared models, etc. The joint learning process can be iteratively repeated until all or part of the model achieves a satisfactory performance level. Joint learning enables devices to protect the privacy and security of local data while supporting collaboration on training and updating all or part of a shared model.

[0163] In some implementations, one or more devices or services may support processes related to the use, maintenance, activation, and reporting of ML models. In some cases, all or part of a dataset or model may be shared across multiple devices, for example, to provide or otherwise enhance or improve processing. In some examples, signaling mechanisms may be used at various nodes in a wireless network to signal the ability to perform specific functions related to the ML model, support for a particular ML model, the ability to collect, create, and transmit training data, or other ML-related capabilities. For example, ML models in wireless communication systems may be used to support decisions related to wireless resource allocation or selection, wireless channel condition estimation, interference mitigation, beam management, positioning accuracy, energy saving, or modulation or decoding schemes. In some implementations, model deployment may occur jointly or individually at various network levels, such as central units (CUs), distributed units (DUs), radio units (RUs), etc.

[0164] Aspects related to core network management for AI used in UE deployment This disclosure provides a framework for managing AI deployed at the UE using network functions of the CN (such as network functions in a data analytics framework). The UE ML model can refer to an ML model that is (or can be) deployed at or on the UE. For example, the ML model is used to predict, infer, encode, and / or decode information associated with the communication link between the UE and the RAN, for example, for CSF (compression, decompression, and / or prediction), beam management, and / or device localization.

[0165] Figure 9 The description outlines the management of AI deployed at UE 904 (e.g., for...). Figure 7 An example AI management framework 900 (described as an ML model 730) is provided. In this example, framework 900 includes a model training function (MTF) 922 (e.g., NWDAF 522) and, for example, via a bus 920 (e.g., ... Figure 5 The data collection and model repository function (DCMRF) 924 (e.g., DCAF 524) communicates with the MTF 922 via the bus 520. In some cases, the frame 900 may include the NEF 918 (e.g., Figure 5 The NEF 918, application server 926, and / or network controller 950 are described herein. As described herein, the NEF 918 controls the accessibility of the MTF 922 and / or DCMRF 924 to network entities outside the CN, such as the application server 926 and / or network controller 950. The NEF 918 can enable certain network entities or functions to access the MTF 922 and / or DCMRF 924, and vice versa. For example, the application server 926 and / or network controller 950 can be allowed to access the MTF 922 and / or DCMRF 924 according to one or more network access policies maintained at the NEF 918.

[0166] In some aspects, UE 904 may include a UE application client 930 and / or a UE management client 932. In some cases, the UE application client 930 may communicate with the UE management client 932 via an application programming interface (API) 934, for example, to implement UE ML operations. As another example, the UE management client 932 may receive a training data collection request from a DCMRF 924. In response to the training data collection request, the UE management client 932 may transmit a data collection request to the UE application client 930. Based on the data collection request, the UE application client 930 provides training data to the UE management client 932 (e.g., ...). Figure 6 The training data 616), and the UE management client 932 forwards the training data to the DCMRF 924 (which can be used as ...). Figure 6The training data source is 606. The UE application client 930 can share the training data with the UE management client 932, which can transfer the training data to the DCMRF 924, as further described herein.

[0167] In some respects, the UE application client 930 may be or include a model inference host (e.g., model inference host 602) deployed at or on the UE 904, and the UE application client 930 may perform ML processing, such as predicting, inferring, encoding and / or decoding information associated with the wireless communication link between the UE 904 and the UE.

[0168] UE application client 930 may obtain and / or request UE ML models from UE management client 932, which can be used as a local UE ML model repository. UE management client 932 may obtain UE ML models from DCMRF 924, as further described herein. In some cases, UE management client 932 may be embedded in or integrated with UE application client 930, and in such cases, API 934 may not be used.

[0169] UE application client 930 can communicate with application server 926 via first interface 936, which may be or include a user (or data) plane tunnel or communication link (e.g., the R8 interface in the core network data analysis framework or a separate interface for UE ML operation and management). Services carried via first interface 936 can be transparently tunneled through a trusted domain without interacting with any control plane entity. Any of various LCM communications can be conveyed between UE application client 930 and application server 926 via first interface 936. For example, application server 926 can transmit LCM communications instructing UE 904 to download, activate, train, reconfigure, and / or deactivate ML models via first interface 936. In some cases, application server 926 can receive requests for downloading, activating, training, reconfiguring, and / or deactivating ML models from UE 904 via first interface 936.

[0170] The UE management client 932 can communicate with the DCMRF 924 via a second interface 936, which may be or include a user (or data) plane tunnel or communication link (e.g., the R2 interface in the core network data analysis framework or a separate interface for UE ML operation and management). The UE management client 932 collects training data at or within the UE 904, and transmits one or more training data reports to the DCMRF 924, for example, via the second interface 938. The training data reports may include training data (which may be sampled at the UE 904) and / or its indications. For example, the training data may include channel state information, beam management information, and / or device location information sampled, measured, and / or determined at the UE 904. In some aspects, the UE management client 932 may use training data collection and / or reporting services (e.g., based on the data collection configuration and / or reporting configuration obtained by the UE 904 from the application server 926 and / or the DCMRF 924). Ndcaf_DataReporting Services and / or network function services specific to UE model data collection and / or reporting, as further described herein.

[0171] In some respects, application server 926 and / or DCMRF 924 may (via corresponding interfaces 936, 938) configure UE 904 to collect training data and / or report training data to DCMRF 924. For example, DCMRF 924 may transmit to UE 904 a configuration indicating that training data is collected and / or that training data is reported to DCMRF 924. The configuration for collecting training data at UE may indicate what information to collect (e.g., sampling at UE) as training data and / or the sampling rate of the training data (e.g., the number of samples sampled per unit of time, such as seconds). For example, training data for AI-based positioning may include the location of the UE, such as the UE's longitude, latitude, and / or altitude / height. For example, UE application client 930 may sample the location of UE 904 at a sampling rate of 1 kHz (or 1,000 samples per second).

[0172] The configuration for reporting training data can indicate when to report training data (e.g., on a periodic, semi-persistent, and / or aperiodic basis) and where to transmit the training data (e.g., to DCMRF 924 and / or application server 926). Application server 926 and / or DCMRF 924 can be configured with a target for reporting training data to the core network. For example, the target for reporting training data can be or include the (fully qualified) domain name and / or IP address of DCMRF 924 (e.g., the address or domain name of DCAF). In some cases, application server 926 and / or DCMRF 924 can be configured with a target for obtaining or accessing the UE ML model. For example, the target for obtaining the UE ML model can be or include the domain name and / or IP address of DCMRF 924 (e.g., the address or domain name of MRF).

[0173] MTF 922 can be or includes application functions of the core network, such as those described in this article. Figure 5 As described. The MTF922 can perform UE ML model training, for example, as described in this paper. Figure 6 The model training host 602 is described herein. In some cases, the MTF 922 may perform UE ML model training for third parties or entities, such as entities that design, manufacture, sell, and / or supply communication equipment (e.g., UE 904) and / or communication hardware (e.g., modems, SoCs, and / or SiPs), as further described herein. The MTF 922 may perform UE ML model training using training data collected at or by the DCMRF 924. In some cases, the training data may include training data collected and / or sampled at one or more UEs. In some aspects, the MTF 922 may be hosted within a trusted domain 940 of the core network. The trusted domain may include a communication network for controlling general communication between network entities of the core network. The trusted domain 940 may be operated and / or controlled by a communication service provider (e.g., a wireless network operator). The MTF 922 may be controlled and / or monitored by a third party via an application server 926, for example, as permitted by the core network operator (e.g., the communication service provider). In some respects, NWDAF (e.g., NWDAF 522) that supports Model Training Logic Functions (MTLF) can perform MTF 922 for training UE ML models as described herein.

