Enhanced collaboration between user devices and networks to facilitate machine learning

Abstract

The present disclosure describes a system, method, and device for collaboration between a user equipment (UE) and a network for machine learning. A Radio Access Network (RAN) Node B device may perform the following steps: sending an indication to the UE device that the Node B device supports machine learning; identifying a service registration received from the UE device indicating that the UE device requests machine learning support from the Node B device; sending a request to the UE device for information related to the UE device, the information being associated with at least one of hardware capabilities or machine learning capabilities of the UE device; identifying the information received from the UE based on the request for information; sending a machine learning configuration to the UE device for use by the UE device; and causing the UE device to send updates to a machine learning configuration for use by the UE device, the machine learning configuration being based on the information.

Description

Technical Field

[0001] Cross - Reference to Related Applications This application claims the benefit of PCT Provisional Application No. PCT / CN2021 / 091798, filed on May 2, 2021, the disclosure of which is incorporated by reference in its entirety as if fully set forth herein.

[0002] Technical Field The present disclosure generally relates to systems and methods for wireless communication, and more particularly, to the cooperation of networks and user equipment devices for machine learning in 5th generation (5G) and 6th generation (6G) communications.

Background Art

[0003] Wireless devices have become widely popular and increasingly use the wireless channel. The 3rd Generation Partnership Project (3GPP (registered trademark)) is developing one or more standards for wireless communication.

Brief Description of the Drawings

[0004] [Figure 1] Illustrative processes for the cooperation of a network and a user equipment device to facilitate machine learning on the user equipment according to some illustrative embodiments of the present disclosure are shown. [Figure 2] A flowchart of an illustrative process for the cooperation of a network and a user equipment device to facilitate machine learning on the user equipment according to one or more illustrative embodiments of the present disclosure is shown. [Figure 3] A network according to one or more illustrative embodiments of the present disclosure is shown. [Figure 4] A wireless network according to one or more illustrative embodiments of the present disclosure is schematically shown. [Figure 5] A block diagram showing components according to one or more illustrative embodiments of the present disclosure.

Modes for Carrying Out the Invention

[0005] The following description and drawings illustrate specific embodiments to enable those skilled in the art to implement them. Other embodiments may incorporate structural, logical, electrical, process, algorithmic, and other modifications. Some parts and features of certain embodiments may be included in or substituted for those of other embodiments. The embodiments described in the claims encompass all available equivalents of those claims.

[0006] Wireless devices can perform measurements defined by technical standards. For cellular telecommunications, standardization bodies such as the 3rd Generation Partnership Programme (3GPP) and the Open Radio Access Network (O-RAN) Alliance define communication technologies, including those for using machine learning with 5th generation (5G) network devices and user equipment (UE) devices.

[0007] In 3GPP RAN3, machine learning applications in self-organized networks (SON) and minimization of drive test (MDT) use cases are being studied as part of improvements to data acquisition for New Radio (NR) and Evolutionary Universal Mobile Communications System Terrestrial Radio Access (E-UTRA) New Radio Dual Connectivity (ENDC). Some machine learning applications in SON focus on network resource optimization and coordination in NG-RAN (next generation radio access network). However, it would be more beneficial for wireless networks if UEs could utilize machine learning results in collaboration with the network or on their own to optimize UE performance.

[0008] Distributed learning and federated learning can be used as frameworks to support collaborative machine learning capabilities between next-generation radio access networks (NG-RANs) and UEs. Examples of using federated learning to support machine learning between NG-RANs and UEs have been proposed. However, unlike machine learning (ML) applications at NG-RAN nodes where ML information / models can be transmitted through the transport network layer (TNL) interface, collaborative machine learning between the network and UEs requires coordination and collaboration through the air interface.

[0009] Therefore, it is important to support machine learning capabilities through air interface signal transmission, including model deployment, model information updates, and machine learning configuration.

[0010] This disclosure provides embodiments relating to collaboration between NG-RAN and UE to support federative / distributed learning, including signal transmission aspects, machine learning-based UE behavior, and configuration / reporting.

[0011] This disclosure provides embodiments / implementations that support machine learning configuration / reporting between NG-RAN and UE via Radio Resource Control (RRC) signaling. The embodiments include a new radio bearer and / or a new system information block (SIB) to carry new messages, including ML configurations, ML reports and information updates, and ML requests. In response to ML reports from the UE, the NG-RAN can also update the ML model according to the confidence level / model bias / variance of the ML results in the reports from the UE and provide it to different UEs. Considering the different roles played by the NG-RAN and UE in the ML framework, this disclosure describes how UE behavior is defined and how its corresponding information is carried through the air interface.

[0012] The ML communication embodiments discussed herein allow for the development of distributed and federative learning by enabling information exchange through an air interface (e.g., RRC signaling). Information exchange through an air interface can effectively facilitate coordination and collaboration between the network and the UE for network performance optimization and resource allocation.

[0013] In one or more embodiments, the NG-RAN (e.g., a gNB device) may generate and send an ML capability instruction to the UE to indicate that the NG-RAN supports ML (e.g., to facilitate ML operation in the UE). The UE may respond with an instruction of interest (e.g., a service registration) indicating which services the UE requests an ML model for use by the UE. The NG-RAN may respond by generating and sending a UE capability query to the UE to request the UE's ML and hardware capabilities, and the UE may respond by providing its ML and hardware capabilities to the NG-RAN. Based on the UE's ML and hardware capabilities, the NG-RAN may generate an ML configuration for the ML model of the service requested by the UE and send the ML model and ML configuration to the UE for use by the UE. Once the UE has implemented the ML model and produced results, the UE may send an ML report to the NG-RAN to report the forecast, results, and action space (e.g., actions the UE will or will request to be performed as a result of the ML results). NG-RAN may provide updated ML models and / or ML configurations, which may be based on a request from the UE or not.

[0014] In one or more embodiments, machine learning models and associated parameter updates (e.g., for federated learning) may have a lower priority compared to messages such as UE capability and measurement reports. A novel wireless bearer is proposed to be used to carry such information between the NG-RAN and the UE.

[0015] A Signaling Radio Bearer (SRB) is defined as a Radio Bearer (RB) used solely for transmitting RRC and NAS messages (e.g., control plane messages). More specifically, the following SRBs are defined: SRB0 is for RRC messages using the CCCH logical channel. SRB1 is for RRC messages (which may include piggybacked NAS messages), as well as NAS messages before the establishment of SRB2, all using the DCCH logical channel. SRB2 is for NAS and RRC messages containing logged measurement information, all using the DCCH logical channel. SRB2 has a lower priority than SRB1 and may be configured by the network after AS security activation. SRB3 is for specific RRC messages when the UE is in (NG)EN-DC or NR-DC, all using the DCCH logical channel. SRB4 is for RRC messages containing application layer measurement reporting information, all using the DCCH logical channel. SRB4 may only be configured after security activation. In one or more embodiments, a new SRB is introduced such that all RRC messages include machine learning model and model parameter information using the DCCH logical channel. SRB5 can only be configured after security activation.

[0016] In one or more embodiments, a new data radio bearer (e.g., a user data plane) may be used to carry machine learning models and parameter updates. The new data radio bearer (e.g., an MLRB) may be dedicated to machine learning and mapping to PDU sessions. Thus, compared to carrying via an SRB, machine learning models transmitted from a higher layer via an MLRB may be intercepted by the NG-RAN. Each ML model / service can be mapped to one dedicated PDU session, and all machine learning models can share the same PDU session containing different QoS flows carrying different ML models. Prioritization between different ML models or between ML models and normal traffic can be handled by QoS requirements or logical channel prioritization.

[0017] In one or more embodiments, three types of UE behavior should be considered under RAN-UE joint machine learning optimization: (1) RRC configuration per RAN configuration. In this mode, UE behavior is the same as the legacy mechanism. The difference is how NG-RAN determines resource allocation and configuration for a given UE. (2) Policy-guided. With policy guidance, the UE makes decisions on its own according to the policy. In this mode, according to the results from the machine learning model (inference output), the UE can determine its actions based on a policy configured by the network (policy guidance configured via RRC configuration). After an action has been taken, the UE may also be required to send a configuration update via RRC signaling. One exemplary use case is beam management. Assuming distributed learning is employed on the UE side, the UE predicts the best received beam index or beamforming matrix according to the inference results. The UE can also receive policy guidance from the network. The guidance policy may include a tolerance for how the UE can adjust its received beam. If the predicted result of the received beam matrix is ​​within the tolerance, the UE can adjust its received beam. (3) Guided by action. The UE sends an action space to the RAN and performs an action according to the feedback from the NG-RAN. In this mode, the UE can either take an action directly from a machine learning model (an action in reinforcement learning), or the UE can decide to take an action according to the prediction result. However, the UE cannot perform these actions on its own because network approval is required. In this case, the UE sends a request message to the NG-RAN indicating the action required on the UE's side. Upon receiving this request, the RAN can decide whether or not to approve the action according to its own resource state. For example, the UE is performing a positioning prediction, and the predicted position indicates that the UE is moving outside the coverage of cell A.Based on the geolocation of other NG-RANs or measurements of neighboring cells, a UE may request a handover to cell B via a handover request (if cell B is performing better than cell A). Once the RAN accepts the handover request, the NG-RAN sends an accepted response to the requesting UE and transfers the corresponding UE context to the target cell.

[0018] In one or more embodiments, before configuring a machine learning setup, the UE is required to send instructions to the NG-RAN indicating whether the NG-RAN can support machine learning. A new field, “ue-capabilityML-RequestFilter,” is introduced in the UECapabilityEnquiry message to request the UE radio access capability to support machine learning. Furthermore, a new field, “ue-capabilityML-Information,” in the UEInformationRequest message is also used to transmit the UE's machine learning capability to the network. This disclosure considers categories of UE capability, namely (1) hardware capability and (2) machine learning capability. Hardware capability is generally used to indicate whether the UE chip hardware can / desirs to support machine learning. Details also include: chip type, maximum battery capacity, current battery status of the UE, batching data size, etc. Regarding machine learning capability, it is used to indicate what types of machine learning models the UE can support (e.g., CNN, RNN, RL, classification, regression, etc.), and for each machine learning model, it includes: maximum model size, training capability (e.g., supported software libraries), inference capability (e.g., supported software libraries), etc. The machine learning capability field is used to indicate the UE's machine learning capability in NG-RAN. The UECapabilityEnquiry message is used to request UE radio access capability for NR and other RATs. Signaling radio bearer: SRB1. RLC-SAP: AM. Logical channel: DCCH. Direction: Network to UE.

[0019] In one or more embodiments, the UECapabilityEnquiry message may be as follows: --ASN1START --TAG-UECAPABILITYENQUIRY-START UECapabilityEnquiry-IEs::= SEQUENCE{ ue-CapabilityML-RequestFilter UE-CapabilityML-RequestFilter, nonCriticalExtension SEQUENCE{} OPTIONAL } -UE-CapabilityMLRequestFilter

[0020] The IE UE-CapabilityMLRequestFilter is used to request filtered UE capabilities. The filter is common to all capability containers being requested.

[0021] In one or more embodiments, the UE-CapabilityMLRequestFilter information element may be as follows: --ASN1START --TAG-UE-CAPABILITYMLREQUESTFILTER-START UE-CapabilityMLRequestFilter::= SEQUENCE { machinelearning-Request SEQUENCE { omitML ENUMERATED{true} OPTIONAL, -- Need N hardwareCapability ENUMERATED{true} OPTIONAL, --Need N modelCapability ENUMERATED{true} OPTIONAL -- Need N } OPTIONAL, --Need N } --TAG-UE-CAPABILITYMLREQUESTFILTER-STOP --ASN1STOP

[0022] Table 1 below provides a description of the UE-CapabilityMLRequestFilter fields. [Table 1]

[0023] In one or more embodiments, the IE UECapabilityInformation message is used to transmit the UE radio access capability requested by the network. Signaling radio bearer: SRB1. RLC-SAP: AM. Logical channel: DCCH. Direction: UE to network.

[0024] In one or more embodiments, the UECapabilityInformation message may be as follows: --ASN1START --TAG-UECAPABILITYINFORMATION-START UECapabilityInformation-IEs::= SEQUENCE{ ue-CapabilityML-Information UE-CapabilityML-Information OPTIONAL, nonCriticalExtension SEQUENCE{} OPTIONAL } --TAG-UECAPABILITYINFORMATION-STOP --ASN1STOP

[0025] In one or more embodiments, the IE UE-CapabilityML-Information information element includes capabilities specific to machine learning.

[0026] In one or more embodiments, the UECapabilityML-Information information element may be as follows: --ASN1START --TAG-UE-CAPABILITYML-INFORMATION-START UE-CapabilityML-Information::= SEQUENCE{ hardwareCapability HardwareCapability, modelCapability ModelCapability } HardwareCapability::= SEQUENCE{ chiptype ENUMERATED{CPU,GPU,ASIC,TPU,others,spare3,spare2,spare1} OPTIONAL, maxPower INETGER(0..100) OPTIONAL, batchsize ENUMERATED{1GB,2GB,4GB,8GB,16GB,spare3,spare2,spare1}, OPTIONAL batterycapacity INTEGER(0...100) OPTIONAL, currentbatterystatus INTEGER(0...100) OPTIONAL, currentbattery INTEGER(0...100) OPTIONAL, } ModelCapability::= SEQUENCE{ modeltype ENUMERATED{CNN,RNN,Regression,Classification,RL,spare3,spare2,spare1} OPTIONAL, modelsize ENUMERATED{1GB,2GB,4GB,8GB,16GB,spare3,spare2,spare1}, OPTIONAL, trainingCapability ENUMERATED{supported} OPTIONAL, inferenceCapability ENUMERATED{supported} OPTIONAL, } --TAG-UE-CAPABILITYML-INFORMATION-STOP --ASN1STOP

[0027] In one or more embodiments, the machine learning configuration may be transmitted from the NG-RAN to the UE via RRC signaling. To support distributed / federal learning between the NG-RAN and the UE, the following should be included in the RRC message: A new message type, "MachineLearningConfiguration," can be used to carry the machine learning configuration. Information from the RAN to the UE includes the machine learning model, configuration (and model parameter updates), etc. Service type: Distributed learning and federal learning have different requirements for the information exchanged between the NG-RAN and the UE. In the case of distributed learning, when the UE independently handles machine learning training and inference (Scenario 1), the NG-RAN does not need to maintain the machine learning model adopted on the UE side, so the UE does not need to synchronize / download the machine learning model with the NG-RAN. Therefore, the NG-RAN does not need to know the machine learning use case on the UE side. For this scenario, the NG-RAN may set the service type to "default". Another scenario for distributed learning (Scenario 2) is that the NG-RAN acts as the machine learning training node and the UE acts as the inference node. A UE can download a model trained by RAN according to its machine learning service type of interest. Upon receiving such a model, the UE can begin inference based on the input data. "Service type" is used to indicate the service from which the machine learning model is trained. Examples include positioning, V2X, channel estimation, etc.

