WO2023014896A1 - User equipment trajectory-assisted handover - Google Patents

User equipment trajectory-assisted handover Download PDF

Info

Publication number
WO2023014896A1
WO2023014896A1 PCT/US2022/039441 US2022039441W WO2023014896A1 WO 2023014896 A1 WO2023014896 A1 WO 2023014896A1 US 2022039441 W US2022039441 W US 2022039441W WO 2023014896 A1 WO2023014896 A1 WO 2023014896A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
network
model
location information
ran
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US2022/039441
Other languages
French (fr)
Inventor
Ziyi Li
Dawei YING
Qian Li
Youn Hyoung Heo
Jaemin HAN
Zongrui DING
Maruti Gupta Hyde
Yi Zhang
Sudeep Palat
Yi Guo
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Intel Corp
Original Assignee
Intel Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Intel Corp filed Critical Intel Corp
Priority to US18/556,421 priority Critical patent/US20240205781A1/en
Priority to EP22853917.7A priority patent/EP4381807A4/en
Publication of WO2023014896A1 publication Critical patent/WO2023014896A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/098—Distributed learning, e.g. federated learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/24—Reselection being triggered by specific parameters
    • H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
    • H04W36/322—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data by location data
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/0005—Control or signalling for completing the hand-off
    • H04W36/0055—Transmission or use of information for re-establishing the radio link
    • H04W36/0058—Transmission of hand-off measurement information, e.g. measurement reports
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/0005—Control or signalling for completing the hand-off
    • H04W36/0055—Transmission or use of information for re-establishing the radio link
    • H04W36/0061—Transmission or use of information for re-establishing the radio link of neighbour cell information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/24—Reselection being triggered by specific parameters
    • H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/34—Reselection control
    • H04W36/36—Reselection control by user or terminal equipment
    • H04W36/362—Conditional handover
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/0005—Control or signalling for completing the hand-off
    • H04W36/0083—Determination of parameters used for hand-off, e.g. generation or modification of neighbour cell lists
    • H04W36/00837—Determination of triggering parameters for hand-off
    • H04W36/008375—Determination of triggering parameters for hand-off based on historical data

Definitions

  • Various embodiments generally may relate to the field of wireless communications.
  • some embodiments may relate to user equipment (UE) trajectory-assisted handovers.
  • some embodiments may include artificial intelligence (Al) or machine learning (ML) to predict UE location information.
  • Al artificial intelligence
  • ML machine learning
  • Figure 1 is a flowchart illustrating an example of a process in accordance with various embodiments.
  • Figure 2 is a flowchart illustrating another example of a process in accordance with various embodiments.
  • Figure 3 illustrates an example of network controls model training and inference at a UE, with no model downloading fromNG-RAN in accordance with various embodiments.
  • Figure 4 illustrates an example of a model for inference downloading to the UE in accordance with various embodiments.
  • Figure 5 illustrates an example of federated learning between the network and the UE in accordance with various embodiments.
  • Figure 6 illustrates an example of how a network decides ML configuration in accordance with various embodiments.
  • Figure 7 illustrates an example of a general procedure of network-oversighted UE's ML training and inference in accordance with various embodiments.
  • Figure 8 illustrates an example of a general procedure of model transferring for UE ML inference in accordance with various embodiments.
  • Figure 9 illustrates an example of a general procedure of model transferring for UE ML training and inference in accordance with various embodiments.
  • Figure 10 illustrates an example of split AI/ML inference in accordance with various embodiments.
  • Figure 11 illustrates an example of AI/ML model downloading over a 5G system in accordance with various embodiments.
  • Figure 12 illustrates an example of federated learning over a 5G system in accordance with various embodiments.
  • Figure 13 illustrates an example of a RAN intelligence functional framework in accordance with various embodiments.
  • Figure 14 illustrates an example of another Functional Framework for RAN Intelligence in accordance with various embodiments.
  • FIG. 15 illustrates a network in accordance with various embodiments.
  • Figure 16 schematically illustrates a wireless network 1600 in accordance with various embodiments.
  • Figure 17 is a block diagram illustrating components, according to some examples of embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein.
  • a machine-readable or computer-readable medium e.g., a non-transitory machine-readable storage medium
  • FIGS 18, 19, and 20 depict examples of procedures for practicing the various embodiments discussed herein.
  • 3GPP SON As introduced above, there is currently standardization activity in 3GPP RAN3 studying mobility optimization that can be enhanced using AI/ML techniques.
  • the legacy approach supported in 3GPP SON is to collect feedback (e.g., ping-pong event, failure rate, etc.) so that the operation administration and maintenance (0AM) system can continuously optimize handover strategy based on failure events.
  • feedback e.g., ping-pong event, failure rate, etc.
  • 3GPP SA5 studied mobility management related issues as a use case in TR 28.809, v. 17.0.0, 2021-04-06, where mobility optimization can be based on resource consumption, inter- gNB beam selection, load balancing, mobility performance, UE trajectory or UE load.
  • the UE location information is provided by the Location Services (LCS), where the location server is resident in the CN.
  • the MDA (management data analytics) function block is responsible for providing ML training and inference service.
  • Al strategies for handover optimization is gaining attention and has also been proposed. However, it is based upon UE's historical record of location information and requires processing a very large amount of information regarding the UE's previous history using its location and can be quite intrusive in terms of UE privacy.
  • the handover strategy at the 0AM level may be generic and cannot adaptively make real-time handover decision based on each UE's information.
  • a more near-term update on the UE's existing location and trajectory information can greatly help to improve the accuracy and predict the target cell, especially in case of the high-speed mobile UEs.
  • location information is reported by the UE to the MDT analytics function as a series of location coordinates as deduced through prior Handovers and some historical UE location data.
  • the future location of UE can further improve the handover successful rate.
  • Various embodiments herein are directed to an AI/ML assisted handover mechanism located at NG-RAN, where the future UE location is used for mobility enhancement.
  • the future UE location may be determined from: 1) UE reported future UE location information 2) current/future positioning information available at the LMF (location management function)/OAM/CN.
  • the network may select the suitable UE(s) to use AI/ML model for handover decision, based on 1) UE information 2) mobility events 3) UE device type/ environment.
  • the described mechanism improves the UE handover success rate with assistance from machine learning-based methods to predict the target cell that has the greatest handover success probability based on the UE's future/near-term location and its trajectory, which can further reduce service interruption at the UE side and reduce network overhead. Since the UE shares the info with the gNB only for the short-term rather than the entire journey, it's easier to preserve UE privacy as a more comprehensive view of UE's location history is not needed to predict the target cell information.
  • the described mechanism also reduces AI/ML workload for highly active UEs in the network by only selecting the UE(s) which can benefit from AI/ML the most (e.g. better output accuracy, better performance improvement, etc).
  • Embodiment 1 LMF predict location/positioning-based mobility handover
  • mobility handover decision is based on the assistance of LMF provided location information.
  • NG-RAN will send measurement report to the LMF, which computes the location of the UE.
  • LMF can predict the UE location information.
  • NG-RAN can optimize the target cell selection and predict the best timing for certain UE to handover, together with UE's load, handover event collection, etc.
  • the positioning AI/ML model is trained and inferenced at the LMF, while the predicted results are sent to RAN over NRPPa or NG interface.
  • the predicted location can also be received from MDA or NWDAF located in 0AM or CN, respectively.
  • Input of the Al-positioning assisted mobility enhancement includes:
  • Predicted location information of UE (based on UE's current speed/location/trajectory info)
  • the network will configure the UE to handover to target cell, together with the best time to perform handover. o the network can also configure this outcome “the group of target cells and the corresponding handover timing, priority” to the UE as CHO (conditional handover) candidate cell(s), the corresponding configuration are sent as conditional handover configuration.
  • the UE can execute CHO configuration and handover to the candidate target cell according to the configured handover execution timing. With the timestamp, UE can stop performing measurements to evaluate conditional execution condition.
  • the CHO execution condition need to be enhanced.
  • the source gNB should also transfer the corresponding predicted handover timing to the candidate target gNB, where Xn interface should be enhanced.
  • UE selects the target cell to handover according to the priority. If handover is fails, UE will continue to select the remaining target cell which has the highest priority until the handover procedure succeeds. If execution timing of candidate cell(s) expires during UE performs handover, UE can delete this candidate cell from the list.
  • the source cell can also send the predicted configuration for the UE based on historical behavior or predicted load information to the target cell.
  • An example of message flows is shown in the flowchart in Figure 1.
  • the NRPPa or NG can be enhanced.
  • the input data are collected from RAN or based on measurement reports.
  • the output data are transmitted from CN or LMF to RAN.
  • NG-RAN sends a predict location request to 0AM (e.g. MDA)/CN (e.g. NWDAF/AMF)/LMF
  • the request can include duration of requested location (e.g. start time stamp + end time stamp/ start time stamp + duration), request location information (relative location, latitude, longitude, altitude, etc), UE ID, etc.
  • LMF/OAM/CN sends the predicted results to NG-RAN, including the requested location information, accuracy of the output, etc.
  • the CHO configuration can also be enhanced.
  • the timing of each candidate cell can be added in the configuration.
  • Example 1 An indication “attemptPredictCondReconfig” is used to indicate UE whether it can use the predicted timing for handover configured by the source cell or not.
  • the UE Indicates the priority of candidate cell.
  • the UE selects target cell for handover from the candidate cell with highest priority. If handover failed, the UE continues to select handover target cell according to the priority until handover succeed.
  • CondTriggerConflg can also be considered as an enhancement for this prediction-based trigger event.
  • An example embodiment with enhancement is shown as below.
  • CondTriggerConfig-r16 SEQUENCE ⁇ condEventld CHOICE ⁇ condEventA3 SEQUENCE ⁇ a3-0ffset MeasTriggerQuantityOffset, hysteresis Hysteresis, timeToTrigger TimeToTrigger condEventA5 SEQUENCE ⁇ a5-Thresholdl MeasTriggerQuantity, a5-Threshold2 MeasTriggerQuantity, hysteresis Hysteresis, timeToTrigger TimeToTrigger
  • Time during of CondEventPredict is the specific criteria for the event to trigger in order to execute the UE handover.
  • the Xn interface HANDOVER REQUEST
  • mobility handover decision is based on the assistance of UE reported current and near-term location information.
  • the handover target cell(s) are predicted based on the joint AI/ML model optimization for load information, UE assistance location information, handover event and configuration selection, etc.
  • the Locationinfo can be enhanced to include the following location information, reporting from UE to the network:
  • location information e.g. GNSS location information
  • location information can either be predicted by UE from AI/ML models or known by UE from positioning function. The uncertainty and confidence are generated based on the results from AI/ML models and positioning algorithms, respectively.
  • An example of a flow chart showing this message flow is shown in Figure 2.
  • ReportConflgNR can be expanded to configure the location report, e.g. "RocationInfoReportConfig"
  • RocationInfoReportConfig An example of this embodiment is shown as below, changes are marked in underline.
  • MeasTriggerQuantityCLI-r16 CHOICE ⁇ srs-RSRP-r16 SRS-RSRP-Range-r16, cli-RSSI-r16 CLI-RSSI-Range-r16
  • Need R includeFutureLocationlnfoList-rxy ENUMERATED ⁇ true ⁇ OPTIONAL.
  • Need R includeBT-Meas-r16 SetupRelease ⁇ BT-NameList-r16 ⁇ OPTIONAL.
  • Need M includeWLAN-Meas-r16 SetupRelease ⁇ WLAN-NameList-r16 ⁇ OPTIONAL.
  • Need M includeSensor-Meas-r16 SetupRelease ⁇ Sensor-NameList-r16 ⁇ OPTIONAL — Need M 1
  • Locationlnfo-r16 SEQUENCE ⁇ commonLocationInfo-r16 CommonLocationInfo-r16 OPTIONAL, bt-Locationlnfo-rl 6 LogMeasResultListBT-r16 OPTIONAL, wlan-LocationInfo-r16 LogMeasResultListWLAN-r16 OPTIONAL, sensor-LocationInfo-r16 Sensor-LocationInfo-r16 OPTIONAL, FutureLocationlnfoList-rxy FutureLocationInfoList-r18 OPTIONAL.
  • FutureLocationlnfoList-rxy SEQUENCE ⁇ future-gnss-TOD-msec-rxy OCTET STRING OPTIONAL. future-locationTimestamp-rxy OCTET STRING OPTIONAL. future-locationCoordinate-rxy OCTET STRING OPTIONAL. future-locationError-rxy OCTET STRING OPTIONAL. future-locationSource-rxy OCTET STRING OPTIONAL. future-velocity Estimate-rxy OCTET STRING OPTIONAL
  • Embodiment 3 Mixed location/positioning-based mobility handover
  • the network can utilize the location prediction results received from upper layer and future/near-term navigation information reported by the UE.
  • UE does not need to report the accurate location (e.g.GNSS, latitude, longitude, etc.), UE only need to report the velocity in near future and direction in the FutureLocationlnfoList. This can help to avoid UE reporting its private information (e.g. accurate location in near future) over the air interface.
  • the accurate location e.g.GNSS, latitude, longitude, etc.
  • UE only need to report the velocity in near future and direction in the FutureLocationlnfoList. This can help to avoid UE reporting its private information (e.g. accurate location in near future) over the air interface.
  • Embodiment 1 The example embodiment of standards impact is similar as Embodiment 1 and Embodiment 2, where NRPPa/NG interface can be enhanced to send predicted location to NG- RAN, and ReportConfigNR, Locationinfo, CondReconfigToAddModList can be enhanced correspondingly.
  • Example Use case Reducing ping-pong event with the assistance of UE location and trajectory prediction
  • AI/ML model is trained to select the best serving cell for a certain UE which is experiencing frequent ping-pong handover between two cells. Based on the predicted location, if AI/ML model predicts a UE is moving slowly from the source cell to the target cell or sits in the cell border for a long time, the network can avoid performing handover of such UE to the target node as it may have leamt that this is a result of current network topology at the current location of the UE. The number of ping-pong events can thus be reduced.
  • Embodiment 5 UE selection for mobility handover optimization
  • Not all UEs may require AI/ML based models to predict successful HOs and the configured NRTs etc. are enough for them. It is useful to know which UEs require additional processing and restrict the use of additional information exchange for such UEs alone. It is proposed that the network can select the suitable UE criteria to use AI/ML model to improve the mobility handover performance (e.g. increase handover successful rate, etc), and to reduce heavy signaling exchanges and heavy processing through AI/ML.
  • the (predicted) location information of the selected UEs is transmitted to the RAN with its UE ID.
  • three types of criteria of UE selection are described.
  • Embodiment 5-1 Based on the statistic of mobility event:
  • the network can select the UE for using AI/ML prediction to improve the Handover success rate.
  • the threshold can also be set by the ratio between above event counts and the corresponding requested handover events.
  • the selected UE ID should be decided by the network.
  • the selected UE ID is transmitted to LMF/MDA/NWDAF to request location information.
  • NG-RAN can configure positioning functionality selectively to the selected UEs (if the UE also have the positioning capability), where only the selected UE need to perform positioning calculation/prediction and report accordingly. Additionally, the above listed events should be collected for each UE.
  • Embodiment 5-2 Based on the UE information and Al/model accuracy measures
  • the accuracy of the AI/ML model outcome highly depends on the coverage of the training data set.
  • the AI/ML model may not be able to make the right/precise decision (as the input is far beyond the data set coverage) and has low accuracy results.
  • the wrong decision will lead to severe service interruption caused by failed handover.
  • the AI/ML based mobility model is used for the UEs with higher accuracy results which may be for more common scenarios for, e.g. moving along the road with a fixed speed/direction, etc.
  • the easier UE's trajectory can be predictable, the higher accuracy of AI/ML generated predicted handover.
  • the network may select the UEbased on following example UE information:
  • Embodiment 5-3 Based on UE device type
  • UE location is easily predictable as the trajectory is following a known map, such as factory, road, indoor, etc.
  • the network can also select the UE according to its device type as the type of UE environment may provide additional contextual information that would help the Al model in doing a more accurate job in predicting it's trajectory.
  • new device type/access category may include (but not limited to):
  • ML Machine learning
  • 3GPP Rel-18 also starts to look into how to adopt AI/ML over air interface, by taking advantage of benefits from AI/ML.
  • functional framework of supporting AI/ML in RAN includes “Data Collection”, “Model Training”, “Model Inference” and “Actor”. It is possible that those functional nodes are deployed at either network side or UE side.
  • Scenario 1 either the network or the UE performs AI/ML training and inference solely and independently, that is “Model Training”, “Model Inference” and “Actor” are locates either at network side or UE side.
  • model transfer over the air interface may not be needed.
  • additional assistance information transferred over the air interface such as measurement reports or data collections.
  • the initial AI/ML models e.g.
  • AI/ML model may not be totally suitable for one UE) running at the UEs are received/configured by the network over certain configurations. Moving further when integration between AI/ML and wireless network becoming tighter, the UE can download well-trained model from the network and perform “Model inference”, when “Model training” at the network side. The UE may even consider federated learning together with the network.
  • Figures 3, 4, and 5 show the three scenarios of AI/ML model at the UE to be oversighted by the network.
  • Figure 3 illustrates an example of model training/inference at the UE side, while UE training/inference policy/configuration is configured by NG-RAN.
  • Figure 4 illustrates an example of model training at the network side (RAN), while model inference at the UE side.
  • Figure 5 illustrates an example of model training and model inference locates at both network side and UE side (e.g. federated learning).
  • the network should be able to control how the UEs perform training and inference. That is, besides the model update policies and other aspects, the network should be able to control the AI/ML training/inference behavior at the UE side, so that it can control the damage in case the training/inference at the UE side is wrong or non-convergence.
  • Various embodiments herein provide mechanisms to support model training and inference at UE side, where the UE AI/ML behavior can be configured/oversighted by the network.
  • UE AI/ML behavior can be configured/oversighted by the network.
  • different scenarios of AI/ML training and inference may be analyzed when they are jointly or separately located at UE and NG-RAN, then describe the configuration and signaling procedure to support the network- oversighted AI/ML at UE, including configuration parameters, conditions of configuration, etc.
  • the described mechanism can support UE to generate the result/actions which can be trusted by the network when using AI/ML at UE side.
  • the described mechanism helps to build a reliable wireless system when using AI/ML at network and UE, especially when UE is allowed to perform model training/re-training.
  • the UE can leam the machine learning capability and network supported models/services enabled by AI/ML through the broadcast information (e.g. MachineLeamingSupport field) via MLCapabilitylndication.
  • the broadcast information e.g. MachineLeamingSupport field
  • UE can decide whether to register or request a ML model from the network based on the received information, and send the corresponding service registration/interest indication message back to the network as model downloading request.
  • the network is responsible for model transfer, update, etc.
  • the UE will not request model from the network, this can either because the UE does not have the ML capability, or because the UE holds the ML model itself.
  • model training/inference configuration is also adoptable when the UE is allowed to perform local model training (e.g. federated learning).
  • the AI/ML model can always be downloaded to the UE according to the UE ML capability and request.
  • ML related information exchanging may be expanded to have a wider scope, where the network can configure UE training and inference, where model downloading from the network is not a constraint.
  • Figure 6 illustrates an example of a process flow for a network to determine ML configuration.
  • Service ID e.g. CSI feedback, positioning, beam management, etc
  • Trainingindicator e.g. CSI feedback, positioning, beam management, etc
  • this field indicates that the UE is allowed to perform training/re-training. If it is not presented, the UE is not configured to further perform training/re-training, either for its own model or model received from network.
  • this field indicates that the UE is allowed to use its own ML model. If it is not presented, the UE should not use its own ML model, even it has the corresponding ML capability.
  • the training configuration is considered per model basis, where the model type is determined by modeltype reported by the UE via UE-CapabilityML-Information.
  • the network can configure the full-set or the subset of the supported model at the UE.
  • the UE reports supported configuration to the network in UE capability “UE- CapabilityML-Information”; network then configure the optimized configuration/threshold based on network training experience and best knowledge of one use case.
  • Model type CNN, RNN, DNN, regression, etc
  • the UE should perform model training according to the following configuration list, then return with the evaluation results through new IE “TrainingReport” in MachineLeamingReport or in MachineLeamingModelUpdateRequest a separate message from UE to network. If the reported evaluation results meet the performance threshold, the network can send an ACK of “TrainingFeedback” to the UE via MachineLeamingConfiguration or in MachineLeamingModelUpdateRequest or a new message from network to UE, indicating UE's trained AI/ML can be used. Otherwise, if UE receives a NACK of “TrainigFeedback” from the network, UE should continue training, and repeat the previous procedure.
  • the configuration may include: o Optimizer
  • Optimizer Algorithm that the UE should use for training optimization include
  • Losses functions/classes that the UE should use for training evaluation include three categories:
  • the metrics is used to judge the performance of the training model.
  • Image segmentation metrics Hinge metrics.
  • This parameter is the lower bound of one AI/ML model need to meet.
  • the UE may stop AI/ML model training when the configured accuracy is met. o Steps per execution
  • network can also configure the wireless KPI (e.g. throughput, SINR, BLER, etc) and AI/ML model KPI (e.g. accuracy level of the model, confidence level of the model, etc) for different use cases.
  • wireless KPI e.g. throughput, SINR, BLER, etc
  • AI/ML model KPI e.g. accuracy level of the model, confidence level of the model, etc
  • ML model related information including service type, report configuration, policy guidance, behavior type, bias threshold and variance threshold for model update request, model itself, model parameter update.
  • the network should also indicate the UE whether it can continue perform model (re-)training for each transferred ML model.
  • the required inputs to the ML model for training and inference should also be configured per model basis.
  • This is used to carry the fields/parameters that are defined as ML model input for model training/inference if UE's ML model is received from the network.
  • This is used to carry the fields/parameters that are defined as ML model output for model training/inference if UE's ML model is received from the network.
  • UE should also send a request or an indicator informing the network whether model downloading/transferring is required or not. This indicator can be sent together with service registration/interest indication message.
  • the RRC impacts may be used used to carry the above signaling and messages.
  • Embodiment 1 No model transfer/ downloading from the network to the UE
  • This embodiment presents a scenario where the model is trained and inferenced at the UE side.
  • the training at the UE side should be oversighted by the network in certain degree, so that the network can trust the result/outcome of the ML inference.
  • FIG. 7 shows an example of a general procedure of network-oversighted UE's ML training and inference.
  • Embodiment 2 UE downloads ML model from the network, while only perform ML inference
  • UE can only perform ML inference to the received ML model from the network.
  • both Set 1 and Set 2 configurations should be configured to the applied UE.
  • An example of this general procedure is shown in Figure 8.