[0174] DCMRF 924 can be or includes application functions of the core network, such as those described in this article. Figure 5As described. In some respects, each of the data collection and model repository functions of DCMRF 924 can be supported by separate application functions. DCMRF 924 can perform training data collection between the UE 904 and the core network, for example, as described in this paper. Figure 6 The data source 606 is described. For example, DCMRF 924 can collect training data from one or more UE 904s. In some respects, DCMRF 924 can be used as a repository for UE ML models, for example, as described in this paper. Figure 6 The model training host 602 is described. The DCMRF 924 allows the UE 904 to access one or more UE ML models, for example, via a second interface 938. The DCMRF 924 can be hosted within a trusted domain 940 of the core network. In some respects, the DCAF (e.g., Figure 5 The DCAF 524 can execute the DCMRF 924 for UEML training data collection and / or UEML model storage / accessibility, as described herein.

[0175] Application server 926 allows third parties to perform and control certain UE ML operations, such as various LCM tasks. The training data collection and / or model training framework described herein enables third parties to participate in and / or control LCM tasks. Application server 926 can configure, control, and / or monitor operations at MTF 922 and / or DCMRF 924. For example, application server 926 can configure or instruct when to perform training data collection, UE model training, and / or UE model deployment at MTF 922 and / or DCMRF 924. Third parties may be associated with communication service providers and / or network operators of the RAN and / or core network. In some respects, application server 926 may be managed or controlled by third parties, such as entities that design, manufacture, sell, and / or supply UEs and / or UE hardware. In this example, application server 926 may be hosted outside trusted domain 940, for example, in untrusted domain 942, where the untrusted domain can be a communication network outside trusted domain 940. In some cases, application server 926 may be hosted within trusted domain 940.

[0176] In some respects, UE ML models can be specific to a particular type of UE. For example, a UE ML model can be specific to the UE vendor and / or the vendor of the modem used at the UE. This is due to factors such as different model training frameworks, different model structures, and / or parameters (e.g., Figure 8 Different UEs can use different ML models due to different hardware, different model processing capabilities (e.g., memory capacity, AI accelerators, etc.), etc.

[0177] For example, a first entity designing modems for certain UEs may provide ML models specifically trained for those UEs, while a second entity designing UE hardware may provide different ML models specifically trained for other UEs. In some respects, ML models may not be cross-compatible between different types of UEs (e.g., different UEs and / or modems). The training data collection and / or model training framework described herein enables entity-specific operations for training UE ML models and deploying UE ML models to specific types of UEs. For example, a first application server hosted by a first entity may monitor and control when the LCM tasks described herein for a first set of UEs (e.g., training data collection, UE ML model training, and / or UE ML deployment) are executed, and a second application server hosted by a second entity may monitor and control when the LCM tasks described herein for a second set of UEs are executed. The first set of UEs may employ modems designed by the first entity, and the second set of UEs may be designed by the second entity.

[0178] In some respects, the network controller 950 may allow an entity (e.g., a communications service provider and / or network operator) to monitor, evaluate, operate, and / or control the model training and / or deployment of the UE ML model, as described herein. For example, the network controller 950 may monitor the performance of the UE ML model deployed at the UE 904, and if this performance fails to meet one or more metrics or thresholds (e.g., a block error rate in communication between the UE 904 and the RAN), the network controller 950 may trigger the MTF 922 to perform model training of the corresponding UE ML model, which may be designed and / or managed by another entity. This other entity may be, or include, entities that design, manufacture, sell, and / or supply the UE and / or UE hardware.

[0179] Although this article describes the communication between the UE and the DCMRF used for data collection and UE model transmission Figure 9 and Figure 10 The examples depicted are for illustrative purposes, but aspects of this disclosure can also be applied to AI management frameworks in which application servers collect training data, perform UE ML model training, and / or transmit the trained UE ML model. For example, the application server may communicate with a UE application client to collect training data and transmit the trained UE ML model.

[0180] Example UE model training in a communication system Figure 10 The process flow 1000 for communication between the first network entity 1002a, the second network entity 1002b, and the UE 1004 in the system is described. In some cases, communication may also occur between the third network entities 1002c.

[0181] In some respects, the first network entity 1002a may be an example of one or more network entities configured to perform, support, and / or manage MTF, as described herein. Figure 9 As described herein, the second network entity 1002b can be an example of one or more network entities configured to perform, support, and / or host DCMRF (e.g., data collection functionality and / or model repository functionality), as described in this document for... Figure 9 The third network entity 1002c can be an example of a network controller 950, which can be configured to perform, support, and / or host operational, administrative, and / or maintenance functions associated with the RAN and / or core network, for example, as described herein. Figure 5 and Figure 9 As described. UE 1004 can be for Figure 1 and Figure 3 UE 104, which is described and depicted Figure 5 UE 504 and / or Figure 9 Example of UE 904. However, in other respects, UE 1004 may be another type of wireless communication device, and any of network entities 1002a to 1002c may be another type of network entity or network node, such as the network entity or network node described herein.

[0182] Optionally, at 1006, the third network entity 1002c can discover the first network entity 1002a and / or the MTF (e.g., MTF 922) hosted by the first network entity 1002a. For example, the third network entity 1002c can discover the first network entity 1002a via an NRF (e.g., NRF516), as described herein. Figure 5 As described.

[0183] In some cases, at point 1008, the first network entity 1002a may obtain instructions from the third network entity 1002c regarding training one or more UE ML models associated with one or more UEs. The first network entity 1002a may obtain instructions for training the UE ML model via messages for network function services (such as messages for NWDAF services and / or UE model training services). As shown in the figure, the message may be… Nnwadaf_Mlmodeltraining_request The first network entity 1002a can, for example, be accessed via " Nnwadaf_Mlmodeltraining_response The message is sent to the third network entity 1002c in response to the instruction to train the UE ML model.

[0184] Instructions for training the UE ML model can configure model training and / or data collection. For example, instructions for training the UE ML model may include one or more of the following: instructions for training data to be collected, instructions for one or more UEs from which some or all of the training data will be collected, or instructions for the ML model to be trained. For example, instructions for training the UE ML model may include a function identifier identifying one or more functions (e.g., CSF, beam management and / or positioning) used by the ML model and / or an analysis identifier identifying one or more ML models that the first network entity 1002a has training capabilities. In some aspects, the analysis identifier may identify the function of the UE ML model (e.g., CSF, beam management and / or positioning). Instructions for training the UE ML model may include the duration of training data collection, the sampling frequency for sampling training data at one or more UEs, the reporting frequency (or period) for reporting training data to the first network entity 1002a, the location from which training data will be collected (e.g., region, location filter, etc.), the total number of UEs from which training data will be collected, and / or one or more characteristics associated with the ML model. In some cases, the UE ML model can be location- or region-specific. Location may correspond to a location filter indicating the coverage area of ​​the RAN, such as one or more network entities (e.g., cell identifiers), one or more tracking areas, one or more PLMN identities, etc. Characteristics associated with the ML model may include parameters that are defined once the UE ML model is trained, such as the expected accuracy of the UE ML model, model parameters (e.g., coefficients and / or weights), model structure (e.g., neural network layers and / or connections), maximum file size of the ML model, expected latency of the UE ML model, etc.