[0028] In one or more embodiments, for federative learning, the same machine learning model is shared between NG-RAN and UE (Scenario 3). The UE may need to register for the machine learning service and receive the ML model from NG-RAN. Thus, NG-RAN may define certain use cases as different "service types" that can be supported for federative learning between NG-RAN and UE.

[0029] In one or more embodiments, the machine learning report configuration field specifies the machine learning to be performed by the UE, as well as configurations such as the machine learning report type or frequency.

[0030] In one or more embodiments, the ML configuration may include machine learning reporting types. In Scenario 1, since the NG-RAN does not have a machine learning model, the UE is only required to transfer results from the machine learning model to the NG-RAN. This field may include measurement predictions, performance feedback, and UE action space (such as handover requests). These reports can be categorized as prediction results and action space.

[0031] In one or more embodiments, for federated learning or model deployment from NG-RAN nodes to UEs, both NG-RAN and UE can maintain the same machine learning model downloaded from NG-RAN during initialization. For federated learning, the UE updates and iterates through its local machine learning model based on its own environment and input / output, and the UE can also report updated machine learning parameters generated by the local nodes to NG-RAN, which in turn can update the centralized model accordingly. This type of reporting can be called model parameter updating.

[0032] In one or more embodiments, for scenarios 2 and 3, in addition to the reporting described above, NG-RAN may also require the UE to report "model bias" and "model variance". Details are as follows:

[0033] In one or more embodiments, there may be an ML reporting cycle and an offset. The reporting cycle is used to indicate how often a UE should report model parameter updates or prediction results to NG-RAN. This cycle can be per UE or per model. The cycle field is used to indicate the cycle of a UE reporting "model parameter updates" or "prediction results" to NG-RAN. If the prediction results require immediate feedback from NG-RAN (such as in case 3 above), the field may also be set to 0. Each cycle corresponds to an offset (given in slot numbers) which indicates the offset of start times between different UEs.

[0034] In one or more embodiments, there may be ML results that store duration and start time. Duration: In addition to the period, for “prediction results” that do not require immediate feedback to NG-RAN, NG-RAN may also configure the UE to have a length of time for which the prediction results should be stored before the period ends. This field is used to indicate how long the report should be stored in the UE. Start time: This field is the start time of the UE for recording the prediction results.

[0035] In one or more embodiments, there may be policy guidance provided to the UE by NG-RAN. The guided policy can be provided by the OAM, CN, or gNB itself and indicates a high level of action that the UE can take according to the output of a machine learning algorithm on the UE side.

[0036] In one or more embodiments, UE behavior may be determined by NG-RAN or by the UE itself. This field is used to indicate to the UE what types of behavior can be considered when inference results are obtained from a machine learning model. Types of UE actions include 1) RRC configuration, 2) Policy Guided, and 3) Action Guided. The UE can determine its behavior according to the configured behavior type.

[0037] In one or more embodiments, the model bias threshold field sets the model bias threshold for the UE, which determines when to trigger a MachineLearningModelUpdateRequest.

[0038] In one or more embodiments, the model distribution threshold field sets the model distribution threshold for when to trigger MachineLearningModelUpdateRequest in the UE.

[0039] In one or more embodiments, if the machine learning model and parameter updates are also carried via the control plane, the following two fields are also considered to be included in MachineLearningConfiguration:

[0040] In one or more embodiments, a machine learning model may be deployed in the UE. This model may also be jointly optimized in the NG-RAN and UE when federated learning is used as the framework in the machine learning-supported NG-RAN network. The model is generated by a training node located in a network node (CU / DU, OAM, or CN) and sent to the UE in a container via a new SRB (for example, as described above) or via an MLRB / DRB in the data plane. The ML model may also be in the form of an identifier for the UE to download the model. Depending on the different UE capabilities, the NG-RAN may assign models with finer granularity.

[0041] In one or more embodiments, the machine learning model field may include other training-related information such as the loss function and the optimizer for training. The machine learning model field is optional for distributed learning frameworks and required for federated frameworks or model deployment from NG-RAN to UE. RRC segmentation is supported if the model size is greater than 8000 bytes.

[0042] In one or more embodiments, the machine learning model parameter update field can carry information used for machine learning model updates and sequential iterations, such as hidden layers, weights, and gradients. Given that machine learning algorithms differ across different use cases and vendors, this information is generated from the network nodes performing the machine learning training and carried within a container passed to the UE.

[0043] In one or more embodiments, the central server in the federated learning framework can be located in the NG-RAN (e.g., CU or DU), OAM, or CN. When the OAM or CN acts as the central server in the federated framework, model parameters are generated by the OAM or CN. A single container is used to transfer these parameters between the network and the UE. In this situation, the RRC layer is transparent to this field when transmitted through the control plane. RRC segmentation is supported when the updated model size is greater than 8000 bytes. For larger data sizes, transmission via DRB / MLRB in the data plane is also considered. When the NG-RAN (e.g., CU or DU) is the central server, this container is formed by the NG-RAN (e.g., CU or DU, depending on which is the training node).

[0044] In one or more embodiments, for an RRC message, MachineLearningConfiguration has three options: (1) a new field in rrcReconfiguration which can be carried in either the otherConfig field or a new dedicated field. An example of a new field in rrcReconfiguration is shown below: RRCReconfiguration-vxyz-IEs::= SEQUENCE{ machineLearningConfiguration MachineLearningConfiguration OPTIONAL } OtherConfig-vxyz-IEs::= SEQUENCE{ machineLearningConfiguration MachineLearningConfiguration OPTIONAL }

[0045] (2) New DL-DCCH-Message. An example of RRC signal transmission from NG-RAN to UE is shown below: -DL-DCCH-Message The DL-DCCH-Message class is a set of RRC messages that can be sent from the network to the UE over the downlink DCCH logical channel. --ASN1START --TAG-DL-DCCH-MESSAGE-START DL-DCCH-Message::= SEQUENCE{ message DL-DCCH-MessageType } DL-DCCH-MessageType::= CHOICE{ c1 CHOICE{ rrcReconfiguration RRCReconfiguration, rrcResume RRCResume, rrcRelease RRCRelease, rrcReestablishment RRCReestablishment, securityModeCommand SecurityModeCommand, dlInformationTransfer DLInformationTransfer, ueCapabilityEnquiry UECapabilityEnquiry, CounterCheck, mobilityFromNRCommand MobilityFromNRCommand, dlDedicatedMessageSegment-r16 DLDedicatedMessageSegment-r16, ueInformationRequest-r16 UEInformationRequest-r16, dlInformationTransferMRDC-r16 DLInformationTransferMRDC-r16, loggedMeasurementConfiguration-r16 LoggedMeasurementConfiguration-r16, machineLearningConfiguration MachineLearningConfiguration, spare2 NULL, spare1 NULL }, messageClassExtension SEQUENCE{} } --TAG-DL-DCCH-MESSAGE-STOP --ASN1STOP

[0046] In one or more embodiments, a new SIB may be used to demonstrate the network's ML capabilities. Aside from machine learning models and machine learning model parameter updates, other fields may also be broadcast to the UE(s) via the new SIB.

[0047] In one or more embodiments, examples of MachineLearningConfiguration messages and information elements are shown below. -MachineLearning Configuration The MachineLearningConfiguration message is used to execute machine learning on the UE side. It is used to transmit machine learning configurations to enable machine learning services for network performance optimization. Signal transmission wireless bearer: SRBX RLC-SAP:AM Logical Channel: DCCH Direction: From network to UE MachineLearningConfiguration messages --ASN1START --TAG-MACHINELEARNINGCONFIGURATION-START MachineLearningConfiguration::=SEQUENCE{ criticalExtensions CHOICE { machineLearningConfiguration MachineLearningConfiguration-IEs, criticalExtensionsFuture SEQUENCE{} } } MachineLearningConfiguration-IEs::= SEQUENCE{ serviceType ENUMERATED{default,positioning,V2X,spare5,spare4,spare3,spare2,spare1} OPTIONAL, machineLearningReportConfiguration MachineLearningReportConfiguration, policyGuidance OCTET STRING (SIZE(1..8000)) OPTIONAL, machineLearningModelUpdate machineLearningModelUpdate, lateNonCriticalExtension OCTET STRING OPTIONAL, nonCriticalExtension SEQUENCE{} OPTIONAL } MachineLearningReportConfiguration::= SEQUENCE{ reportType ENUMERATED{prediction-result,action-space,model-update,model-bias,model-variance,spare3,spare2,spare1}, behaviorType ENUMERATED{rrc-config,policy-guided,action-guided,spare4,space3,space2,space1}, reportPeriodicityAndOffset MachinelearningPeriodicityAndOffset, reportDurationAndStarttime MachineLearningDurationAndStarttime, predictionResultConfiguration PredictionResultConfiguration, actionSpaceConfiguration ActionSpaceConfiguration, … } MachineLearningModelUpdate::= SEQUENCE{ machineLearningModel OCTET STRING(SIZE(1..8000)) OPTIONAL, modelUpdate OCTET STRING(SIZE(1..8000)) OPTIONAL, … } MachinelearningPeriodicityAndOffset::= SEQUENCE{ slots4 INTEGER(0..3), slots5 INTEGER(0..4), slots8 INTEGER(0..7), slots10 INTEGER(0..9), slots16 INTEGER(0..15), slots20 INTEGER(0..19), slots40 INTEGER(0..39), slots80 INTEGER(0..79), slots160 INTEGER(0..159), slots320 INTEGER(0..319) } --TAG-MACHINELEARNINGCONFIGURATION-STOP --ASN1STOP

[0048] Table 2 below describes the MachineLearningConfiguration field. [Table 2]

[0049] In one or more embodiments, ML reports may be transmitted from the UE to the NG-RAN via RRC signaling. Information uploaded / reported from the UE to the RAN includes machine learning model parameter updates, prediction results, action space, and feedback (e.g., model performance feedback and / or wireless feedback (e.g., system KPIs, e.g., throughput / latency)) via a new message type "MachineLearningReport". The report type depends on the machine learning report type configuration received from the NG-RAN. RRC segmentation is supported if the report size is greater than 8000 bytes. The service type field is used to report registered machine learning services to the NG-RAN. If the report type is configured as "Machine Learning Model Parameter Update", the UE reports the corresponding updated parameters, which are updated by itself according to the environment and local training. If the report type is configured as "Prediction Results", the UE reports the predicted values, which are the results of the machine learning model. The predicted values ​​may differ among different machine learning use cases, some examples of which may be channel matrix prediction, CSI prediction, location prediction, etc. This field may be carried in the new message "MachineLearningReport". Alternatively, if there is a corresponding measurement report or information sent from the UE to the RAN, the prediction results may also be included in the same message / field as the legacy report, but with a separate IE, especially for the predicted values. Current / expected feedback is also reported along with the prediction results. Expected feedback describes how the UE's performance would look if the corresponding prediction results were used by the UE.

[0050] In one or more embodiments, in accordance with the output of a machine learning algorithm, the UE may also determine its own actions in accordance with guidance policies received from the NG-RAN (Embodiment 4) or direct results from the machine learning algorithm. If the reporting type is configured as “Action Space”, the UE sends its action space to the NG-RAN to determine how the UE should act. An action space can be either a request or a report. For example, in a mobility improvement use case, in accordance with predicted channel quality and positioning, the UE may request a handover to a neighboring cell. In this case, the UE sends its preferred neighboring cell identification information and related information to the NG-RAN requesting the handover. The preferred neighboring cell identification information and handover request in this example can be called an action space. Current / expected feedback is also reported along with the prediction results. Expected feedback describes how the UE's performance would look if the corresponding action were taken by the UE.

[0051] In one or more embodiments, when the input data does not fit well to the model inferred / trained on the UE side, the UE calculates the model's bias and reports it to NG-RAN. The value can be classified into different levels, such as poor bias, moderately poor bias, no bias, etc.

[0052] In one or more embodiments, depending on different input data and environments, the inference model or trained model may not fit the input data very well. Therefore, this field is used to report the variance of the machine learning model.

[0053] In one or more embodiments, a confidence level is introduced to indicate how reliable the network is in reporting from the UE.

[0054] An example of RRC signal transmission from the UE to the NG-RAN is shown below. The DCCH-Message class is a set of RRC messages that can be sent from the UE to the network over the uplink DCCH logical channel: --ASN1START --TAG-UL-DCCH-MESSAGE-START UL-DCCH-Message::= SEQUENCE{ message UL-DCCH-MessageType } UL-DCCH-MessageType::= CHOICE{ c1 CHOICE { measurementReport MeasurementReport, rrcReconfigurationComplete RRCReconfigurationComplete, rrcSetupComplete RRCSetupComplete, rrcReestablishmentComplete RRCReestablishmentComplete, rrcResumeComplete RRCResumeComplete, securityModeComplete SecurityModeComplete, securityModeFailure SecurityModeFailure, ulInformationTransfer ULInformationTransfer, locationMeasurementIndication LocationMeasurementIndication, ueCapabilityInformation UECapabilityInformation, counterCheckResponse CounterCheckResponse, ueAssistanceInformation UEAssistanceInformation, failureInformation FailureInformation, ulInformationTransferMRDC ULInformationTransferMRDC, scgFailureInformation SCGFailureInformation, scgFailureInformationEUTRA SCGFailureInformationEUTRA }, messageClassExtension CHOICE { c2 CHOICE { ulDedicatedMessageSegment-r16 ULDedicatedMessageSegment-r16, dedicatedSIBRequest-r16 DedicatedSIBRequest-r16, mcgFailureInformation-r16 MCGFailureInformation-r16, ueInformationResponse-r16 UEInformationResponse-r16, sidelinkUEInformationNR-r16 SidelinkUEInformationNR-r16, ulInformationTransferIRAT-r16 ULInformationTransferIRAT-r16, iabOtherInformation-r16 IABOtherInformation-r16, machineLearningReport MachineLearningReport, spare8 NULL, spare7 NULL, spare6 NULL, spare5 NULL, spare4 NULL, spare3 NULL, spare2 NULL, spare1 NULL }, messageClassExtensionFuture-r16 SEQUENCE{} } } --TAG-UL-DCCH-MESSAGE-STOP --ASN1STOP

[0055] In one or more embodiments, the MachineLearningReport message is used to indicate measurement results. Signal-transmitting radio bearer: SRBX. RLC-SAP: AM. Logical channel: DCCH. Direction: UE to network. The MachineLearningReport message may look like this: --ASN1START --TAG-MACHINELEARNINGREPORT-START MachineLearningReport::= SEQUENCE { criticalExtensions CHOICE { machineLearningReport MachineLearningReport-IEs, criticalExtensionsFuture SEQUENCE {} } } MachineLearningReport-IEs::= SEQUENCE{ serviceType ENUMERATED{default, positioning, V2X, spare5, spare4, spare3, spare2, spare1} OPTIONAL, modelUpdate OCTET STRING(SIZE(1..8000)) OPTIONAL, modelBias ENUMERATED{poor-bias, less-poor, no-bias, spare5, spare4, spare3, spare2, spare1} OPTIONAL, modelVariance INTEGER(0,...,100) OPTIONAL, predictionResult PredictionResult OPTIONAL, confidenceLevel ENUMERATED{0,1,2,...,10} OPTIONAL, modelPerformanceFeedback ENUMERATED{good, medium, poor} OPTIONAL, userPerformanceFeedback ENUMERATED{increase,decrease,fair} OPTIONAL, lateNonCriticalExtension OCTET STRING OPTIONAL, nonCriticalExtension SEQUENCE{} OPTIONAL } --TAG-MEASUREMENTREPORT-STOP --ASN1STOP

[0056] User Performance Feedback can be further broken down into IE components such as throughput, latency, and QoS.