  • Embodiment 3 UE downloads ML model from the network, performing both ML (re-)training and ML inference
  • the learning rate Floating point value.
  • the learning rate Floating point value.
  • the learning rate Floating point value.
  • Adadelta The name of Adadelta. “Adadelta”.
  • the learning rate Floating point value.
  • Adagrad The name of Adagrad. “Adagrad”.
  • the learning rate Floating point value.
  • the learning rate Floating point value.
  • the learning rate Floating point value.
  • Adagrad The name of Adagrad. “Adagrad”.
  • the learning rate Floating point value.
  • 3 GPP RAN3 started to study Al intelligence in RAN network in Rel-17, where RAN intelligence functional framework is discussed, including mapping of AI/ML functionality into existing NG-RAN architecture.
  • RAN intelligence will first address use cases related to SON (self-organizing network), for example, load balancing, network energy saving, and mobility optimization.
  • 3GPP Rel-18 will also investigate how to adopt AI/ML over air interface.
  • the 5G system can at least support three types of AI/ML operations: AI/ML operation splitting between AI/ML endpoints; AI/ML model/data distribution and sharing over 5G system; and Distributed/Federated Learning over 5G system.
  • FIG. 10 An example of a scheme of split AI/ML inference is depicted in Figure 10.
  • the AI/ML operation/model is split into multiple parts according to the current task and environment. The intention is to offload the computation-intensive, energy-intensive parts to network endpoints, whereas leave the privacy-sensitive and delay-sensitive parts at the end device.
  • the device executes the operation/model up to a specific part/layer and then sends the intermediate data to the network endpoint.
  • the network endpoint executes the remaining parts/layers and feeds the inference results back to the device.
  • Multifunctional mobile terminals might need to switch the AI/ML model in response to task and environment variations.
  • the condition of adaptive model selection is that the models to be selected are available for the mobile device.
  • Online model distribution e.g., new model downloading
  • an AI/ML model can be distributed from a NW endpoint to the devices when they need it to adapt to the changed AI/ML tasks and environments. For this purpose, the model performance at the UE needs to be monitored constantly.
  • FIG. 12 An example of a scheme of Federated Learning (FL) is depicted as in Figure 12.
  • the cloud server trains a global model by aggregating local models partially-trained by each end devices.
  • a UE performs the training based on the model downloaded from the Al server using the local training data.
  • the UE reports the interim training results to the cloud server via 5G UL channels.
  • the server aggregates the interim training results from the UEs and updates the global model.
  • the updated global model is then distributed back to the UEs and the UEs can perform the training for the next iteration.
  • a functional framework of supporting AI/ML in RAN includes “Data Collection”, “Model Training”, “Model Inference” and “Actor”. Where to realize each AI/ML functionality highly depends on use cases to address and various requirements related to data collection and performance (e.g., latency, real-time, etc).
  • 3GPP SAI initiated the study of traffic characteristics and performance requirements for AI/ML model transfer in 5GS, which also support model downloading and update from network to UE (see e.g., 3GPP TR 22.874 V18.0.1 (2021-06-24) (“[TR22874]”)).
  • FIG. 13 An example of a RAN intelligence functional framework is shown in Figure 13, where model deployment/update from “Model training” to “Model inference” are issues for future study.
  • FIG 14. An example of another Functional Framework for RAN Intelligence is shown in Figure 14.
  • the present disclosure provides a mechanism to update AI/ML models (e.g., for inferences, predictions, federated learning, etc.) between network nodes (e.g., from CN and/or 0AM to RAN) or between network and UE (e.g., from CN, 0AM, and/or RAN to UE).
  • the AI/ML model update can be triggered based on following conditions: 1) timer (e.g., expiration of a timer); 2) loss weight of model outputs; 3) accuracy level of AI/ML model; 4) probabilistic of AI/ML model generated output; 5) performance feedback; 6) per request from NG-RAN or UE; and/or 7) trigger/indication from CN/OAM or NG-RAN.
  • the mechanisms discussed herein maintain a good performance AI/ML model by updating models between network node(s) and/or UE(s) and between different network nodes according to certain rules, trigger conditions, etc. These mechanisms can also help avoid unnecessary model updates, which also helps to save radio resources for model exchanging and conserve computational resources.
  • AI/ML includes processing of data collection, model training, model validation and model inference. Normally, a well-trained model is trained based on massive input data, which also need to be validated based on another set of data which is not used for training. In wireless system, the environment (including traffic, radio, accessed number of UE, etc) may be various. It's hard to train a comprehensive AI/ML model which can suit for all scenarios just based on several rounds of training or initial offline training. To maintain a good accuracy level of prediction results for which network can trust, it is necessary for the network node/UE with model inference functionality to get an up-to-date AI/ML model which can accommodate to the changes in the environment.
  • the network node/UE with model training functionality should continuously train AI/ML based on feedbacks collected from Actor (which is distributed in the network, e.g., UE, gNB-DU, gNB-CU, etc), regardless of online (e.g., RL) or offline (e.g., supervised learning, unsupervised learning) training mechanism used.
  • Actor which is distributed in the network, e.g., UE, gNB-DU, gNB-CU, etc
  • online e.g., RL
  • offline e.g., supervised learning, unsupervised learning
  • model update mechanism Several issues should be considered when defining model update mechanism:
  • AI/ML model The purpose of updating AI/ML model is to provide better performance and output with higher accuracy to the system.
  • One deployed model may be outdated or not suitable for current environment. This requires the Model training function either to update the model based on performance feedback from Actor or based on other mechanisms.
  • Model update and synchronization for federated learning When model training and model inference are deployed at different locations in the network, federated learning can be considered.
  • one central trained model is located at central server (e.g., CN or RAN, respectively) and another local retrained model is located at local nodes (e.g., RAN or UE, respectively). Certain interaction and synchronization of trained models between central server and local nodes are required.
  • the central server may also selectively update the model to the local nodes, considering workload or other impacts.
  • the present disclosure provides various embodiments to support model update between network nodes with model training (e.g., OAM/CN for model deployment to RAN, RAN for model deployment to UE) and network nodes/UE with model inference (e.g., RAN for model deployment from CN/OAM, UE for model deployment from RAN).
  • model training e.g., OAM/CN for model deployment to RAN, RAN for model deployment to UE
  • model inference e.g., RAN for model deployment from CN/OAM, UE for model deployment from RAN.
  • the model update is triggered from model training (e.g., OAM/CN or RAN) based on the configured timer.
  • model training e.g., OAM/CN or RAN
  • the timer can either be configured and running at network nodes with model training or network nodes/UE with model inference.
  • the timer can be used to: indicate the duration of one AI/ML model's update/re-train; the model is retrained based on data collected during this period; and/or trigger model update and send updated AI/ML model to network nodes/UE with model inference.
  • the timer can be used to: trigger model update and send model update request message from NG-RAN to OAM/CN or from UE to NG- RAN.
  • the timer is configured from CN/OAM to NG-RAN via NG interface or El interface; or the timer is configured from NG-RAN to UE via Uu interface as part of model configuration in Machine Learning Configuration message.
  • model training side and/or model inference side embodiments are shown below:
  • MachineLeamingConfiguration SEQUENCE ⁇ criticalExtensions CHOICE ⁇ machineLeamingConfiguration MachineLeamingConfiguration-IEs, machineLeamingModelUpdateConfiguration MachineLeamingModelUpdateConfiguration, criticalExtensionsFuture SEQUENCE ⁇
  • MachineLeamingModelUpdateConfiguration SEQUENCE ⁇ modelUpateConfigToRemoveList ModelUpateConfigToRemoveList
  • modelUpdateConfigToAddModList ModelUpdateConfigToAddModList OPTIONAL modelUpdateConfigToAddModList OPTIONAL
  • ModelUpdateReconfigToAddModList-rxy SEQUENCE (SIZE (L. maxNrofModel- rxy)) OF ModelUpdateConfigToAddMod-rxy ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY, updateTimer ENUMERATED ⁇ sf50, sflOO, sf200,.., infinity ⁇ OPTIONAL,
  • ModelUpdateConfigToAddModList concerns a list of model update configurations to add or modify, with for each entry the modelldentifier and the associated trigger condition/threshold. ’ Network may configure or set different timer according to different use cases or to different models for different use cases.
  • a central server (CN/OAM or RAN) is allowed to set the same timer to all local nodes (RAN or UE) to support synchronized model training. If the central server (CN/OAM or RAN) sets different timers to different local nodes (RAN or UE), asynchronized model training is considered for federated averaging.
  • a model ID is required to be encoded in the model update message from OAM/CN to RAN or in RRC reconfiguration (fromNG-RAN to UE).
  • the model update can be triggered by accuracy level of one AI/ML model.
  • the accuracy level can be calculated by comparing predicted results with the real value at the model training node during validation or testing. If the model accuracy level is higher than that of the pervious deployed model, the model training node (e.g., CN/OAM or RAN) can provide this updated model to model inference (RAN or UE).
  • the model training node may also calculate the accuracy level gain comparing accuracy level of current model and previous one. Model update is only triggered if the accuracy gain is higher than the threshold.
  • the model update can be triggered by confidence level of one AI/ML model. Confidence level is calculated at the model inference node, by comparing prediction results and the real value (real future input to model inference). If the confidence level is lower than certain threshold, a model update should be triggered to avoid using the wrong/inaccurate predicted results/actions. This could also be triggered by either network or UE.
  • the OAM/CN or the NG-RAN configures a threshold of confidence level to NG-RAN or UE via AI/ML model update configuration (as part of model management). If the confidence level calculated by network nodes/UE for model inference (e.g., RAN or UE) is lower than the confidence level threshold, the RAN/UE should send a model update request to the network nodes with model training functionality (e.g., CN/OAM or RAN). Upon receiving the model update request, an updated model should be sent to the requested node.
  • the confidence level is between [0,1], here, for example, an integer value is used to represent 10 times of exact loss value.
  • model update configuration from NG-RAN to UE is shown as below:
  • network nodes with model training can trigger model update based on received confidence level report from model inference (e.g., RAN or UE). If the confidence level in the report is lower than the value expected at the model training, network nodes can send anew up-to-date model to its model inference node (e.g., RAN or UE).
  • model inference e.g., RAN or UE
  • the network may configure or set different confidence level threshold according to different use cases. For example, UE trajectory normally requires a high accuracy, while some non-real-time prediction can have a relax threshold.
  • the model update is triggered by performance feedback (e.g., UE measurement report, performance indicator from network node or its neighbor nodes, etc).
  • performance feedback e.g., UE measurement report, performance indicator from network node or its neighbor nodes, etc.
  • the model will only be updated when the performance feedback shows a poor performance or performance down-gradation.
  • the Actor who performs predicted results/actions provides performance feedback (e.g., UE measurement report, SON/MDT measurement report, etc) to the network.
  • performance feedback e.g., UE measurement report, SON/MDT measurement report, etc
  • the network nodes with model training can provide an updated AI/ML model to network nodes/UE with model inference to optimize the predicted results.
  • the Actor can compare its previous performance and its performance after taking predicted results/actions. If the performance (e.g., throughput, SINR, energy cost, etc) is down-graded, the Actor can directly send a simple feedback indicator “performance down-graded” to the network node with model training. Upon receiving this “downgraded” feedback indicator, OAM/CN or RAN can provide an updated model to the model inference node, e.g., RAN or UE, respectively.
  • This “performance indicator” in Uu interface is shown as below:
  • MachineLeamingReport-IEs SEQUENCE ⁇ serviceType ENUMERATED ⁇ default, positioning, V2X, spare5, spared, spare3, spare2, spare 1 ⁇ OPTIONAL, modelUpdate OCTET STRING (SIZE(L.8000)) OPTIONAL, modelBias ENUMERATED ⁇ poor-bias, less-poor, no-bias, spare5, spared, spare3, spare2, spare 1 ⁇ OPTIONAL, modelvariance INTEGER (0, ... , 100) OPTIONAL, predictionResult PredictionResult OPTIONAL, confidenceLevel ENUMERATED ⁇ 0, 1 , 2, ... , 10 ⁇ OPTIONAL, performanceindicator ENUMERATED ⁇ improved, downgraded, nochange ⁇ OPTIONAL, lateNonCriticalExtension OCTET STRING
  • This performance indicator can also be added in UE MDT measurement report, L2 measurement report, etc.
  • the performance feedback to trigger model update can be different across use cases.
  • this could be handover failure rate or other handover events; for energy saving, this could be an increase of energy consumption of ES-Cell or neighbor cell; for load balancing, it could be the network node's average packet delay/throughput or a percentage of UE's packet delay/throughput; for CSI feedback, it could be system or UE throughput.
  • the model update is triggered based on the model loss.
  • This embodiment is mainly suitable for federated learning, e.g., between OAM/CN and NG-RAN; between NG-RAN and UE.
  • a certain threshold of loss is configured by central node during model deployment and configuration.
  • a fix number of epochs or a given period is configured together with loss.
  • the output of loss function may be various based on the exact algorithm, the example integer can also be replaced by dB or other unit per use case.
  • the loss gain is the relative percentage of loss improvement compared with last calculated loss of the model. Normally, the gain value is between [0,1], here an integer value is used to represent 10 times of exact gain value.
  • the local node When the loss of a trained model at UE is larger than the configured loss threshold, the local node (RAN or UE) should trigger a request to the central server (CN/OAM or RAN) via NG/E1 or Uu interface (e.g., Machine Learning Model Update Request), asking for a model update.
  • CN/OAM or RAN e.g., Machine Learning Model Update Request
  • ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY, lossStartTimeOffset ENUMERATED ⁇ sflO, sf20, sf50,.. ⁇ OPTIONAL, lossStartEpoch ENUMERATED ⁇ 10, 20, 30,.. ⁇ OPTIONAL, lossThreshold ENUMERATED ⁇ 1, 2, 3,.., 10 ⁇ OPTIONAL, lossGain ENUMERATED ⁇ 1, 2, 3,.., 10 ⁇ OPTIONAL,
  • the model update can also be managed by a central server, if the loss weights of local nodes (e.g., RAN or UE) are reported to central node (e.g., OAM/CN or RAN). Based on received model loss weight, the central node may know whether this AI/ML model need to be refined/updated.
  • the loss in the model report can also be replaced by some categorized level of loss (e.g., high loss, low loss, etc.). Based on the report, whether to update the model or not is decided by the central server - The model update may only be triggered if the central node receives a high loss in the model report from local nodes.
  • MachineLeamingReport-IEs SEQUENCE ⁇ serviceType ENUMERATED ⁇ default, positioning, V2X, spare5, spared, spare3, spare2, sparel ⁇ OPTIONAL, modelUpdate OCTET STRING (SIZE(L.8OOO)) OPTIONAL, modelBias ENUMERATED ⁇ poor-bias, less-poor, no-bias, spare5, spared, spare3, spare2, sparel ⁇ OPTIONAL, modelvariance INTEGER (0, ... , 100) OPTIONAL, predictionResult PredictionResult OPTIONAL, confidenceLevel ENUMERATED ⁇ 0, 1 , 2, ... , 10 ⁇ OPTIONAL, loss INTEGER (0, 1, .., 9, 10)
  • model variance and model bias can also be used as trigger event for model update.
  • One AI/ML model may also be configured with multiple trigger condition.
  • a modelUpdateTriggerCategory may include all possibilities of a trigger condition. An example is shown below:
  • ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY
  • modelUpdateTriggerCategory ENUMERATED ⁇ timer, accuracylevel, confidencelevel, performancefeedback, loss, spare3, spare2, sparel ⁇ OPTIONAL
  • modelUpdateTriggerToAddModList ModelUpdateTriggerToAddModList
  • ModelUpdateTriggerToAddModList-rxy :: SEQUENCE (SIZE (1.. maxNrofModelUpdateCat-rxy)) OF ModelUpdateTriggerToAddMod-rxy
  • ModelUpdateTriggerToAddMod updateTimer ENUMERATED ⁇ sflO, sf20, sf50,.., infinity ⁇ OPTIONAL, confidenceLevelThreshold ENUMERATED ⁇ 1, 2, 3,.., 10 ⁇ OPTIONAL, lossStartTimeOffset ENUMERATED ⁇ sflO, s£20, sf50,.. ⁇ OPTIONAL, lossStartEpoch ENUMERATED ⁇ 10, 20, 30,.. ⁇ OPTIONAL, lossThreshold ENUMERATED ⁇ 1, 2, 3,.., 10 ⁇ OPTIONAL, lossGain ENUMERATED ⁇ 1, 2, 3,.., 10 ⁇ OPTIONAL,
  • FIGS 15-17 illustrate various systems, devices, and components that may implement aspects of disclosed embodiments.
  • FIG. 15 illustrates a network 1500 in accordance with various embodiments.
  • the network 1500 may operate in a manner consistent with 3GPP technical specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard and the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like.
  • the network 1500 may include a UE 1502, which may include any mobile or non-mobile computing device designed to communicate with a RAN 1504 via an over-the-air connection.
  • the UE 1502 may be, but is not limited to, a smartphone, tablet computer, wearable computer device, desktop computer, laptop computer, in-vehicle infotainment, in-car entertainment device, instrument cluster, head-up display device, onboard diagnostic device, dashtop mobile equipment, mobile data terminal, electronic engine management system, electron! c/engine control unit, electronic/engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked appliance, machine-type communication device, M2M or D2D device, loT device, etc.
  • the network 1500 may include a plurality of UEs coupled directly with one another via a sidelink interface.
  • the UEs may be M2M/D2D devices that communicate using physical sidelink channels such as, but not limited to, PSBCH, PSDCH, PSSCH, PSCCH, PSFCH, etc.
  • the UE 1502 may additionally communicate with an AP 1506 via an over-the-air connection.
  • the AP 1506 may manage a WLAN connection, which may serve to offload some/all network traffic from the RAN 1504.
  • the connection between the UE 1502 and the AP 1506 may be consistent with any IEEE 802.11 protocol, wherein the AP 1506 could be a wireless fidelity (Wi-Fi®) router.
  • the UE 1502, RAN 1504, and AP 1506 may utilize cellular- WLAN aggregation (for example, LWA/LWIP).
  • Cellular- WLAN aggregation may involve the UE 1502 being configured by the RAN 1504 to utilize both cellular radio resources and WLAN resources.
  • the RAN 1504 may include one or more access nodes, for example, AN 1508.
  • AN 1508 may terminate air-interface protocols for the UE 1502 by providing access stratum protocols including RRC, PDCP, RLC, MAC, and LI protocols. In this manner, the AN 1508 may enable data/voice connectivity between CN 1520 and the UE 1502.
  • the AN 1508 may be implemented in a discrete device or as one or more software entities running on server computers as part of, for example, a virtual network, which may be referred to as a CRAN or virtual baseband unit pool.
  • the AN 1508 be referred to as a BS, gNB, RAN node, eNB, ng-eNB, NodeB, RSU, TRxP, TRP, etc.
  • the AN 1508 may be a macrocell base station or a low power base station for providing femtocells, picocells or other like cells having smaller coverage areas, smaller user capacity, or higher bandwidth compared to macrocells.
  • the RAN 1504 may be coupled with one another via an X2 interface (if the RAN 1504 is an LTE RAN) or an Xn interface (if the RAN 1504 is a 5G RAN).
  • the X2/Xn interfaces which may be separated into control/user plane interfaces in some embodiments, may allow the ANs to communicate information related to handovers, data/context transfers, mobility, load management, interference coordination, etc.
  • the ANs of the RAN 1504 may each manage one or more cells, cell groups, component carriers, etc. to provide the UE 1502 with an air interface for network access.
  • the UE 1502 may be simultaneously connected with a plurality of cells provided by the same or different ANs of the RAN 1504.
  • the UE 1502 and RAN 1504 may use carrier aggregation to allow the UE 1502 to connect with a plurality of component carriers, each corresponding to a Pcell or Scell.
  • a first AN may be a master node that provides an MCG and a second AN may be secondary node that provides an SCG.
  • the first/second ANs may be any combination of eNB, gNB, ng-eNB, etc.
  • the RAN 1504 may provide the air interface over a licensed spectrum or an unlicensed spectrum.
  • the nodes may use LAA, eLAA, and/or feLAA mechanisms based on CA technology with PCells/Scells.
  • the nodes Prior to accessing the unlicensed spectrum, the nodes may perform medium/carrier-sensing operations based on, for example, a listen-before-talk (LBT) protocol.
  • LBT listen-before-talk
  • the UE 1502 or AN 1508 may be or act as a RSU, which may refer to any transportation infrastructure entity used for V2X communications.
  • An RSU may be implemented in or by a suitable AN or a stationary (or relatively stationary) UE.
  • An RSU implemented in or by: a UE may be referred to as a “UE-type RSU”; an eNB may be referred to as an “eNB-type RSU”; a gNB may be referred to as a “gNB-type RSU”; and the like.
  • an RSU is a computing device coupled with radio frequency circuitry located on a roadside that provides connectivity support to passing vehicle UEs.
  • the RSU may also include internal data storage circuitry to store intersection map geometry, traffic statistics, media, as well as applications/software to sense and control ongoing vehicular and pedestrian traffic.
  • the RSU may provide very low latency communications required for high speed events, such as crash avoidance, traffic warnings, and the like. Additionally or alternatively, the RSU may provide other cellular/WLAN communications services.
  • the components of the RSU may be packaged in a weatherproof enclosure suitable for outdoor installation, and may include a network interface controller to provide a wired connection (e.g., Ethernet) to a traffic signal controller or a backhaul network.
  • the RAN 1504 may be an LTE RAN 1510 with eNBs, for example, eNB 1512.
  • the LTE RAN 1510 may provide an LTE air interface with the following characteristics: SCS of 15 kHz; CP-OFDM waveform for DL and SC-FDMA waveform for UL; turbo codes 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; and CRS for cell search and initial acquisition, channel quality measurements, and channel estimation for coherent demodulation/detection at the UE.
  • the LTE air interface may operating on sub-6 GHz bands.
  • the RAN 1504 may be an NG-RAN 1514 with gNBs, for example, gNB 1516, or ng-eNBs, for example, ng-eNB 1518.
  • the gNB 1516 may connect with 5G-enabled UEs using a 5G NR interface.
  • the gNB 1516 may connect with a 5G core through an NG interface, which may include an N2 interface or an N3 interface.
  • the ng-eNB 1518 may also connect with the 5G core through an NG interface, but may connect with a UE via an LTE air interface.
  • the gNB 1516 and the ng-eNB 1518 may connect with each other over an Xn interface.