[0185] At 1010, first network entity 1002a transmits a request to second network entity 1002b to collect training data for a UE ML model. First network entity 1002a may determine the configuration for collecting training data, and the request may include such a configuration. For example, the request may include any information described herein with respect to the instructions for training the UE ML model at 1008. In some aspects, the request to collect training data may include an event identifier associated with collecting training data. The event identifier may indicate any information described herein with respect to the instructions for training the UE ML model at 1008. In some aspects, the request to collect training data may include an event open subscription request associated with a data collection function.

[0186] At 1012, the second network entity 1002b transmits to the UE 1004 an instruction for the UE 1004 to report training data for the UE ML model to the second network entity 1002b (or any other network entity, such as application server 926). In some respects, the second network entity 1002b may communicate with the UE 1004 via the second interface 938. For example, the instruction to report training data may indicate the training data to be collected at the UE, the sampling frequency, and / or the reporting frequency.

[0187] At 1014, UE 1004 transmits one or more training data reports to the second network entity 1002b. In some cases, UE 1004 may transmit training data reports periodically, for example, for online training or batch training as described herein. In other cases, UE 1004 may transmit training data reports in response to a criterion or event, such as a request for training data and / or performance metrics associated with the UE ML model. UE 1004 may transmit training data reports via the second interface 938. It should be noted that the second network entity 1002b may collect training data from a single UE or multiple UEs.

[0188] At 1016, the second network entity 1002b transmits training data to the first network entity 1002a, which may be partially collected from UE 1004 at 1014. In some cases, the second network entity 1002b may transmit training data periodically, for example, for online training or batch training as described herein. In other cases, the second network entity 1002b may transmit training data in response to a criterion or event, such as a request for training data.

[0189] At point 1018, the first network entity 1002a performs model training on one or more UE ML models (such as ANN 800). In some cases, the first network entity 1002a may perform online training and / or batch training of the UE ML model. For example, the first network entity 1002a may adjust parameters that affect the functionality of artificial neurons and layers of an ANN 800 as described herein. It should be noted that other training techniques may be employed in addition to or in place of the ANN training operation.

[0190] In some cases, at 1020, the first network entity 1002a may notify the third network entity 1002c that training for at least one of the UE ML models is complete. The first network entity 1002a may indicate characteristics associated with the UE ML model. For example, characteristics may include the accuracy of the UE ML model, model parameters (e.g., coefficients and / or weights), model structure (e.g., neural network layers and / or connections), file size of the ML model, expected latency of the UE ML model, etc. For example, the first network entity 1002a may transmit messages for NWDAF services, such as as " Nnwadaf_Mlmodeltraining_notify ".

[0191] First network entity 1002a and / or third network entity 1002c may assign usage parameters to the UE ML model, such as the effective duration of the UE ML model (e.g., effective time or duration) and / or the effective range of the UE ML model (e.g., effective area). The effective time may indicate when the UE ML model expires or how long the UE ML model can be used at the UE, for example, 3 hours, 1 day, 14 days, or 28 days. Such a time may indicate when the expected updated UE ML model will be trained and / or deployed to the UE. The effective area may define the location where the UE ML model can be used, for example, because the UE ML model is specific to a particular area (e.g., trained using data collected in that area). The effective area may be indicated as a RAN coverage area, for example, in terms of one or more network entities (e.g., cell identifiers), one or more tracking areas, one or more PLMN identities, one or more mobile network codes (MNCs), one or more mobile country codes (MCCs), etc. The effective area may be indicated as a geographic area, for example, in terms of one or more longitudes, latitudes, altitudes, heights, and / or boundaries or perimeters defined by such coordinates.

[0192] The third network entity 1002c may assign an identifier to the trained UE ML model, and the third network entity 1002c may notify the first network entity 1002a of the identifier, and vice versa. In some cases, the third network entity 1002c may transmit an instruction to the first network entity 1002a to transfer the trained UE ML model to the UE 1004.

[0193] In some cases, at 1022, the first network entity 1002a may transmit an instruction to the second network entity 1002b to transfer the trained UE ML model to the UE 1004. The first network entity 1002a may transfer the UE ML model to the second network entity 1002b, or provide the second network entity 1002b with access to the UE ML model.

[0194] At 1024, the second network entity 1002b transmits a UE ML model and / or a configuration for reproducing the UE ML model at UE 1004 to UE 1004. For example, this configuration may indicate parameters for the UE ML model, including, for example, coefficients and / or weights of artificial neurons, the number of layers for the artificial neurons, etc. In some aspects, the second network entity 1002b may transmit to UE 1004 an instruction to download the UE ML model, for example, from the second network entity 1002b and / or another network entity (such as application server 926). In response to a successful download of the UE ML model, UE 1004 may transmit an acknowledgment to the second network entity 1002b regarding the successful receipt of the UE ML model at UE 1004.

[0195] At point 1026, the first network entity 1002a obtains an indication from the second network entity 1002b that the UE ML model has been successfully transmitted to UE 1004. In some respects, the UE ML model can be activated at UE 1004, for example, to predict, infer, encode, and / or decode information associated with the communication link between UE 1004 and the RAN, such as as described herein. Figures 6 to 8 As described.

[0196] The core network management technologies for AI in UE deployments described in this paper provide efficient, dynamic, and / or cost-effective solutions for managing ML models for UE deployments. For example, enhancing the CN's data analytics framework to support the management of ML models for UE deployments allows network operators to implement AI-based wireless communications without significant changes to the CN and / or radio access networks. Therefore, the data analytics framework for managing ML models for UE deployments avoids costly and time-consuming hardware and / or software upgrades to wireless communication systems. In some respects, the data analytics framework for managing ML models for UE deployments facilitates the assignment of LCM tasks to various entities, such as network operators and / or entities that design, manufacture, sell, and / or supply UEs and / or UE hardware.

[0197] Example operation of communication equipment Figure 11 A method 1100 for wireless communication by a device, such as an MTF, NWDAF, and / or MTLF, or a network entity configured to perform, support, and / or host an MTF, NWDAF, and / or MTLF, is shown, as for Figure 5 , Figure 9 and Figure 10 The subject of discussion.

[0198] Method 1100 begins at box 1105, where instructions for training the ML model for deployment at the UE are obtained. In some cases, the MTF may obtain instructions for training the ML model from another network entity, such as the network controller, as described in this paper. Figure 10 As described. In some cases, the MTF may determine to train the ML model, for example, in response to the performance of the ML model failing to meet one or more performance criteria or thresholds. In some cases, the MTF may train the ML model periodically (e.g., for batch training) and / or in response to the availability of new training data (e.g., for online training). In some aspects, box 1105 includes instructions to obtain the training ML model from a second network entity.

[0199] Method 1100 then proceeds to block 1110, where a request for training data associated with one or more UEs is transmitted to a first entity (e.g., DCAF and / or DCMRF). In some aspects, the MTF may communicate control plane services with the first network entity via a public service bus (such as bus 920). In some aspects, method 1100 includes communicating control plane services with the first network entity via a public service bus (e.g., public bus 520).

[0200] Method 1100 then proceeds to box 1115, where training data is obtained from the first network entity, for example, as described in this paper for... Figure 9 and Figure 10 As described. In some respects, training data includes information sampled at one or more UEs.

[0201] Method 1100 then proceeds to box 1120, wherein the ML model is trained at least in part based on training data, for example, as described herein with respect to 9 and 10. In some respects, box 1120 includes training the ML model to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or localization.