[0057] Table 3 below describes the MachineLearningConfiguration field. [Table 3]

[0058] In one or more embodiments, the RAN and UE are coordinated with a federated learning framework or model deployed from the NG-RAN to the UE for ML model update request and response exchange. When federated learning is considered, each UE trains a machine learning model locally. However, after some time (for example, before a UE needs to report a "machine learning model parameter update"), some machine learning model may not converge in some UE(s), or its prediction results may be outside the acceptable range according to policy. In this case, the UE can send a "MachineLearningModelUpdateRequest" to the network requesting a parameter update to the corresponding machine learning model. Upon receiving this request, the network sends a "MachineLearningModelUpdateResponse" message, which includes "MachineLearningModelUpdate" and the corresponding service type.

[0059] In one or more embodiments, UE-selective training / ML model updating may be possible. A "ConfidenceLevel" is introduced to indicate how much the network trusts the model updated by the UE or the prediction results obtained from the UE. According to the "confidentialLevel" reported by the UE, the NG-RAN can selectively update the machine learning model for different UEs. The NG-RAN may prioritize updating the machine learning model for UEs with lower confidenceLevel rates.

[0060] In one or more embodiments, the network's ML capability may be indicated by a new SIB. This new SIB (e.g., SIBX) contains information related to machine learning. It may include: network machine learning capability, machine learning services that the network can provide, and for each service, the information includes required software libraries, required machine learning models, required memory size (e.g., in MB / kB), and other information (e.g., machine learning configuration). An example of this new SIB is shown below: -SIBX SIBX5 contains information related to machine learning. --ASN1START --TAG-SIB15-START SIB15-rxyz::= SEQUENCE{ MachineLearningSupport ENUMERATED{true,false}, mlServiceInfoList MLServiceInfoList OPTIONAL, (machineLearningConfiguration MachineLearningConfiguration OPTIONAL,) ... } MLServiceInfoList::= SEQUENCE(SIZE(1..maxMLSerivce)) OF MLServiceInfo MLServiceInfo::= SEQUENCE{ fieldServiceId MlServiceId, memoryRequirement INTEGER (1,..,1000), requiredMlModelList MLModelList, requiredMlSoftwareLibList MLModelSoftwareLibList, ... } MLModelList::= SEQUENCE(SIZE(1..maxModel)) OF MLModel MLModelSoftwareLibList::= SEQUENCE(SIZE(1..maxSoftLib)) OF MLSoftwareLib MLModel::= ENMERATED{CNN, RNN, NN, DQN, RandomForest, …} MLSoftwareLib::= ENUMERATED{numpy, scipy, …} -- TAG-SIB15-STOP -- ASN1STOP

[0061] In one or more embodiments, there may be a service registration / interest instruction message. Upon receiving a list of services provided by the network (carried in the network capability instruction in Embodiment 8), the UE(s) send the service types / IDs of their interest to the network and request such services(s). If the UE is interested and requests to receive machine learning models from the network, the UE sends the requested service IDs / service types to the network.

[0062] In one or more embodiments, there may be an action feedback message. In the case of action-guided RAN-UE machine learning collaboration, instead of sending the RRC configuration to the UE, the network may instead send a new message carrying action feedback. This can reduce the reconfiguration of such UEs and reduce the message size through the air interface. This feedback may be either action approval or action rejection.

[0063] The above description is illustrative and not intended to be limiting. Numerous other examples, configurations, processes, algorithms, etc., may exist, some of which are described in more detail below. Illustrative embodiments are described with reference to the accompanying drawings.

[0064] Figure 1 illustrates an exemplary process 100 for network and user device collaboration to facilitate machine learning on user devices, according to some exemplary embodiments of the present disclosure.

[0065] Referring to Figure 1, process 100 may include a network device 102 (e.g., gNB) and the UE device 104, which cooperate to allow the UE device 104 to perform ML. The network device 102 may generate and transmit an ML capability instruction 106 to indicate that it can provide ML models and configurations for one or more services and can cooperate with the UE device 104 to provide the relevant ML models, configurations, and other data to allow the UE device 104 to perform ML. The UE device 104 may respond by generating and transmitting a service registration 108 (e.g., an interest instruction) to indicate to the network device 102 that the UE device 104 is interested in ML for one or more of the services advertised by the network device 102. The network device 102 may generate and transmit a UE capability query 110 to the UE device 104 to request the UE device 104 to provide information about the UE's capabilities (e.g., ML capabilities and / or hardware capabilities). The UE device 104 may respond by generating and sending UE capability information 112 (for example, indicating ML and / or hardware capabilities). Based on the UE capability information 112, the network device 102 may select or generate an ML configuration 114 for the UE device 104 and send the ML configuration 114 to the UE device. For example, based on any requested service in a service registration 108, the network device 102 may select an ML model for the service and an ML configuration for that ML model (for example, an ML model / configuration that uses more or less resources depending on the UE capability information 112).

[0066] Still referring to Figure 1, after receiving the ML model and ML configuration 114, the UE device 104 may execute the ML model based on the ML configuration 114. After the UE device 104 executes the ML model and generates the corresponding output, the UE device 104 may generate an ML report 118 and send it to the network device 102. The ML report 118 may show the predictions preceding the ML execution 116, the results of the ML execution 116 (e.g., the actual output of the ML execution 116), and the requested or selected actions for the UE device 104 to perform based on the ML execution 116 (e.g., the action space). The ML model and / or configuration may be trained and / or updated by the network device 102 and / or the UE device 104. When the ML model and / or configuration is updated by the network device 102, the UE device 104 may periodically and optionally generate an ML model update request 120 to request the update and send it to the network device 102. The network device 102 may optionally generate and transmit a model update response 122 having the updated ML model and configuration in response to an ML model update request 120, or without an update request from the UE device 104.

[0067] In one or more embodiments, the ML capability instruction 106 may be sent via SIB.

[0068] In one or more embodiments, the ML configuration 114 and the ML report 118 may be transmitted using a new SRB or a new data radio bearer.

[0069] In one or more embodiments, the network device 102 can initiate the ML configuration procedure by sending an ML configuration message 114.

[0070] The UE device 104 may include any suitable processor-driven device, including but not limited to mobile or non-mobile devices, such as static devices. For example, UE device 104 includes personal computers (PCs), wearable wireless devices (e.g., bracelets, watches, glasses, rings, etc.), desktop computers, mobile computers, laptop computers, ultrabook® computers, notebook computers, tablet computers, server computers, handheld computers, handheld devices, Internet of Things (IoT) devices, sensor devices, PDA devices, handheld PDA devices, onboard devices, offboard devices, hybrid devices (e.g., combining cellular phone functionality with PDA device functionality), consumer devices, vehicle devices, non-vehicle devices, mobile or portable devices, non-mobile or non-portable devices, mobile phones, cellular phones, PCS devices, PDA devices incorporating wireless communication devices, mobile or portable GPS devices, DVB devices, relatively small computing devices, non-desktop computers, context-aware devices, video devices, audio devices, A / V devices, set-top boxes (STBs), Blu-ray disc (BD) players, BD recorders, digital video disc (DVD) players, high-definition (HD) DVD players, DVD recorders, and HD This may include DVD recorders, personal video recorders (PVRs), broadcast HD receivers, video sources, audio sources, video syncs, audio syncs, stereo tuners, broadcast radio receivers, flat panel displays, personal media players (PMPs), digital video cameras (DVCs), digital audio players, speakers, audio receivers, game devices, data sources, data syncs, digital still cameras (DSCs), media players, smartphones, televisions, music players, and the like.Other devices, including smart devices such as lamps, environmental control systems, automotive components, household components, and electrical appliances, may also be included in this list.

[0071] As used herein, the term “Internet of Things (IoT) device” is used to refer to any object (e.g., appliance, sensor, etc.) that has an addressable interface (e.g., Internet Protocol (IP) address, Bluetooth® identifier (ID), Near Field Communication (NFC) ID, etc.) and can transmit information to one or more other devices via a wired or wireless connection. IoT devices may have passive communication interfaces such as quick response (QR) codes, radio frequency identification (RFID) tags, NFC tags, or active communication interfaces such as modems, transceivers, transmitters and receivers. IoT devices may have a specific set of attributes (e.g., device state or status such as whether the IoT device is on or off, open or closed, idle or active, available or busy for task execution, cooling or heating capabilities, environmental monitoring or recording capabilities, light emission capabilities, sound emission capabilities, etc.) that are incorporated into and / or controlled / monitored by a central processing unit (CPU), microprocessor, ASIC, etc., and can be configured for connection to an IoT network such as a local ad-hoc network or the Internet. For example, IoT devices may include, but are not limited to, refrigerators, toasters, ovens, microwave ovens, freezers, dishwashers, dishes, hand tools, washing machines, clothes dryers, furnaces, air conditioners, thermostats, televisions, lighting fixtures, vacuum cleaners, sprinklers, electric meters, gas meters, etc., as long as they are equipped with an addressable communication interface for communicating with an IoT network. IoT devices may also include mobile phones, desktop computers, laptop computers, tablet computers, personal digital assistants (PDAs), etc. Thus, an IoT network may consist of a combination of devices that typically do not have internet connectivity (e.g., dishwashers) and "legacy" internet-accessible devices (e.g., laptops or desktop computers, mobile phones, etc.).

[0072] Any of the UE device 104 and network device 102 may include one or more communication antennas. These one or more communication antennas may be any suitable type of antenna corresponding to the communication protocols used by the UE device 104 and network device 102. Some non-exclusive examples of suitable communication antennas include 3GPP antennas, directional antennas, omnidirectional antennas, dipole antennas, folded dipole antennas, patch antennas, multiple input multiple output (MIMO) antennas, omnidirectional antennas, and quasi-omnidirectional antennas. These one or more communication antennas may be communicatively coupled to a wireless component to transmit and / or receive signals, such as communication signals, to and from the UE device 104 and network device 102.

[0073] Figure 2 shows a flowchart of an exemplary process 200 for network and user device collaboration to facilitate machine learning on user device, according to one or more exemplary embodiments of the present disclosure.

[0074] In block 202, a device (for example, network device 102 in Figure 1, NG-RAN 314 in Figure 3) can generate an instruction (for example, ML capability instruction 106 in Figure 1) indicating that the network device supports ML and send it to a UE device (for example, UE device 104 in Figure 1).

[0075] In block 204, the device may identify a service registration received from a UE device (for example, service registration 108 in Figure 1). A service registration may indicate a request for an ML model and configuration for one or more services supported by the device.

[0076] In block 206, the device may generate and send an information request (e.g., UE capability query 110 in Figure 1) to the UE device to request the ML and / or hardware capabilities of the UE device. Based on this information, the device may select / generate ML configurations, ML parameters, policies, etc., and send them to the UE device for use by the UE device. For example, a given service may have one or more ML models with different parameters that can be selected based on the ML and / or hardware capabilities of the UE device.

[0077] In block 208, the device may identify information received from the UE device (for example, UE capability information 112 in Figure 1). Based on which models and configurations the UE device supports, the device may select a model and configuration that meets the hardware capabilities of the UE device.

[0078] In block 210, the device may generate / select a machine learning configuration and an ML model (e.g., ML configuration 114 in Figure 1) based on information received from the UE device and send it to the UE device. The device may optionally update the ML model and / or configuration in response to requests received from the UE device and / or in response to ML reports provided by the UE device (e.g., showing what the results of running the ML model were compared to predicted results). The device may provide policies that govern the UE device's decisions regarding the use and results of the ML model, and / or receive responses to requests received from the UE device (e.g., confirmation or rejection) regarding what actions the UE device may take based on the results of running the ML model.

[0079] It should be understood that the above explanation is for illustrative purposes only and not intended to be limiting.

[0080] Figure 3 shows a network 300 according to one or more exemplary embodiments of the present disclosure.

[0081] Network 300 may operate in a manner consistent with the 3GPP technical specifications for LTE or 5G / NR systems. However, exemplary embodiments are not limited in this respect, and the embodiments described may be applied to other networks that benefit from the principles described herein, such as future 3GPP systems.

[0082] Network 300 may include UE 302, which may include any mobile or non-mobile computing device designed to communicate with RAN 304 via an over-the-air connection. UE 302 may be communicatively coupled with RAN 304 by a Uu interface. UE 302 may include, but is not limited to, smartphones, tablet computers, wearable computing devices, desktop computers, laptop computers, automotive infotainment, automotive entertainment devices, instrument clusters, head-up display devices, onboard diagnostic devices, dashboard mobile devices, mobile data terminals, electronic engine management systems, electronic / engine control units, electronic / engine control modules, embedded systems, sensors, microcontrollers, control modules, engine management systems, network appliances, machine-type communication devices, M2M or D2D devices, IoT devices, etc.

[0083] In some embodiments, the network 300 may include multiple UEs directly coupled to one another via a sidelink interface. The UEs may be M2M / D2D devices that communicate using physical sidelink channels such as PSBCH, PSDCH, PSSCH, PSCCH, and PSFCH.

[0084] In some embodiments, UE 302 may further communicate with AP 306 via an over-the-air connection. AP 306 may manage the WLAN connection, which may offload some / all network traffic from RAN 304. The connection between UE 302 and AP 306 may be compatible with any IEEE 802.11 protocol, and AP 306 may be a Wireless Fidelity (Wi-Fi®) router. In some embodiments, UE 302, RAN 304, and AP 306 may utilize cellular-WLAN aggregation (e.g., LWA / LWIP). Cellular-WLAN aggregation may involve UE 302 being configured by RAN 304 to utilize both cellular radio resources and WLAN resources.

[0085] RAN 304 may include one or more access nodes, such as AN 308. AN 308 may terminate the air interface protocol for UE 302 by providing access layer protocols including RRC, PDCP, RLC, MAC, and L1 protocols. In this way, AN 308 can enable data / voice connectivity between CN 320 and UE 302. In some embodiments, AN 308 may be implemented as one or more software entities operating on a discrete device or on a server computer as part of a virtual network which may be called CRAN or virtual baseband unit pool. AN 308 may be referred to as BS, gNB, RAN node, eNB, ng-eNB, NodeB, RSU, TRxP, TRP, etc. AN 308 may be a macrocell base station, or a low-power base station for providing a femtocell, picocell, or other similar cell with a smaller coverage area, smaller user capacity, or higher bandwidth compared to a macrocell.

[0086] In embodiments where RAN 304 includes multiple ANs, they may be coupled to each other via an X2 interface (if RAN 304 is an LTE RAN) or an Xn interface (if RAN 304 is a 5G RAN). In some embodiments, the X2 / Xn interface may be separated into a control / user plane interface, which may allow ANs to communicate information related to handover, data / context transfer, mobility, load management, interference coordination, etc.