  • the NG interface may be split into two parts, an NG user plane (NG-U) interface, which carries traffic data between the nodes of the NG-RAN 1514 and a UPF 1548 (e.g., N3 interface), and an NG control plane (NG-C) interface, which is a signaling interface between the nodes of the NG-RAN1514 and an AMF 1544 (e.g., N2 interface).
  • NG-U NG user plane
  • N-C NG control plane
  • the NG-RAN 1514 may provide a 5G-NR air interface with the following characteristics: variable SCS; CP-OFDM for DL, CP-OFDM and DFT-s-OFDM for UL; polar, repetition, 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 a CRS, but may use PBCH DMRS for PBCH demodulation; PTRS for phase tracking for PDSCH; and tracking reference signal for time tracking.
  • the 5G-NR air interface may operating on FR1 bands that include sub-6 GHz bands or FR2 bands that include bands from 24.25 GHz to 52.6 GHz.
  • the 5G-NR air interface may include an SSB that is an area of a downlink resource grid that includes PSS/SSS/PBCH.
  • the 5G-NR air interface may utilize BWPs for various purposes.
  • BWP can be used for dynamic adaptation of the SCS.
  • the UE 1502 can be configured with multiple BWPs where each BWP configuration has a different SCS. When a BWP change is indicated to the UE 1502, the SCS of the transmission is changed as well.
  • Another use case example of BWP is related to power saving.
  • multiple BWPs can be configured for the UE 1502 with different amount of frequency resources (for example, PRBs) to support data transmission under different traffic loading scenarios.
  • a BWP containing a smaller number of PRBs can be used for data transmission with small traffic load while allowing power saving at the UE 1502 and in some cases at the gNB 1516.
  • a BWP containing a larger number of PRBs can be used for scenarios with higher traffic load.
  • the RAN 1504 is communicatively coupled to CN 1520 that includes network elements to provide various functions to support data and telecommunications services to customers/subscribers (for example, users of UE 1502).
  • the components of the CN 1520 may be implemented in one physical node or separate physical nodes.
  • NFV may be utilized to virtualize any or all of the functions provided by the network elements of the CN 1520 onto physical compute/storage resources in servers, switches, etc.
  • a logical instantiation of the CN 1520 may be referred to as a network slice, and a logical instantiation of a portion of the CN 1520 may be referred to as a network sub-slice.
  • the CN 1520 may be an LTE CN 1522, which may also be referred to as an EPC.
  • the LTE CN 1522 may include MME 1524, SGW 1526, SGSN 1528, HSS 1530, PGW 1532, and PCRF 1534 coupled with one another over interfaces (or “reference points”) as shown. Functions of the elements of the LTE CN 1522 may be briefly introduced as follows.
  • the MME 1524 may implement mobility management functions to track a current location of the UE 1502 to facilitate paging, bearer activation/deactivation, handovers, gateway selection, authentication, etc.
  • the SGW 1526 may terminate an SI interface toward the RAN and route data packets between the RAN and the LTE CN 1522.
  • the SGW 1526 may be a local mobility anchor point for inter-RAN node handovers and also may provide an anchor for inter-3 GPP mobility. Other responsibilities may include lawful intercept, charging, and some policy enforcement.
  • the SGSN 1528 may track a location of the UE 1502 and perform security functions and access control. In addition, the SGSN 1528 may perform inter-EPC node signaling for mobility between different RAT networks; PDN and S-GW selection as specified by MME 1524; MME selection for handovers; etc.
  • the S3 reference point between the MME 1524 and the SGSN 1528 may enable user and bearer information exchange for inter-3 GPP access network mobility in idle/active states.
  • the HSS 1530 may include a database for network users, including subscription-related information to support the network entities' handling of communication sessions.
  • the HSS 1530 can provide support for routing/roaming, authentication, authorization, naming/addressing resolution, location dependencies, etc.
  • An S6a reference point between the HSS 1530 and the MME 1524 may enable transfer of subscription and authentication data for authenticating/ authorizing user access to the LTE CN 1520.
  • the PGW 1532 may terminate an SGi interface toward a data network (DN) 1536 that may include an application/content server 1538.
  • the PGW 1532 may route data packets between the LTE CN 1522 and the data network 1536.
  • the PGW 1532 may be coupled with the SGW 1526 by an S5 reference point to facilitate user plane tunneling and tunnel management.
  • the PGW 1532 may further include a node for policy enforcement and charging data collection (for example, PCEF).
  • the SGi reference point between the PGW 1532 and the data network 15 36 may be an operator external public, a private PDN, or an intra-operator packet data network, for example, for provision of IMS services.
  • the PGW 1532 may be coupled with a PCRF 1534 via a Gx reference point.
  • the PCRF 1534 is the policy and charging control element of the LTE CN 1522.
  • the PCRF 1534 may be communicatively coupled to the app/content server 1538 to determine appropriate QoS and charging parameters for service flows.
  • the PCRF 1532 may provision associated rules into a PCEF (via Gx reference point) with appropriate TFT and QCI.
  • the CN 1520 may be a 5GC 1540.
  • the 5GC 1540 may include an AUSF 1542, AMF 1544, SMF 1546, UPF 1548, NSSF 1550, NEF 1552, NRF 1554, PCF 1556, UDM 1558, and AF 1560 coupled with one another over interfaces (or “reference points”) as shown.
  • Functions of the elements of the 5GC 1540 may be briefly introduced as follows.
  • the AUSF 1542 may store data for authentication of UE 1502 and handle authentication- related functionality.
  • the AUSF 1542 may facilitate a common authentication framework for various access types.
  • the AUSF 1542 may exhibit an Nausf service-based interface.
  • the AMF 1544 may allow other functions of the 5GC 1540 to communicate with the UE 1502 and the RAN 1504 and to subscribe to notifications about mobility events with respect to the UE 1502.
  • the AMF 1544 may be responsible for registration management (for example, for registering UE 1502), connection management, reachability management, mobility management, lawful interception of AMF-related events, and access authentication and authorization.
  • the AMF 1544 may provide transport for SM messages between the UE 1502 and the SMF 1546, and act as a transparent proxy for routing SM messages.
  • AMF 1544 may also provide transport for SMS messages between UE 1502 and an SMSF.
  • AMF 1544 may interact with the AUSF 1542 and the UE 1502 to perform various security anchor and context management functions.
  • AMF 1544 may be a termination point of a RAN CP interface, which may include or be an N2 reference point between the RAN 1504 and the AMF 1544; and the AMF 1544 may be a termination point of NAS (Nl) signaling, and perform NAS ciphering and integrity protection.
  • AMF 1544 may also support NAS signaling with the UE 1502 over an N3 IWF interface.
  • the SMF 1546 may be responsible for SM (for example, session establishment, tunnel management between UPF 1548 and AN 1508); UE IP address allocation and management (including optional authorization); selection and control of UP function; configuring traffic steering at UPF 1548 to route traffic to proper destination; termination of interfaces toward policy control functions; controlling part of policy enforcement, charging, and QoS; lawful intercept (for SM events and interface to LI system); termination of SM parts of NAS messages; downlink data notification; initiating AN specific SM information, sent via AMF 1544 over N2 to AN 1508; and determining SSC mode of a session.
  • SM may refer to management of a PDU session, and a PDU session or “session” may refer to a PDU connectivity service that provides or enables the exchange of PDUs between the UE 1502 and the data network 1536.
  • the UPF 1548 may act as an anchor point for intra-RAT and inter-RAT mobility, an external PDU session point of interconnect to data network 1536, and a branching point to support multi -homed PDU session.
  • the UPF 1548 may also perform packet routing and forwarding, perform packet inspection, enforce the user plane part of policy rules, lawfully intercept packets (UP collection), perform traffic usage reporting, perform QoS handling for a user plane (e.g., packet filtering, gating, UL/DL rate enforcement), perform uplink traffic verification (e.g., SDF- to-QoS flow mapping), transport level packet marking in the uplink and downlink, and perform downlink packet buffering and downlink data notification triggering.
  • UPF 1548 may include an uplink classifier to support routing traffic flows to a data network.
  • the NSSF 1550 may select a set of network slice instances serving the UE 1502.
  • the NSSF 1550 may also determine allowed NSSAI and the mapping to the subscribed S-NSSAIs, if needed.
  • the NSSF 1550 may also determine the AMF set to be used to serve the UE 1502, or a list of candidate AMFs based on a suitable configuration and possibly by querying the NRF 1554.
  • the selection of a set of network slice instances for the UE 1502 may be triggered by the AMF 1544 with which the UE 1502 is registered by interacting with the NSSF 1550, which may lead to a change of AMF.
  • the NSSF 1550 may interact with the AMF 1544 via an N22 reference point; and may communicate with another NSSF in a visited network via an N31 reference point (not shown). Additionally, the NSSF 1550 may exhibit an Nnssf service-based interface.
  • the NEF 1552 may securely expose services and capabilities provided by 3GPP network functions for third party, internal exposure/re-exposure, AFs (e.g., AF 1560), edge computing or fog computing systems, etc.
  • the NEF 1552 may authenticate, authorize, or throttle the AFs.
  • NEF 1552 may also translate information exchanged with the AF 1560 and information exchanged with internal network functions. For example, the NEF 1552 may translate between an AF-Service-Identifier and an internal 5GC information.
  • NEF 1552 may also receive information from other NFs based on exposed capabilities of other NFs. This information may be stored at the NEF 1552 as structured data, or at a data storage NF using standardized interfaces. The stored information can then be re-exposed by the NEF 1552 to other NFs and AFs, or used for other purposes such as analytics. Additionally, the NEF 1552 may exhibit an Nnef servicebased interface.
  • the NRF 1554 may support service discovery functions, receive NF discovery requests from NF instances, and provide the information of the discovered NF instances to the NF instances. NRF 1554 also maintains information of available NF instances and their supported services. As used herein, the terms “instantiate,” “instantiation,” and the like may refer to the creation of an instance, and an “instance” may refer to a concrete occurrence of an object, which may occur, for example, during execution of program code. Additionally, the NRF 1554 may exhibit the Nnrf service-based interface.
  • the PCF 1556 may provide policy rules to control plane functions to enforce them, and may also support unified policy framework to govern network behavior.
  • the PCF 1556 may also implement a front end to access subscription information relevant for policy decisions in a UDR of the UDM 1558.
  • the PCF 1556 exhibit an Npcf service-based interface.
  • the UDM 1558 may handle subscription-related information to support the network entities' handling of communication sessions, and may store subscription data of UE 1502. For example, subscription data may be communicated via an N8 reference point between the UDM 1558 and the AMF 1544.
  • the UDM 1558 may include two parts, an application front end and a UDR.
  • the UDR may store subscription data and policy data forthe UDM 1558 and the PCF 1556, and/or structured data for exposure and application data (including PFDs for application detection, application request information for multiple UEs 1502) for the NEF 1552.
  • the Nudr service-based interface may be exhibited by the UDR 221 to allow the UDM 1558, PCF 1556, and NEF 1552 to access a particular set of the stored data, as well as to read, update (e.g., add, modify), delete, and subscribe to notification of relevant data changes in the UDR.
  • the UDM may include a UDM- FE, which is in charge of processing credentials, location management, subscription management and so on. Several different front ends may serve the same user in different transactions.
  • the UDM-FE accesses subscription information stored in the UDR and performs authentication credential processing, user identification handling, access authorization, registration/mobility management, and subscription management.
  • the UDM 1558 may exhibit the Nudm service-based interface.
  • the AF 1560 may provide application influence on traffic routing, provide access to NEF, and interact with the policy framework for policy control.
  • the 5GC 1540 may enable edge computing by selecting operator/3 rd party services to be geographically close to a point that the UE 1502 is attached to the network. This may reduce latency and load on the network.
  • the 5GC 1540 may select a UPF 1548 close to the UE 1502 and execute traffic steering from the UPF 1548 to data network 1536 via the N6 interface. This may be based on the UE subscription data, UE location, and information provided by the AF 1560. In this way, the AF 1560 may influence UPF (re)selection and traffic routing.
  • the network operator may permit AF 1560 to interact directly with relevant NFs. Additionally, the AF 1560 may exhibit an Naf service-based interface.
  • the data network 1536 may represent various network operator services, Internet access, or third party services that may be provided by one or more servers including, for example, application/content server 1538.
  • FIG 16 schematically illustrates a wireless network 1600 in accordance with various embodiments.
  • the wireless network 1600 may include a UE 1602 in wireless communication with an AN 1604.
  • the UE 1602 and AN 1604 may be similar to, and substantially interchangeable with, like-named components described elsewhere herein.
  • the UE 1602 may be communicatively coupled with the AN 1604 via connection 1606.
  • the connection 1606 is illustrated as an air interface to enable communicative coupling, and can be consistent with cellular communications protocols such as an LTE protocol or a 5GNR protocol operating at mmWave or sub-6GHz frequencies.
  • the UE 1602 may include a host platform 1608 coupled with a modem platform 1610.
  • the host platform 1608 may include application processing circuitry 1612, which may be coupled with protocol processing circuitry 1614 of the modem platform 1610.
  • the application processing circuitry 1612 may run various applications for the UE 1602 that source/sink application data.
  • the application processing circuitry 1612 may further implement one or more layer operations to transmit/receive application data to/from a data network. These layer operations may include transport (for example UDP) and Internet (for example, IP) operations
  • the protocol processing circuitry 1614 may implement one or more of layer operations to facilitate transmission or reception of data over the connection 1606.
  • the layer operations implemented by the protocol processing circuitry 1614 may include, for example, MAC, RLC, PDCP, RRC and NAS operations.
  • the modem platform 1610 may further include digital baseband circuitry 1616 that may implement one or more layer operations that are “below” layer operations performed by the protocol processing circuitry 1614 in a network protocol stack. These operations may include, for example, PHY operations including one or more of HARQ-ACK functions, scrambling/descrambling, encoding/decoding, layer mapping/de-mapping, modulation symbol mapping, received symbol/bit metric determination, multi-antenna port precoding/decoding, which may include one or more of space-time, space-frequency or spatial coding, reference signal generation/detection, preamble sequence generation and/or decoding, synchronization sequence generation/detection, control channel signal blind decoding, and other related functions.
  • PHY operations including one or more of HARQ-ACK functions, scrambling/descrambling, encoding/decoding, layer mapping/de-mapping, modulation symbol mapping, received symbol/bit metric determination, multi-antenna port precoding/decoding, which may
  • the modem platform 1610 may further include transmit circuitry 1618, receive circuitry 1620, RF circuitry 1622, and RF front end (RFFE) 1624, which may include or connect to one or more antenna panels 1626.
  • the transmit circuitry 1618 may include a digital-to-analog converter, mixer, intermediate frequency (IF) components, etc.
  • the receive circuitry 1620 may include an analog-to-digital converter, mixer, IF components, etc.
  • the RF circuitry 1622 may include a low-noise amplifier, a power amplifier, power tracking components, etc.
  • RFFE 1624 may include filters (for example, surface/bulk acoustic wave filters), switches, antenna tuners, beamforming components (for example, phase-array antenna components), etc.
  • transmit/receive components may be specific to details of a specific implementation such as, for example, whether communication is TDM or FDM, in mmWave or sub-6 gHz frequencies, etc.
  • the transmit/receive components may be arranged in multiple parallel transmit/receive chains, may be disposed in the same or different chips/modules, etc.
  • the protocol processing circuitry 1614 may include one or more instances of control circuitry (not shown) to provide control functions for the transmit/receive components.
  • a UE reception may be established by and via the antenna panels 1626, RFFE 1624, RF circuitry 1622, receive circuitry 1620, digital baseband circuitry 1616, and protocol processing circuitry 1614.
  • the antenna panels 1626 may receive a transmission from the AN 1604 by receive-beamforming signals received by a plurality of antennas/antenna elements of the one or more antenna panels 1626.
  • a UE transmission may be established by and via the protocol processing circuitry 1614, digital baseband circuitry 1616, transmit circuitry 1618, RF circuitry 1622, RFFE 1624, and antenna panels 1626.
  • the transmit components of the UE 1604 may apply a spatial filter to the data to be transmitted to form a transmit beam emitted by the antenna elements of the antenna panels 1626.
  • the AN 1604 may include a host platform 1628 coupled with a modem platform 1630.
  • the host platform 1628 may include application processing circuitry 1632 coupled with protocol processing circuitry 1634 of the modem platform 1630.
  • the modem platform may further include digital baseband circuitry 1636, transmit circuitry 1638, receive circuitry 1640, RF circuitry 1642, RFFE circuitry 1644, and antenna panels 1646.
  • the components of the AN 1604 may be similar to and substantially interchangeable with like- named components of the UE 1602.
  • the components of the AN 1608 may perform various logical functions that include, for example, RNC functions such as radio bearer management, uplink and downlink dynamic radio resource management, and data packet scheduling.
  • Figure 17 is a block diagram illustrating components, according to some example embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein.
  • Figure 17 shows a diagrammatic representation of hardware resources 1700 including one or more processors (or processor cores) 1710, one or more memory /storage devices 1720, and one or more communication resources 1730, each of which may be communicatively coupled via a bus 1740 or other interface circuitry.
  • a hypervisor 1702 may be executed to provide an execution environment for one or more network slices/sub-slices to utilize the hardware resources 1700.
  • the processors 1710 may include, for example, a processor 1712 and a processor 1714.
  • the processors 1710 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 radiofrequency integrated circuit (RFIC), another processor (including those discussed herein), or any suitable combination thereof.
  • CPU central processing unit
  • RISC reduced instruction set computing
  • CISC complex instruction set computing
  • GPU graphics processing unit
  • DSP such as a baseband processor, an ASIC, an FPGA, a radiofrequency integrated circuit (RFIC), another processor (including those discussed herein), or any suitable combination thereof.
  • the memory /storage devices 1720 may include main memory, disk storage, or any suitable combination thereof.
  • the memory /storage devices 1720 may include, but are not limited to, any type of volatile, non-volatile, or semi-volatile memory such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), Flash memory, solid-state storage, etc.
  • DRAM dynamic random access memory
  • SRAM static random access memory
  • EPROM erasable programmable read-only memory
  • EEPROM electrically erasable programmable read-only memory
  • Flash memory solid-state storage, etc.
  • the communication resources 1730 may include interconnection or network interface controllers, components, or other suitable devices to communicate with one or more peripheral devices 1704 or one or more databases 1706 or other network elements via a network 1708.
  • the communication resources 1730 may include wired communication components (e.g., for coupling via USB, Ethernet, etc.), cellular communication components, NFC components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components, and other communication components.
  • Instructions 1750 may comprise software, a program, an application, an applet, an app, or other executable code for causing at least any of the processors 1710 to perform any one or more of the methodologies discussed herein.
  • the instructions 1750 may reside, completely or partially, within at least one of the processors 1710 (e.g., within the processor's cache memory), the memory /storage devices 1720, or any suitable combination thereof.
  • any portion of the instructions 1750 may be transferred to the hardware resources 1700 from any combination of the peripheral devices 1704 or the databases 1706. Accordingly, the memory of processors 1710, the memory/storage devices 1720, the peripheral devices 1704, and the databases 1706 are examples of computer-readable and machine-readable media.
  • the electronic device(s), network(s), system(s), chip(s) or component(s), or portions or implementations thereof, of Figures 15-17, or some other figure herein may be configured to perform one or more processes, techniques, or methods as described herein, or portions thereof.
  • the process 1800 may include, at 1805, Receiving, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE).
  • the process further includes, at 1810, retrieving UE measurement report information from a memory.
  • the process further includes, at 1815, determining, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on the UE measurement report information and the request for location information, wherein the location information for the UE includes an indication of a predicted location for the UE.
  • the process further includes, at 1820, encoding a message for transmission to the gNB that includes an indication of the determined location information for the UE.
  • the process 1900 includes, at 1905, receiving, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE).
  • the process further includes, at 1910, determining, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on UE measurement report information and the request for location information wherein the location information for the UE includes an indication of a predicted location for the UE.
  • the process further includes, at 1915, encoding a message for transmission to the gNB that includes an indication of the determined location information for the UE.
  • the process 2000 includes, at 2005, encoding, for transmission to a network function, a request for location information for a user equipment (UE).
  • the process further includes, at 2010, receiving, from the network function, the location information for the UE, wherein the location information for the UE includes an indication of a predicted location for the UE.
  • the process further includes, at 2015, determining configuration information that includes an indication of a combination of candidate handover cells and respective priorities for the candidate handover cells.
  • the process further includes, at 2020, encoding, for transmission to the UE, a message that includes the configuration information.
  • At least one of the components set forth in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and/or methods as set forth in the example section below.
  • the baseband circuitry as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below.
  • circuitry associated with a UE, base station, network element, etc. as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below in the example section.
  • Example 1 may include NG-RAN can generate a group of predicted candidate target cells for handover according to the predicted UE location information.
  • Example 2 may include the NG-RAN in example 1 or some other example herein, wherein requests UE current/future location information from OAM/CN or LMF.
  • Example 3 may include the NG-RAN in example 1 or some other example herein, wherein receives UE current/future location information from UE report.
  • Example 4 may include the OAM/CN/LMF sends the predicted future UE's location information to NG-RAN in example 1.
  • Example 5 may include the group of target cells in example 1 or some other example herein, wherein includes handover configuration, handover execution timestamp and priority of each cell.
  • Example 6 may include the source gNB in example 1 or some other example herein, wherein sends the handover request together with the handover execution timestamp to target candidate gNBs.
  • Example 7 may include the source gNB in example 1 or some other example herein, wherein configures UE with the combination/list of cells with handover execution timestamp and the corresponding priority.
  • Example 8 may include the UE in example 7 or some other example herein wherein selects handover target cell according to the priority and execute handover at the execution timestamp.
  • Example 9 may include the UE in example 7 or some other example herein, wherein may delete the candidate cell from the list if the execution timestamp expires.
  • Example 10 may include the UE in example 7 or some other example herein, wherein may continuously select the highest priority candidate cell from the remaining list if handover is failed.
  • Example 11 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to request future location information as in example 2 according to UE speed.
  • Example 12 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on statistic of mobility events.
  • Example 13 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on UE information.
  • Example 14 may include NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on UE device type or environment.
  • Example 15 may include if UE is predicted with slow moving, sit in the cell border, the source gNB in example 1 decides not to perform handover.
  • Example 16 may include the network in example 1 or some other example herein, wherein collects and reports the mobility events for each UE independently.
  • Example 17 may include a method of a NG-RAN node, the method comprising: receiving predicted UE location information associated with a UE; and generating a set of one or more predicted candidate target cells for the UE based on the predicted UE location information.
  • Example 18 may include the method of example 17 or some other example herein, wherein the predicted UE location information is received from a 0AM, a CN, or an LMF.