[0202] Method 1100 then proceeds to box 1125, where an indication of training completion of the ML model is transmitted to the second network entity, for example, as described in this paper for... Figure 10 As described. In some aspects, the indication of training completion of the ML model includes one or more of a function identifier that identifies the ML model or an analysis identifier associated with the ML model. In some aspects, box 1125 includes transmitting the indication of training completion of the ML model via a network function service message on a public service bus.

[0203] Method 1100 then proceeds to block 1130, wherein the trained ML model is transmitted to the first network entity and / or an instruction is given to transmit the trained ML model to the UE.

[0204] In some aspects, instructions for training an ML model include one or more of the following: instructions for training data to be collected, instructions for one or more UEs from which some or all of the training data will be collected, or instructions for the ML model to be trained. In some aspects, instructions for training an ML model include one or more of the following: a function identifier that identifies one or more functions used by the ML model; an analysis identifier that identifies one or more ML models that the device has training capabilities for; the duration of training data collection; the sampling frequency for sampling training data at one or more UEs; the reporting frequency for reporting training data to the device; the location from which training data will be collected; the total number of one or more UEs from which training data will be collected; or one or more characteristics associated with the ML model.

[0205] In some aspects, a request for training data includes one or more of the following: an instruction for training data to be collected, an instruction for one or more UEs from which some or all of the training data is to be collected, or an instruction for an ML model to be trained. In some aspects, a request for training data includes one or more of the following: an event identifier associated with collecting training data; a function identifier associated with the ML model; the duration of collecting training data; a sampling frequency for sampling training data at one or more UEs; a reporting frequency for reporting training data to the device; the location from which training data is to be collected; the total number of one or more UEs from which training data is to be collected; or one or more features associated with the ML model. In some aspects, a request for training data includes an event open subscription request.

[0206] In some aspects, method 1100 also includes obtaining an identifier for the ML model and an instruction to transmit the trained ML model to the UE from a second network entity, wherein block 1130 includes transmitting the trained ML model to a first network entity in response to obtaining the instruction to transmit the trained ML model to the UE.

[0207] In some respects, the apparatus is configured to perform NWDAF; the first network entity is configured to perform DCAF; and the second network entity includes a network controller.

[0208] In some respects, method 1100 or any aspect thereof may be made by means of a device (such as...) Figure 14 The communication device 1400 performs the method, which includes various components operable to, configured to, or adapted to perform the method 1100. The communication device 1400 is described in more detail below.

[0209] It should be noted that Figure 11This is merely one example of a method, and other methods that include fewer, additional, or alternative operations may also be consistent with this disclosure.

[0210] Figure 12 A method 1200 for wireless communication by a device, such as DCAF and / or DCMR, or a network entity configured to perform, support, and / or host DCAF and / or DCMRF, is shown, as for Figure 5 , Figure 9 and Figure 10 The subject of discussion.

[0211] Method 1200 begins at box 1205, wherein a first request for training data associated with one or more UEs is obtained from a first network entity (e.g., MTF and / or NWDAF).

[0212] Method 1200 then proceeds to block 1210, where a second request for training data is transmitted to one or more UEs. In some respects, the DCMRF transmits the second request in response to receiving the first request.

[0213] Method 1200 then proceeds to box 1215, where training data is obtained from one or more UEs, for example, as described in this paper for... Figure 9 and Figure 10 As described. In some respects, training data includes information sampled at one or more UEs.

[0214] Method 1200 then proceeds to box 1220, where training data is transmitted to the first network entity.

[0215] Method 1200 then proceeds to block 1225, wherein an ML model trained at least in part based on training data and an instruction to transmit the ML model to at least one UE are obtained from a first network entity. In some respects, the ML model is trained to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or positioning.

[0216] Method 1200 then proceeds to block 1230, where an ML model is transmitted to at least one UE, for example, as described herein for... Figure 9 and Figure 10 As described.

[0217] In some aspects, a first request for training data includes one or more of the following: an instruction for training data to be collected, an instruction for one or more UEs from which some or all of the training data is to be collected, or an instruction for an ML model to be trained. In some aspects, a first request for training data includes one or more of the following: an event identifier associated with collecting training data; a function identifier associated with the ML model; the duration of collecting training data; a sampling frequency for sampling training data at one or more UEs; a reporting frequency for reporting training data to the device; the location from which training data is to be collected; or the total number of one or more UEs from which training data is to be collected.

[0218] In some aspects, a second request for training data includes one or more of the following: an instruction for training data to be collected or an instruction for an ML model to be trained. In some aspects, a second request for training data includes one or more of the following: an event identifier associated with collecting training data; a function identifier associated with an ML model; the duration of collecting training data; a sampling frequency for sampling training data at one or more UEs; a reporting frequency for reporting training data to a device; or the location from which training data is to be collected.

[0219] In some respects, the device is configured to perform, support, and / or manage DCAF and / or DCMRF; and the first network entity is configured to perform, support, and / or manage MTF, NWDAF, and / or MTLF.

[0220] In some respects, method 1200 also includes via a public service bus (such as...) Figure 9 The common bus 920 communicates control plane services with the first network entity.

[0221] In some respects, method 1200 also includes communicating with one or more UEs via a user plane communication link (or user / data plane tunnel).

[0222] In some respects, method 1200 or any aspect thereof may be made possible by means of a device (such as...) Figure 15 The communication device 1500 performs the method, which includes various components operable to, configured to, or adapted to perform the method 1200. The communication device 1500 is described in more detail below.

[0223] It should be noted that Figure 12 This is merely one example of a method, and other methods that include fewer, additional, or alternative operations may also be consistent with this disclosure.

[0224] Figure 13 It shows a device (such as) Figure 1 and Figure 3Method 1300 for wireless communication of UE 104.

[0225] Method 1300 begins at block 1305, wherein a request for training data associated with the device is obtained from a first network entity (e.g., DCAF and / or DCMRF). In some aspects, the device may communicate with the first network entity via a data plane tunnel, as described herein. Figure 9 As described. In some aspects, method 1300 also includes communicating with a first network entity via a user plane communication link.

[0226] Method 1300 then proceeds to block 1310, where training data is transmitted to the first network entity. In some respects, the training data includes information sampled at the device.

[0227] Method 1300 then proceeds to box 1315, where an ML model trained at least in part based on the training data is obtained from the first network entity.

[0228] Method 1300 then proceeds to block 1320, wherein communication with the second network entity is based at least in part on an ML model. In some respects, block 1320 includes using the ML model to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or localization.

[0229] In some aspects, a request for training data includes one or more of the following: instructions for training data to be collected or instructions for an ML model to be trained. In some aspects, a request for training data includes one or more of the following: an event identifier associated with the collection of training data; a function identifier associated with the ML model; the duration of the training data collection; a sampling frequency for sampling the training data at the device; a reporting frequency for reporting the training data to a first network entity; or the location from which the training data is to be collected.

[0230] In some respects, the first network entity is configured to perform DCAF; and the first network entity is configured to perform NWDAF.

[0231] In some respects, box 1305 is included in the first client (e.g., Figure 9 The method 1300 further includes obtaining a request for training data from a UE management client 932; the method 1300 also includes transmitting the request for training data to a second client (e.g., UE management client 932). Figure 9 (UE application client 930); method 1300 further includes transmitting training data from a second client to a first client; and block 1310 includes transmitting training data from the first client to a first network entity.

[0232] In some respects, method 1300 also includes a first client (e.g., Figure 9 The method 1300 also includes obtaining instructions to download the ML model from a second client (e.g., the UE management client 932). Figure 9 The UE application client 930) transmits an instruction to download the ML model, wherein box 1315 includes downloading the ML model to the second client at least in part based on the instruction to download the ML model.