[0087] Each AN of RAN 304 can manage one or more cells, cell groups, component carriers, etc., to provide an air interface for network access to UE 302. UE 302 may be simultaneously connected to multiple cells provided by the same or different ANs of RAN 304. For example, UE 302 and RAN 304 may use carrier aggregation that allows UE 302 to connect to multiple component carriers corresponding to Pcells or Scells, respectively. In a dual-connection scenario, the first AN may be a master node providing an MCG, and the second AN may be a secondary node providing an SCG. The first / second ANs can be any combination of eNBs, gNBs, ng-eNBs, etc.

[0088] RAN 304 can provide an air interface through licensed or unlicensed spectra. To operate in unlicensed spectra, nodes may use LAA, eLAA, and / or feLAA mechanisms based on CA technology using PCell / SCell. Prior to accessing unlicensed spectra, nodes may perform medium / carrier sensing operations based, for example, on a listen-before-talk (LBT) protocol.

[0089] In a V2X scenario, UE 302 or AN 308 can be or may act as an RSU, referring to any transport infrastructure entity used for V2X communication. An RSU may be implemented in or by a suitable AN or stationary (or relatively stationary) UE. An RSU implemented in or by a UE may be called a “UE-type RSU,” an “eNB-type RSU” in the case of an eNB, a “gNB-type RSU” in the case of a gNB, and so on. In one example, an RSU is a computing device coupled to a radio frequency circuit located on the roadside, providing connectivity support to a passing vehicle UE. An RSU may also include internal data storage circuitry for storing intersection map geometry, traffic statistics, media, and applications / software for sensing and controlling oncoming vehicle and pedestrian traffic. An RSU can provide very low-latency communication required for high-speed events such as collision avoidance and traffic warnings. Additionally or alternatively, an RSU may provide other cellular / WLAN communication services. The components of the RSU may be packaged in a weatherproof enclosure suitable for outdoor installation and may include a network interface controller for providing wired connectivity (e.g., Ethernet®) to a traffic signal controller or backhaul network.

[0090] In some embodiments, RAN 304 may be an LTE RAN 310 having an eNB, for example, eNB 312. The LTE RAN 310 can provide an LTE air interface having the following characteristics: 15kHz SCS; CP-OFDM waveform for DL ​​and SC-FDMA waveform for UL; turbo coding for data and TBCC for control, etc. The LTE air interface may rely on CSI-RS for CSI acquisition and beam management; PDSCH / PDCCH DMRS for PDSCH / PDCCH demodulation; CRS for cell discovery and initial acquisition, channel quality measurement, and channel estimation for coherent demodulation / detection at UE. The LTE air interface may operate in the sub-6GHz band.

[0091] In some embodiments, the RAN 304 may be an NG-RAN 314 having a gNB, for example, gNB 316, or an ng-eNB, for example, ng-eNB 318. The gNB 316 can connect to a 5G-enabled UE using a 5G NR interface. The gNB 316 can connect to the 5G core through an NG interface, which may include an N2 interface or an N3 interface. The ng-eNB 318 can also connect to the 5G core through an NG interface, but can also connect to the UE via an LTE air interface. The gNB 316 and ng-eNB 318 can connect to each other through an Xn interface.

[0092] In some embodiments, the NG interface may be divided into two parts: an NG user plane (NG-U) interface (e.g., N3 interface) that carries traffic data between the nodes of the NG-RAN 314 and the UPF 348, and an NG control plane (NG-C) interface (e.g., N2 interface) that is a signaling interface between the nodes of the NG-RAN 314 and the AMF 344.

[0093] NG-RAN 314 can provide a 5G-NR air interface having the following characteristics: variable SCS; CP-OFDM for DL, CP-OFDM and DFT-s-OFDM for UL; polar, repetitive, simplex, and Reed-Muller codes for control, and LDPC for data. The 5G-NR air interface may rely on CSI-RS, PDSCH / PDCCH DMRS, similar to the LTE air interface. The 5G-NR air interface may not use CRS, but may use PBCH DMRS for PBCH demodulation; PTRS for phase tracking for PDSCH; and a tracking reference signal for time tracking. The 5G-NR air interface may operate on the FR1 band, including the sub-6 GHz band, or the FR2 band, including the 24.25 GHz to 52.6 GHz band. The 5G-NR air interface may include SSB, which is the area of ​​the downlink resource grid including PSS / SSS / PBCH.

[0094] In some embodiments, a 5G-NR air interface can utilize BWPs for various purposes. For example, BWPs can be used for dynamic adaptation of SCSs. For instance, UE 302 may consist of multiple BWPs, each with a different SCS. When a BWP change is indicated to UE 302, the SCS of the transmission is also changed. Another example of a use case for BWPs relates to power saving. In particular, multiple BWPs can be configured for UE 302 with different amounts of frequency resources (e.g., PRBs) to support data transmission under different traffic load scenarios. BWPs with fewer PRBs can be used for data transmission with low traffic loads, allowing power saving in UE 302 and possibly in gNB 316. BWPs with more PRBs can be used for scenarios with higher traffic loads.

[0095] RAN 304 is telecommunicatively coupled to CN 320, which contains network elements for providing various functions to customers / subscribers (e.g., users of UE 302) to support data and telecommunications services. The components of CN 320 may be implemented on one physical node or on separate physical nodes. In some embodiments, NFV may be used to virtualize some or all of the functions provided by the network elements of CN 320 onto physical computing / storage resources such as servers, switches, etc. Logical instantiations of CN 320 may be referred to as network slices, and some logical instantiations of CN 320 may be referred to as network subslices.

[0096] In some embodiments, CN 320 may be LTE CN 322, sometimes referred to as EPC. LTE CN 322 may include MME 324, SGW 326, SGSN 328, HSS 330, PGW 332, and PCRF 334 coupled to one another through an interface (or “reference point”) as shown. The functions of the elements of LTE CN 322 can be briefly described below.

[0097] The MME 324 can implement mobility management capabilities to track the current location of the UE 302, facilitating paging, bearer activation / deactivation, handover, gateway selection, authentication, and more.

[0098] SGW 326 terminates the S1 interface toward the RAN and can route data packets between the RAN and the LTE CN 322. SGW 326 may also be a local mobility anchor point for inter-RAN node handover and may provide an anchor for 3GPP inter-mobility. Other roles may include lawful interception, billing, and some policy enforcement.

[0099] The SGSN 328 can track the location of the UE 302 and perform security functions and access control. Furthermore, the SGSN 328 can perform inter-EPC node signaling for mobility between different RAT networks; PDN and S-GW selection specified by the MME 324; MME selection for handover, etc. An S3 reference point between the MME 324 and the SGSN 328 can enable user and bearer information exchange for inter-3GPP access network mobility in idle / active states.

[0100] The HSS 330 may include a database of network users, including subscription-related information to support the handling of communication sessions by network entities. The HSS 330 can provide support for routing / roaming, authentication, authorization, naming / address resolution, location dependency, etc. The S6a reference point between the HSS 330 and the MME 324 may enable the transfer of subscription and authentication data for authenticating / authorizing user access to the LTE CN 320.

[0101] PGW 332 may terminate an SGi interface toward a data network (DN) 336, which may include an application / content server 338. PGW 332 may route data packets between the LTE CN 322 and the data network 336. PGW 332 may be coupled with SGW 326 by an S5 reference point to facilitate user plane tunneling and tunnel management. PGW 332 may further include nodes (e.g., PCEFs) for policy enforcement and billing data collection. In addition, the SGi reference point between PGW 332 and the data network 336 may be an external public, private PDN, or intra-operator packet data network, for example, for providing IMS services. PGW 332 may be coupled with PCRF 334 via a Gx reference point.

[0102] PCRF 334 is the policy and billing control element of LTE CN 322. PCRF 334 may be telecommunically coupled to the app / content server 338 to determine appropriate QoS and billing parameters for the service flow. PCRF 332 may provision associated rules with appropriate TFT and QCI to the PCEF (via the Gx reference point).

[0103] In some embodiments, CN 320 may be 5GC 340. 5GC 340 may include AUSF 342, AMF 344, SMF 346, UPF 348, NSSF 350, NEF 352, NRF 354, PCF 356, UDM 358, AF 360, and LMF 362 coupled to one another through interfaces (or “reference points”) as shown. The functions of the elements of 5GC 340 can be briefly described below.

[0104] The AUSF 342 may store data for authentication of the UE 302 and handle authentication-related functions. The AUSF 342 can facilitate a common authentication framework for various access types. In addition to communicating with other elements of the 5GC 340 through reference points as illustrated, the AUSF 342 may present a Nausf service-based interface.

[0105] The AMF 344 may allow other functions of the 5GC 340 to communicate with the UE 302 and RAN 304 and subscribe to notifications about mobility events concerning the UE 302. The AMF 344 may be responsible for registration management (e.g., for registering the UE 302), connectivity management, reachability management, mobility management, lawful interception of AMF-related events, and access authentication and authorization. The AMF 344 may provide transport for SM messages between the UE 302 and the SMF 346 and act as a transparent proxy for routing SM messages. The AMF 344 may also provide transport for SMS messages between the UE 302 and the SMSF. The AMF 344 can interact with the AUSF 342 and UE 302 to perform various security anchor and context management functions. Furthermore, the AMF 344 may include, or be the N2 reference point between the RAN 304 and the AMF 344, and may also be the termination point of the RAN CP interface, and may also be the termination point of NAS(N1) signaling, enabling NAS encryption and integrity protection. The AMF 344 may also support NAS signaling with the UE 302 via the N3 IWF interface.

[0106] SMF 346 may be responsible for SM (e.g., session establishment, tunnel management between UPF 348 and AN 308); UE IP address allocation and management (including optional authorization); selection and control of UP functions; configuration of traffic steering in UPF 348 for routing traffic to appropriate destinations; termination of interfaces to policy control functions; control of policy enforcement, billing, and some QoS; lawful interception (of SM events and interfaces to L1 systems); termination of SM portions of NAS messages; downlink data notification; initiation of AN-specific SM information sent to AN 308 on N2 via AMF 344; and determination of the session's SSC mode. SM may refer to the management of PDU sessions, and PDU sessions or “session” may refer to PDU connectivity services that provide or enable the exchange of PDUs between UE 302 and data network 336.

[0107] The UPF 348 can act as an anchor point for intra-RAT and inter-RAT mobility, an external PDU session point for interconnection to the data network 336, and a branch point to support multi-homed PDU sessions. The UPF 348 can also perform packet routing and forwarding, perform packet inspection, enforce the user plane portion of policy rules, lawfully intercept packets (UP collection), perform traffic usage reporting, perform QoS processing for the user plane (e.g., packet filtering, gating, UL / DL rate enforcement), perform uplink traffic verification (e.g., SDF-to-QoS flow mapping), perform transport-level packet marking on uplink and downlink, and perform downlink packet buffering and downlink data notification triggers. The UPF 348 may include an uplink classifier to support routing of traffic flows to the data network.

[0108] The NSSF 350 may select a set of network slice instances to serve the UE 302. The NSSF 350 may also determine, if necessary, the acceptable NSSAIs and their mappings to the subscribed S-NSSAIs. The NSSF 350 may also determine a set of AMFs, or a list of candidate AMFs, to be used to serve the UE 302, based on a preferred configuration and possibly by querying the NRF 354. The selection of a set of network slice instances for the UE 302 may also be triggered by the AMF 344 to which the UE 302 registers, by interacting with the NSSF 350, which may result in a change of AMF. The NSSF 350 may interact with the AMF 344 via the N22 reference point, or communicate with another NSSF in the visited network via the N31 reference point (not shown). In addition, the NSSF 350 may provide an Nnssf service-based interface.

[0109] NEF 352 can securely expose services and capabilities provided by 3GPP network functions for third parties, internal exposure / re-exposure, AFs (e.g., AF 360), edge computing, or fog computing systems. In such embodiments, NEF 352 can authenticate, authorize, or throttle AFs. NEF 352 can also translate information exchanged with AF 360 and information exchanged with internal network functions. For example, NEF 352 can translate between AF service identifiers and internal 5 GC information. NEF 352 can also receive information from other NFs based on the exposed capabilities of those NFs. This information may be stored in NEF 352 as structured data or in a data storage NF using a standardized interface. The stored information can then be re-exposed by NEF 352 to other NFs and AFs, or used for other purposes such as analysis. Furthermore, NEF 352 can represent Nnef service-based interfaces.

[0110] The NRF 354 supports service discovery functionality, receiving NF discovery requests from NF instances and providing NF instances with information about discovered NF instances. The NRF 354 also maintains information about available NF instances and their supported services. As used herein, terms such as “instantiate” and “instantiation” may refer to the creation of an instance, while “instance” may refer to the specific occurrence of an object that may occur, for example, during the execution of program code. Furthermore, the NRF 354 can represent Nnrf service-based interfaces.

[0111] PCF 356 may provide policy rules to control plane functions that enforce them, and may also support a unified policy framework to govern network behavior. PCF 356 may also implement a front-end for accessing subscription information related to policy decisions within the UDR of UDM 358. In addition to communicating with functions through reference points as illustrated, PCF 356 exhibits an Npcf service-based interface.

[0112] UDM 358 can process subscription-related information to support the processing of network entities in a communication session and can store subscription data for UE 302. For example, subscription data may be communicated via an N8 reference point between UDM 358 and AMF 344. UDM 358 can include two parts: an application frontend and a UDR. The UDR can store subscription and policy data for UDM 358 and PCF 356, and / or structured data for publication and application data for NEF 352 (including PFD for application discovery and application request information for multiple UE 302). The Nudr service-based interface is indicated by UDR 321 to allow UDM 358, PCF 356, and NEF 352 to access specific sets of stored data, as well as to read, update (e.g., add, modify), delete, and subscribe to notifications of relevant data changes within the UDR. The UDM may include a UDM-FE responsible for credential processing, location management, subscription management, etc. Several different frontends can serve the same user in different transactions. The UDM-FE accesses subscription information stored in the UDR and performs authentication credential processing, user identification processing, access authorization, registration / mobility management, and subscription management. In addition to communicating with other NFs through a reference point as shown in the diagram, the UDM 358 may represent a Nudm service-based interface.

[0113] AF 360 provides application influence on traffic routing, offers access to NEF, and can interact with the policy framework for policy control.

[0114] In some embodiments, the 5GC 340 may enable edge computing by selecting operator / third-party services to be geographically closer to the point where the UE 302 is attached to the network. This can reduce latency and load on the network. To provide an edge computing implementation, the 5GC 340 may select a UPF 348 closer to the UE 302 and perform traffic steering from the UPF 348 to the data network 336 via the N6 interface. This may be based on UE subscription data, UE location, and information provided by the AF 360. In this way, the AF 360 may influence UPF (re)selection and traffic routing. When the AF 360 is considered a trusted entity based on the operator deployment, the network operator may allow the AF 360 to interact directly with the relevant NF. In addition, the AF 360 may represent a NAF service-based interface.