  • Example 19 may include the method of example 17 or some other example herein, wherein the predicted UE location information is received from the UE.
  • Example 20 may include the method of example 17-19 or some other example herein, further comprising encoding, for transmission, a request for the predicted UE location information, wherein the predicted UE location information is received responsive to the request.
  • Example 21 may include the method of example 17-20 or some other example herein, further comprising generating a handover configuration that includes the set of predicted candidate target cells and a handover execution timestamp and/or a priority of individual cells of the set of predicted candidate target cells.
  • Example 22 may include the method of example 21 or some other example herein, further comprising encoding the handover configuration for transmission to the UE.
  • Example 23 may include the method of example 17-22 or some other example herein, wherein the NG-RAN node is a source gNB, and wherein the method further comprises encoding a handover request for transmission to a target gNB, wherein the handover request includes a handover execution timestamp.
  • Example 24 may include the method of example 17-23 or some other example herein, further comprising sending a request to the UE for the predicted UE location information based on a speed of the UE a statistic of mobility events, a UE device type, and/or a UE environment.
  • Example 25 may include the method of example 17-24 or some other example herein, wherein the predicted UE location information is determined based on an artificial intelligence and/or machine learning (AI/ML) model.
  • AI/ML artificial intelligence and/or machine learning
  • Example 26 may include a method of a UE, the method comprising: receiving a handover configuration that includes one or more predicted target cells and at least one of a handover execution timestamp or a priority for respective cells of the one or more predicted target cells; and executing a handover to a target cell of the one or more predicted target cells based on the handover configuration.
  • Example 27 may include the method of example 26 or some other example herein, wherein the handover is executed at a time indicated by the handover execution timestamp associated with the target cell.
  • Example 28 may include the method of example 26-27 or some other example herein, further comprising selecting the target cell from the one or more predicted target cells based on the respective priorities.
  • Example 29 may include the method of example 26-28 or some other example herein, further comprising removing a first predicted target cell of the one or more predicted target cells from consideration for the handover based on expiration of the handover execution timestamp associated with the first predicted target cell.
  • Example 30 may include the method of example 26-29 or some other example herein, further comprising, upon failure of the handover to the target cell, selecting another target cell from the one or more predicted candidate target cells based on the associated priority.
  • Example 31 may include the method of example 26-30 or some other example herein, further comprising obtaining predicted UE location information for the UE and providing the predicted UE location information to a gNB.
  • Example 32 may include the method of example 31 or some other example herein, wherein the handover configuration is based on the predicted UE location information.
  • Example 33 may include the method of example 31-32 or some other example herein, wherein the predicted UE location information is obtained and/or provided responsive to a request from the gNB.
  • Example 34 may include the method of example 31-33 or some other example herein, wherein the predicted UE location information is obtained using an artificial intelligence and/or machine learning (AI/ML) model.
  • AI/ML artificial intelligence and/or machine learning
  • Example QI may include NG-RAN node should configure ML model training and inference configuration for defining how UE trains the AI/ML model, e.g. sending Model training configuration to the UE.
  • Example Q2 may include NG-RAN node in Example QI or some other example herein, wherein should configure whether the UE can perform training/re-training via RRC signaling, e.g. Training Indicator.
  • Example Q3 may include NG-RAN node in Example QI or some other example herein, wherein should configure whether the UE can use its private AI/ML model via RRC signaling, e.g. an indicator of private model training.
  • Example Q4 may include NG-RAN node in Example QI or some other example herein, wherein should configure model training configuration to the UE, including optimizer, loss (function), metrics, loss weight, accuracy level, steps per execution, aggregate bits, etc.
  • Example Q5 may include the UE in Example QI or some other example herein, wherein should report the evaluation results, where model training is following configuration in Example 4, through RRC messages (e.g. via Training Report in Machine learning Report or Model update request, etc).
  • Example Q6 may include NG-RAN node in Example QI or some other example herein, wherein further indicates the UE in Example 5 whether the model is well-trained and can be used for predict! on/decision making by sending a training feedback via RRC message.
  • Example Q7 may include upon receiving an ACK training feedback in Example Q6 or some other example herein, wherein from NG-RAN node, the UE in Example Q5 can use the model trained/retrained by itself for decision making/prediction.
  • Example Q8 may include upon receiving a NACK training feedback in Example 6 or some other example herein, wherein from NG-RAN node, the UE in Example Q5 need to further train/re-train the AI/ML model, and repeat example Q5 and example Q6.
  • Example Q9 may include NG-RAN node in Example 1 or some other example herein, wherein may also configure the model training threshold, indicating when the AI/ML is well- trained, to the UE together with model training configuration in Example Q4.
  • Example Q10 may include the training threshold is met, the UE in Example QI or some other example herein, wherein should provide a training complete feedback to the network.
  • Example QI 1 may include the UE may request model downloading from the NG-RAN node in Example 1 or some other example herein, wherein by indicating the interested service or requested model to the network via RRC message.
  • Example Q12 may include theNG-RAN node in Example Qll or some other example herein, wherein should also configure the required input/output of AI/ML model when UE in Example QI requests model download from itself.
  • Example Q13 may include the UE in Example Qll or some other example herein, wherein may further re-train the AI/ML model downloaded from NG-RAN, following Example QI to Example Q10.
  • Example Q14 may include a method of a UE, the method comprising: receiving, from a NG-RAN node, a model training and inference configuration to indicate information for the UE to train an artificial intelligence (Al) and/or machine learning (ML) (AI/ML) model; and training the AI/ML model based on the configuration.
  • Al artificial intelligence
  • ML machine learning
  • Example Q15 may include the method of Example Q14 or some other example herein, further comprising receiving, from the NG-RAN node, a training indicator to indicate whether the UE is permitted to perform training/re-training of the AI/ML model.
  • Example Q16 may include method of Example Q14-Q15 or some other example herein, further comprising receiving, from the NG-RAN node an indication of whether the UE is permitted to use a private AI/ML model.
  • Example QI 7 may include the method of example Q14-Q16 or some other example herein, wherein the configuration, the training indicator, and/or the indicator are received via RRC signaling.
  • Example A01 includes a method comprising: model update between network nodes triggered by one of or multiple trigger conditions (e.g., timer, accuracy level, confidence level, performance feedback, and loss of AI/ML model).
  • trigger conditions e.g., timer, accuracy level, confidence level, performance feedback, and loss of AI/ML model.
  • Example A02 includes a new message is used to configure machine learning update configuration from RAN to UE.
  • Example A03 includes a new message is used to configure machine learning update configuration from OAM/CN to RAN.
  • Example A04 includes the method of example A01 and/or some other example(s) herein, wherein the timer is configured by the network nodes who deploys AI/ML model to other network nodes, and upon the configured timer is approaching, UE or RAN may send a model update request to RAN or CN/OAM, respectively, and CN/OAM or RAN will then provide an updated model to the corresponding network node.
  • Example A05 includes the method of example A01 and/or some other example(s) herein, wherein the timer is maintained by the network nodes who deploys AI/ML model, and upon the timer is approaching, CN/OAM or RAN will provide an updated model to the RAN or UE, respectively.
  • Example A06 includes the method of example A01 and/or some other example(s) herein, wherein the timer is maintained by the network nodes who deploys AI/ML model to other network nodes, and the CN/OAM or RAN may trigger model update based on the observed accuracy level.
  • Example A07 includes the method of example A01 and/or some other example(s) herein, wherein the confidence level is configured by the network nodes who deploys AI/ML model to other network nodes, and the UE or RAN may trigger model update request to RAN or UE.
  • Example A08 includes the method of example A01 and/or some other example(s) herein, wherein the confidence level is reported by UE or RAN to RAN or CN/OAM, respectively, and the network may update the AI/ML model if confidence level is lower than certain threshold.
  • Example A09 includes the method of examples A01-A08 and/or some other example(s) herein, wherein the UE or RAN may trigger model update request to RAN or UE.
  • Example A10 includes the method of examples A01-A09 and/or some other example(s) herein, wherein the UE or RAN reports the performance indicator to RAN or OAM/CN (respectively), indicating performance improvement or down-gradation, and RAN or OAM/CN may send an updated AI/ML model if the performance is downgraded.
  • Example All includes the method of examples A01-A10 and/or some other example(s) herein, wherein the UE or RAN may trigger model update request to RAN or UE based on performance observation of its own.
  • Example Al 2 includes the method of example A01 and/or some other example(s) herein, wherein the loss is reported by UE or RAN to RAN or CN/OAM, respectively, and the network may update the AI/ML model if loss is higher than certain threshold.
  • Example B01 includes amethod comprising: triggering an artificial inteligence (Al) and/or machine learning (ML) model update between one or more network nodes in response to one or more trigger conditions.
  • Al artificial inteligence
  • ML machine learning
  • Example B02 includes the method of example B01 and/or some other example(s) herein, wherein the one or more trigger conditions include expiration of a timer, a accuracy level, a confidence level, performance feedback, and loss of an AI/ML model.
  • Example B03 includes the method of examples B01-B02 and/or some other example(s) herein, further comprising: sending a message to configure AI/ML model update configuration from a radio access network (RAN) to a user equipment (UE).
  • RAN radio access network
  • UE user equipment
  • Example B04 includes the method of examples B01-B03 and/or some other example(s) herein, further comprising: sending a message to configure AI/ML model update configuration from an operations, administrations, and management function (0AM) and/or core network (CN) to a RAN.
  • AI/ML model update configuration from an operations, administrations, and management function (0AM) and/or core network (CN) to a RAN.
  • CN core network
  • Example B05 includes the method of examples B01-B04 and/or some other example(s) herein, wherein a timer is configured by the one or more network nodes who deploy the AI/ML model to one or more other network nodes.
  • Example B06 includes the method of example B05 and/or some other example(s) herein, further comprising: as the timer approaches expiration or upon expiration of the timer, sending the AI/ML model update request to a RAN, a CN, or 0AM.
  • Example B07 includes the method of example B06 and/or some other example(s) herein, further comprising: obtaining an updated AI/ML model from the CN, the 0AM, or the RAN.
  • Example B08 includes the method of examples B06-B07 and/or some other example(s) herein, wherein the CN, the 0AM, and/or the RAN trigger the AI/ML model update based on an observed accuracy level and/or in response to one or more trigger conditions.
  • Example B09 includes the method of examples B01-B08 and/or some other example(s) herein, wherein a confidence level is configured by the one or more network nodes who deploy AI/ML model to other network nodes, and a UE or a RAN triggers the AI/ML model update request to the RAN or the UE.
  • Example BIO includes the method of examples B01-B09 and/or some other example(s) herein, wherein the confidence level is reported by a UE or a RAN to the RAN, the CN, and/or the 0AM, and the method comprises: updating the AI/ML model if the confidence level is lower than a threshold confidence level.
  • Example Bl l includes the method of examples B01-B10 and/or some other example(s) herein, wherein the UE or the RAN trigger the AI/ML model update request to the RAN or the UE.
  • Example B12 includes the method of examples B01-B11 and/or some other example(s) herein, wherein a UE or a RAN reports the performance indicator to the RAN, the CN, and/or the 0AM, and the method comprises: indicating performance improvement or down-gradations, and sending or obtaining an updated AI/ML model if the performance is downgraded from the RAN, the CN, and/or the 0AM.
  • Example B13 includes the method of examples B01-B12 and/or some other example(s) herein, wherein the UE or the RAN trigger the AI/ML model update request to the RAN or the UE based on performance observation of its own.
  • Example B14 includes the method of examples BO 1 -Bl 3 and/or some other example(s) herein, wherein the loss is reported by the UE or the RAN to the RAN, the CN, and/or the 0AM, and the method comprises: updating the AI/ML model if the loss is higher than a threshold amount of loss.
  • Example B15 includes the method of examples B01-B14 and/or some other example(s) herein, wherein the method is performed by a UE, a RAN, a RAN node, a CN function, and/or an 0 AM.
  • Example XI includes an apparatus comprising: memory to store user equipment (UE) measurement report information; and processing circuitry, coupled with the memory, to: receive, from a next-generation NodeB (gNB), a request for location information for the UE; retrieve the UE measurement report information from the memory; determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on the UE measurement report information and the request for location information, wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
  • gNB next-generation NodeB
  • Al positioning artificial intelligence
  • ML machine learning
  • Example X2 includes the apparatus of example XI or some other example herein, wherein the UE measurement report information includes historical and current information for the UE.
  • Example X3 includes the apparatus of example X2 or some other example herein, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
  • Example X4 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
  • Example X5 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
  • Example X6 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE.
  • Example X7 includes the apparatus of any of examples XI -X6 or some other example herein, wherein the processing circuitry is to implement a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
  • NWDAAF network data analytics function
  • AM operation administration and maintenance
  • LMF location management function
  • Example X8 includes one or more computer-readable media storing instructions that, when executed by one or more processors, cause a network function to: receive, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE); determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on UE measurement report information and the request for location information wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
  • gNB next-generation NodeB
  • UE user equipment
  • ML machine learning
  • Example X9 includes the one or more computer-readable media of example X8 or some other example herein, wherein the UE measurement report information includes historical and current information for the UE.
  • Example XI 0 includes the one or more computer-readable media of example X9 or some other example herein, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
  • Example XI 1 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
  • Example XI 2 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
  • Example XI 3 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE.
  • Example XI 4 includes the one or more computer-readable media of any of examples X8- XI 3 or some other example herein, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
  • NWDAAF network data analytics function
  • AM operation administration and maintenance
  • LMF location management function
  • Example XI 5 includes one or more computer-readable media storing instructions that, when executed by one or more processors, cause a next-generation NodeB (gNB) to: encode, for transmission to a network function, a request for location information for a user equipment (UE); receive, from the network function, the location information for the UE, wherein the location information for the UE includes an indication of a predicted location for the UE; determine configuration information that includes an indication of a combination of candidate handover cells and respective priorities for the candidate handover cells; and encode, for transmission to the UE, a message that includes the configuration information.
  • gNB next-generation NodeB
  • Example XI 6 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein determining the configuration information includes determining whether the UE should start a handover.
  • Example XI 7 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein the configuration information includes an indication of a target cell from the candidate handover cells to perform a handover, and timing information for performing the handover.
  • Example XI 8 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein the configuration information includes conditional handover (CHO) information.
  • the configuration information includes conditional handover (CHO) information.
  • Example XI 9 includes the one or more computer-readable media of example XI 8 or some other example herein, wherein the CHO information includes an indication of conditional handover execution timing.
  • Example X20 includes the one or more computer-readable media of example XI 8 or some other example herein, wherein the CHO information includes an indication of timing information for each candidate cell in the combination of candidate handover cells.
  • Example X21 includes the one or more computer-readable media of examples X15-X20 or some other example herein, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
  • NWDAAF network data analytics function
  • AM operation administration and maintenance
  • LMF location management function
  • Example Z01 may include an apparatus comprising means to perform one or more elements of a method described in or related to any of examples 1-X21, or any other method or process described herein.
  • Example Z02 may include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of a method described in or related to any of examples 1- X21, or any other method or process described herein.
  • Example Z03 may include an apparatus comprising logic, modules, or circuitry to perform one or more elements of a method described in or related to any of examples 1- X21, or any other method or process described herein.
  • Example Z04 may include a method, technique, or process as described in or related to any of examples 1- X21, or portions or parts thereof.
  • Example Z05 may include an apparatus comprising: one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
  • Example Z06 may include a signal as described in or related to any of examples 1- X21, or portions or parts thereof.
  • Example Z07 may include a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
  • PDU protocol data unit
  • Example Z08 may include a signal encoded with data as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
  • Example Z09 may include a signal encoded with a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
  • PDU protocol data unit
  • Example Z10 may include an electromagnetic signal carrying computer-readable instructions, wherein execution of the computer-readable instructions by one or more processors is to cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
  • Example Zll may include a computer program comprising instructions, wherein execution of the program by a processing element is to cause the processing element to carry out the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
  • Example Z12 may include a signal in a wireless network as shown and described herein.
  • Example Z13 may include a method of communicating in a wireless network as shown and described herein.
  • Example Z14 may include a system for providing wireless communication as shown and described herein.
  • Example Z15 may include a device for providing wireless communication as shown and described herein. Any of the above-described examples may be combined with any other example (or combination of examples), unless explicitly stated otherwise.
  • the foregoing description of one or more implementations provides illustration and description, but is not intended to be exhaustive or to limit the scope of embodiments to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of various embodiments.
  • Gateway Function Premise Measurement CHF Charging 50 Equipment 85 CSI-RS CSI
  • CSI-RSRP CSI CID Cell-ID
  • Indicator received power CIM Common 55
  • CPU CSI processing 90
  • CSI-RSRQ CSI Information Model unit Central reference signal
  • CIR Carrier to Processing Unit received quality Interference Ratio
  • C/R CSI-SINR CSI CK Cipher Key
  • C ommand/Respo signal-to-noise and CM Connection 60 nse field bit 95 interference ratio Management
  • Conditional Access Network Multiple Access Mandatory Cloud RAN CSMA/CA CSMA CM AS Commercial CRB Common with collision Mobile Alert Service 65 Resource Block 100 avoidance CMD Command CRC Cyclic CSS Common Search CMS Cloud Redundancy Check Space, Cell- specific Management System CRI Channel-State Search Space CO Conditional Information Resource CTF Charging Optional 70 Indicator, CSI-RS 105 Trigger Function CTS Clear-to-Send DSL Domain Specific 70 ECSP Edge
  • EREG enhanced REG Channel/Half Programmable Gate enhanced resource 55 rate 90
  • E-UTRA Evolved FDD Frequency GGSN Gateway GPRS
  • NodeB Number Access and Backhaul distributed unit HHO Hard Handover ICIC Inter-Cell GNSS Global HLR Home Location Interference Navigation Satellite 55 Register 90 Coordination
  • IMC IMS Credentials ISDN Integrated 85 ksps kilo-symbols per
  • Machine Management Entity MN Master Node MSIN Mobile Station NE-DC NR-E- MNO Mobile Identification 70 UTRA Dual Network Operator Number Connectivity MO Measurement MSISDN Mobile NEF Network Object, Mobile 40 Subscriber ISDN Exposure Function
  • NPUSCH wake-up signal Primary CC
  • NS Network Service Packet Data Convergence PNFD Physical 70 PSCCH Physical Protocol layer Network Function Sidelink Control PDCCH Physical Descriptor Channel Downlink Control PNFR Physical PSSCH Physical Channel 40 Network Function Sidelink Shared PDCP Packet Data Record 75 Channel Convergence Protocol POC PTT over PSCell Primary SCell PDN Packet Data Cellular PSS Primary Network, Public PP, PTP Point-to- Synchronization
  • RIV Resource Compression Radio Network indicator value Temporary SDAP Service Data SI System Identity Adaptation Protocol, Information S-TMSI SAE Service Data SI-RNTI System Temporary Mobile Adaptation Information RNTI
  • SAPD Service Access Description Protocol Package Point Descriptor SDSF Structured Data SL Sidelink SAPI Service Access Storage Function SLA Service Level Point Identifier 50 SDT Small Data 85 Agreement SCC Secondary Transmission SM Session Component Carrier, SDU Service Data Management Secondary CC Unit SMF Session
  • SCell Secondary Cell SEAF Security Anchor Management Function
  • SCEF Service 55 Function 90 SMS Short Message Capability Exposure SeNB secondary eNB Service Function
  • Protocol 70 S-GW Serving Gateway 105 Scheduling SQN Sequence Signal based Signal to TCP Transmission number Noise and Interference 70 Communication
  • UDP User Datagram 55 search space 90 VPLMN Visited Protocol UTRA UMTS Public Land Mobile
  • circuitry refers to, is part of, or includes hardware components such as an electronic circuit, a logic circuit, a processor (shared, dedicated, or group) and/or memory (shared, dedicated, or group), an Application Specific Integrated Circuit (ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-capacity PLD (HCPLD), a structured ASIC, or a programmable SoC), digital signal processors (DSPs), etc., that are configured to provide the described functionality.
  • FPD field-programmable device
  • FPGA field-programmable gate array
  • PLD programmable logic device
  • CPLD complex PLD
  • HPLD high-capacity PLD
  • DSPs digital signal processors
  • the circuitry may execute one or more software or firmware programs to provide at least some of the described functionality.
  • the term “circuitry” may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) with the program code used to carry out the functionality of that program code. In these embodiments, the combination of hardware elements and program code may be referred to as a particular type of circuitry.
  • processor circuitry refers to, is part of, or includes circuitry capable of sequentially and automatically carrying out a sequence of arithmetic or logical operations, or recording, storing, and/or transferring digital data.
  • Processing circuitry may include one or more processing cores to execute instructions and one or more memory structures to store program and data information.
  • processor circuitry may refer to one or more application processors, one or more baseband processors, a physical central processing unit (CPU), a single-core processor, a dual-core processor, a triple-core processor, a quad-core processor, and/or any other device capable of executing or otherwise operating computerexecutable instructions, such as program code, software modules, and/or functional processes.
  • Processing circuitry may include more hardware accelerators, which may be microprocessors, programmable processing devices, or the like.
  • the one or more hardware accelerators may include, for example, computer vision (CV) and/or deep learning (DL) accelerators.
  • CV computer vision
  • DL deep learning
  • application circuitry and/or “baseband circuitry” may be considered synonymous to, and may be referred to as, “processor circuitry.”
  • interface circuitry refers to, is part of, or includes circuitry that enables the exchange of information between two or more components or devices.
  • interface circuitry may refer to one or more hardware interfaces, for example, buses, I/O interfaces, peripheral component interfaces, network interface cards, and/or the like.
  • user equipment or “UE” as used herein refers to a device with radio communication capabilities and may describe a remote user of network resources in a communications network.
  • user equipment or “UE” may be considered synonymous to, 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.
  • user equipment or “UE” may include any type of wireless/wired device or any computing device including a wireless communications interface.
  • network element refers to physical or virtualized equipment and/or infrastructure used to provide wired or wireless communication network services.
  • network element may be considered synonymous to and/or referred to as a networked computer, networking hardware, network equipment, network node, router, switch, hub, bridge, radio network controller, RAN device, RAN node, gateway, server, virtualized VNF, NFVI, and/or the like.
  • computer system refers to any type interconnected electronic devices, computer devices, or components thereof. Additionally, the term “computer system” and/or “system” may refer to various components of a computer that are communicatively coupled with one another. Furthermore, the term “computer system” and/or “system” may refer to multiple computer devices and/or multiple computing systems that are communicatively coupled with one another and configured to share computing and/or networking resources.
  • appliance refers to a computer device or computer system with program code (e.g., software or firmware) that is specifically designed to provide a specific computing resource.