[0233] In some respects, method 1300 also includes a first client (e.g., Figure 9 The method 1300 obtains an instruction to download the ML model from the first network entity to the first client at least in part based on the instruction to download the ML model. In some aspects, method 1300 also includes transferring the ML model from the first client to a second client (e.g., ...). Figure 9 UE application client 930).

[0234] In some respects, method 1300 or any aspect thereof may be made by means of a device (such as...) Figure 16 The communication device 1600 performs the method, which includes various components operable to, configured to, or adapted to perform the method 1300. The communication device 1600 is described in more detail below.

[0235] It should be noted that Figure 13 This is merely one example of a method, and other methods that include fewer, additional, or alternative operations may also be consistent with this disclosure.

[0236] Example communication device Figure 14 Various aspects of the example communication device 1400 are depicted. In some aspects, the communication device 1400 is a network entity, such as MTF, NWDAF and / or MTLF, or a network entity configured to perform, support and / or manage MTF, NWDAF and / or MTLF, as for Figure 5 , Figure 9 and Figure 10 The subject of discussion.

[0237] Communication device 1400 includes a processing system 1405 coupled to a transceiver 1465 (e.g., a transmitter and / or receiver) and / or a network interface 1475. Transceiver 1465 is configured to transmit and receive signals for communication device 1400 via antenna 1470, such as various signals as described herein. Network interface 1475 is configured to transmit via a communication link (such as those described herein for...) Figure 2The described backhaul link, midhaul link, and / or fronthaul link acquire and transmit signals for the communication device 1400. The processing system 1405 can be configured to perform the processing functions of the communication device 1400, including processing signals received by the communication device 1400 and / or to be transmitted by the communication device.

[0238] Processing system 1405 includes one or more processors 1410. In various aspects, the one or more processors 1410 can represent, for example, for... Figure 3 The described receiver processor 338, transmitter processor 320, TX MIMO processor 330, and / or controller / processor 340 are one or more of these. One or more processors 1410 are coupled to a computer-readable medium / memory 1435 via a bus 1460. In some aspects, the computer-readable medium / memory 1435 is configured to store instructions (e.g., computer-executable code) that, when executed by one or more processors 1410, enable one or more processors 1410 to execute and perform the desired actions. Figure 11 The described method 1100 or any aspect thereof, including regarding Figure 11 Any additional operations described. Note that references to the processor of the communication device 1400 performing the function may include one or more processors of the communication device 1400, such as those performing the function in a distributed manner.

[0239] In the depicted example, computer-readable medium / memory 1435 stores code 1440 for acquisition, code 1445 for transmission, code 1450 for training, and code 1455 for communication. Processing of codes 1440 to 1455 enables communication device 1400 to execute and perform actions for... Figure 11 The method 1100 described or any aspect thereof.

[0240] One or more processors 1410 include circuitry configured to implement (e.g., execute) code stored in computer-readable medium / memory 1435, including acquisition circuitry 1415, transmission circuitry 1420, training circuitry 1425, and communication circuitry 1430. Processing using circuitry 1415 to 1430 enables communication device 1400 to perform and allow the communication device to perform [specific actions / requirements]. Figure 11 The method 1100 described or any aspect thereof.

[0241] More generally, components used for communication, sending, transmitting, or outputting for transmission may include Figure 3The BS102 illustrated includes a transceiver 332, an antenna 334, a transmit processor 320, a TX MIMO processor 330, and / or a controller / processor 340. Figure 14 The transceiver 1465 and / or antenna 1470 of the communication device 1400 in the middle. Figure 14 One or more processors 1410 of the communication device 1400. Components for transmitting, receiving, or acquiring may include... Figure 3 The BS 102 illustrated includes transceiver 332, antenna 334, receiver processor 338, and / or controller / processor 340. Figure 14 The transceiver 1465 and / or antenna 1470 of the communication device 1400 in the middle. Figure 14 One or more processors 1410 of the communication device 1400 in the middle.

[0242] Figure 15 Various aspects of the example communication device 1500 are depicted. In some aspects, the communication device 1500 is a network entity, such as DCAF and / or DCMRF, or a network entity configured to perform, support, and / or host DCAF and / or DCMR, as for Figure 5 , Figure 9 and Figure 10 The subject of discussion.

[0243] Communication device 1500 includes a processing system 1505 coupled to a transceiver 1555 (e.g., a transmitter and / or receiver) and / or a network interface 1565. Transceiver 1555 is configured to transmit and receive signals for communication device 1500 via antenna 1560, such as various signals as described herein. Network interface 1565 is configured to transmit and receive signals for communication device 1500 via a communication link (such as those described herein). Figure 2 The described backhaul link, midhaul link, and / or fronthaul link acquire and transmit signals for communication device 1500. Processing system 1505 can be configured to perform processing functions of communication device 1500, including processing signals received by and / or to be transmitted by communication device 1500.

[0244] Processing system 1505 includes one or more processors 1510. In various aspects, the one or more processors 1510 can represent, as for... Figure 3The described receiver processor 338, transmitter processor 320, TX MIMO processor 330, and / or controller / processor 340 are one or more of these. One or more processors 1510 are coupled to a computer-readable medium / memory 1530 via a bus 1550. In some aspects, the computer-readable medium / memory 1530 is configured to store instructions (e.g., computer-executable code) that, when executed by one or more processors 1510, enable one or more processors 1510 to execute and perform the desired actions. Figure 12 The described method 1200 or any aspect thereof, including regarding Figure 12 Any additional operations described. Note that references to the processor of the communication device 1500 performing the function may include one or more processors of the communication device 1500, such as those performing the function in a distributed manner.

[0245] In the depicted example, computer-readable medium / memory 1530 stores code 1535 for acquisition, code 1540 for transmission, and code 1545 for communication. Processing of codes 1535 to 1545 enables communication device 1500 to execute and perform actions related to... Figure 12 The method 1200 described or any aspect thereof.

[0246] One or more processors 1510 include circuitry configured to implement (e.g., execute) code stored in computer-readable medium / memory 1530, the circuitry including circuitry 1515 for acquisition, circuitry 1520 for transmission, and circuitry 1525 for communication. Processing using circuitry 1515 to 1525 enables communication device 1500 to perform and allow the communication device to perform [specific actions / requirements]. Figure 12 The method 1200 described or any aspect thereof.

[0247] More generally, components used for communication, sending, transmitting, or outputting for transmission may include Figure 3 The BS102 illustrated includes a transceiver 332, an antenna 334, a transmit processor 320, a TX MIMO processor 330, and / or a controller / processor 340. Figure 15 The transceiver 1555 and / or antenna 1560 of the communication equipment 1500 in the middle. Figure 15 One or more processors 1510 of the communication device 1500. Components for transmitting, receiving, or acquiring may include... Figure 3 The BS 102 illustrated includes transceiver 332, antenna 334, receiver processor 338, and / or controller / processor 340. Figure 15The transceiver 1555 and / or antenna 1560 of the communication equipment 1500 in the middle. Figure 15 One or more processors 1510 of the communication device 1500 in the middle.

[0248] Figure 16 Various aspects of the example communication device 1600 are described. In some aspects, the communication device 1600 is user equipment, as described above. Figure 1 and Figure 3 The UE 104 described.