[0115] The data network 336 may represent various network operator services, internet access, or third-party services that may be provided by one or more servers, including, for example, an application / content server 338.

[0116] The LMF 562 can receive measurement information (e.g., measurement reports) from the NG-RAN 314 and / or UE 302 via the AMF 344. The LMF 362 may use the measurement information to determine the device's position for indoor and / or outdoor positioning.

[0117] Figure 4 schematically shows a wireless network 400 according to one or more exemplary embodiments of the present disclosure.

[0118] The wireless network 400 may include a UE 402 that is wirelessly communicating with AN 404. UE 402 and AN 404 are similar to and substantially interchangeable components of similar names described elsewhere in this specification.

[0119] UE 402 can be communicatively coupled with AN 404 via connection 406. Connection 406 is shown as an air interface to enable communicative coupling and can be compatible with cellular communication protocols such as LTE or 5G NR protocols operating on mm wave or sub-6GHz frequencies.

[0120] The UE 402 may include a host platform 408 coupled to a modem platform 410. The host platform 408 may include an application processing circuit 412 that can be coupled to a protocol processing circuit 414 of the modem platform 410. The application processing circuit 412 can run various applications for the UE 402 to source / sink application data. The application processing circuit 412 may further implement one or more layer operations to send / receive application data to / from a data network. These layer operations may include transport (e.g., UDP) and internet (e.g., IP) operations.

[0121] The protocol processing circuit 414 may implement one or more layer operations to facilitate the transmission or reception of data through connection 406. The layer operations implemented by the protocol processing circuit 414 may include, for example, MAC, RLC, PDCP, RRC, and NAS operations.

[0122] The modem platform 410 may further include a digital baseband circuit 416 that can implement one or more layer operations, which are “lower” layer operations performed by the protocol processing circuit 414 in the network protocol stack. These operations may include PHY operations, for example, one or more of the following: HARQ-ACK functionality, scrambling / descrambling, encoding / decoding, layer mapping / demapping, modulation symbol mapping, received symbol / bit metric determination, multi-antenna port pre-coding / decoding which may include one or more of spatial time, spatial frequency, or spatial coding, reference signal generation / detection, preamble sequence generation and / or decoding, synchronous sequence generation / detection, control channel signal blind decoding, and other related functions.

[0123] The modem platform 410 may further include a transmitting circuit 418, a receiving circuit 420, an RF circuit 422, and an RF front end (RFFE) 424, the RFFE of which may include or be connected to one or more antenna panels 426. In short, the transmitting circuit 418 may include a digital-to-analog converter, a mixer, an intermediate frequency (IF) component, etc.; the receiving circuit 420 may include an analog-to-digital converter, a mixer, an IF component, etc.; the RF circuit 422 may include a low-noise amplifier, a power amplifier, a power tracking component, etc.; the RFFE 424 may include filters (e.g., surface / volume acoustic wave filters), switches, an antenna tuner, a beamforming component (e.g., a phase array antenna component), etc. The selection and arrangement of components for the transmitting circuit 418, receiving circuit 420, RF circuit 422, RFFE 424, and antenna panel 426 (collectively referred to as the “transmitting and receiving components”) may be specific to particular implementation details, such as whether the communication is TDM or FDM at millimeter-wave or sub-6 GHz frequencies. In some embodiments, the transmitting and receiving components may be arranged in multiple parallel transmit and receive chains, or they may be located on the same or different chips / modules, and so on.

[0124] In some embodiments, the protocol processing circuit 414 may include one or more instances of a control circuit (not shown) that provides control functions for the transmit / receive components.

[0125] UE reception may be established by and through the antenna panel 426, RFFE 424, RF circuit 422, receiving circuit 420, digital baseband circuit 416, and protocol processing circuit 414. In some embodiments, the antenna panel 426 may receive transmissions from AN 404 by receiving beamforming signals received by multiple antennas / antenna elements of one or more antenna panels 426.

[0126] UE transmission can be established by and through the protocol processing circuit 414, the digital baseband circuit 416, the transmitting circuit 418, the RF circuit 422, the RFFE 424, and the antenna panel 426. In some embodiments, the transmitting component of UE 404 may apply a spatial filter to the data to be transmitted in order to form a transmit beam emitted by the antenna elements of the antenna panel 426.

[0127] Similar to UE 402, AN 404 may include a host platform 428 coupled to a modem platform 430. The host platform 428 may include an application processing circuit 432 coupled to the protocol processing circuit 434 of the modem platform 430. The modem platform may further include a digital baseband circuit 436, a transmit circuit 438, a receive circuit 440, an RF circuit 442, an RFFE circuit 444, and an antenna panel 446. The components of AN 404 are similar to and substantially interchangeable with similarly named components of UE 402. In addition to performing data transmission / reception as described above, the components of AN 408 can perform a variety of logical functions, including RNC functions such as radio bearer management, uplink and downlink dynamic radio resource management, and data packet scheduling.

[0128] Figure 5 is a block diagram 500 showing components according to one or more exemplary embodiments of the present disclosure.

[0129] The components may be capable of reading instructions from a machine-readable medium or a computer-readable medium (e.g., a non-temporary machine-readable storage medium) and executing any one or more of the methods described herein. Specifically, Figure 5 shows a schematic diagram of hardware resources including one or more processors (or processor cores) 510, one or more memory / storage devices 520, and one or more communication resources 530, each of which may be communicatively coupled via a bus 540 or other interface circuitry. In embodiments where node virtualization (e.g., NFV) is utilized, a hypervisor 502 may be executed to provide an execution environment in which one or more network slices / subslice utilize the hardware resources.

[0130] The processor 510 may include, for example, processors 512 and 514. The processor 510 may be, for example, a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a DSP such as a baseband processor, an ASIC, an FPGA, a radio frequency integrated circuit (RFIC), another processor (including those described herein), or any suitable combination thereof.

[0131] The memory / storage device 520 may include main memory, disk storage, or any suitable combination thereof. The memory / storage device 520 may include, but is not limited to, any type of volatile, non-volatile, or semi-volatile memory, such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or solid-state storage.

[0132] The communication resource 530 may include interconnects or network interface controllers, components, or other suitable devices for communicating with one or more peripheral devices 504 or one or more databases 506 or other network elements via the network 508. For example, the communication resource 530 may include wired communication components (for coupling via USB, Ethernet®, etc.), cellular communication components, NFC components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components, and other communication components.

[0133] Instruction 550 may include other executable code causing at least one of the following to perform one or more of the methods described herein: software, programs, applications, applets, apps, or processor 510. Instruction 550 may reside entirely or partially in at least one of the following: processor 510 (for example, in the processor's cache memory), memory / storage device 520, or any preferred combination thereof. Furthermore, any part of instruction 550 may be transferred to hardware resources from any combination of peripheral device 504 or database 506. Thus, the memory of processor 510, memory / storage device 520, peripheral device 504, and database 506 are examples of computer-readable and machine-readable media.

[0134] In one or more embodiments, at least one of the components shown in one or more of the aforementioned figures may be configured to perform one or more operations, techniques, processes, and / or methods as described in the following exemplary section. For example, the baseband circuit described above in relation to one or more of the aforementioned figures may be configured to operate according to one or more of the examples described below. In another example, the circuit related to the UE, base station, network element, etc., described above in relation to one or more of the aforementioned figures may be configured to operate according to one or more of the examples described in the following exemplary section.

[0135] The term “exemplary” is used herein to mean “serving as an example, case, or illustration.” Any embodiment described herein as “exemplary” should not necessarily be construed as being preferable or advantageous to other embodiments. The terms “computing device,” “user device,” “communication station,” “station,” “handheld device,” “mobile device,” “wireless device,” and “user equipment” (UE) as used herein refer to wireless communication devices such as cellular telephones, smartphones, tablets, netbooks, wireless terminals, laptop computers, femtocells, high data rate (HDR) subscriber stations, access points, printers, point-of-sale devices, access terminals, or other personal communications system (PCS) devices. Devices may be mobile or stationary.

[0136] As used herein, the term “communicate” is intended to include transmitting, receiving, or both transmitting and receiving. This may be particularly useful in claims when describing an organization of data transmitted by one device and received by another device, but where only the functionality of one of those devices is required to infringe the claim. Similarly, a bidirectional exchange of data between two devices (both devices transmitting and receiving during the exchange) may be described as “communicating” when only the functionality of one of those devices is claimed. As used herein with respect to wireless communication signals, the term “communicate” includes transmitting a wireless communication signal and / or receiving a wireless communication signal. For example, a wireless communication unit capable of communicating a wireless communication signal may include a wireless transmitter for transmitting a wireless communication signal to at least one other wireless communication unit and / or a wireless communication receiver for receiving a wireless communication signal from at least one other wireless communication unit.

[0137] Where used herein, unless otherwise specified, the use of ordinal adjectives such as "first," "second," "third," etc., to describe a common object merely indicates that different instances of a similar object are being referred to, and is not intended to imply that the objects described in this way must be in a given order in time, space, ranking, or any other way.

[0138] As used herein, the term “access point” (AP) may refer to a fixed station. Access points may also be referred to as access nodes, base stations, evolved node B (enode B), or any other similar terms known in the art. Access terminals may also be referred to as mobile stations, user equipment (UE), wireless communication devices, or any other similar terms known in the art. The embodiments disclosed herein generally relate to wireless networks. Some embodiments may relate to wireless networks operating in accordance with one of the IEEE 802.11 standards.

[0139] Some embodiments can be used in conjunction with various devices and systems, such as personal computers (PCs), desktop computers, mobile computers, laptop computers, notebook computers, tablet computers, server computers, handheld computers, handheld devices, personal digital assistant (PDA) devices, handheld PDA devices, onboard devices, offboard devices, hybrid devices, vehicle devices, non-vehicle devices, mobile or portable devices, consumer devices, non-mobile or non-portable devices, wireless communication stations, wireless communication devices, wireless access points (APs), wired or wireless routers, wired or wireless modems, video devices, audio devices, audio-video (A / V) devices, wired or wireless networks, wireless area networks, wireless video area networks (WVANs), local area networks (LANs), wireless LANs (WLANs), personal area networks (PANs), wireless PANs (WPANs), and the like.

[0140] Some embodiments may be used in connection with one-way and / or two-way wireless communication systems, cellular radiotelephone communication systems, mobile phones, cellular phones, wireless phones, personal communication system (PCS) devices, PDA devices incorporating wireless communication devices, mobile or portable global positioning system (GPS) devices, devices incorporating GPS receivers or transceivers or chips, devices incorporating RFID elements or chips, multiple input multiple output (MIMO) transceivers or devices, single input multiple output (SIMO) transceivers or devices, multiple input single output (MISO) transceivers or devices, devices having one or more internal and / or external antennas, digital video broadcast (DVB) devices or systems, multistandard wireless devices or systems, wired or wireless handheld devices, such as smartphones, wireless application protocol (WAP) devices, and the like.

[0141] Some embodiments include, for example, radio frequency (RF), infrared (IR), frequency division multiplexing (FDM), quadrature FDM (OFDM), time division multiplexing (TDM), time division multiple access (TDMA), extended TDMA (E-TDMA), general-purpose packet radio service (GPRS), extended GPRS, code division multiple access (CDMA), broadband CDMA (WCDMA®), and CDMA. This includes 2000, single-carrier CDMA, multi-carrier CDMA, multi-carrier modulation (MDM), discrete multitone (DMT), Bluetooth®, Global Positioning System (GPS), Wi-Fi, Wi-Max, ZigBee, ultra-wideband (UWB), Global Mobile Communication System (GSM®), 2G, 2.5G, 3G, 3.5G, 4G, fifth-generation (5G) mobile networks, 3GPP, Long-Term Evolution (LTE), LTE Advanced, and GSM® Evolution Enhanced Data Rate (EDGE). Other embodiments may be used in a variety of other devices, systems, and / or networks.