  • program code e.g., software or firmware
  • a “virtual appliance” is a virtual machine image to be implemented by a hypervisor-equipped device that virtualizes or emulates a computer appliance or otherwise is dedicated to provide a specific computing resource.
  • resource refers to a physical or virtual device, a physical or virtual component within a computing environment, and/or a physical or virtual component within a particular device, such as computer devices, mechanical devices, memory space, processor/CPU time, processor/CPU usage, processor and accelerator loads, hardware time or usage, electrical power, input/output operations, ports or network sockets, channel/link allocation, throughput, memory usage, storage, network, database and applications, workload units, and/or the like.
  • a “hardware resource” may refer to compute, storage, and/or network resources provided by physical hardware element(s).
  • a “virtualized resource” may refer to compute, storage, and/or network resources provided by virtualization infrastructure to an application, device, system, etc.
  • network resource or “communication resource” may refer to resources that are accessible by computer devices/systems via a communications network.
  • system resources may refer to any kind of shared entities to provide services, and may include computing and/or network resources. System resources may be considered as a set of coherent functions, network data objects or services, accessible through a server where such system resources reside on a single host or multiple hosts and are clearly identifiable.
  • channel refers to any transmission medium, either tangible or intangible, which is used to communicate data or a data stream.
  • channel may be synonymous with and/or equivalent to “communications channel,” “data communications channel,” “transmission channel,” “data transmission channel,” “access channel,” “data access channel,” “link,” “data link,” “carrier,” “radiofrequency carrier,” and/or any other like term denoting a pathway or medium through which data is communicated.
  • link refers to a connection between two devices through a RAT for the purpose of transmitting and receiving information.
  • instantiate refers to the creation of an instance.
  • An “instance” also refers to a concrete occurrence of an object, which may occur, for example, during execution of program code.
  • Coupled may mean two or more elements are in direct physical or electrical contact with one another, may mean that two or more elements indirectly contact each other but still cooperate or interact with each other, and/or may mean that one or more other elements are coupled or connected between the elements that are said to be coupled with each other.
  • directly coupled may mean that two or more elements are in direct contact with one another.
  • communicatively coupled may mean that two or more elements may be in contact with one another by a means of communication including through a wire or other interconnect connection, through a wireless communication channel or link, and/or the like.
  • information element refers to a structural element containing one or more fields.
  • field refers to individual contents of an information element, or a data element that contains content.
  • SMTC refers to an SSB-based measurement timing configuration configured by SSB-MeasurementTimingConflguration.
  • SSB refers to an SS/PBCH block.
  • Primary Cell refers to the MCG cell, operating on the primary frequency, in which the UE either performs the initial connection establishment procedure or initiates the connection re-establishment procedure.
  • Primary SCG Cell refers to the SCG cell in which the UE performs random access when performing the Reconfiguration with Sync procedure for DC operation.
  • Secondary Cell refers to a cell providing additional radio resources on top of a Special Cell for a UE configured with CA.
  • Secondary Cell Group refers to the subset of serving cells comprising the PSCell and zero or more secondary cells for a UE configured with DC.
  • Secondary Cell refers to the primary cell for a UE in RRC CONNECTED not configured with CA/DC there is only one serving cell comprising of the primary cell.
  • serving cell refers to the set of cells comprising the Special Cell(s) and all secondary cells for a UE in RRC_CONNECTED configured with CA/.
  • Special Cell refers to the PCell of the MCG or the PSCell of the SCG for DC operation; otherwise, the term “Special Cell” refers to the Pcell.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

Systems, apparatuses, methods, and computer-readable media are provided for user equipment (UE) trajectory-assisted handovers. In particular, some embodiments may include artificial intelligence (AI) or machine learning (ML) to predict UE location information. Other embodiments may be described and/or claimed.

Description

USER EQUIPMENT TRAJECTORY-ASSISTED HANDOVER
CROSS REFERENCE TO RELATED APPLICATIONS
The present application claims priority to International Application No.
PCT/CN2021/110778, which was filed August 5, 2021; U.S. Provisional Patent Application No. 63/236,456, which was filed August 24, 2021; and to U.S. Provisional Patent Application No. 63/270,445, which was filed October 21, 2021.
FIELD
Various embodiments generally may relate to the field of wireless communications. For example, some embodiments may relate to user equipment (UE) trajectory-assisted handovers. In particular, some embodiments may include artificial intelligence (Al) or machine learning (ML) to predict UE location information.
BACKGROUND
There is currently standardization activity in 3GPP RAN3 studying mobility optimization that can be enhanced using artificial intelligence/machine learning (AI/ML) techniques. The motivation is to guarantee service continuity during mobility by minimizing call drops, radio link failures (RLFs), unnecessary handovers, ping-pong events, etc., especially when a user equipment (UE) is moving at high-speed or when the UE is frequently performing handover procedures across different target nodes. Embodiments of the present disclosure address these and other issues.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments will be readily understood by the following detailed description in conjunction with the accompanying drawings. To facilitate this description, like reference numerals designate like structural elements. Embodiments are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings.
Figure 1 is a flowchart illustrating an example of a process in accordance with various embodiments.
Figure 2 is a flowchart illustrating another example of a process in accordance with various embodiments.
Figure 3 illustrates an example of network controls model training and inference at a UE, with no model downloading fromNG-RAN in accordance with various embodiments.
Figure 4 illustrates an example of a model for inference downloading to the UE in accordance with various embodiments.
Figure 5 illustrates an example of federated learning between the network and the UE in accordance with various embodiments. Figure 6 illustrates an example of how a network decides ML configuration in accordance with various embodiments.
Figure 7 illustrates an example of a general procedure of network-oversighted UE's ML training and inference in accordance with various embodiments.
Figure 8 illustrates an example of a general procedure of model transferring for UE ML inference in accordance with various embodiments.
Figure 9 illustrates an example of a general procedure of model transferring for UE ML training and inference in accordance with various embodiments.
Figure 10 illustrates an example of split AI/ML inference in accordance with various embodiments.
Figure 11 illustrates an example of AI/ML model downloading over a 5G system in accordance with various embodiments.
Figure 12 illustrates an example of federated learning over a 5G system in accordance with various embodiments.
Figure 13 illustrates an example of a RAN intelligence functional framework in accordance with various embodiments.
Figure 14 illustrates an example of another Functional Framework for RAN Intelligence in accordance with various embodiments.
Figure 15 illustrates a network in accordance with various embodiments.
Figure 16 schematically illustrates a wireless network 1600 in accordance with various embodiments.
Figure 17 is a block diagram illustrating components, according to some examples of embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein.
Figures 18, 19, and 20 depict examples of procedures for practicing the various embodiments discussed herein.
DETAILED DESCRIPTION
The following detailed description refers to the accompanying drawings. The same reference numbers may be used in different drawings to identify the same or similar elements. In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular structures, architectures, interfaces, techniques, etc. in order to provide a thorough understanding of the various aspects of various embodiments. However, it will be apparent to those skilled in the art having the benefit of the present disclosure that the various aspects of the various embodiments may be practiced in other examples that depart from these specific details. In certain instances, descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the various embodiments with unnecessary detail. For the purposes of the present document, the phrase “A or B” means (A), (B), or (A and B).
As introduced above, there is currently standardization activity in 3GPP RAN3 studying mobility optimization that can be enhanced using AI/ML techniques. The legacy approach supported in 3GPP SON is to collect feedback (e.g., ping-pong event, failure rate, etc.) so that the operation administration and maintenance (0AM) system can continuously optimize handover strategy based on failure events. However, it is very challenging and costly to develop an optimal handover scheme based on trial-and-error methodology.
3GPP SA5 studied mobility management related issues as a use case in TR 28.809, v. 17.0.0, 2021-04-06, where mobility optimization can be based on resource consumption, inter- gNB beam selection, load balancing, mobility performance, UE trajectory or UE load. It is noticed that the UE location information is provided by the Location Services (LCS), where the location server is resident in the CN. The MDA (management data analytics) function block is responsible for providing ML training and inference service. The use of Al strategies for handover optimization is gaining attention and has also been proposed. However, it is based upon UE's historical record of location information and requires processing a very large amount of information regarding the UE's previous history using its location and can be quite intrusive in terms of UE privacy. This is because the methods require the data to be collected within a single central location where the UE's location information is combined to create various possible trajectories to find the optimal target cells. Given that currently, it is difficult to track UEs across different cells, this means UE's location information is somehow tagged along with UE's ID, which can be highly intrusive and having it present in a single central location presents a security risk. Furthermore, past record while indicative cannot be fully trusted to make the right prediction on which target cell to pick if the UE's route is not the same this time.
In addition, the handover strategy at the 0AM level may be generic and cannot adaptively make real-time handover decision based on each UE's information. To further help optimize handover mechanism for a specific UE, it is essential to optimize mobility management at the RAN level as well.
Instead of tracking a UE's long-term history of location info, a more near-term update on the UE's existing location and trajectory information can greatly help to improve the accuracy and predict the target cell, especially in case of the high-speed mobile UEs. In logged MDT, location information is reported by the UE to the MDT analytics function as a series of location coordinates as deduced through prior Handovers and some historical UE location data. Hence, by knowing the future location of UE (either report from UE or predicted by the network) can further improve the handover successful rate.
Various embodiments herein are directed to an AI/ML assisted handover mechanism located at NG-RAN, where the future UE location is used for mobility enhancement. The future UE location may be determined from: 1) UE reported future UE location information 2) current/future positioning information available at the LMF (location management function)/OAM/CN.
Moreover, in some embodiments, the network may select the suitable UE(s) to use AI/ML model for handover decision, based on 1) UE information 2) mobility events 3) UE device type/ environment.
The described mechanism improves the UE handover success rate with assistance from machine learning-based methods to predict the target cell that has the greatest handover success probability based on the UE's future/near-term location and its trajectory, which can further reduce service interruption at the UE side and reduce network overhead. Since the UE shares the info with the gNB only for the short-term rather than the entire journey, it's easier to preserve UE privacy as a more comprehensive view of UE's location history is not needed to predict the target cell information.
The described mechanism also reduces AI/ML workload for highly active UEs in the network by only selecting the UE(s) which can benefit from AI/ML the most (e.g. better output accuracy, better performance improvement, etc).
Embodiment 1: LMF predict location/positioning-based mobility handover
In this embodiment, mobility handover decision is based on the assistance of LMF provided location information. Following the positioning procedure, NG-RAN will send measurement report to the LMF, which computes the location of the UE. Based on the stored measurement reports, LMF can predict the UE location information. Based on the predicted result of UE location and UE trajectory, NG-RAN can optimize the target cell selection and predict the best timing for certain UE to handover, together with UE's load, handover event collection, etc.
The positioning AI/ML model is trained and inferenced at the LMF, while the predicted results are sent to RAN over NRPPa or NG interface. The predicted location can also be received from MDA or NWDAF located in 0AM or CN, respectively.
Input of the Al-positioning assisted mobility enhancement includes:
Predicted location information of UE (based on UE's current speed/location/trajectory info)
(predicted) load of UE Target cell current/predicted resource status Current/predicted resource status of its own Output of the Al -positioning assisted mobility enhancement includes:
Whether the UE should start handover? the list of target cells with priority and the timing of handover o if UE should handover to another target cell within a threshold of timing, the network will configure the UE to handover to target cell, together with the best time to perform handover. o the network can also configure this outcome “the group of target cells and the corresponding handover timing, priority” to the UE as CHO (conditional handover) candidate cell(s), the corresponding configuration are sent as conditional handover configuration. The UE can execute CHO configuration and handover to the candidate target cell according to the configured handover execution timing. With the timestamp, UE can stop performing measurements to evaluate conditional execution condition. The CHO execution condition need to be enhanced. o The source gNB should also transfer the corresponding predicted handover timing to the candidate target gNB, where Xn interface should be enhanced. o UE selects the target cell to handover according to the priority. If handover is fails, UE will continue to select the remaining target cell which has the highest priority until the handover procedure succeeds. If execution timing of candidate cell(s) expires during UE performs handover, UE can delete this candidate cell from the list.
Predicted configuration or load information
To help the target cell make better decision on the resource allocation, when UE is going to handover to the target cell, the source cell can also send the predicted configuration for the UE based on historical behavior or predicted load information to the target cell. An example of message flows is shown in the flowchart in Figure 1.
To support the prediction location information transfer, the NRPPa or NG can be enhanced. The input data are collected from RAN or based on measurement reports. The output data are transmitted from CN or LMF to RAN.
NG-RAN sends a predict location request to 0AM (e.g. MDA)/CN (e.g. NWDAF/AMF)/LMF, the request can include duration of requested location (e.g. start time stamp + end time stamp/ start time stamp + duration), request location information (relative location, latitude, longitude, altitude, etc), UE ID, etc.
LMF/OAM/CN sends the predicted results to NG-RAN, including the requested location information, accuracy of the output, etc.
Further to support candidate cell configuration, the CHO configuration can also be enhanced. The timing of each candidate cell can be added in the configuration.
An example embodiment with enhancement to the CondReconfigToAddModList IE is shown as below. Changes are marked in underline.
Example 1: An indication “attemptPredictCondReconfig” is used to indicate UE whether it can use the predicted timing for handover configured by the source cell or not.
CondReconfigToAddModList information element
- ASN1 START
- TAG-CONDRECONFIGTOADDMODLIST-START CondReconfigToAddModList-r16 ::= SEQUENCE (SIZE (1.. maxNrofCondCells- r!6)) OF CondReconfigToAddMod-r16 CondReconfigToAddMod-r16 ::= SEQUENCE { condReconfigId-r16 CondReconfigId-r16, condExecutionCond-r16 SEQUENCE (SIZE (1..2)) OF Measld OPTIONAL, - Cond condReconfigAdd condRRCReconfig-r16 OCTET STRING (CONTAINING RRCReconfiguration) OPTIONAL, — Cond condReconfigAdd condExecutionTime-rxy8 INTEGER (Q.,5119) OPTIONAL. — Need N condExecutionPrioritv-rxy INTEGER (L. maxNrofCondCell-r16) OPTIONAL. —Need N
- TAG-CONDRECONFIGTOADDMODLIST-STOP
- ASN1STOP
Figure imgf000008_0001
condExecutionPrioritv
Indicates the priority of candidate cell. The UE selects target cell for handover from the candidate cell with highest priority. If handover failed, the UE continues to select handover target cell according to the priority until handover succeed.
Example 2. Alternatively, CondTriggerConflg can also be considered as an enhancement for this prediction-based trigger event. An example embodiment with enhancement is shown as below.
ReportConfigNR information element
CondTriggerConfig-r16 ::= SEQUENCE { condEventld CHOICE { condEventA3 SEQUENCE { a3-0ffset MeasTriggerQuantityOffset, hysteresis Hysteresis, timeToTrigger TimeToTrigger condEventA5 SEQUENCE { a5-Thresholdl MeasTriggerQuantity, a5-Threshold2 MeasTriggerQuantity, hysteresis Hysteresis, timeToTrigger TimeToTrigger
CondEventPredict SEQUENCE I
> timeToTrigger TimeToTrigger
L
Figure imgf000009_0001
rsType-r16 NR-RS-Type,
Figure imgf000009_0002
timeToTrisser
Time during of CondEventPredict is the specific criteria for the event to trigger in order to execute the UE handover. To support the predicted configuration or load information transfer, the Xn interface (HANDOVER REQUEST) can be enhanced.
An example embodiment with enhancement is shown as below. Changes are marked in underline. HANDOVER REQUEST ACKNOWLEDGE
Figure imgf000010_0001
Embodiment 2: UE based location/positioning-based mobility handover
In this embodiment, mobility handover decision is based on the assistance of UE reported current and near-term location information. The handover target cell(s) are predicted based on the joint AI/ML model optimization for load information, UE assistance location information, handover event and configuration selection, etc.
To support this functionality, the Locationinfo can be enhanced to include the following location information, reporting from UE to the network:
- near-term/future Latitude, longitude (mandatory);
- near-term/future Velocity (mandatory);
- near-term/future Altitude (conditional on availability);
- near-term/future Uncertainty (conditional on availability);
- near-term/future Confidence (conditional on availability);
- near-term/future Direction (conditional on availability).
- near-term/future Timestamp (mandatory)
This could be a location information list which includes the future/predicted locations corresponding to the multiple time points in the near-term.
If location information (e.g. GNSS location information) can either be predicted by UE from AI/ML models or known by UE from positioning function. The uncertainty and confidence are generated based on the results from AI/ML models and positioning algorithms, respectively. An example of a flow chart showing this message flow is shown in Figure 2.
ReportConflgNR can be expanded to configure the location report, e.g. "RocationInfoReportConfig" An example of this embodiment is shown as below, changes are marked in underline.
ReportConflgNR information element
- ASN1 START
- TAG-REPORTCONFIGNR-START
MeasTriggerQuantityCLI-r16 ::= CHOICE { srs-RSRP-r16 SRS-RSRP-Range-r16, cli-RSSI-r16 CLI-RSSI-Range-r16
MeasReportQuantityCLI-r16 ::= ENUMERATED {srs-rsrp, cli-rssi} LocationlnfoReportConfig ::= SEQUENCE { reportinterval _ Reportinterval. reportAmount _ ENUMERATED {rl. r2, r4, r8, r!6, r32, r64, infinity}. includeCommonLocationlnfo-rl 6 ENUMERATED {true} OPTIONAL. — Need R includeFutureLocationlnfoList-rxy ENUMERATED {true} OPTIONAL. — Need R includeBT-Meas-r16 SetupRelease {BT-NameList-r16} OPTIONAL. — Need M includeWLAN-Meas-r16 SetupRelease {WLAN-NameList-r16} OPTIONAL. — Need M includeSensor-Meas-r16 SetupRelease {Sensor-NameList-r16} OPTIONAL — Need M 1
- TAG-REPORTCONFIGNR-STOP
- ASN1STOP
Locationinfo information element
- ASN1 START
- TAG-LOCATIONINFO-START Locationlnfo-r16 ::= SEQUENCE { commonLocationInfo-r16 CommonLocationInfo-r16 OPTIONAL, bt-Locationlnfo-rl 6 LogMeasResultListBT-r16 OPTIONAL, wlan-LocationInfo-r16 LogMeasResultListWLAN-r16 OPTIONAL, sensor-LocationInfo-r16 Sensor-LocationInfo-r16 OPTIONAL, FutureLocationlnfoList-rxy FutureLocationInfoList-r18 OPTIONAL.
- TAG-LOCATIONINFO-STOP
- ASN1STOP
FutureLocationlnfoList information element
- ASN1 START
- TAG-FUTURELOCATIONINFOLIST-START FutureLocationlnfoList-rxy :==
FutureLocationlnfoList-rxy ::= SEQUENCE { future-gnss-TOD-msec-rxy OCTET STRING OPTIONAL. future-locationTimestamp-rxy OCTET STRING OPTIONAL. future-locationCoordinate-rxy OCTET STRING OPTIONAL. future-locationError-rxy OCTET STRING OPTIONAL. future-locationSource-rxy OCTET STRING OPTIONAL. future-velocity Estimate-rxy OCTET STRING OPTIONAL
1
- TAG-FUTURELOCATIONINFO-STOP
- ASN1STOP
Besides, similar as embodiment 1, CHO configuration can be enhanced similarly.
Embodiment 3: Mixed location/positioning-based mobility handover
In this embodiment, the network can utilize the location prediction results received from upper layer and future/near-term navigation information reported by the UE. In this scenario, UE does not need to report the accurate location (e.g.GNSS, latitude, longitude, etc.), UE only need to report the velocity in near future and direction in the FutureLocationlnfoList. This can help to avoid UE reporting its private information (e.g. accurate location in near future) over the air interface.
The example embodiment of standards impact is similar as Embodiment 1 and Embodiment 2, where NRPPa/NG interface can be enhanced to send predicted location to NG- RAN, and ReportConfigNR, Locationinfo, CondReconfigToAddModList can be enhanced correspondingly.
Example Use case: Reducing ping-pong event with the assistance of UE location and trajectory prediction
In this embodiment, AI/ML model is trained to select the best serving cell for a certain UE which is experiencing frequent ping-pong handover between two cells. Based on the predicted location, if AI/ML model predicts a UE is moving slowly from the source cell to the target cell or sits in the cell border for a long time, the network can avoid performing handover of such UE to the target node as it may have leamt that this is a result of current network topology at the current location of the UE. The number of ping-pong events can thus be reduced.
Embodiment 5: UE selection for mobility handover optimization Not all UEs may require AI/ML based models to predict successful HOs and the configured NRTs etc. are enough for them. It is useful to know which UEs require additional processing and restrict the use of additional information exchange for such UEs alone. It is proposed that the network can select the suitable UE criteria to use AI/ML model to improve the mobility handover performance (e.g. increase handover successful rate, etc), and to reduce heavy signaling exchanges and heavy processing through AI/ML.
The (predicted) location information of the selected UEs is transmitted to the RAN with its UE ID. In this embodiment, three types of criteria of UE selection are described.
Embodiment 5-1, Based on the statistic of mobility event:
- Number of successful handover preparations
- Number of failed handover preparations
- Number of successful handover resource allocation
- Number of failed handover resource allocation
- Number of successful handover execution
- Number of failed handover execution
If a UE is lower than a certain threshold (for a successful event count) or higher than a certain threshold (for a failed event count), the network can select the UE for using AI/ML prediction to improve the Handover success rate.
Alternatively, the threshold can also be set by the ratio between above event counts and the corresponding requested handover events. o support this functionality, the selected UE ID should be decided by the network. For embodiment 1, the selected UE ID is transmitted to LMF/MDA/NWDAF to request location information. For embodiment 2, NG-RAN can configure positioning functionality selectively to the selected UEs (if the UE also have the positioning capability), where only the selected UE need to perform positioning calculation/prediction and report accordingly. Additionally, the above listed events should be collected for each UE.
Embodiment 5-2, Based on the UE information and Al/model accuracy measures
The accuracy of the AI/ML model outcome highly depends on the coverage of the training data set.. Sometimes, for certain environment or scenario, the AI/ML model may not be able to make the right/precise decision (as the input is far beyond the data set coverage) and has low accuracy results. In mobility use cases, the wrong decision will lead to severe service interruption caused by failed handover. Hence in the trigger condition for the gNB to seek additional Al support, a measure of past accuracy result for these scenarios along with the UE information could be incorporated. In this case, it is proposed the AI/ML based mobility model is used for the UEs with higher accuracy results which may be for more common scenarios for, e.g. moving along the road with a fixed speed/direction, etc. The easier UE's trajectory can be predictable, the higher accuracy of AI/ML generated predicted handover. The network may select the UEbased on following example UE information:
UE's current/future speed under a threshold
UE's current/future speed is not changed for a certain time
UE's current/future moving direction is unchanged for a certain time
Embodiment 5-3, Based on UE device type
For some environments, UE location is easily predictable as the trajectory is following a known map, such as factory, road, indoor, etc. In this case, the network can also select the UE according to its device type as the type of UE environment may provide additional contextual information that would help the Al model in doing a more accurate job in predicting it's trajectory.