[0249] Communication device 1600 includes a processing system 1605 coupled to a transceiver 1665 (e.g., a transmitter and / or receiver). Transceiver 1665 is configured to transmit and receive signals for communication device 1600 via antenna 1670, such as the various signals described herein. Processing system 1605 may be configured to perform processing functions of communication device 1600, including processing signals received by and / or to be transmitted by communication device 1600.

[0250] Processing system 1605 includes one or more processors 1610. In various aspects, the one or more processors 1610 can represent, for example, for... Figure 3 The described receiver processor 358, transmitter processor 364, TX MIMO processor 366, and / or controller / processor 380 are one or more of these. One or more processors 1610 are coupled to a computer-readable medium / memory 1635 via a bus 1660. In some aspects, the computer-readable medium / memory 1635 is configured to store instructions (e.g., computer-executable code) that, when executed by one or more processors 1610, enable one or more processors 1610 to execute and perform the desired actions. Figure 13 The described method 1300 or any aspect thereof, including regarding Figure 13 Any additional operations described. Note that references to processors performing the functions of communication device 1600 may include one or more processors, such as performing the functions of communication device 1600 in a distributed manner.

[0251] In the depicted example, computer-readable medium / memory 1635 stores code 1640 for obtaining, code 1645 for transmitting, code 1650 for conveying, and code 1655 for transmitting. Processing of codes 1640 to 1655 enables communication device 1600 to perform and allows the communication device to perform actions related to... Figure 13 The method described 1300 or any aspect thereof.

[0252] One or more processors 1610 include circuitry configured to implement (e.g., execute) code stored in computer-readable medium / memory 1635, including acquisition circuitry 1615, transmission circuitry 1620, communication circuitry 1625, and transfer circuitry 1630. Processing using circuitry 1615 to 1630 enables communication device 1600 to perform and allow the communication device to perform [specific actions / requirements]. Figure 13 The method described 1300 or any aspect thereof.

[0253] More generally, components used for communication, sending, transmitting, or outputting for transmission may include Figure 3 The UE104 illustrated includes a transceiver 354, an antenna 352, a transmit processor 364, a TX MIMO processor 366, and / or a controller / processor 380. Figure 16 The transceiver 1665 and / or antenna 1670 of the communication device 1600 in the middle. Figure 16 One or more processors 1610 of the communication device 1600. Components for communicating, receiving, or acquiring may include... Figure 3 The UE 104 illustrated includes a transceiver 354, an antenna 352, a receiver processor 358, and / or a controller / processor 380. Figure 16 The transceiver 1665 and / or antenna 1670 of the communication device 1600 in the middle. Figure 16 One or more processors 1610 of the communication device 1600 in the middle.

[0254] Example Terms Specific implementation examples are described in the following numbered clauses: Clause 1: A method for wireless communication by a device, the method comprising: obtaining an instruction to train an ML model for deployment at a UE; transmitting to a first network entity a request for training data associated with one or more UEs; obtaining the training data from the first network entity; training the ML model at least in part based on the training data; transmitting to a second network entity an instruction regarding the completion of training of the ML model; and transmitting to the first network entity the trained ML model and an instruction to transmit the trained ML model to the UE.

[0255] Clause 2: The method according to Clause 1, wherein training the ML model includes training the ML model to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or localization.

[0256] Clause 3: The method according to any one of Clauses 1 to 2, wherein the training data includes information sampled at the one or more UEs.

[0257] Clause 4: The method according to any one of Clauses 1 to 3, wherein obtaining the instruction for training the ML model includes obtaining the instruction for training the ML model from the second network entity.

[0258] Clause 5: The method according to Clause 4, wherein the instruction for training the ML model includes one or more of the following: an instruction for the training data to be collected, an instruction for the one or more UEs to collect some or all of the training data from there, or an instruction for the ML model to be trained.

[0259] Clause 6: The method according to Clause 4, wherein the instruction for training the ML model includes one or more of the following: a function identifier that identifies one or more functions used by the ML model; an analysis identifier that identifies one or more ML models that the device has training capabilities for; the duration of collecting the training data; a sampling frequency for sampling the training data at the one or more UEs; a reporting frequency for reporting the training data to the device; a location from which the training data is to be collected; the total number of the one or more UEs from which the training data is to be collected; or one or more characteristics associated with the ML model.

[0260] Clause 7: The method according to any one of Clauses 1 to 6, wherein the request for the training data includes one or more of the following: an instruction for the training data to be collected, an instruction for one or more UEs to collect some or all of the training data from there, or an instruction for the ML model to be trained.

[0261] Clause 8: The method according to any one of Clauses 1 to 7, wherein the request for the training data includes one or more of the following: an event identifier associated with collecting the training data; a function identifier associated with the ML model; the duration of collecting the training data; a sampling frequency for sampling the training data at the one or more UEs; a reporting frequency for reporting the training data to the device; the location from which the training data is to be collected; the total number of the one or more UEs from which the training data is to be collected; or one or more features associated with the ML model.

[0262] Clause 9: The method according to any one of Clauses 1 to 8, wherein the request for the training data includes an event open subscription request.

[0263] Clause 10: The method according to any one of Clauses 1 to 9, wherein the indication of the completion of training of the ML model includes one or more of a functional identifier identifying the ML model or an analysis identifier associated with the ML model.

[0264] Clause 11: The method according to any one of Clauses 1 to 10, the method further comprising: obtaining from the second network entity an identifier for the ML model and the instruction to transmit the trained ML model to the UE, wherein transmitting the trained ML model includes transmitting the trained ML model to the first network entity in response to obtaining the instruction to transmit the trained ML model to the UE.

[0265] Clause 12: The method according to any one of Clauses 1 to 11, wherein: the apparatus is configured to perform NWDAF; the first network entity is configured to perform DCAF; and the second network entity includes a network controller.

[0266] Clause 13: The method according to any one of Clauses 1 to 12 further includes: communicating control plane services to the first network entity via a public service bus.

[0267] Clause 14: The method according to any one of Clauses 1 to 13, wherein transmitting the indication of training completion of the ML model comprises transmitting the indication of training completion of the ML model via a network function service message on a public service bus.

[0268] Clause 15: A method for wireless communication by a device, the method comprising: obtaining from a first network entity a first request for training data associated with one or more UEs; transmitting to the one or more UEs a second request for the training data; obtaining from the one or more UEs the training data; transmitting the training data to the first network entity; obtaining from the first network entity a machine learning (ML) model trained at least in part based on the training data and an instruction to transmit the ML model to at least one UE; and transmitting the ML model to the at least one UE.

[0269] Clause 16: The method described in Clause 15, wherein the ML model is trained to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or localization.

[0270] Clause 17: The method according to any one of Clauses 15 to 16, wherein the training data includes information sampled at the one or more UEs.

[0271] Clause 18: The method according to any one of Clauses 15 to 17, wherein the first request for the training data includes one or more of the following: an instruction for the training data to be collected, an instruction for the one or more UEs to collect some or all of the training data from there, or an instruction for the ML model to be trained.

[0272] Clause 19: The method according to any one of Clauses 15 to 18, wherein the first request for the training data includes one or more of the following: an event identifier associated with collecting the training data; a function identifier associated with the ML model; the duration of collecting the training data; a sampling frequency for sampling the training data at the one or more UEs; a reporting frequency for reporting the training data to the device; the location from which the training data is to be collected; or the total number of the one or more UEs from which the training data is to be collected.

[0273] Clause 20: The method according to any one of Clauses 15 to 19, wherein the second request for the training data includes one or more of the following: an instruction for the training data to be collected or an instruction for the ML model to be trained.