[0142] Various embodiments are described below. Example 1 is a system of a radio access network (RAN) node B device for facilitating machine learning operations on a user equipment (UE) device, the system having a processing circuit coupled to storage, the processing circuit comprising: causing the node B device to send an instruction to the UE device that the node B device supports machine learning; identifying a service registration received from the UE device indicating that the UE device requests machine learning support and a machine learning model from the node B device; causing the node B device to send a request for information relating to the UE device, the information being associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; identifying the information received from the UE device based on the request for the information; causing the node B device to send the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, the machine learning model, machine learning model parameters, and machine learning configuration being based on the information; The system may be configured such that the node B device performs the step of causing the UE device to send updates to the machine learning model, the machine learning model parameters, or the machine learning configuration for use by the UE device. Example 2 may be a system described in Example 1 and / or any other example herein, wherein the machine learning model, the machine learning parameters, and the machine learning configuration are transmitted using a signal-transmitting radio bearer. Example 3 may be a system described in Example 1 and / or any other example herein, wherein the machine learning model, the machine learning parameters, and the machine learning configuration are transmitted using a wireless bearer dedicated to machine learning transmissions. Example 4 is a system described in Example 2 or 3 and / or any other embodiment of this specification, wherein the processing circuit may further be configured to cause the node B device to transmit a radio resource control (RRC) configuration to the UE device relating to controlling the machine learning operation of the UE device using the machine learning configuration. Example 5 is a system described in Example 1 and / or any other embodiment of this specification, wherein setting the time period further includes the following, and the processing circuit further: The node B device is configured to cause the UE device to send a policy to the UE device that includes an action for the UE device to perform, the policy is associated with the machine learning configuration, and the policy may allow the UE device to select from a plurality of actions based on the policy. Example 6 is a system described in Example 1 and / or any other embodiment of this specification, wherein the processing circuit is further configured to perform: a step of identifying a request received from the UE for taking action based on the results of the UE device's use of the machine learning model; and a step of causing the node B device to send a response to the UE device based on the request, wherein the response confirms or rejects the action. Example 7 is a system described in Example 1 and / or any other embodiment of this specification, wherein the hardware capability includes an indication of whether the processor of the UE device supports machine learning, and further includes at least one of the following: processor type, maximum battery capacity, current battery state of the UE device, or batching data size associated with the UE device. Example 8 is a system described in Example 1 and / or any other example herein, wherein the machine learning capability includes the type of machine learning model supported by the UE device, and further includes at least one of the following: maximum machine learning model size, supported libraries, machine learning model training capability, or machine learning model inference capability. Example 9 may be a system described in Example 1 and / or any other example herein, wherein the RRC message includes the machine learning configuration. Example 10 may be a system described in Example 1 and / or any other embodiment of the Spec., wherein the RRC message further includes a service type indicator indicating that the UE device is not required to synchronize the machine learning configuration with the Node B device, and the UE device is permitted to train the machine learning model associated with the action space of the machine learning configuration. Example 11 may be a system described in Example 9 and / or any other example herein, wherein the RRC message further includes a service type indicator indicating the service to which the machine learning configuration is associated and that the Node B device is training a machine learning model associated with the machine learning configuration. Example 12 may be a system described in Example 9 and / or any other embodiment of this specification, wherein the RRC message further includes a service type indicator indicating that the UE device needs to register with the Node B device for a service associated with the machine learning configuration. Example 13 is a system described in Example 1 and / or any other embodiment of this specification, wherein the processing circuit is further configured to identify a machine learning report received from the UE device, the machine learning report may include a measurement prediction related to the machine learning configuration, performance feedback related to the machine learning configuration, and a requested action for the UE to perform based on the performance feedback. Example 14 is a system described in Example 1 and / or any other embodiment of this specification, wherein the processing circuit is further configured to identify a second machine learning configuration received from the UE device, and the second machine learning configuration includes updates to the machine learning configuration. Example 15 may be a system described in Example 1 and / or any other embodiment of the Spec. wherein the processing circuit is further configured to perform: a step of determining at least one of machine learning model bias, machine learning model variance, machine learning model confidence level, or feedback relating to the use of the machine learning model; a step of generating an update to the machine learning configuration for use by the UE device based on the at least one of the machine learning model bias, machine learning model variance, machine learning model confidence level, or feedback; and a step of causing the node B device to transmit the update to the machine learning configuration to the UE device. Example 16 is a system described in Example 1 and / or any other embodiment of this specification, wherein the processing circuit is further configured to: identify an update request received from the UE device requesting an updated machine learning configuration, and generating the update to the machine learning configuration may be a system based on the update request. Example 17 may be a system described in Example 1 and / or any other embodiment of this specification, wherein the system information block includes the instruction that the node B device supports machine learning, and the instruction that the node B device supports machine learning includes an instruction for machine learning capabilities and an instruction for machine learning services provided by the node B device. Example 18 may be a system described in Example 17 and / or any other example herein, wherein the service registration includes a requested machine learning service from among the machine learning services. Example 19 is a system described in Example 1 and / or any other example herein, wherein the machine learning configuration may include the service type of the machine learning model, a model bias threshold, a model variance threshold, instructions on whether the machine learning model is maintained and trained by the Node B device or the UE device, and instructions for the machine learning report configuration. Example 20 is a system described in Example 19 and / or any other example herein, wherein the machine learning report configuration includes instructions for a machine learning report, a report period and an offset, and a memory duration and start time for the machine learning results. Example 21 is a computer-readable storage medium having instructions, wherein the instructions cause a processing circuit of a Wireless Access Network (RAN) Node B device to: when the processing circuit executes the instructions: the Node B device to send an instruction to a User Equipment (UE) device that the Node B device supports machine learning; the Node B device to identify a service registration received from the UE device indicating that the UE device requests machine learning support and a machine learning model from the Node B device; the Node B device to send a request for information relating to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; the Node B device to identify the information received from the UE based on the request for information; and the Node B device to send the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; The node B device may be a computer-readable storage medium that causes the node B device to perform the step of causing the UE device to send updates to the machine learning model, the machine learning model parameters, or the machine learning configuration for use by the UE device. Example 22 is a computer-readable medium as described in Example 21 and / or any other example herein, wherein the machine learning model, the machine learning parameters, and the machine learning configuration may be transmitted using a wireless bearer dedicated to machine learning transmissions. Example 23 is a method for facilitating machine learning operations on a user equipment (UE) device, comprising: a step of: a processing circuit in a radio access network (RAN) node B device causing the node B device to transmit an instruction to the UE device that the node B device supports machine learning; a step of the processing circuit identifying a service registration received from the UE device indicating that the UE device requests machine learning support and a machine learning model from the node B device; a step of the processing circuit causing the node B device to transmit a request for information relating to the UE device to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; a step of the processing circuit identifying the information received from the UE based on the request for the information; a step of the processing circuit causing the node B device to transmit the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; The method may include the step of causing the node B device to send updates to the machine learning model, the machine learning model parameters, or the machine learning configuration to the UE device for use by the UE device. Embodiment 24 may be an apparatus having means for performing the following steps: causing a node B device to send an instruction to a UE device that the node B device supports machine learning; identifying a service registration received from the UE device indicating that the UE device requests machine learning support and a machine learning model from the node B device; causing the node B device to send a request for information relating to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; identifying the information received from the UE based on the request for the information; causing the node B device to send the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; and causing the node B device to send an update to the machine learning model, machine learning model parameters, or machine learning configuration to the UE device for use by the UE device. Example 25 may include one or more non-temporary computer-readable media containing instructions that cause an electronic device to execute one or more elements of any other method or process described herein, or any other method described herein, when an instruction is executed by one or more processors of the electronic device. Example 26 may include an apparatus having logic, modules, and / or circuits that perform one or more elements of any other method or process described herein, or any other method or process described herein. Example 27 may include any method, technique, or process described in or related to any one of Examples 1 through 24, or any part or portion thereof. Example 28 may include a device having one or more processors and one or more computer-readable media containing instructions, which, when executed by the one or more processors, cause the one or more processors to execute a method, technique, or process, or part thereof, described in or related to any one of Examples 1 to 24. Example 29 may include a method of communication in a wireless network as shown and described herein. Example 30 may include a system for providing wireless communication as shown and described herein. Example 31 may include a device for providing wireless communication as shown and described herein.

[0143] The embodiments provided herein are disclosed in particular in the appended claims directed to methods, storage media, devices, and computer program products, and any feature referred to in one claim category, for example, a method, may also be claimed in another claim category, for example, a system. Dependencies or references in the appended claims are selected for formal reasons only. However, any subject matter arising from intentional references to any prior claims (in particular, multiple dependencies) may also be claimed. Thus, regardless of the dependencies selected in the appended claims, any combination of claims and their features is disclosed and can be claimed. The subject matter that can be claimed includes not only the combinations of features described in the appended claims, but also any other combination of features in the claims, and each feature referred to in the claims may be combined with any other feature or combination of features in the claims. Furthermore, any embodiment and feature described or depicted herein may be claimed in a separate claim and / or in any combination with any embodiment or feature described or depicted herein, or with any of the features of the appended claims.

[0144] The above description of one or more implementations is illustrative and descriptive, but is not intended to be exhaustive or to limit the scope of embodiments to the exact forms disclosed. Modifications and variations are possible in light of the above teachings or can be obtained from the implementation of various embodiments.

[0145] Certain aspects of this disclosure have been described above with reference to block diagrams and flow diagrams of systems, methods, devices, and / or computer program products in various implementation forms. It will be understood that one or more blocks in block diagrams and flow diagrams, as well as combinations of blocks within block diagrams and flow diagrams, can each be implemented by computer executable program instructions. Similarly, some blocks in block diagrams and flow diagrams may not necessarily be executed in the order presented, or may not be executed at all, depending on the implementation.

[0146] These computer executable program instructions may be loaded onto a special-purpose computer or other specific machine, processor, or other programmable data processing device, thereby generating a specific machine such that the instructions executed on the computer, processor, or other programmable data processing device create means for implementing one or more functions specified in one or more blocks of a flowchart. These computer program instructions may be stored in a computer-readable storage medium or memory on which the instructions can be directed to a computer or other programmable data processing device to function in a specific manner, thereby generating a product that includes instruction means for implementing one or more functions specified in one or more blocks of a flowchart. As an example, certain implementations can provide a computer program product comprising a computer-readable storage medium on which computer-readable program code or program instructions are implemented, the computer-readable program code being adapted to be executed to implement one or more functions specified in one or more blocks of a flowchart. Computer program instructions can also be loaded onto a computer or other programmable data processing device to generate a computer implementation process by executing a set of operating elements or steps on the computer or other programmable device, so that the instructions executed on the computer or other programmable device provide elements or steps for implementing a function specified in one or more blocks of a flowchart.

[0147] Therefore, the blocks in block diagrams and flowcharts support combinations of means for performing a specified function, combinations of elements or steps for performing a specified function, and means of program instructions for performing a specified function. It will also be understood that each block in block diagrams and flowcharts, as well as combinations of blocks in block diagrams and flowcharts, may be implemented by a special-purpose hardware-based computer system that performs a specified function, element or step, or combination of special-purpose hardware and computer instructions.

[0148] In particular, conditional language such as “can,” “has been,” “has been,” or “may,” unless otherwise stated or understood in the context in which they are used, is generally intended to convey that one implementation may include a certain feature, element, and / or behavior, while other implementations may not. Thus, such conditional language is not generally intended to imply that a feature, element, and / or behavior is required in some way for one or more implementations, or that one or more implementations necessarily include logic for determining whether these features, elements, and / or behaviors are included or performed in any particular implementation, with or without user input or prompting.

[0149] It will be apparent that many modifications and other implementations of the disclosure described herein have an interest in the teachings presented in the foregoing description and the relevant drawings. Therefore, it should be understood that this disclosure should not be limited to any specific implementation described herein, and that modifications and other implementations are intended to be included within the scope of the appended claims. Certain terms are used herein, but they are used in a general and descriptive sense only and are not intended to be limiting.

[0150] For the purposes of this specification, the following terms and definitions are applicable to the examples and embodiments discussed herein.

[0151] As used herein, the term “circuit” refers to, is part of, or includes hardware components configured to provide the functionality described, such as electronic circuits, logic circuits, processors (shared, dedicated, or group) and / or memory (shared, dedicated, or group), application-specific integrated circuits (ASICs), field-programmable devices (FPDs) (e.g., field-programmable gate arrays (FPGAs), programmable logic devices (PLDs), composite PLDs (CPLDs), high-performance PLDs (HCPLDs), constructed ASICs, or programmable SoCs), and digital signal processors (DSPs). In some embodiments, a circuit may run one or more software or firmware programs to provide at least some of the functionality described. The term “circuit” may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) having program code used to perform the functionality of the program code. In these embodiments, a combination of hardware elements and program code may be referred to as a particular type of circuit.

[0152] As used herein, the term “processor circuit” refers to, is part of, or includes a circuit capable of sequentially and automatically performing a series of arithmetic or logical operations, or recording, storing, and / or transferring digital data. A processing circuit may include one or more processing cores for executing instructions and one or more memory structures for storing program and data information. The term “processor circuit” may also refer to one or more application processors, one or more baseband processors, physical central processing units (CPUs), single-core processors, dual-core processors, triple-core processors, quad-core processors, and / or any other devices capable of executing or otherwise operating computer-executable instructions, such as program code, software modules, and / or functional processes. A processing circuit may include more hardware accelerators, such as microprocessors and programmable processing devices. One or more hardware accelerators may include, for example, computer vision (CV) and / or deep learning (DL) accelerators. The terms “application circuit” and / or “baseband circuit” may be considered synonymous with “processor circuit” and may also be referred to as “processor circuit.”

[0153] As used herein, the term “interface circuit” refers to, is part of, or includes a circuit that enables the exchange of information between two or more components or devices. The term “interface circuit” may refer to one or more hardware interfaces, such as a bus, an I / O interface, a peripheral component interface, a network interface card, and / or others.

[0154] As used herein, the terms “User Equipment” or “UE” refer to a device having wireless communication capabilities and may represent a remote user of network resources within a communication network. The terms “User Equipment” or “UE” may be considered synonymous with, and may be referred to as, client, mobile, mobile device, mobile terminal, user terminal, mobile unit, mobile station, mobile user, subscriber, user, remote station, access agent, user agent, receiver, radio equipment, reconfigurable radio equipment, reconfigurable mobile device, etc. Furthermore, the terms “User Equipment” or “UE” may include any computing device including any type of wireless / wired device or wireless communication interface.

[0155] As used herein, the term “Network Element” refers to physical or virtualized equipment and / or infrastructure used to provide wired or wireless communication network services. The term “Network Element” may be considered synonymous with, and / or referred to as, networked computers, networking hardware, network equipment, network nodes, routers, switches, hubs, bridges, wireless network controllers, RAN devices, RAN nodes, gateways, servers, virtualized VNFs, NFVIs, and / or others.

[0156] As used herein, the term “computer system” refers to any type of interconnected electronic devices, computer devices, or components thereof. Furthermore, the terms “computer system” and / or “system” may refer to various components of a computer that are interconnected by communication. Additionally, the terms “computer system” and / or “system” may refer to multiple computer devices and / or multiple computing systems that are interconnected by communication and configured to share computing and / or networking resources.

[0157] As used herein, terms such as “appliance” and “computer appliance” refer to a computer device or computer system having program code (e.g., software or firmware) specifically designed to provide a particular computing resource. A “virtual appliance” is a virtual machine image implemented by a hypervisor-based device that virtualizes or emulates a computer appliance, or is otherwise dedicated in a way to provide a particular computing resource.

[0158] As used herein, the term “resource” refers to physical or virtual devices, physical or virtual components, and / or physical or virtual components within a particular device in a computing environment, including computer devices, mechanical devices, memory space, processor / CPU time, processor / CPU usage, processor and accelerator load, hardware time or usage, power, input / output operation, ports or network sockets, channel / link allocation, throughput, memory usage, storage, networks, databases and applications, and workload units. “Hardware resources” may refer to computing, storage, and / or network resources provided by a physical hardware element(s). “Virtualized resources” may refer to computing, storage, and / or network resources provided to applications, devices, systems, etc., by a virtualization infrastructure. The term “network resources” or “communication resources” may refer to resources accessible by computer devices / systems over a communication network. The term “system resources” may refer to any kind of shared entity providing services, and may include computing and / or network resources. System resources can be considered as a set of coherent functions, network data objects, or services that reside on a single host or multiple hosts and are accessible via a clearly identifiable server.

[0159] As used herein, the term “channel” refers to any tangible or intangible transmission medium used to communicate data or data streams. The term “channel” may be synonymous and / or equivalent to any other similar term indicating a path or medium through which data is communicated, such as “communication channel,” “data communication channel,” “transmission channel,” “data transmission channel,” “access channel,” “data access channel,” “link,” “data link,” “carrier,” “radio frequency carrier,” and / or any other similar term. Furthermore, as used herein, the term “link” refers to a connection between two devices via a RAT for the purpose of transmitting and receiving information.

[0160] As used herein, terms such as "instantiate" and "instantiate" refer to the creation of an instance. An "instance" can also refer to the specific occurrence of an object, for example, during the execution of program code.

[0161] The terms “coupled” and “communicationally coupled” are used herein together with their derivatives. The term “coupled” may mean that two or more elements are in direct physical or electrical contact with each other, that two or more elements are indirectly in contact with each other but still cooperate or interact with each other, and / or that one or more other elements are coupled or connected between elements said to be coupled to each other. The term “directly coupled” may mean that two or more elements are in direct contact with each other. The term “communicationally coupled” may mean that two or more elements are in contact with each other by means of communication, such as through wires or other interconnections, through wireless communication channels or links.

[0162] The term "information element" refers to a structural element that contains one or more fields. The term "field" refers to the individual contents of an information element, or a data element that contains contents.

[0163] Unless otherwise used herein, terms, definitions, and abbreviations may be consistent with those defined in 3GPP TR 21.905 v16.0.0 (2019-06) and / or any other 3GPP standard. For the purposes of this paper, the following abbreviations (shown in Table 4) may apply to the examples and embodiments discussed herein.