To support this functionality, UE need to report its device type/access category when accessing to the network, new device type/access category may include (but not limited to):
Factory
Vehicle
Indoor device
Mobile phone
Rural area
Outdoor agricultural
NETWORK-OVERSIGHTED AI/ML TRAINING AND INFERENCE
Applications of Machine learning (ML) for RAN intelligence have been widely studied in academia and standardization groups, such as O-RAN, 3GPP, etc. 3GPP Rel-18 also starts to look into how to adopt AI/ML over air interface, by taking advantage of benefits from AI/ML. As defined in TR 37.817, v. 0.1.0, 2021-01-04 (from RAN3 SI “enhancement for data collection for NR and EN-DCoversig”), functional framework of supporting AI/ML in RAN includes “Data Collection”, “Model Training”, “Model Inference” and “Actor”. It is possible that those functional nodes are deployed at either network side or UE side.
Regarding to AI/ML over air interface, in general, the involvement of network and UE in the AI/ML procedure can be considered with two scenarios. Scenario 1 : either the network or the UE performs AI/ML training and inference solely and independently, that is “Model Training”, “Model Inference” and “Actor” are locates either at network side or UE side. In this scenario, model transfer over the air interface may not be needed. For majority scenarios, it can be achieved by implementation or with additional assistance information transferred over the air interface, such as measurement reports or data collections. In the second scenario, to bring more visibility to the network and let the network understand the behavior/knowledge of UE's AI/ML model, it is also possible that the initial AI/ML models (e.g. AI/ML model may not be totally suitable for one UE) running at the UEs are received/configured by the network over certain configurations. Moving further when integration between AI/ML and wireless network becoming tighter, the UE can download well-trained model from the network and perform “Model inference”, when “Model training” at the network side. The UE may even consider federated learning together with the network.
Figures 3, 4, and 5 show the three scenarios of AI/ML model at the UE to be oversighted by the network. In particular, Figure 3 illustrates an example of model training/inference at the UE side, while UE training/inference policy/configuration is configured by NG-RAN. Figure 4 illustrates an example of model training at the network side (RAN), while model inference at the UE side. Figure 5 illustrates an example of model training and model inference locates at both network side and UE side (e.g. federated learning).
To achieve a more understandable AI/ML behavior at the UEs and trust the result/ outcome generated by AI/ML model at the UE side, the network should be able to control how the UEs perform training and inference. That is, besides the model update policies and other aspects, the network should be able to control the AI/ML training/inference behavior at the UE side, so that it can control the damage in case the training/inference at the UE side is wrong or non-convergence.
Various embodiments herein provide mechanisms to support model training and inference at UE side, where the UE AI/ML behavior can be configured/oversighted by the network. To generate a trustable and reliable results from AI/ML model located at UE side, different scenarios of AI/ML training and inference may be analyzed when they are jointly or separately located at UE and NG-RAN, then describe the configuration and signaling procedure to support the network- oversighted AI/ML at UE, including configuration parameters, conditions of configuration, etc.
The described mechanism can support UE to generate the result/actions which can be trusted by the network when using AI/ML at UE side. The described mechanism helps to build a reliable wireless system when using AI/ML at network and UE, especially when UE is allowed to perform model training/re-training.
To allow the network to control the ML training and inference at the UE side, consider two-levels of information which need to be exchanged between the network and the UE: 1) whether require model downloading from the network, and 2) network-oversighted UE's model training/inference configuration.
In some embodiments, to exchange the first level of information (e.g. request to download model from the network or not), the UE can leam the machine learning capability and network supported models/services enabled by AI/ML through the broadcast information (e.g. MachineLeamingSupport field) via MLCapabilitylndication. For those services where the network can offer the AI/ML models, UE can decide whether to register or request a ML model from the network based on the received information, and send the corresponding service registration/interest indication message back to the network as model downloading request. In this scenario, the network is responsible for model transfer, update, etc. However, it is also possible that the UE will not request model from the network, this can either because the UE does not have the ML capability, or because the UE holds the ML model itself.
The same model training/inference configuration is also adoptable when the UE is allowed to perform local model training (e.g. federated learning).
In some embodiments, the AI/ML model can always be downloaded to the UE according to the UE ML capability and request. In the present disclosure, ML related information exchanging may be expanded to have a wider scope, where the network can configure UE training and inference, where model downloading from the network is not a constraint.
As illustrated below in Table 1, six scenarios are identified to help the network to determine what configuration(s) the network need to send to the UE. Figure 6 illustrates an example of a process flow for a network to determine ML configuration.
Moreover, two sets of ML configurations are described to be send over air interface via RRC signaling.
Set 1 : UE training configuration (ModelTrainingConfiguration)
Model ID
Service ID (e.g. CSI feedback, positioning, beam management, etc) Trainingindicator
If this field is present, it indicates that the UE is allowed to perform training/re-training. If it is not presented, the UE is not configured to further perform training/re-training, either for its own model or model received from network.
PrivateModelTraining
If this field is present, it indicates that the UE is allowed to use its own ML model. If it is not presented, the UE should not use its own ML model, even it has the corresponding ML capability.
The training configuration is considered per model basis, where the model type is determined by modeltype reported by the UE via UE-CapabilityML-Information. The network can configure the full-set or the subset of the supported model at the UE.
The UE reports supported configuration to the network in UE capability “UE- CapabilityML-Information”; network then configure the optimized configuration/threshold based on network training experience and best knowledge of one use case. Model type (CNN, RNN, DNN, regression, etc)
Model configuration
The UE should perform model training according to the following configuration list, then return with the evaluation results through new IE “TrainingReport” in MachineLeamingReport or in MachineLeamingModelUpdateRequest a separate message from UE to network. If the reported evaluation results meet the performance threshold, the network can send an ACK of “TrainingFeedback” to the UE via MachineLeamingConfiguration or in MachineLeamingModelUpdateRequest or a new message from network to UE, indicating UE's trained AI/ML can be used. Otherwise, if UE receives a NACK of “TrainigFeedback” from the network, UE should continue training, and repeat the previous procedure.
An alternative can be NW configure the metrics together with the corresponding threshold for UE. Once all thresholds are met, UE can declare that the training is done via new IE “TrainingComplete” via MachineLeamingReport or a separate message from UE to network.
The configuration may include: o Optimizer
Optimizer Algorithm that the UE should use for training optimization include
SGD, RMSprop, Adam, Adadelta, Adagrad, Adamax, Nadam, Ftrl
Besides the algorithm configuration, the following two parameters should also be set to help UE get convergence to the training model:
Learning rate; — the learning rate of the optimizer
Weights. — the state of the optimizer. o Loss
Compute the quantity that a model should seek to minimize during training.
Losses functions/classes that the UE should use for training evaluation include three categories:
Probabilistic losses;
Regression losses;
Hinge losses for maximum-margin classification
}, where
Probabilistic losses Binary Cross entropy ;
CategoricalCrossentropy;
SparseCategoricalCrossentropy;
Poisson;
KLDivergence;
Regression losses
MeanS quaredError ;
MeanAbsoluteError;
MeanAbsolutePercentageError;
MeanS quaredLogarithmi cError;
CosineSimilarity;
Huber;
LogCosh;
And Hinge losses for maximum-margin classification
Hinge;
SquaredHinge;
CategoricalHinge; o Metrics
The metrics is used to judge the performance of the training model.
Six metrics are described to be included in the configuration:
Accuracy metrics;
Probabilistic metrics;
Regression metrics;
Classification metrics;
Image segmentation metrics; Hinge metrics.
}, where
Accuracy metrics
Accuracy;
Binary Accuracy; CategoricalAccuracy TopKCategorical Accuracy SparseTopKCategoricalAccuracy
Probabilistic metrics
BinaryCrossentropy CategoricalCrossentropy SparseCategoricalCrossentropy KLDivergence
Poisson
Regression Metrics
MeanS quaredError RootMeanSquaredError MeanAbsoluteError MeanAbsolutePercentageError MeanS quaredLogarithmi cError CosineSimilarity
LogCoshError
Classification metrics
AUC
Precision
Recall
TruePositives TrueNegatives FalsePositives
FalseNegatives
PrecisionAtRecall
SensitivityAtSpecificity
Specificity AtSensitivity
Image Segmentation metrics
MeanloU
}, and
Hinge metrics
Hinge
SquaredHinge
CategoricalHinge o Loss weight
Optional list or dictionary specifying scalar coefficients to weight the loss contributions of different model outputs. o Accuracy level
This parameter is the lower bound of one AI/ML model need to meet. The UE may stop AI/ML model training when the configured accuracy is met. o Steps per execution
The number of batches to run during each call.
For federated learning between network and UE, the following parameters may also need to be configured: o Aggregate bit
Data that each UE can return to the base station after local retraining
Additionally, network can also configure the wireless KPI (e.g. throughput, SINR, BLER, etc) and AI/ML model KPI (e.g. accuracy level of the model, confidence level of the model, etc) for different use cases.
Set 2: ML model related information, including service type, report configuration, policy guidance, behavior type, bias threshold and variance threshold for model update request, model itself, model parameter update. Moreover, recalling that the UE may download ML model from the network, while only perform ML inference. The network should also indicate the UE whether it can continue perform model (re-)training for each transferred ML model.
Additionally, the required inputs to the ML model for training and inference should also be configured per model basis.
Model input
This is used to carry the fields/parameters that are defined as ML model input for model training/inference if UE's ML model is received from the network.
Model output
This is used to carry the fields/parameters that are defined as ML model output for model training/inference if UE's ML model is received from the network.
Moreover, UE should also send a request or an indicator informing the network whether model downloading/transferring is required or not. This indicator can be sent together with service registration/interest indication message.
The RRC impacts may be used used to carry the above signaling and messages.
Embodiment 1 : No model transfer/ downloading from the network to the UE
This embodiment presents a scenario where the model is trained and inferenced at the UE side. To help the network understand the behavior at the UE side, the training at the UE side should be oversighted by the network in certain degree, so that the network can trust the result/outcome of the ML inference.
To support the above scenario, Set 1 configuration should be sent to the UE who are going to perform ML training and inference independently. Figure 7 shows an example of a general procedure of network-oversighted UE's ML training and inference.
Embodiment 2: UE downloads ML model from the network, while only perform ML inference
In this embodiment, UE can only perform ML inference to the received ML model from the network. To support it, both Set 1 and Set 2 configurations should be configured to the applied UE. An example of this general procedure is shown in Figure 8.
Embodiment 3: UE downloads ML model from the network, performing both ML (re-)training and ML inference
Similar as Embodiment 2, ML model is transferred from the network to the UE, while the UE is allowed to perform (re-training) to the received model according to the training configuration. Both Set 1 and Set 2 configuration needs to be supported. The procedure supporting embodiment 3 is shown in Figure 9. The configuration sets and the freedom of UE performing ML training and inference is summarized below in Table 1. TABLE 1: NW/UE training/inference configuration selection
A: ML capability available at the network?
B: Does UE require to download ML model from the network?
C: ML capability available at the UE side?
Figure imgf000023_0001
ANNEX
• SGD
• Name
The name of SGD. “sgd”. • Learning rate
The learning rate. Floating point value.
• Momentum A parameter that accelerates gradient descent in the relevant direction and dampens oscillations.
• Nesterov
Indicate whether to apply Nesterov momentum. Boolean {True, False}.
• RMSprop
• Name
The name of RMSprop. “RMSprop”.
• Learning rate
The learning rate. Floating point value.
• Rho
Discounting factor for the history/coming gradient.
• Momentum
A parameter that accelerates gradient descent in the relevant direction and dampens oscillations.
• Epsilon
A small constant for numerical stability.
• Centered
Indicate whether gradients are normalized by the estimated variance of the gradient (true) or by the uncentered second moment (false). Boolean {True, False}.
• Adam
• Name
The name of Adam. “Adam”.
• Learning rate
The learning rate. Floating point value.
• Beta l
The exponential decay rate for the 1st moment estimates.
• Beta_2
The exponential decay rate for the 2nd moment estimates.
• Epsilon
A small constant for numerical stability.
• amsgrad
Indicate whether to apply AMSGrad variant of this algorithm. Boolean {True, False}.
• Adadelta • Name
The name of Adadelta. “Adadelta”.
• Learning rate
The learning rate. Floating point value.
• rho
The decay rate. Range:
• Epsilon
A small constant for numerical stability.
• Adagrad
• Name
The name of Adagrad. “Adagrad”.
• Learning rate
The learning rate. Floating point value.
• Initial accumulator value
Starting value for the accumulators (per parameter momentum values). Range:
• Epsilon
A small constant for numerical stability.
• Adamax
• Name
The name of Adamax. “Adamax”.
• Learning rate
The learning rate. Floating point value.
• Beta l
The exponential decay rate for the 1st moment estimates.
• Beta_2
The exponential decay rate for the exponentially weighted infinity norm.
• Epsilon
A small constant for numerical stability.
• Nadam
• Name
The name of Nadam. “Nadam”.
• Learning rate
The learning rate. Floating point value.
• Beta l The exponential decay rate for the 1st moment estimates. Range:
• Beta_2
The exponential decay rate for the exponentially weighted infinity norm.
• Epsilon
A small constant for numerical stability.
• Ftrl
• Name
The name of Adagrad. “Adagrad”.
• Learning rate
The learning rate. Floating point value.
• Learning rate power
Controls how the learning rate decreases during training. Float value. Range:
• Initial accumulator value
The starting value for accumulator.
• LI regularization strength
Float value.
• L2 regularization strength Stabilization penalty.
• L2 shrinkage regularization strength
A magnitude penalty, that will only happen on the active weights.
• beta
Float value.
AI/ML Model Updates in an NG-RAN Network
3 GPP RAN3 started to study Al intelligence in RAN network in Rel-17, where RAN intelligence functional framework is discussed, including mapping of AI/ML functionality into existing NG-RAN architecture. In 3GPP RAN3, RAN intelligence will first address use cases related to SON (self-organizing network), for example, load balancing, network energy saving, and mobility optimization. 3GPP Rel-18 will also investigate how to adopt AI/ML over air interface.
In mobile communications systems, mobile devices (e.g., smartphones, automotive, robots, etc.) are increasingly replacing conventional algorithms (e.g., speech recognition, image recognition, video processing, etc.) with AI/ML models to enable applications. The 5G system (5GS) can at least support three types of AI/ML operations: AI/ML operation splitting between AI/ML endpoints; AI/ML model/data distribution and sharing over 5G system; and Distributed/Federated Learning over 5G system.
An example of a scheme of split AI/ML inference is depicted in Figure 10. The AI/ML operation/model is split into multiple parts according to the current task and environment. The intention is to offload the computation-intensive, energy-intensive parts to network endpoints, whereas leave the privacy-sensitive and delay-sensitive parts at the end device. The device executes the operation/model up to a specific part/layer and then sends the intermediate data to the network endpoint. The network endpoint executes the remaining parts/layers and feeds the inference results back to the device.
An example of a scheme of AI/ML model distribution is depicted in Figure 11. Multifunctional mobile terminals might need to switch the AI/ML model in response to task and environment variations. The condition of adaptive model selection is that the models to be selected are available for the mobile device. However, given the fact that the AI/ML models are becoming increasingly diverse, and with the limited storage resource in a UE, it can be determined to not pre-load all candidate AI/ML models on-board. Online model distribution (e.g., new model downloading) is needed, in which an AI/ML model can be distributed from a NW endpoint to the devices when they need it to adapt to the changed AI/ML tasks and environments. For this purpose, the model performance at the UE needs to be monitored constantly.
An example of a scheme of Federated Learning (FL) is depicted as in Figure 12. The cloud server trains a global model by aggregating local models partially-trained by each end devices. Within each training iteration, a UE performs the training based on the model downloaded from the Al server using the local training data. Then the UE reports the interim training results to the cloud server via 5G UL channels. The server aggregates the interim training results from the UEs and updates the global model. The updated global model is then distributed back to the UEs and the UEs can perform the training for the next iteration.
As discussed in 3GPP TR 37.817 vO.l.O (2021-01-04) (“[TR37817]”) (from RAN3 SI “enhancement for data collection for NR and EN-DC”), a functional framework of supporting AI/ML in RAN includes “Data Collection”, “Model Training”, “Model Inference” and “Actor”. Where to realize each AI/ML functionality highly depends on use cases to address and various requirements related to data collection and performance (e.g., latency, real-time, etc).
Previous proposals have considered different deployment types of AI/ML model training and model inference functionalities between CN, RAN and UE. In some deployment options, one network node/UE may deploy an AI/ML model to another one for model inference or re-training purpose. Additionally, 3GPP SAI initiated the study of traffic characteristics and performance requirements for AI/ML model transfer in 5GS, which also support model downloading and update from network to UE (see e.g., 3GPP TR 22.874 V18.0.1 (2021-06-24) (“[TR22874]”)).
An example of a RAN intelligence functional framework is shown in Figure 13, where model deployment/update from “Model training” to “Model inference” are issues for future study. An example of another Functional Framework for RAN Intelligence is shown in Figure 14.
The accuracy of AI/ML models highly depends on training data (e.g., data that is used to train an AI/ML model), where different scenarios and events should be considered as part of the training process. Hence, it is important to monitor the performance of models deployed in the "Model inference" and to timely provide an update if necessary, so that the accuracy of generated outputs from “Model inference” can be guaranteed with best performance all the time. However, none of the current work in 3GPP (SAI, SA2, SA5 and RAN3) started to discuss on how to update the deployed AI/ML model(s) over the system (from CN/ operations, administrations, and management function (0 AM) to RAN or from RAN to UE).
The present disclosure provides a mechanism to update AI/ML models (e.g., for inferences, predictions, federated learning, etc.) between network nodes (e.g., from CN and/or 0AM to RAN) or between network and UE (e.g., from CN, 0AM, and/or RAN to UE). The AI/ML model update can be triggered based on following conditions: 1) timer (e.g., expiration of a timer); 2) loss weight of model outputs; 3) accuracy level of AI/ML model; 4) probabilistic of AI/ML model generated output; 5) performance feedback; 6) per request from NG-RAN or UE; and/or 7) trigger/indication from CN/OAM or NG-RAN.
The mechanisms discussed herein maintain a good performance AI/ML model by updating models between network node(s) and/or UE(s) and between different network nodes according to certain rules, trigger conditions, etc. These mechanisms can also help avoid unnecessary model updates, which also helps to save radio resources for model exchanging and conserve computational resources.
1. AI/ML Model Update Mechanisms
AI/ML includes processing of data collection, model training, model validation and model inference. Normally, a well-trained model is trained based on massive input data, which also need to be validated based on another set of data which is not used for training. In wireless system, the environment (including traffic, radio, accessed number of UE, etc) may be various. It's hard to train a comprehensive AI/ML model which can suit for all scenarios just based on several rounds of training or initial offline training. To maintain a good accuracy level of prediction results for which network can trust, it is necessary for the network node/UE with model inference functionality to get an up-to-date AI/ML model which can accommodate to the changes in the environment.
Hence, the network node/UE with model training functionality should continuously train AI/ML based on feedbacks collected from Actor (which is distributed in the network, e.g., UE, gNB-DU, gNB-CU, etc), regardless of online (e.g., RL) or offline (e.g., supervised learning, unsupervised learning) training mechanism used. The updated/re-trained/re-tuned AI/ML model should be further deployed to network nodes/UE with model inference functionality.
Several issues should be considered when defining model update mechanism:
1. Considering AI/ML model normally has a large size, it is important to avoid unnecessary deployment and frequent model update
2. The purpose of updating AI/ML model is to provide better performance and output with higher accuracy to the system. One deployed model may be outdated or not suitable for current environment. This requires the Model training function either to update the model based on performance feedback from Actor or based on other mechanisms.
3. Model update and synchronization for federated learning. When model training and model inference are deployed at different locations in the network, federated learning can be considered. In previous proposals, one central trained model is located at central server (e.g., CN or RAN, respectively) and another local retrained model is located at local nodes (e.g., RAN or UE, respectively). Certain interaction and synchronization of trained models between central server and local nodes are required. For federated learning, the central server may also selectively update the model to the local nodes, considering workload or other impacts.
The present disclosure provides various embodiments to support model update between network nodes with model training (e.g., OAM/CN for model deployment to RAN, RAN for model deployment to UE) and network nodes/UE with model inference (e.g., RAN for model deployment from CN/OAM, UE for model deployment from RAN).
1.1. Timer Based Model Update Trigger
In this embodiment, the model update is triggered from model training (e.g., OAM/CN or RAN) based on the configured timer. The timer can either be configured and running at network nodes with model training or network nodes/UE with model inference.
1.1.1. Model Training Side
If the timer is running at model training side, the timer can be used to: indicate the duration of one AI/ML model's update/re-train; the model is retrained based on data collected during this period; and/or trigger model update and send updated AI/ML model to network nodes/UE with model inference. 1.1.2. Model Inference Side
If the timer is running at model inference side, the timer can be used to: trigger model update and send model update request message from NG-RAN to OAM/CN or from UE to NG- RAN. Under this scenario, the timer is configured from CN/OAM to NG-RAN via NG interface or El interface; or the timer is configured from NG-RAN to UE via Uu interface as part of model configuration in Machine Learning Configuration message.
An example for the model training side and/or model inference side embodiments is shown below:
MachineLeamingConfiguration ::= SEQUENCE { criticalExtensions CHOICE { machineLeamingConfiguration MachineLeamingConfiguration-IEs, machineLeamingModelUpdateConfiguration MachineLeamingModelUpdateConfiguration, criticalExtensionsFuture SEQUENCE {}
MachineLeamingModelUpdateConfiguration ::= SEQUENCE { modelUpateConfigToRemoveList ModelUpateConfigToRemoveList
OPTIONAL, modelUpdateConfigToAddModList ModelUpdateConfigToAddModList OPTIONAL, ModelUpdateReconfigToAddModList-rxy ::= SEQUENCE (SIZE (L. maxNrofModel- rxy)) OF ModelUpdateConfigToAddMod-rxy ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY, updateTimer ENUMERATED {sf50, sflOO, sf200,.., infinity} OPTIONAL,
ModelUpdateConfigToAddModList
The IE ModelUpdateConfigToAddModList concerns a list of model update configurations to add or modify, with for each entry the modelldentifier and the associated trigger condition/threshold. ’ Network may configure or set different timer according to different use cases or to different models for different use cases.