[0274] Clause 21: The method according to any one of Clauses 15 to 20, wherein the second request for the training data includes one or more of the following: an event identifier associated with collecting the training data; a function identifier associated with the ML model; the duration of collecting the training data; a sampling frequency for sampling the training data at the one or more UEs; a reporting frequency for reporting the training data to the device; or the location from which the training data is to be collected.

[0275] Clause 22: The method according to any one of Clauses 15 to 21, wherein: the apparatus is configured to perform DCAF; and the first network entity is configured to perform NWDAF.

[0276] Clause 23: The method according to any one of Clauses 15 to 22 further includes: communicating control plane services to the first network entity via a public service bus.

[0277] Clause 24: The method according to any one of Clauses 15 to 23 further includes: communicating with the one or more UEs via a user plane communication link.

[0278] Clause 25: A method for wireless communication by a device, the method comprising: obtaining a request for training data associated with the device from a first network entity; transmitting the training data to the first network entity; obtaining an ML model trained at least in part based on the training data from the first network entity; and communicating with a second network entity at least in part based on the ML model.

[0279] Clause 26: The method according to Clause 25, wherein communicating with the second network entity includes using the ML model to predict one or more attributes associated with one or more of the following: channel state feedback, beamforming, or positioning.

[0280] Clause 27: The method according to any one of Clauses 25 to 26, wherein the training data includes information sampled at the device.

[0281] Clause 28: The method according to any one of Clauses 25 to 27, wherein the request for the training data includes one or more of the following: an instruction for the training data to be collected or an instruction for the ML model to be trained.

[0282] Clause 29: The method according to any one of Clauses 25 to 28, wherein the request for the training data includes one or more of the following: an event identifier associated with collecting the training data; a function identifier associated with the ML model; the duration of collecting the training data; a sampling frequency for sampling the training data at the device; a reporting frequency for reporting the training data to the first network entity; or the location from which the training data is to be collected.

[0283] Clause 30: The method according to any one of Clauses 25 to 29, wherein: the first network entity is configured to perform DCAF; and the first network entity is configured to perform NWDAF.

[0284] Clause 31: The method according to any one of Clauses 25 to 30 further includes: communicating with the first network entity via a user plane communication link.

[0285] Clause 32: The method according to any one of Clauses 25 to 31, wherein: obtaining the request includes obtaining the request for the training data at a first client; the method further includes transmitting the request for the training data to a second client; the method further includes transmitting the training data from the second client to the first client; and transmitting the training data includes transmitting the training data from the first client to the first network entity.

[0286] Clause 33: The method according to any one of Clauses 25 to 32, the method further comprising: obtaining an instruction to download the ML model at a first client; and transmitting the instruction to download the ML model to a second client, wherein obtaining the ML model includes downloading the ML model to the second client at least in part based on the instruction to download the ML model.

[0287] Clause 34: The method according to any one of Clauses 25 to 33, the method further comprising: obtaining an instruction at a first client to download the ML model, wherein obtaining the ML model includes downloading the ML model from the first network entity to the first client based at least in part on the instruction to download the ML model; and transferring the ML model from the first client to a second client.

[0288] Clause 35: One or more apparatuses comprising: one or more memories including executable instructions; and one or more processors configured to execute the executable instructions and cause the one or more apparatuses to perform the method according to any one of Clauses 1 to 34.

[0289] Clause 36: One or more apparatuses, said one or more apparatuses comprising components for performing the method according to any one of Clauses 1 to 34.

[0290] Clause 37: One or more non-transitory computer-readable media, the one or more non-transitory computer-readable media comprising executable instructions that, when executed by one or more processors of one or more devices, cause the one or more devices to perform the method according to any one of Clauses 1 to 34.

[0291] Clause 38: One or more computer program products embodied on one or more computer-readable storage media, the one or more computer-readable storage media including code for performing the method according to any one of Clauses 1 to 34.

[0292] Additional Notes The foregoing description is provided to enable any person skilled in the art to practice the various aspects described herein. The examples discussed herein do not limit the scope, applicability, or aspects set forth in the claims. Various modifications to these aspects will be apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects. For example, the function and arrangement of the elements discussed may be changed without departing from the scope of this disclosure. Various processes or components may be omitted, substituted, or added as appropriate in the various examples. For example, the described methods may be performed in a different order than described, and various actions may be added, omitted, or combined. Furthermore, features described in some examples may be combined in some other examples. For example, any number of aspects set forth herein may be used to implement an apparatus or practice. Moreover, the scope of this disclosure is intended to cover such apparatuses or methods practiced using other structures, functionalities, or structures and functionalities that complement or replace the various aspects of this disclosure set forth herein. It should be understood that any aspect of the disclosure herein may be embodied by one or more elements of the claims.

[0293] The various exemplary logic blocks, modules, and circuits described in this disclosure can be implemented or executed using a general-purpose processor, AI processor, digital signal processor (DSP), ASIC, field-programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic element, discrete hardware component, or any combination thereof designed to perform the functions described herein. While the general-purpose processor may be a microprocessor, in alternative embodiments, the processor may be any commercially available processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors working in conjunction with a DSP core, a system-on-a-chip (SoC), or any other such configuration.

[0294] As used in this article, the phrase “at least one of the items” refers to any combination of these items, including a single member. As an example, “at least one of a, b, or c” is intended to cover: a, b, c, ab, ac, bc, and abc, as well as any combination with multiple identical elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbb, cc, and ccc, or any other ordering of a, b, and c).

[0295] As used herein, the term "determine" encompasses a wide variety of actions. For example, "determine" can include calculation, operation, processing, deduction, investigation, lookup (e.g., searching in a table, database, or other data structure), assertion, and so on. Additionally, "determine" can include receiving (e.g., receiving information), accessing (e.g., accessing data in memory), and so on. Furthermore, "determine" can include parsing, selecting, picking, building, and so on.

[0296] As used herein, unless otherwise stated, “coupled to” and “coupled with” generally encompass both direct and indirect coupling (e.g., including intermediate aspects of coupling). For example, stating that a processor is coupled to memory allows for direct coupling or coupling via an intermediate aspect such as a bus.

[0297] The methods disclosed herein include one or more actions for implementing the methods. These actions may be interchanged without departing from the scope of the claims. In other words, unless a specific order of actions is specified, the order and / or use of a particular action may be modified without departing from the scope of the claims. Furthermore, the various operations of the methods described above may be performed by any suitable component capable of performing the corresponding function. This component may include various hardware and / or software components and / or modules, including but not limited to circuits, application-specific integrated circuits (ASICs), or processors.

[0298] The following claims are not intended to be limited to the aspects shown herein, but should be given the full scope consistent with the language of the claims. References to singular elements are not intended to mean “only one” (unless specifically stated as “only one”), but rather “one or more”. Unless otherwise specified, definite articles (e.g., “the” or “described”) subsequently used with an element (e.g., “processor”) are not intended to give that element a singular meaning (e.g., “only one”). For example, unless otherwise specified, references to elements (e.g., “processor”, “controller”, “memory”, “transceiver”, “antenna”, “the processor”, “the controller”, “the memory”, “the transceiver”, “the antenna”, etc.) should be understood to refer to one or more elements (e.g., “one or more processors”, “one or more controllers”, “one or more memories”, “a plurality of transceivers”, etc.). The terms “set” and “group” are intended to include one or more elements and may be used interchangeably with “one or more”. In the case of references to one or more elements performing a function (e.g., steps of a method), one element may perform all the functions, or more than one element may collectively perform those functions. When more than one element performs these functions together, each function does not need to be performed by every single element (e.g., different functions can be performed by different elements), and / or each function does not need to be performed by only one element as a whole (e.g., different elements can perform different sub-functions of a function). Similarly, when referring to one or more elements configured to cause another element (e.g., a device) to perform a function, one element may be configured to cause another element to perform all functions, or more than one element may be jointly configured to cause another element to perform these functions. Unless otherwise specifically stated, the term "some" refers to one or more. All structural and functional equivalents of the various aspects described throughout this disclosure that are currently or hereafter known to those skilled in the art are intended to be covered by the claims. Furthermore, nothing disclosed herein is intended to be offered to the public, whether or not such disclosure is explicitly recited in the claims.