[0164] Table 4: Abbreviations 3GPP Third Generation Partnership Project 4G Fourth Generation 5G Fifth Generation 5GC 5G Core network AC Application Client ACK (Acknowledgement) ACID Application Client Identification AF Application Function AM Acknowledged Mode AMBR Aggregate Maximum Bit Rate AMF Access and Mobility Management Function AN Access Network ANR Automatic Neighbor Relation AP (Application Protocol), Antenna Port, Access Point API Application Programming Interface APN (Access Point Name) ARP Allocation and Retention Priority ARQ Automatic Repeat Request Automatic repeat request AS Access Stratum Access Layer ASP Application Service Provider ASN.1 Abstract Syntax Notation One AUSF Authentication Server Function AWGN Additive White Gaussian Noise BAP (Backhaul Adaptation Protocol) BCH Broadcast Channel BER (Bit Error Ratio) BFD Beam Failure Detection BLER Block Error Rate BPSK (Binary Phase Shift Keying) - 2-state phase shift keying BRAS Broadband Remote Access Server BSS Business Support System BS Base Station BSR Buffer Status Report BW Bandwidth BWP Bandwidth Part C-RNTI Cell Radio Network Temporary Identity CA (Carrier Aggregation), Certification Authority CAPEX CAPital EXpenditure Capital Expenditure CBRA Contention-Based Random Access CC Component Carrier, Country Code, Cryptographic Checksum CCA Clear Channel Assessment (Available Channel Assessment) CCE Control Channel Element CCCH Common Control Channel CE Coverage Enhancement CDM (Content Delivery Network) CDMA Code-Division Multiple Access CFRA Contention Free Random Access CG Cell Group CGF Charging Gateway Function CHF Charging Function CI cell characteristics [identification information] CID Cell-ID Cell ID (e.g., positioning method) CIM Common Information Model CIR Carrier-to-Interference Ratio CK Cipher Key CM Connection Management, Conditional Mandatory CMAS Commercial Mobile Alert Service CMD Command CMS Cloud Management System CO Conditional Optional CoMP Coordinated Multi-Point CORESET Control Resource Set Control resource set COTS Commercial Off-The-Shelf CP Control Plane, Cyclic Prefix, Connection Point CPD Connection Point Descriptor CPE Customer Premise Equipment CPICH Common Pilot Channel CQI Channel Quality Indicator CPU (CSI processing unit), Central Processing Unit (CSI processing unit) C / R Command / Response field bit CRAN (Cloud Radio Access Network) CRB Common Resource Block CRC Cyclic Redundancy Check Cyclic Redundancy Check CRI Channel-State Information Resource Indicator, CSI-RS Resource Indicator C-RNTI Cell RNTI CS Circuit Switched CSAR Cloud Service Archive CSI Channel-State Information CSI-IM CSI Interference Measurement CSI Interference Measurement CSI-RS CSI Reference Signal CSI reference signal CSI-RSRP CSI reference signal received power CSI reference signal received power CSI-RSRQ CSI reference signal received quality CSI reference signal received quality CSI-SINR CSI signal-to-noise and interference ratio CSMA Carrier Sense Multiple Access CSMA / CA CSMA with collision avoidance CSS Common Search Space, Cell-specific Search Space CTF Charging Trigger Function Charge Trigger Function CTS Clear-to-Send Clear to Send CW Codeword Codeword CWS Contention Window Size Contention Window Size D2D Device-to-Device Device-to-Device DC Dual Connectivity Dual Connectivity, Direct Current Direct Current DCI Downlink Control Information Downlink Control Information DF Deployment Flavour Deployment Flavour DL Downlink Downlink DMTF Distributed Management Task Force Distributed Management Task Force DPDK Data Plane Development Kit Data Plane Development Kit DM-RS, DMRS Demodulation Reference Signal Demodulation Reference Signal DN Data network Data network DNN Data Network Name Data Network Name DNAI Data Network Access Identifier Data Network Access Identifier DRB Data Radio Bearer Data Radio Bearer DRS Discovery Reference Signal Discovery Reference Signal DRX Discontinuous Reception Discontinuous Reception DSL Domain Specific Language Domain Specific Language, Digital Subscriber Line Digital Subscriber Line DSLAM DSL Access Multiplexer DSL Access Multiplexer DwPTS Downlink Pilot Time Slot E-LAN ​​Ethernet Local Area Network E2E End-to-End ECCA extended clear channel assessment, extended CCA ECCE Enhanced Control Channel Element ED Energy Detection EDGE Enhanced Datarates for GSM Evolution EAS Edge Application Server EASID (Edge Application Server Identification) ECS Edge Configuration Server ECSP (Edge Computing Service Provider) EDN (Edge Data Network) EEC Edge Enabler Client EECID Edge Enabler Client Identification EES Edge Enabler Server EESID: Edge Enabler Server Identification EHE Edge Hosting Environment EGMF Exposure Governance Table Management Function EGPRS Enhanced GPRS Enhanced GPRS EIR Equipment Identity Register eLAA (enhanced Licensed Assisted Access) EM, Element Manager eMBB Enhanced Mobile Broadband EMS Element Management System eNB evolved NodeB, E-UTRAN NodeB EN-DC E-UTRA-NR Dual Connectivity EPC Evolved Packet Core EPDCCH enhanced PDCCH, enhanced Physical Downlink Control Channel EPRE: Energy per resource element EPS Evolved Packet System EREG (enhanced REG), enhanced resource element groups ETSI (European Telecommunications Standards Institute) ETWS Earthquake and Tsunami Warning System eUICC embedded UICC, embedded Universal Integrated Circuit Card E-UTRA Evolved UTRA Evolved UTRA E-UTRAN Evolved UTRAN EV2X Improved V2X F1AP F1 Application Protocol F1-C F1 Control Plane Interface F1-U F1 User Plane Interface FACCH Fast Associated Control Channel FACCH / F Fast Associated Control Channel / Full rate FACCH / H Fast Associated Control Channel / Half rate FACH (Forward Access Channel) FAUSCH Fast Uplink Signalling Channel FB Functional Block FBI Feedback Information FCC (Federal Communications Commission) FCCH Frequency Correction CHannel Frequency correction channel FDD Frequency Division Duplex Frequency division duplex FDM Frequency Division Multiplex Frequency division multiplexing FDMA Frequency Division Multiple Access Frequency division multiple access FE Front End Front end FEC Forward Error Correction Forward error correction FFS For Further Study For further study FFT Fast Fourier Transformation Fast Fourier transform feLAA further enhanced Licensed Assisted Access Further enhanced licensed assisted access, further enhanced LAA, additional enhanced LAA FN Frame Number Frame number FPGA Field-Programmable Gate Array Field-programmable gate array FR Frequency Range Frequency range FQDN Fully Qualified Domain Name Fully qualified domain name G-RNTI GERAN Radio Network Temporary Identity GERAN radio network temporary identity GERAN GSM EDGE RAN GSM EDGE RAN, GSM EDGE Radio Access Network GSM EDGE radio access network GGSN Gateway GPRS Support Node Gateway GPRS support node GLONASS GLObal'naya NAvigatsionnaya Sputnikovaya Sistema (English: Global Navigation Satellite System) gNB Next Generation NodeB gNB-CU (gNB-centralized unit), Next Generation NodeB Centralized unit gNB-DU (gNB-distributed unit), Next Generation NodeB distributed unit GNSS Global Navigation Satellite System GPRS General Packet Radio Service GSM Global System for Mobile Communications, Groupe Spécial Mobile GSM Alliance GTP GPRS Tunneling Protocol GTP-U GPRS Tunneling Protocol for User Plane GTS Go To Sleep Signal (Sleep Transition Signal related to WUS) GUMMEI: Globally Unique MME Identifier GUTI: Globally Unique Temporary UE Identity HARQ Hybrid ARQ, Hybrid Automatic Repeat Request HANDO handover HFN HyperFrame Number HHO Hard Handover HLR Home Location Register HN Home Network Home Network HO Handover HPLMN Home Public Land Mobile Network HSDPA (High Speed ​​Downlink Packet Access) HSN Hopping Sequence Number HSPA High Speed ​​Packet Access HSS Home Subscriber Server HSUPA High Speed ​​Uplink Packet Access HTTP (Hypertext Transfer Protocol) HTTPS (Hypertext Transfer Protocol Secure) is a secure hypertext transfer protocol (HTTPS stands for http / 1.1 over SSL, i.e., port 443). I-Block Information Block ICCID (Integrated Circuit Card Identification) IAB Integrated Access and Backhaul ICIC Inter-Cell Interference Coordination ID Identity, Identifier Identification information, Identifier IDFT Inverse Discrete Fourier Transform IE Information Element IBE In-Band Emission IEEE Institute of Electrical and Electronics Engineers IEI Information Element Identifier IEIDL Information Element Identifier Data Length IETF Internet Engineering Task Force IF Infrastructure IM Interference Measurement, Intermodulation, IP Multimedia IMC IMS Credentials IMS Credentials IMEI (International Mobile Equipment Identity) IMGI International mobile group identity IMPI IP Multimedia Private Identity IMPU IP Multimedia Public Identity IMS IP Multimedia Subsystem IMSI International Mobile Subscriber Identity IoT (Internet of Things) IP Internet Protocol IPsec IP Security, Internet Protocol Security IP-CAN IP-Connectivity Access Network IP-M IP Multicast IP Multicast IPv4 Internet Protocol Version 4 IPv6 Internet Protocol Version 6 IR Infrared IS In Sync Syncing IRP Integration Reference Point ISDN Integrated Services Digital Network ISIM IM Services Identity Module ISO International Organization for Standardization ISP (Internet Service Provider) IWF Interworking Function I-WLAN Interworking WLAN Constraint length of the convolutional code, USIM individual key kB Kilobyte (1000 bytes) kbps: kilobits per second Kc Ciphering key Encryption key Individual subscriber authentication key KPI Key Performance Indicator KQI Key Quality Indicator KSI Key Set Identifier ksps kilo-symbols per second KVM (Kernel Virtual Machine) L1 Layer 1 (physical layer) L1-RSRP Layer 1 reference signal received power L2 Layer 2 (data link layer) L3 Layer 3 (Network Layer) LAA Licensed Assisted Access LAN (Local Area Network) LADN (Local Area Data Network) LBT Listen Before Talk LCM (Life Cycle Management) LCR Low Chip Rate LCS Location Services LCID: Logical Channel ID LI Layer Indicator LLC Logical Link Control, Low Layer Compatibility LPLMN Local PLMN Local PLMN LPP LTE Positioning Protocol LSB (Least Significant Bit) LTE Long Term Evolution LWA LTE-WLAN aggregation LWIP LTE / WLAN Radio Level Integration with IPsec Tunnel: LTE / WLAN radio level integration via IPsec tunnel. LTE Long Term Evolution M2M Machine-to-Machine MAC Medium Access Control (in the context of protocol layering) MAC Message authentication code (in the context of security / encryption) MAC-A MAC used for authentication and key agreement (in the context of TSG T WG3) MAC-I MAC used for data integrity: MAC used for data integrity in signaling messages (in the context of TSG T WG3) MANO Management and Orchestration MBMS (Multimedia Broadcast and Multicast Service) MBSFN Multimedia Broadcast Multicast Service Single Frequency Network MCC Mobile Country Code MCG Master Cell Group MCOT Maximum Channel Occupancy Time MCS Modulation and coding scheme MDAF Management Data Analytics Function MDAS Management Data Analytics Service Minimization of Drive Tests (MDT) ME Mobile Equipment Mobile Devices MeNB master eNB Master eNB MER Message Error Ratio MGL Measurement Gap Length MGRP Measurement Gap Repetition Period MIB Master Information Block, Management Information Base MIMO Multiple Input Multiple Output MLC Mobile Location Centre MM Mobility Management Mobility Management MME Mobility Management Entity MN Master Node MNO Mobile Network Operator MO: Measurement Object, Mobile Originated. MPBCH MTC Physical Broadcast Channel MPDCCH MTC Physical Downlink Control Channel MPDSCH MTC Physical Downlink Shared Channel MPRACH MTC Physical Random Access Channel MPUSCH MTC Physical Uplink Shared Channel MPLS (MultiProtocol Label Switching) MS Mobile Station Mobile station MSB (Most Significant Bit) MSC Mobile Switching Centre MSI Minimum System Information, MCH Scheduling Information MSID Mobile Station Identifier MSIN Mobile Station Identification Number MSISDN Mobile Subscriber ISDN Number MT Mobile Terminated, Mobile Termination MTC Machine-Type Communications mMTC massive MTC, massive Machine-Type Communications Mechanical large-scale machine-type communication MU-MIMO (Multi-User MIMO) MWUS MTC wake-up signal, MTC WUS MTC WUS NACK (Negative Acknowledgement) NAI (Network Access Identifier) NAS Non-Access Stratum, Non-Access Stratum layer NCT Network Connectivity Topology NC-JT Non-Coherent Joint Transmission NEC Network Capability Exposure: Network Function Disclosure NE-DC NR-E-UTRA Dual Connectivity NEF Network Exposure Function (Network Exposure Function) NF Network Function NFP Network Forwarding Path NFPD Network Forwarding Path Descriptor NFV (Network Functions Virtualization) NFVI NFV Infrastructure NFV Infrastructure NFVO NFV orchestrator NFV orchestrator NG Next Generation Next Generation, Next Gen NGEN-DC NG-RAN E-UTRA-NR Dual Connectivity NM Network Manager NMS Network Management System N-PoP Network Point of Presence NMIB, N-MIB, Narrowband MIB NPBCH (Narrowband Physical Broadcast Channel) NPDCCH Narrowband Physical Downlink Control Channel NPDSCH Narrowband Physical Downlink Shared Channel NPRACH Narrowband Physical Random Access Channel NPUSCH Narrowband Physical Uplink Shared Channel NPSS Narrowband Primary Synchronization Signal NSSS Narrowband Secondary Synchronization Signal NR (New Radio), Neighbor Relation NRF NF Repository Function NRS Narrowband Reference Signal NS Network Service NSA Non-Standalone operation mode NSD (Network Service Descriptor) NSR Network Service Record NSSAI Network Slice Selection Assistance Information S-NNSAI Single-NSSAI Single NSSAI NSSF Network Slice Selection Function NW Network NWUS (Narrowband WUS) NZP Non-Zero Power O&M Operation and Maintenance ODU2 Optical Channel Data Unit - Type 2 OFDM (Orthogonal Frequency Division Multiplexing) OFDMA (Orthogonal Frequency Division Multiple Access) OOB (Out-of-band) OOS (Out of Sync) OPEX: Operating Expenses OSI Other System Information OSS Operations Support System OTA over-the-air PAPR (Peak-to-Average Power Ratio) PAR (Peak to Average Ratio) PBCH (Physical Broadcast Channel) PC Power Control, Personal Computer PCC Primary Component Carrier, Primary CC PCell Primary Cell Main cell PCI Physical Cell ID, Physical Cell Identity PCEF Policy and Charging Enforcement Function PCF Policy Control Function PCRF Policy Control and Charging Rules Function PDCP (Packet Data Convergence Protocol) and Packet Data Convergence Protocol layer. PDCCH Physical Downlink Control Channel PDCP (Packet Data Convergence Protocol) PDN (Packet Data Network), Public Data Network PDSCH (Physical Downlink Shared Channel) PDU Protocol Data Unit PEI Permanent Equipment Identifiers PFD Packet Flow Description P-GW PDN Gateway PHICH Physical hybrid-ARQ indicator channel PHY Physical layer PLMN Public Land Mobile Network PIN (Personal Identification Number) PM Performance Measurement Performance measurement PMI Precoding Matrix Indicator PNF (Physical Network Function) PNFD (Physical Network Function Descriptor) PNFR (Physical Network Function Record) POC PTT over Cellular PP, PTP: Point-to-Point PPP (Point-to-Point Protocol) PRACH Physical RACH Physical RACH PRB Physical resource block PRG Physical resource block group ProSe (Proximity Services), Proximity-Based Service PRS Positioning Reference Signal PRR Packet Reception Radio PS Packet Services PSBCH Physical Sidelink Broadcast Channel PSDCH Physical Sidelink Downlink Channel PSCCH Physical Sidelink Control Channel PSSCH Physical Sidelink Shared Channel PSCell Primary SCell Primary SCell PSS Primary Synchronization Signal Primary synchronization signal PSTN Public Switched Telephone Network PT-RS Phase-tracking reference signal Phase-tracking reference signal PTT Push-to-Talk PUCCH Physical Uplink Control Channel PUSCH Physical Uplink Shared Channel QAM (Quadrature Amplitude Modulation) QCI QoS class of identifier QCL Quasi co-location (quasi-co-location) QFI QoS Flow ID, QoS Flow Identifier QoS (Quality of Service) QPSK Quadrature (Quaternary) Phase Shift Keying QZSS Quasi-Zenith Satellite System RA-RNTI (Random Access RNTI) RAB Radio Access Bearer, Random Access Burst RACH Random Access Channel RADIUS Remote Authentication Dial In User Service RAN Radio Access Network RAND: Random number (used for authentication) RAR (Random Access Response) RAT Radio Access Technology RAU Routing Area Update RB Resource block, Radio Bearer RBG Resource block group REG Resource Element Group Rel Release REQ REQuest request RF Radio Frequency RI Rank Indicator RIV Resource indicator value RL Radio Link RLC Radio Link Control, Radio Link Control layer RLC AM RLC Acknowledged Mode RLC UM RLC Unacknowledged Mode RLF Radio Link Failure RLM Radio Link Monitoring RLM-RS Reference Signal for RLM RM Registration Management RMC Reference Measurement Channel RMSI (Remaining MSI), Remaining Minimum System Information RN Relay Node RNC Radio Network Controller RNL (Radio Network Layer) RNTI (Radio Network Temporary Identifier) ROHC Robust Header Compression RRC Radio Resource Control Radio resource control, Radio Resource Control layer Wireless resource control layer RRM Radio Resource Management Radio resource management RS Reference Signal Reference signal RSRP Reference Signal Received Power RSRQ Reference Signal Received Quality RSSI Received Signal Strength Indicator RSU Road Side Unit RSTD Reference Signal Time Difference RTP (Real Time Protocol) RTS Ready-To-Send Ready to send RTT (Round Trip Time) Rx Reception, Receiving Receiver S1AP S1 Application Protocol S1-MME S1 for the control plane S1-U S1 for the user plane S-GW Serving Gateway S-RNTI SRNC Radio Network Temporary Identity S-TMSI SAE Temporary Mobile Station Identifier SAE Temporary Mobile Station Identifier SA Standalone operation mode SAE System Architecture Evolution SAP Service Access Point SAPD Service Access Point Descriptor SAPI Service Access Point Identifier SCC Secondary Component Carrier, Secondary CC SCell Secondary Cell SCEF Service Capability Exposure Function SC-FDMA (Single Carrier Frequency Division Multiple Access) SCG Secondary Cell Group SCM Security Context Management SCS Subcarrier Spacing SCTP Stream Control Transmission Protocol SDAP (Service Data Adaptation Protocol) - Service Data Adaptation Protocol Layer SDL Supplementary Downlink SDNF (Structured Data Storage Network Function) SDP Session Description Protocol SDSF (Structured Data Storage Function) SDU Service Data Unit SEAF Security Anchor Function SeNB secondary eNB SEPP Security Edge Protection Proxy SFI Slot Format Indication SFTD (Space-Frequency Time Diversity), SFN (Space-Frequency Network) and frame timing difference SFN System Frame Number SgNB Secondary gNB Secondary gNB SGSN Serving GPRS Support Node S-GW Serving Gateway SI System Information SI-RNTI System Table RNTI System Information RNTI SIB System Information Block SIM Subscriber Identity Module SIP Session Initiation Protocol SiP System in Package SL Sidelink SLA (Service Level Agreement) SM Session Management SMF Session Management Function SMS Short Message Service SMSF SMS Function SMTC SSB-based Measurement Timing Configuration SN Secondary Node, Sequence Number SoC (System on Chip) SON Self-Organizing Network SpCell Special Cell SP-CSI-RNTI Semi-Persistent CSI RNTI SPS Semi-Persistent Scheduling SQN Sequence number SR Scheduling Request SRB Signaling Radio Bearer SRS Sounding Reference Signal Detection reference signal SS Synchronization Signal Synchronization signal SSB Synchronization Signal Block SSID (Service Set Identifier) SS / PBCH Block SS / PBCH Block SSBRI SS / PBCH Block Resource Indicator, Syncheronization Signal Block Resource Indicator SSC Session and Service Continuity SS-RSRP Synchronization Signal based Reference Signal Received Power SS-RSRQ Synchronization Signal based Reference Signal Received Quality SS-SINR Synchronization Signal-based Signal-to-Noise and Interference Ratio SSS Secondary Synchronization Signal SSSG Search Space Set Group SSSIF Search Space Set Indicator SST Slice / Service Types SU-MIMO (Single User MIMO) SUL Supplementary Uplink TA Timing Advance, Tracking Area TAC Tracking Area Code TAG Timing Advance Group TAI Tracking Area Identity TAU Tracking Area Update TB Transport Block TBS Transport Block Size TBD To Be Defined TCI Transmission Configuration Indicator TCP Transmission Communication Protocol TDD (Time Division Duplex) TDRA Time Domain Resource Allocation TDM Time Division Multiplexing TDMA (Time Division Multiple Access) TE Terminal Equipment TEID: Tunnel End Point Identifier TFT Traffic Flow Template TMSI (Temporary Mobile Subscriber Identity) TNL (Transport Network Layer) TPC Transmit Power Control TPMI Transmitted Precoding Matrix Indicator TR Technical Report Technical report TRP, TRxP Transmission Reception Point TRS Tracking Reference Signal TRx Transceiver TS Technical Specifications, Technical Standards TTI Transmission Time Interval Tx Transmission, Transmitting, Transmission, Transmitter U-RNTI UTRAN Radio Network Temporary Identity UART Universal Asynchronous Receiver and Transmitter UCI Uplink Control Information UE User Equipment UDM Unified Data Management: Centralized Data Management UDP User Datagram Protocol UDSF (Unstructured Data Storage Network Function) UICC Universal Integrated Circuit Card UL Uplink UM Unacknowledged Mode (No Acknowledgment Response Mode) UML (Unified Modeling Language) UMTS Universal Mobile Telecommunications System UP User Plane UPF User Plane Function URI Uniform Resource Identifier URL Uniform Resource Locator URLLC Ultra-Reliable and Low Latency USB Universal Serial Bus USIM Universal Subscriber Identity Module USS UE-specific search space UTRA UMTS Terrestrial Radio Access UTRAN (Universal Terrestrial Radio Access Network) UwPTS Uplink Pilot Time Slot V2I Vehicle-to-Infrastructure V2P Vehicle-to-Pedestrian V2V Vehicle-to-Vehicle V2X Vehicle-to-everything VIM Virtualized Infrastructure Manager VL Virtual Link, VLAN (Virtual LAN), Virtual Local Area Network VM (Virtual Machine) VNF (Virtualized Network Function) VNF Forwarding Graph VNFFGD VNF Forwarding Graph Descriptor VNFM VNF Manager VNF Manager VoIP (Voice-over-IP, Voice-over-Internet Protocol) VPLMN Visited Public Land Mobile Network VPN (Virtual Private Network) VRB (Virtual Resource Block) WiMAX Worldwide Interoperability for Microwave Access WLAN (Wireless Local Area Network) WMAN Wireless Metropolitan Area Network WPAN Wireless Personal Area Network X2-C X2-Control plane X2-U X2-User plane XML eXtensible Markup Language XRES Expected User Response XOR eXclusive OR exclusive OR ZC Zadoff-Chu ZP Zero Po Zero Power