To support federated learning, a central server (CN/OAM or RAN) is allowed to set the same timer to all local nodes (RAN or UE) to support synchronized model training. If the central server (CN/OAM or RAN) sets different timers to different local nodes (RAN or UE), asynchronized model training is considered for federated averaging.
Together with the updated model, a model ID is required to be encoded in the model update message from OAM/CN to RAN or in RRC reconfiguration (fromNG-RAN to UE).
1.2. Accuracy level of AI/ML model Trigger
In this embodiment, the model update can be triggered by accuracy level of one AI/ML model. The accuracy level can be calculated by comparing predicted results with the real value at the model training node during validation or testing. If the model accuracy level is higher than that of the pervious deployed model, the model training node (e.g., CN/OAM or RAN) can provide this updated model to model inference (RAN or UE). The model training node may also calculate the accuracy level gain comparing accuracy level of current model and previous one. Model update is only triggered if the accuracy gain is higher than the threshold.
1.3. Confidence level of AI/ML model Trigger
In this embodiment, the model update can be triggered by confidence level of one AI/ML model. Confidence level is calculated at the model inference node, by comparing prediction results and the real value (real future input to model inference). If the confidence level is lower than certain threshold, a model update should be triggered to avoid using the wrong/inaccurate predicted results/actions. This could also be triggered by either network or UE.
1.3.1. Configured Confidence Threshold
In this embodiment, the OAM/CN or the NG-RAN configures a threshold of confidence level to NG-RAN or UE via AI/ML model update configuration (as part of model management). If the confidence level calculated by network nodes/UE for model inference (e.g., RAN or UE) is lower than the confidence level threshold, the RAN/UE should send a model update request to the network nodes with model training functionality (e.g., CN/OAM or RAN). Upon receiving the model update request, an updated model should be sent to the requested node. Normally, the confidence level is between [0,1], here, for example, an integer value is used to represent 10 times of exact loss value.
An example of model update configuration from NG-RAN to UE is shown as below:
ModelUpdateConfigT o AddMod modelldentifier Modelidentifier MANDATORY, confidenceLevelThreshold ENUMERATED {1, 2, 3,.., 10} OPTIONAL,
1.3.2. Network Node Triggering
In this embodiment, network nodes with model training (e.g., CN/OAM or RAN) can trigger model update based on received confidence level report from model inference (e.g., RAN or UE). If the confidence level in the report is lower than the value expected at the model training, network nodes can send anew up-to-date model to its model inference node (e.g., RAN or UE).
The network may configure or set different confidence level threshold according to different use cases. For example, UE trajectory normally requires a high accuracy, while some non-real-time prediction can have a relax threshold.
1.4. Performance Feedback Trigger
In this embodiment, the model update is triggered by performance feedback (e.g., UE measurement report, performance indicator from network node or its neighbor nodes, etc). The model will only be updated when the performance feedback shows a poor performance or performance down-gradation.
In some embodiments, the Actor who performs predicted results/actions provides performance feedback (e.g., UE measurement report, SON/MDT measurement report, etc) to the network. Based on the performance feedback, the network nodes with model training can provide an updated AI/ML model to network nodes/UE with model inference to optimize the predicted results.
Additionally or alternatively, the Actor can compare its previous performance and its performance after taking predicted results/actions. If the performance (e.g., throughput, SINR, energy cost, etc) is down-graded, the Actor can directly send a simple feedback indicator “performance down-graded” to the network node with model training. Upon receiving this “downgraded” feedback indicator, OAM/CN or RAN can provide an updated model to the model inference node, e.g., RAN or UE, respectively. An example of this “performance indicator” in Uu interface is shown as below:
MachineLeamingReport-IEs ::= SEQUENCE { serviceType ENUMERATED {default, positioning, V2X, spare5, spared, spare3, spare2, spare 1} OPTIONAL, modelUpdate OCTET STRING (SIZE(L.8000)) OPTIONAL, modelBias ENUMERATED {poor-bias, less-poor, no-bias, spare5, spared, spare3, spare2, spare 1} OPTIONAL, modelvariance INTEGER (0, ... , 100) OPTIONAL, predictionResult PredictionResult OPTIONAL, confidenceLevel ENUMERATED {0, 1 , 2, ... , 10} OPTIONAL, performanceindicator ENUMERATED {improved, downgraded, nochange} OPTIONAL, lateNonCriticalExtension OCTET STRING
OPTIONAL, nonCriticalExtension SEQUENCE}} OPTIONAL
This performance indicator can also be added in UE MDT measurement report, L2 measurement report, etc.
The performance feedback to trigger model update can be different across use cases.
For example, for handover mobility use case, this could be handover failure rate or other handover events; for energy saving, this could be an increase of energy consumption of ES-Cell or neighbor cell; for load balancing, it could be the network node's average packet delay/throughput or a percentage of UE's packet delay/throughput; for CSI feedback, it could be system or UE throughput.
1.5. Model Loss Trigger
In this embodiment, the model update is triggered based on the model loss. This embodiment is mainly suitable for federated learning, e.g., between OAM/CN and NG-RAN; between NG-RAN and UE. A certain threshold of loss is configured by central node during model deployment and configuration. However, when a model is first deployed to a node (RAN or UE), it is still possible it may not be converged and need some time to reduce model loss by taking several rounds of re-training. Hence, it is also proposed a fix number of epochs or a given period is configured together with loss.
The output of loss function may be various based on the exact algorithm, the example integer can also be replaced by dB or other unit per use case.
Alternatively, to avoid specifying algorithms, a normalized value is proposed to calculate the loss gain. The loss gain is the relative percentage of loss improvement compared with last calculated loss of the model. Normally, the gain value is between [0,1], here an integer value is used to represent 10 times of exact gain value.
1.5.1. RAN-UE Model Update
When the loss of a trained model at UE is larger than the configured loss threshold, the local node (RAN or UE) should trigger a request to the central server (CN/OAM or RAN) via NG/E1 or Uu interface (e.g., Machine Learning Model Update Request), asking for a model update. An example of RAN-UE model update configuration triggered by loss is shown as below: ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY, lossStartTimeOffset ENUMERATED {sflO, sf20, sf50,..} OPTIONAL, lossStartEpoch ENUMERATED {10, 20, 30,..} OPTIONAL, lossThreshold ENUMERATED {1, 2, 3,.., 10} OPTIONAL, lossGain ENUMERATED {1, 2, 3,.., 10} OPTIONAL,
1.5.2. Central Node
Additionally or alternatively, the model update can also be managed by a central server, if the loss weights of local nodes (e.g., RAN or UE) are reported to central node (e.g., OAM/CN or RAN). Based on received model loss weight, the central node may know whether this AI/ML model need to be refined/updated. The loss in the model report can also be replaced by some categorized level of loss (e.g., high loss, low loss, etc.). Based on the report, whether to update the model or not is decided by the central server - The model update may only be triggered if the central node receives a high loss in the model report from local nodes.
An example of loss report to be added in an ML report is shown below:
MachineLeamingReport-IEs ::= SEQUENCE { serviceType ENUMERATED {default, positioning, V2X, spare5, spared, spare3, spare2, sparel} OPTIONAL, modelUpdate OCTET STRING (SIZE(L.8OOO)) OPTIONAL, modelBias ENUMERATED {poor-bias, less-poor, no-bias, spare5, spared, spare3, spare2, sparel} OPTIONAL, modelvariance INTEGER (0, ... , 100) OPTIONAL, predictionResult PredictionResult OPTIONAL, confidenceLevel ENUMERATED {0, 1 , 2, ... , 10} OPTIONAL, loss INTEGER (0, 1, .., 9, 10)
OPTIONAL, lateNonCriticalExtension OCTET STRING
OPTIONAL, nonCriticalExtension SEQUENCE {} OPTIONAL
Instead of model loss, the model variance and model bias can also be used as trigger event for model update. 1.6. Multiple Triggers
One AI/ML model may also be configured with multiple trigger condition. In this embodiment, a modelUpdateTriggerCategory may include all possibilities of a trigger condition. An example is shown below:
ModelUpdateConfigT o AddMod modelldentifier Modelldentifier MANDATORY, modelUpdateTriggerCategory ENUMERATED {timer, accuracylevel, confidencelevel, performancefeedback, loss, spare3, spare2, sparel} OPTIONAL, modelUpdateTriggerToAddModList ModelUpdateTriggerToAddModList
ModelUpdateTriggerToAddModList-rxy ::= SEQUENCE (SIZE (1.. maxNrofModelUpdateCat-rxy)) OF ModelUpdateTriggerToAddMod-rxy
ModelUpdateTriggerToAddMod updateTimer ENUMERATED {sflO, sf20, sf50,.., infinity} OPTIONAL, confidenceLevelThreshold ENUMERATED {1, 2, 3,.., 10} OPTIONAL, lossStartTimeOffset ENUMERATED {sflO, s£20, sf50,..} OPTIONAL, lossStartEpoch ENUMERATED {10, 20, 30,..} OPTIONAL, lossThreshold ENUMERATED {1, 2, 3,.., 10} OPTIONAL, lossGain ENUMERATED {1, 2, 3,.., 10} OPTIONAL,
SYSTEMS AND IMPLEMENTATIONS
Figures 15-17 illustrate various systems, devices, and components that may implement aspects of disclosed embodiments.
Figure 15 illustrates a network 1500 in accordance with various embodiments. The network 1500 may operate in a manner consistent with 3GPP technical specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard and the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like. The network 1500 may include a UE 1502, which may include any mobile or non-mobile computing device designed to communicate with a RAN 1504 via an over-the-air connection. The UE 1502 may be, but is not limited to, a smartphone, tablet computer, wearable computer device, desktop computer, laptop computer, in-vehicle infotainment, in-car entertainment device, instrument cluster, head-up display device, onboard diagnostic device, dashtop mobile equipment, mobile data terminal, electronic engine management system, electron! c/engine control unit, electronic/engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked appliance, machine-type communication device, M2M or D2D device, loT device, etc.
In some embodiments, the network 1500 may include a plurality of UEs coupled directly with one another via a sidelink interface. The UEs may be M2M/D2D devices that communicate using physical sidelink channels such as, but not limited to, PSBCH, PSDCH, PSSCH, PSCCH, PSFCH, etc.
In some embodiments, the UE 1502 may additionally communicate with an AP 1506 via an over-the-air connection. The AP 1506 may manage a WLAN connection, which may serve to offload some/all network traffic from the RAN 1504. The connection between the UE 1502 and the AP 1506 may be consistent with any IEEE 802.11 protocol, wherein the AP 1506 could be a wireless fidelity (Wi-Fi®) router. In some embodiments, the UE 1502, RAN 1504, and AP 1506 may utilize cellular- WLAN aggregation (for example, LWA/LWIP). Cellular- WLAN aggregation may involve the UE 1502 being configured by the RAN 1504 to utilize both cellular radio resources and WLAN resources.
The RAN 1504 may include one or more access nodes, for example, AN 1508. AN 1508 may terminate air-interface protocols for the UE 1502 by providing access stratum protocols including RRC, PDCP, RLC, MAC, and LI protocols. In this manner, the AN 1508 may enable data/voice connectivity between CN 1520 and the UE 1502. In some embodiments, the AN 1508 may be implemented in a discrete device or as one or more software entities running on server computers as part of, for example, a virtual network, which may be referred to as a CRAN or virtual baseband unit pool. The AN 1508 be referred to as a BS, gNB, RAN node, eNB, ng-eNB, NodeB, RSU, TRxP, TRP, etc. The AN 1508 may be a macrocell base station or a low power base station for providing femtocells, picocells or other like cells having smaller coverage areas, smaller user capacity, or higher bandwidth compared to macrocells.
In embodiments in which the RAN 1504 includes a plurality of ANs, they may be coupled with one another via an X2 interface (if the RAN 1504 is an LTE RAN) or an Xn interface (if the RAN 1504 is a 5G RAN). The X2/Xn interfaces, which may be separated into control/user plane interfaces in some embodiments, may allow the ANs to communicate information related to handovers, data/context transfers, mobility, load management, interference coordination, etc.
The ANs of the RAN 1504 may each manage one or more cells, cell groups, component carriers, etc. to provide the UE 1502 with an air interface for network access. The UE 1502 may be simultaneously connected with a plurality of cells provided by the same or different ANs of the RAN 1504. For example, the UE 1502 and RAN 1504 may use carrier aggregation to allow the UE 1502 to connect with a plurality of component carriers, each corresponding to a Pcell or Scell. In dual connectivity scenarios, a first AN may be a master node that provides an MCG and a second AN may be secondary node that provides an SCG. The first/second ANs may be any combination of eNB, gNB, ng-eNB, etc.
The RAN 1504 may provide the air interface over a licensed spectrum or an unlicensed spectrum. To operate in the unlicensed spectrum, the nodes may use LAA, eLAA, and/or feLAA mechanisms based on CA technology with PCells/Scells. Prior to accessing the unlicensed spectrum, the nodes may perform medium/carrier-sensing operations based on, for example, a listen-before-talk (LBT) protocol.
In V2X scenarios the UE 1502 or AN 1508 may be or act as a RSU, which may refer to any transportation infrastructure entity used for V2X communications. An RSU may be implemented in or by a suitable AN or a stationary (or relatively stationary) UE. An RSU implemented in or by: a UE may be referred to as a “UE-type RSU”; an eNB may be referred to as an “eNB-type RSU”; a gNB may be referred to as a “gNB-type RSU”; and the like. In one example, an RSU is a computing device coupled with radio frequency circuitry located on a roadside that provides connectivity support to passing vehicle UEs. The RSU may also include internal data storage circuitry to store intersection map geometry, traffic statistics, media, as well as applications/software to sense and control ongoing vehicular and pedestrian traffic. The RSU may provide very low latency communications required for high speed events, such as crash avoidance, traffic warnings, and the like. Additionally or alternatively, the RSU may provide other cellular/WLAN communications services. The components of the RSU may be packaged in a weatherproof enclosure suitable for outdoor installation, and may include a network interface controller to provide a wired connection (e.g., Ethernet) to a traffic signal controller or a backhaul network.
In some embodiments, the RAN 1504 may be an LTE RAN 1510 with eNBs, for example, eNB 1512. The LTE RAN 1510 may provide an LTE air interface with the following characteristics: SCS of 15 kHz; CP-OFDM waveform for DL and SC-FDMA waveform for UL; turbo codes 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; and CRS for cell search and initial acquisition, channel quality measurements, and channel estimation for coherent demodulation/detection at the UE. The LTE air interface may operating on sub-6 GHz bands.
In some embodiments, the RAN 1504 may be an NG-RAN 1514 with gNBs, for example, gNB 1516, or ng-eNBs, for example, ng-eNB 1518. The gNB 1516 may connect with 5G-enabled UEs using a 5G NR interface. The gNB 1516 may connect with a 5G core through an NG interface, which may include an N2 interface or an N3 interface. The ng-eNB 1518 may also connect with the 5G core through an NG interface, but may connect with a UE via an LTE air interface. The gNB 1516 and the ng-eNB 1518 may connect with each other over an Xn interface.
In some embodiments, the NG interface may be split into two parts, an NG user plane (NG-U) interface, which carries traffic data between the nodes of the NG-RAN 1514 and a UPF 1548 (e.g., N3 interface), and an NG control plane (NG-C) interface, which is a signaling interface between the nodes of the NG-RAN1514 and an AMF 1544 (e.g., N2 interface).
The NG-RAN 1514 may provide a 5G-NR air interface with the following characteristics: variable SCS; CP-OFDM for DL, CP-OFDM and DFT-s-OFDM for UL; polar, repetition, 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 a CRS, but may use PBCH DMRS for PBCH demodulation; PTRS for phase tracking for PDSCH; and tracking reference signal for time tracking. The 5G-NR air interface may operating on FR1 bands that include sub-6 GHz bands or FR2 bands that include bands from 24.25 GHz to 52.6 GHz. The 5G-NR air interface may include an SSB that is an area of a downlink resource grid that includes PSS/SSS/PBCH.
In some embodiments, the 5G-NR air interface may utilize BWPs for various purposes. For example, BWP can be used for dynamic adaptation of the SCS. For example, the UE 1502 can be configured with multiple BWPs where each BWP configuration has a different SCS. When a BWP change is indicated to the UE 1502, the SCS of the transmission is changed as well. Another use case example of BWP is related to power saving. In particular, multiple BWPs can be configured for the UE 1502 with different amount of frequency resources (for example, PRBs) to support data transmission under different traffic loading scenarios. A BWP containing a smaller number of PRBs can be used for data transmission with small traffic load while allowing power saving at the UE 1502 and in some cases at the gNB 1516. A BWP containing a larger number of PRBs can be used for scenarios with higher traffic load.
The RAN 1504 is communicatively coupled to CN 1520 that includes network elements to provide various functions to support data and telecommunications services to customers/subscribers (for example, users of UE 1502). The components of the CN 1520 may be implemented in one physical node or separate physical nodes. In some embodiments, NFV may be utilized to virtualize any or all of the functions provided by the network elements of the CN 1520 onto physical compute/storage resources in servers, switches, etc. A logical instantiation of the CN 1520 may be referred to as a network slice, and a logical instantiation of a portion of the CN 1520 may be referred to as a network sub-slice.
In some embodiments, the CN 1520 may be an LTE CN 1522, which may also be referred to as an EPC. The LTE CN 1522 may include MME 1524, SGW 1526, SGSN 1528, HSS 1530, PGW 1532, and PCRF 1534 coupled with one another over interfaces (or “reference points”) as shown. Functions of the elements of the LTE CN 1522 may be briefly introduced as follows.
The MME 1524 may implement mobility management functions to track a current location of the UE 1502 to facilitate paging, bearer activation/deactivation, handovers, gateway selection, authentication, etc.
The SGW 1526 may terminate an SI interface toward the RAN and route data packets between the RAN and the LTE CN 1522. The SGW 1526 may be a local mobility anchor point for inter-RAN node handovers and also may provide an anchor for inter-3 GPP mobility. Other responsibilities may include lawful intercept, charging, and some policy enforcement.
The SGSN 1528 may track a location of the UE 1502 and perform security functions and access control. In addition, the SGSN 1528 may perform inter-EPC node signaling for mobility between different RAT networks; PDN and S-GW selection as specified by MME 1524; MME selection for handovers; etc. The S3 reference point between the MME 1524 and the SGSN 1528 may enable user and bearer information exchange for inter-3 GPP access network mobility in idle/active states.
The HSS 1530 may include a database for network users, including subscription-related information to support the network entities' handling of communication sessions. The HSS 1530 can provide support for routing/roaming, authentication, authorization, naming/addressing resolution, location dependencies, etc. An S6a reference point between the HSS 1530 and the MME 1524 may enable transfer of subscription and authentication data for authenticating/ authorizing user access to the LTE CN 1520.
The PGW 1532 may terminate an SGi interface toward a data network (DN) 1536 that may include an application/content server 1538. The PGW 1532 may route data packets between the LTE CN 1522 and the data network 1536. The PGW 1532 may be coupled with the SGW 1526 by an S5 reference point to facilitate user plane tunneling and tunnel management. The PGW 1532 may further include a node for policy enforcement and charging data collection (for example, PCEF). Additionally, the SGi reference point between the PGW 1532 and the data network 15 36 may be an operator external public, a private PDN, or an intra-operator packet data network, for example, for provision of IMS services. The PGW 1532 may be coupled with a PCRF 1534 via a Gx reference point.
The PCRF 1534 is the policy and charging control element of the LTE CN 1522. The PCRF 1534 may be communicatively coupled to the app/content server 1538 to determine appropriate QoS and charging parameters for service flows. The PCRF 1532 may provision associated rules into a PCEF (via Gx reference point) with appropriate TFT and QCI.
In some embodiments, the CN 1520 may be a 5GC 1540. The 5GC 1540 may include an AUSF 1542, AMF 1544, SMF 1546, UPF 1548, NSSF 1550, NEF 1552, NRF 1554, PCF 1556, UDM 1558, and AF 1560 coupled with one another over interfaces (or “reference points”) as shown. Functions of the elements of the 5GC 1540 may be briefly introduced as follows.
The AUSF 1542 may store data for authentication of UE 1502 and handle authentication- related functionality. The AUSF 1542 may facilitate a common authentication framework for various access types. In addition to communicating with other elements of the 5GC 1540 over reference points as shown, the AUSF 1542 may exhibit an Nausf service-based interface.
The AMF 1544 may allow other functions of the 5GC 1540 to communicate with the UE 1502 and the RAN 1504 and to subscribe to notifications about mobility events with respect to the UE 1502. The AMF 1544 may be responsible for registration management (for example, for registering UE 1502), connection management, reachability management, mobility management, lawful interception of AMF-related events, and access authentication and authorization. The AMF 1544 may provide transport for SM messages between the UE 1502 and the SMF 1546, and act as a transparent proxy for routing SM messages. AMF 1544 may also provide transport for SMS messages between UE 1502 and an SMSF. AMF 1544 may interact with the AUSF 1542 and the UE 1502 to perform various security anchor and context management functions. Furthermore, AMF 1544 may be a termination point of a RAN CP interface, which may include or be an N2 reference point between the RAN 1504 and the AMF 1544; and the AMF 1544 may be a termination point of NAS (Nl) signaling, and perform NAS ciphering and integrity protection. AMF 1544 may also support NAS signaling with the UE 1502 over an N3 IWF interface.
The SMF 1546 may be responsible for SM (for example, session establishment, tunnel management between UPF 1548 and AN 1508); UE IP address allocation and management (including optional authorization); selection and control of UP function; configuring traffic steering at UPF 1548 to route traffic to proper destination; termination of interfaces toward policy control functions; controlling part of policy enforcement, charging, and QoS; lawful intercept (for SM events and interface to LI system); termination of SM parts of NAS messages; downlink data notification; initiating AN specific SM information, sent via AMF 1544 over N2 to AN 1508; and determining SSC mode of a session. SM may refer to management of a PDU session, and a PDU session or “session” may refer to a PDU connectivity service that provides or enables the exchange of PDUs between the UE 1502 and the data network 1536.
The UPF 1548 may act as an anchor point for intra-RAT and inter-RAT mobility, an external PDU session point of interconnect to data network 1536, and a branching point to support multi -homed PDU session. The UPF 1548 may also perform packet routing and forwarding, perform packet inspection, enforce the user plane part of policy rules, lawfully intercept packets (UP collection), perform traffic usage reporting, perform QoS handling for a user plane (e.g., packet filtering, gating, UL/DL rate enforcement), perform uplink traffic verification (e.g., SDF- to-QoS flow mapping), transport level packet marking in the uplink and downlink, and perform downlink packet buffering and downlink data notification triggering. UPF 1548 may include an uplink classifier to support routing traffic flows to a data network.