Claims

1. An apparatus configured for wireless communication, the apparatus comprising: One or more memory units; and One or more processors, said one or more processors coupled to said one or more memories, said one or more processors being configured to cause the device to: Obtain instructions for training machine learning (ML) models for deployment at user equipment (UE); Send a request to the first network entity for training data associated with one or more UEs; The training data is obtained from the first network entity; The ML model is trained at least in part based on the training data; Send an indication to the second network entity that the training of the ML model is complete; as well as The trained ML model and an instruction to transmit the trained ML model to the UE are transmitted to the first network entity.

2. The apparatus of claim 1, wherein, in order to train the ML model, the one or more processors are configured to cause the apparatus to train the ML model to predict one or more attributes associated with one or more of the following: channel state feedback, beamforming, or localization.

3. The apparatus of claim 1, wherein the training data includes information sampled at the one or more UEs.

4. The apparatus of claim 1, wherein, in order to obtain the instruction for training the ML model, the one or more processors are configured to cause the apparatus to: obtain the instruction for training the ML model from the second network entity.

5. The apparatus of claim 4, wherein the instruction for training the ML model comprises one or more of the following: an instruction for the training data to be collected, an instruction for the one or more UEs to collect some or all of the training data from the training data, or an instruction for the ML model to be trained.

6. The apparatus of claim 1, wherein the request for the training data includes one or more of the following: an instruction for the training data to be collected, an instruction for one or more UEs to collect some or all of the training data from the training data, or an instruction for the ML model to be trained.

7. The apparatus of claim 1, wherein the request for the training data includes an event open subscription request.

8. The apparatus of claim 1, wherein the indication of training completion of the ML model includes one or more of a functional identifier identifying the ML model or an analysis identifier associated with the ML model.

9. The apparatus according to claim 1, wherein: The one or more processors are configured to further enable the device to obtain from the second network entity an identifier for the ML model and an instruction to transmit the trained ML model to the UE; and In order to transmit the trained ML model, the one or more processors are configured to cause the device to transmit the trained ML model to the first network entity in response to receiving the instruction to transmit the trained ML model to the UE.

10. The apparatus according to claim 1, wherein: The device is configured to perform network data analysis function (NWDAF); The first network entity is configured to perform a Data Collection Application Function (DCAF); and The second network entity includes a network controller.

11. The apparatus of claim 1, wherein the one or more processors are configured to further enable the apparatus to communicate control plane services with the first network entity via a public service bus.

12. The apparatus of claim 1, wherein, in order to transmit the indication of training completion of the ML model, the one or more processors are configured to further cause the apparatus to transmit the indication of training completion of the ML model via a network function service message on a public service bus.

13. An apparatus configured for wireless communication, the apparatus comprising: One or more memory units; and One or more processors, said one or more processors coupled to said one or more memories, said one or more processors being configured to cause the device to: Obtain a first request from a first network entity for training data associated with one or more user equipment (UEs); Transmit a second request for the training data to the one or more UEs; The training data is obtained from the one or more UEs; The training data is transmitted to the first network entity; Obtain from the first network entity a machine learning (ML) model trained at least in part based on the training data and an instruction to transmit the ML model to at least one UE; as well as The ML model is transmitted to the at least one UE.

14. The apparatus of claim 13, wherein the ML model is trained to predict one or more properties associated with one or more of the following: channel state feedback, beamforming, or localization.

15. The apparatus of claim 13, wherein the training data includes information sampled at the one or more UEs.

16. The apparatus of claim 13, wherein the first request for the training data comprises one or more of the following: an instruction for the training data to be collected, an instruction for the one or more UEs to collect some or all of the training data from there, or an instruction for the ML model to be trained.

17. The apparatus of claim 13, wherein the second request for the training data includes one or more of the following: an instruction for the training data to be collected or an instruction for the ML model to be trained.

18. The apparatus according to claim 13, wherein: The device is configured to perform a Data Collection Application Function (DCAF); and The first network entity is configured to perform network data analysis function (NWDAF).

19. The apparatus of claim 13, wherein the one or more processors are configured to enable the apparatus to communicate control plane services with the first network entity via a public service bus.

20. The apparatus of claim 13, wherein the one or more processors are configured to enable the apparatus to communicate with the one or more UEs via a user plane communication link.

21. An apparatus configured for wireless communication, the apparatus comprising: One or more memory units; and One or more processors, said one or more processors coupled to said one or more memories, said one or more processors being configured to cause the device to: A request for training data associated with the device is obtained from a first network entity; The training data is transmitted to the first network entity; Obtain a machine learning (ML) model trained at least in part based on the training data from the first network entity; and The second network entity communicates at least in part based on the ML model.

22. The apparatus of claim 21, wherein, in order to communicate with the second network entity, the one or more processors are configured to cause the apparatus to use the ML model to predict one or more attributes associated with one or more of the following: channel state feedback, beamforming, or positioning.

23. The apparatus of claim 21, wherein the training data includes information sampled at the apparatus.

24. The apparatus of claim 21, wherein the request for the training data includes one or more of the following: an instruction for the training data to be collected or an instruction for the ML model to be trained.

25. The apparatus according to claim 21, wherein: The first network entity is configured to perform a Data Collection Application Function (DCAF); and The first network entity is configured to perform network data analysis function (NWDAF).

26. The apparatus of claim 21, wherein the one or more processors are configured to enable the apparatus to communicate with the first network entity via a user plane communication link.

27. The apparatus according to claim 21, wherein: In order to obtain the request, the one or more processors are configured to cause the device to obtain the request for the training data at a first client; The one or more processors are configured to cause the device to transmit the request for the training data to a second client; The one or more processors are configured to cause the device to transmit the training data from the second client to the first client; and In order to transmit the training data, the one or more processors are configured to cause the device to transmit the training data from the first client to the first network entity.

28. The apparatus according to claim 21, wherein: The one or more processors are configured to enable the device to receive an instruction at a first client to download the ML model; The one or more processors are configured to cause the device to transmit the instruction to a second client to download the ML model; In order to obtain the ML model, the one or more processors are configured to cause the device to download the ML model to the second client, at least in part, based on the instruction to download the ML model.

29. The apparatus according to claim 21, wherein: The one or more processors are configured to enable the device to receive an instruction at a first client to download the ML model; To obtain the ML model, the one or more processors are configured to cause the device to download the ML model from the first network entity to the first client, at least in part, based on the instruction to download the ML model; and The one or more processors are configured to cause the device to transfer the ML model from the first client to the second client.

30. A method for wireless communication by a device, the method comprising: Obtain instructions for training machine learning (ML) models for deployment at user equipment (UE); Send a request to the first network entity for training data associated with one or more UEs; The training data is obtained from the first network entity; The ML model is trained at least in part based on the training data; Send an indication to the second network entity that the training of the ML model is complete; as well as The trained ML model and an instruction to transmit the trained ML model to the UE are transmitted to the first network entity.

Images