Claims

1. A system for a radio access network (RAN) node B device for facilitating machine learning operations on user equipment (UE) devices, the system having a processing circuit coupled to storage, the processing circuit being: The step of causing the node B device to send an instruction to the UE device that the node B device supports machine learning; A step of identifying a service registration received from the UE device, indicating that the UE device requests machine learning support and machine learning models from the node B device; A step of causing the node B device to send a request for information relating to the UE device to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; The steps include: identifying the information received from the UE device based on the request for the aforementioned information; A step in which the node B device transmits the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; The node B device is configured to perform the step of causing the UE device to send updates to the machine learning model, the machine learning model parameters, or the machine learning configuration for use by the UE device. system.

2. The system according to claim 1, wherein the machine learning model, the machine learning model parameters, and the machine learning configuration are transmitted using a signal-transmitting wireless bearer.

3. The system according to claim 1, wherein the machine learning model, the machine learning model parameters, and the machine learning configuration are transmitted using a wireless bearer dedicated to machine learning transmission.

4. The aforementioned processing circuit further: The node B device is further configured to cause the UE device to transmit a radio resource control (RRC) configuration related to controlling the machine learning operation of the UE device using the machine learning configuration. The system according to claim 2 or 3.

5. The aforementioned processing circuit further: The node B device is configured to send a policy to the UE device that includes an action for the UE device to perform, and the policy is associated with the machine learning configuration. The aforementioned policy allows the UE device to select from a number of actions based on the policy. The system according to claim 1.

6. The aforementioned processing circuit further: A step of identifying a request received from the UE for taking action based on the results of the UE device's use of the machine learning model; The node B device is configured to perform the steps of causing the UE device to send a response based on the request, wherein the response confirms or rejects the action. The system according to claim 1.

7. The system according to claim 1, wherein the hardware capability includes an indication of whether the processor of the UE device supports machine learning, and further includes at least one of the processor type, maximum battery capacity, the current battery state of the UE device, or the batching data size associated with the UE device.

8. The system according to claim 1, wherein the machine learning capability includes the types of machine learning models supported by the UE device, and further includes at least one of the maximum machine learning model size, supported libraries, machine learning model training capability, or machine learning model inference capability.

9. The system according to claim 1, wherein the RRC message includes the machine learning configuration.

10. The system according to claim 9, further comprising a service type indicator indicating that the RRC message is not required to synchronize the machine learning configuration with the node B device, and the UE device is permitted to train the machine learning model associated with the action space of the machine learning configuration.

11. The system according to claim 9, wherein the RRC message further includes a service type indicator indicating the service to which the machine learning configuration is associated, and the Node B device is training a machine learning model associated with the machine learning configuration.

12. The system according to claim 9, wherein the RRC message further includes a service type indicator indicating that the UE device needs to register with the Node B device for a service associated with the machine learning configuration.

13. The aforementioned processing circuit further: The system is configured to identify machine learning reports received from the UE device, the machine learning reports including measurement predictions related to the machine learning configuration, performance feedback related to the machine learning configuration, and requested actions for the UE to perform based on the performance feedback. The system according to claim 1.

14. The aforementioned processing circuit further: It is configured to identify a second machine learning configuration received from the UE device, wherein the second machine learning configuration includes an update to the machine learning configuration. The system according to claim 1.

15. The aforementioned processing circuit further: A step of determining at least one of the following: machine learning model bias, machine learning model variance, machine learning model confidence level, or feedback related to the use of the machine learning model; A step of generating an update to the machine learning configuration for use by the UE device based on at least one of the machine learning model bias, machine learning model variance, machine learning model confidence level, or feedback; The node B device is configured to perform the step of causing the UE device to send the update to the machine learning configuration. The system according to claim 1.

16. The aforementioned processing circuit further: It is configured to identify update requests received from the aforementioned UE device that request updated machine learning configurations, Generating the update to the machine learning configuration is based on the update request, The system according to claim 15.

17. The system according to claim 1, wherein the system information block includes the instruction that the node B device supports machine learning, and the instruction that the node B device supports machine learning includes an instruction for machine learning capabilities and an instruction for machine learning services provided by the node B device.

18. The system according to claim 17, wherein the service registration includes a requested machine learning service among the machine learning services.

19. The system according to claim 1, wherein the machine learning configuration includes the service type of the machine learning model, a model bias threshold, a model variance threshold, an instruction on whether the machine learning model is maintained and trained by the node B device or the UE device, and an instruction on the machine learning report configuration.

20. The system according to claim 19, wherein the machine learning report configuration includes a machine learning report, a report period and offset, and instructions for the duration and start time of memory for the machine learning results.

21. A computer-readable storage medium having instructions, wherein the instructions are transmitted to a processing circuit of a wireless access network (RAN) node B device when the processing circuit executes the instructions: The steps include: causing the node B device to send an instruction to the user device (UE) device that the node B device supports machine learning; A step of identifying a service registration received from the UE device, indicating that the UE device requests machine learning support and machine learning models from the node B device; A step of causing the node B device to send a request for information relating to the UE device to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; The steps include: identifying the information received from the UE based on the request for the aforementioned information; A step in which the node B device transmits the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; The node B device is instructed to perform the step of sending updates to the machine learning model, the machine learning model parameters, or the machine learning configuration to the UE device for use by the UE device. A computer-readable storage medium.

22. The computer-readable storage medium according to claim 21, wherein the machine learning model, the machine learning model parameters, and the machine learning configuration are transmitted using a wireless bearer dedicated to machine learning transmission.

23. A method for facilitating machine learning operations on user equipment (UE) devices, The process involves a processing circuit in a Wireless Access Network (RAN) Node B device causing the Node B device to send an instruction to the UE device that the Node B device supports machine learning; The processing circuit includes the steps of: identifying a service registration received from the UE device indicating that the UE device requests machine learning support and a machine learning model from the node B device; The processing circuit causes the node B device to send a request for information related to the UE device to the UE device, wherein the information is associated with at least one of the hardware capabilities or machine learning capabilities of the UE device; The processing circuit includes the steps of: identifying the information received from the UE based on the request for the information; The processing circuit causes the node B device to transmit the machine learning model, machine learning model parameters, and machine learning configuration to the UE device for use by the UE device, wherein the machine learning model, machine learning model parameters, and machine learning configuration are based on the information; The processing circuit includes the step of causing the node B device to send updates to the machine learning model, the machine learning model parameters, or the machine learning configuration to the UE device for use by the UE device, method.

24. A computer-readable storage medium having instructions for performing the method described in claim 23.

25. An apparatus configured to perform the method described in claim 23.

Images