The NSSF 1550 may select a set of network slice instances serving the UE 1502. The NSSF 1550 may also determine allowed NSSAI and the mapping to the subscribed S-NSSAIs, if needed. The NSSF 1550 may also determine the AMF set to be used to serve the UE 1502, or a list of candidate AMFs based on a suitable configuration and possibly by querying the NRF 1554. The selection of a set of network slice instances for the UE 1502 may be triggered by the AMF 1544 with which the UE 1502 is registered by interacting with the NSSF 1550, which may lead to a change of AMF. The NSSF 1550 may interact with the AMF 1544 via an N22 reference point; and may communicate with another NSSF in a visited network via an N31 reference point (not shown). Additionally, the NSSF 1550 may exhibit an Nnssf service-based interface.
The NEF 1552 may securely expose services and capabilities provided by 3GPP network functions for third party, internal exposure/re-exposure, AFs (e.g., AF 1560), edge computing or fog computing systems, etc. In such embodiments, the NEF 1552 may authenticate, authorize, or throttle the AFs. NEF 1552 may also translate information exchanged with the AF 1560 and information exchanged with internal network functions. For example, the NEF 1552 may translate between an AF-Service-Identifier and an internal 5GC information. NEF 1552 may also receive information from other NFs based on exposed capabilities of other NFs. This information may be stored at the NEF 1552 as structured data, or at a data storage NF using standardized interfaces. The stored information can then be re-exposed by the NEF 1552 to other NFs and AFs, or used for other purposes such as analytics. Additionally, the NEF 1552 may exhibit an Nnef servicebased interface.
The NRF 1554 may support service discovery functions, receive NF discovery requests from NF instances, and provide the information of the discovered NF instances to the NF instances. NRF 1554 also maintains information of available NF instances and their supported services. As used herein, the terms “instantiate,” “instantiation,” and the like may refer to the creation of an instance, and an “instance” may refer to a concrete occurrence of an object, which may occur, for example, during execution of program code. Additionally, the NRF 1554 may exhibit the Nnrf service-based interface.
The PCF 1556 may provide policy rules to control plane functions to enforce them, and may also support unified policy framework to govern network behavior. The PCF 1556 may also implement a front end to access subscription information relevant for policy decisions in a UDR of the UDM 1558. In addition to communicating with functions over reference points as shown, the PCF 1556 exhibit an Npcf service-based interface.
The UDM 1558 may handle subscription-related information to support the network entities' handling of communication sessions, and may store subscription data of UE 1502. For example, subscription data may be communicated via an N8 reference point between the UDM 1558 and the AMF 1544. The UDM 1558 may include two parts, an application front end and a UDR. The UDR may store subscription data and policy data forthe UDM 1558 and the PCF 1556, and/or structured data for exposure and application data (including PFDs for application detection, application request information for multiple UEs 1502) for the NEF 1552. The Nudr service-based interface may be exhibited by the UDR 221 to allow the UDM 1558, PCF 1556, and NEF 1552 to access a particular set of the stored data, as well as to read, update (e.g., add, modify), delete, and subscribe to notification of relevant data changes in the UDR. The UDM may include a UDM- FE, which is in charge of processing credentials, location management, subscription management and so on. Several different front ends may serve the same user in different transactions. The UDM-FE accesses subscription information stored in the UDR and performs authentication credential processing, user identification handling, access authorization, registration/mobility management, and subscription management. In addition to communicating with other NFs over reference points as shown, the UDM 1558 may exhibit the Nudm service-based interface.
The AF 1560 may provide application influence on traffic routing, provide access to NEF, and interact with the policy framework for policy control.
In some embodiments, the 5GC 1540 may enable edge computing by selecting operator/3rd party services to be geographically close to a point that the UE 1502 is attached to the network. This may reduce latency and load on the network. To provide edge-computing implementations, the 5GC 1540 may select a UPF 1548 close to the UE 1502 and execute traffic steering from the UPF 1548 to data network 1536 via the N6 interface. This may be based on the UE subscription data, UE location, and information provided by the AF 1560. In this way, the AF 1560 may influence UPF (re)selection and traffic routing. Based on operator deployment, when AF 1560 is considered to be a trusted entity, the network operator may permit AF 1560 to interact directly with relevant NFs. Additionally, the AF 1560 may exhibit an Naf service-based interface.
The data network 1536 may represent various network operator services, Internet access, or third party services that may be provided by one or more servers including, for example, application/content server 1538.
Figure 16 schematically illustrates a wireless network 1600 in accordance with various embodiments. The wireless network 1600 may include a UE 1602 in wireless communication with an AN 1604. The UE 1602 and AN 1604 may be similar to, and substantially interchangeable with, like-named components described elsewhere herein.
The UE 1602 may be communicatively coupled with the AN 1604 via connection 1606. The connection 1606 is illustrated as an air interface to enable communicative coupling, and can be consistent with cellular communications protocols such as an LTE protocol or a 5GNR protocol operating at mmWave or sub-6GHz frequencies.
The UE 1602 may include a host platform 1608 coupled with a modem platform 1610. The host platform 1608 may include application processing circuitry 1612, which may be coupled with protocol processing circuitry 1614 of the modem platform 1610. The application processing circuitry 1612 may run various applications for the UE 1602 that source/sink application data. The application processing circuitry 1612 may further implement one or more layer operations to transmit/receive application data to/from a data network. These layer operations may include transport (for example UDP) and Internet (for example, IP) operations
The protocol processing circuitry 1614 may implement one or more of layer operations to facilitate transmission or reception of data over the connection 1606. The layer operations implemented by the protocol processing circuitry 1614 may include, for example, MAC, RLC, PDCP, RRC and NAS operations.
The modem platform 1610 may further include digital baseband circuitry 1616 that may implement one or more layer operations that are “below” layer operations performed by the protocol processing circuitry 1614 in a network protocol stack. These operations may include, for example, PHY operations including one or more of HARQ-ACK functions, scrambling/descrambling, encoding/decoding, layer mapping/de-mapping, modulation symbol mapping, received symbol/bit metric determination, multi-antenna port precoding/decoding, which may include one or more of space-time, space-frequency or spatial coding, reference signal generation/detection, preamble sequence generation and/or decoding, synchronization sequence generation/detection, control channel signal blind decoding, and other related functions.
The modem platform 1610 may further include transmit circuitry 1618, receive circuitry 1620, RF circuitry 1622, and RF front end (RFFE) 1624, which may include or connect to one or more antenna panels 1626. Briefly, the transmit circuitry 1618 may include a digital-to-analog converter, mixer, intermediate frequency (IF) components, etc.; the receive circuitry 1620 may include an analog-to-digital converter, mixer, IF components, etc.; the RF circuitry 1622 may include a low-noise amplifier, a power amplifier, power tracking components, etc.; RFFE 1624 may include filters (for example, surface/bulk acoustic wave filters), switches, antenna tuners, beamforming components (for example, phase-array antenna components), etc. The selection and arrangement of the components of the transmit circuitry 1618, receive circuitry 1620, RF circuitry 1622, RFFE 1624, and antenna panels 1626 (referred generically as “transmit/receive components”) may be specific to details of a specific implementation such as, for example, whether communication is TDM or FDM, in mmWave or sub-6 gHz frequencies, etc. In some embodiments, the transmit/receive components may be arranged in multiple parallel transmit/receive chains, may be disposed in the same or different chips/modules, etc.
In some embodiments, the protocol processing circuitry 1614 may include one or more instances of control circuitry (not shown) to provide control functions for the transmit/receive components.
A UE reception may be established by and via the antenna panels 1626, RFFE 1624, RF circuitry 1622, receive circuitry 1620, digital baseband circuitry 1616, and protocol processing circuitry 1614. In some embodiments, the antenna panels 1626 may receive a transmission from the AN 1604 by receive-beamforming signals received by a plurality of antennas/antenna elements of the one or more antenna panels 1626.
A UE transmission may be established by and via the protocol processing circuitry 1614, digital baseband circuitry 1616, transmit circuitry 1618, RF circuitry 1622, RFFE 1624, and antenna panels 1626. In some embodiments, the transmit components of the UE 1604 may apply a spatial filter to the data to be transmitted to form a transmit beam emitted by the antenna elements of the antenna panels 1626.
Similar to the UE 1602, the AN 1604 may include a host platform 1628 coupled with a modem platform 1630. The host platform 1628 may include application processing circuitry 1632 coupled with protocol processing circuitry 1634 of the modem platform 1630. The modem platform may further include digital baseband circuitry 1636, transmit circuitry 1638, receive circuitry 1640, RF circuitry 1642, RFFE circuitry 1644, and antenna panels 1646. The components of the AN 1604 may be similar to and substantially interchangeable with like- named components of the UE 1602. In addition to performing data transmission/reception as described above, the components of the AN 1608 may perform various logical functions that include, for example, RNC functions such as radio bearer management, uplink and downlink dynamic radio resource management, and data packet scheduling.
Figure 17 is a block diagram illustrating components, according to some example embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein. Specifically, Figure 17 shows a diagrammatic representation of hardware resources 1700 including one or more processors (or processor cores) 1710, one or more memory /storage devices 1720, and one or more communication resources 1730, each of which may be communicatively coupled via a bus 1740 or other interface circuitry. For embodiments where node virtualization (e.g., NFV) is utilized, a hypervisor 1702 may be executed to provide an execution environment for one or more network slices/sub-slices to utilize the hardware resources 1700.
The processors 1710 may include, for example, a processor 1712 and a processor 1714. The processors 1710 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 radiofrequency integrated circuit (RFIC), another processor (including those discussed herein), or any suitable combination thereof.
The memory /storage devices 1720 may include main memory, disk storage, or any suitable combination thereof. The memory /storage devices 1720 may include, but are not limited to, any type of volatile, non-volatile, or semi-volatile memory such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), Flash memory, solid-state storage, etc.
The communication resources 1730 may include interconnection or network interface controllers, components, or other suitable devices to communicate with one or more peripheral devices 1704 or one or more databases 1706 or other network elements via a network 1708. For example, the communication resources 1730 may include wired communication components (e.g., for coupling via USB, Ethernet, etc.), cellular communication components, NFC components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components, and other communication components.
Instructions 1750 may comprise software, a program, an application, an applet, an app, or other executable code for causing at least any of the processors 1710 to perform any one or more of the methodologies discussed herein. The instructions 1750 may reside, completely or partially, within at least one of the processors 1710 (e.g., within the processor's cache memory), the memory /storage devices 1720, or any suitable combination thereof. Furthermore, any portion of the instructions 1750 may be transferred to the hardware resources 1700 from any combination of the peripheral devices 1704 or the databases 1706. Accordingly, the memory of processors 1710, the memory/storage devices 1720, the peripheral devices 1704, and the databases 1706 are examples of computer-readable and machine-readable media.
EXAMPLE PROCEDURES
In some embodiments, the electronic device(s), network(s), system(s), chip(s) or component(s), or portions or implementations thereof, of Figures 15-17, or some other figure herein, may be configured to perform one or more processes, techniques, or methods as described herein, or portions thereof.
One such process is depicted in Figure 18, which may be performed by a network function such as a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF). For example, the process 1800 may include, at 1805, Receiving, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE). The process further includes, at 1810, retrieving UE measurement report information from a memory. The process further includes, at 1815, determining, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on the UE measurement report information and the request for location information, wherein the location information for the UE includes an indication of a predicted location for the UE. The process further includes, at 1820, encoding a message for transmission to the gNB that includes an indication of the determined location information for the UE.
Another such process is illustrated in Figure 19. In this example, the process 1900 includes, at 1905, receiving, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE). The process further includes, at 1910, determining, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on UE measurement report information and the request for location information wherein the location information for the UE includes an indication of a predicted location for the UE. The process further includes, at 1915, encoding a message for transmission to the gNB that includes an indication of the determined location information for the UE.
Another such process is illustrated in Figure 20, which may be performed by a gNB. In this example, the process 2000 includes, at 2005, encoding, for transmission to a network function, a request for location information for a user equipment (UE). The process further includes, at 2010, receiving, from the network function, the location information for the UE, wherein the location information for the UE includes an indication of a predicted location for the UE. The process further includes, at 2015, determining configuration information that includes an indication of a combination of candidate handover cells and respective priorities for the candidate handover cells. The process further includes, at 2020, encoding, for transmission to the UE, a message that includes the configuration information.
For one or more embodiments, at least one of the components set forth in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and/or methods as set forth in the example section below. For example, the baseband circuitry as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below. For another example, circuitry associated with a UE, base station, network element, etc. as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below in the example section.
EXAMPLES
Example 1 may include NG-RAN can generate a group of predicted candidate target cells for handover according to the predicted UE location information.
Example 2 may include the NG-RAN in example 1 or some other example herein, wherein requests UE current/future location information from OAM/CN or LMF.
Example 3 may include the NG-RAN in example 1 or some other example herein, wherein receives UE current/future location information from UE report.
Example 4 may include the OAM/CN/LMF sends the predicted future UE's location information to NG-RAN in example 1.
Example 5 may include the group of target cells in example 1 or some other example herein, wherein includes handover configuration, handover execution timestamp and priority of each cell.
Example 6 may include the source gNB in example 1 or some other example herein, wherein sends the handover request together with the handover execution timestamp to target candidate gNBs.
Example 7 may include the source gNB in example 1 or some other example herein, wherein configures UE with the combination/list of cells with handover execution timestamp and the corresponding priority.
Example 8 may include the UE in example 7 or some other example herein wherein selects handover target cell according to the priority and execute handover at the execution timestamp.
Example 9 may include the UE in example 7 or some other example herein, wherein may delete the candidate cell from the list if the execution timestamp expires. Example 10 may include the UE in example 7 or some other example herein, wherein may continuously select the highest priority candidate cell from the remaining list if handover is failed.
Example 11 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to request future location information as in example 2 according to UE speed.
Example 12 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on statistic of mobility events.
Example 13 may include the NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on UE information.
Example 14 may include NG-RAN in example 1 or some other example herein, wherein selects UE(s) to perform AI/ML based mobility enhancement as in example 2 based on UE device type or environment.
Example 15 may include if UE is predicted with slow moving, sit in the cell border, the source gNB in example 1 decides not to perform handover.
Example 16 may include the network in example 1 or some other example herein, wherein collects and reports the mobility events for each UE independently.
Example 17 may include a method of a NG-RAN node, the method comprising: receiving predicted UE location information associated with a UE; and generating a set of one or more predicted candidate target cells for the UE based on the predicted UE location information.
Example 18 may include the method of example 17 or some other example herein, wherein the predicted UE location information is received from a 0AM, a CN, or an LMF.
Example 19 may include the method of example 17 or some other example herein, wherein the predicted UE location information is received from the UE.
Example 20 may include the method of example 17-19 or some other example herein, further comprising encoding, for transmission, a request for the predicted UE location information, wherein the predicted UE location information is received responsive to the request.
Example 21 may include the method of example 17-20 or some other example herein, further comprising generating a handover configuration that includes the set of predicted candidate target cells and a handover execution timestamp and/or a priority of individual cells of the set of predicted candidate target cells. Example 22 may include the method of example 21 or some other example herein, further comprising encoding the handover configuration for transmission to the UE.
Example 23 may include the method of example 17-22 or some other example herein, wherein the NG-RAN node is a source gNB, and wherein the method further comprises encoding a handover request for transmission to a target gNB, wherein the handover request includes a handover execution timestamp.
Example 24 may include the method of example 17-23 or some other example herein, further comprising sending a request to the UE for the predicted UE location information based on a speed of the UE a statistic of mobility events, a UE device type, and/or a UE environment.
Example 25 may include the method of example 17-24 or some other example herein, wherein the predicted UE location information is determined based on an artificial intelligence and/or machine learning (AI/ML) model.
Example 26 may include a method of a UE, the method comprising: receiving a handover configuration that includes one or more predicted target cells and at least one of a handover execution timestamp or a priority for respective cells of the one or more predicted target cells; and executing a handover to a target cell of the one or more predicted target cells based on the handover configuration.
Example 27 may include the method of example 26 or some other example herein, wherein the handover is executed at a time indicated by the handover execution timestamp associated with the target cell.
Example 28 may include the method of example 26-27 or some other example herein, further comprising selecting the target cell from the one or more predicted target cells based on the respective priorities.
Example 29 may include the method of example 26-28 or some other example herein, further comprising removing a first predicted target cell of the one or more predicted target cells from consideration for the handover based on expiration of the handover execution timestamp associated with the first predicted target cell.
Example 30 may include the method of example 26-29 or some other example herein, further comprising, upon failure of the handover to the target cell, selecting another target cell from the one or more predicted candidate target cells based on the associated priority.
Example 31 may include the method of example 26-30 or some other example herein, further comprising obtaining predicted UE location information for the UE and providing the predicted UE location information to a gNB. Example 32 may include the method of example 31 or some other example herein, wherein the handover configuration is based on the predicted UE location information.
Example 33 may include the method of example 31-32 or some other example herein, wherein the predicted UE location information is obtained and/or provided responsive to a request from the gNB.
Example 34 may include the method of example 31-33 or some other example herein, wherein the predicted UE location information is obtained using an artificial intelligence and/or machine learning (AI/ML) model.
Example QI may include NG-RAN node should configure ML model training and inference configuration for defining how UE trains the AI/ML model, e.g. sending Model training configuration to the UE.
Example Q2 may include NG-RAN node in Example QI or some other example herein, wherein should configure whether the UE can perform training/re-training via RRC signaling, e.g. Training Indicator.
Example Q3 may include NG-RAN node in Example QI or some other example herein, wherein should configure whether the UE can use its private AI/ML model via RRC signaling, e.g. an indicator of private model training.
Example Q4 may include NG-RAN node in Example QI or some other example herein, wherein should configure model training configuration to the UE, including optimizer, loss (function), metrics, loss weight, accuracy level, steps per execution, aggregate bits, etc.
Example Q5 may include the UE in Example QI or some other example herein, wherein should report the evaluation results, where model training is following configuration in Example 4, through RRC messages (e.g. via Training Report in Machine learning Report or Model update request, etc).
Example Q6 may include NG-RAN node in Example QI or some other example herein, wherein further indicates the UE in Example 5 whether the model is well-trained and can be used for predict! on/decision making by sending a training feedback via RRC message.
Example Q7 may include upon receiving an ACK training feedback in Example Q6 or some other example herein, wherein from NG-RAN node, the UE in Example Q5 can use the model trained/retrained by itself for decision making/prediction.
Example Q8 may include upon receiving a NACK training feedback in Example 6 or some other example herein, wherein from NG-RAN node, the UE in Example Q5 need to further train/re-train the AI/ML model, and repeat example Q5 and example Q6. Example Q9 may include NG-RAN node in Example 1 or some other example herein, wherein may also configure the model training threshold, indicating when the AI/ML is well- trained, to the UE together with model training configuration in Example Q4.
Example Q10 may include the training threshold is met, the UE in Example QI or some other example herein, wherein should provide a training complete feedback to the network.
Example QI 1 may include the UE may request model downloading from the NG-RAN node in Example 1 or some other example herein, wherein by indicating the interested service or requested model to the network via RRC message.
Example Q12 may include theNG-RAN node in Example Qll or some other example herein, wherein should also configure the required input/output of AI/ML model when UE in Example QI requests model download from itself.
Example Q13 may include the UE in Example Qll or some other example herein, wherein may further re-train the AI/ML model downloaded from NG-RAN, following Example QI to Example Q10.
Example Q14 may include a method of a UE, the method comprising: receiving, from a NG-RAN node, a model training and inference configuration to indicate information for the UE to train an artificial intelligence (Al) and/or machine learning (ML) (AI/ML) model; and training the AI/ML model based on the configuration.
Example Q15 may include the method of Example Q14 or some other example herein, further comprising receiving, from the NG-RAN node, a training indicator to indicate whether the UE is permitted to perform training/re-training of the AI/ML model.
Example Q16 may include method of Example Q14-Q15 or some other example herein, further comprising receiving, from the NG-RAN node an indication of whether the UE is permitted to use a private AI/ML model.
Example QI 7 may include the method of example Q14-Q16 or some other example herein, wherein the configuration, the training indicator, and/or the indicator are received via RRC signaling.
Example A01 includes a method comprising: model update between network nodes triggered by one of or multiple trigger conditions (e.g., timer, accuracy level, confidence level, performance feedback, and loss of AI/ML model).
Example A02 includes a new message is used to configure machine learning update configuration from RAN to UE.
Example A03 includes a new message is used to configure machine learning update configuration from OAM/CN to RAN. Example A04 includes the method of example A01 and/or some other example(s) herein, wherein the timer is configured by the network nodes who deploys AI/ML model to other network nodes, and upon the configured timer is approaching, UE or RAN may send a model update request to RAN or CN/OAM, respectively, and CN/OAM or RAN will then provide an updated model to the corresponding network node.
Example A05 includes the method of example A01 and/or some other example(s) herein, wherein the timer is maintained by the network nodes who deploys AI/ML model, and upon the timer is approaching, CN/OAM or RAN will provide an updated model to the RAN or UE, respectively.
Example A06 includes the method of example A01 and/or some other example(s) herein, wherein the timer is maintained by the network nodes who deploys AI/ML model to other network nodes, and the CN/OAM or RAN may trigger model update based on the observed accuracy level.
Example A07 includes the method of example A01 and/or some other example(s) herein, wherein the confidence level is configured by the network nodes who deploys AI/ML model to other network nodes, and the UE or RAN may trigger model update request to RAN or UE.
Example A08 includes the method of example A01 and/or some other example(s) herein, wherein the confidence level is reported by UE or RAN to RAN or CN/OAM, respectively, and the network may update the AI/ML model if confidence level is lower than certain threshold.
Example A09 includes the method of examples A01-A08 and/or some other example(s) herein, wherein the UE or RAN may trigger model update request to RAN or UE.
Example A10 includes the method of examples A01-A09 and/or some other example(s) herein, wherein the UE or RAN reports the performance indicator to RAN or OAM/CN (respectively), indicating performance improvement or down-gradation, and RAN or OAM/CN may send an updated AI/ML model if the performance is downgraded.
Example All includes the method of examples A01-A10 and/or some other example(s) herein, wherein the UE or RAN may trigger model update request to RAN or UE based on performance observation of its own.
Example Al 2 includes the method of example A01 and/or some other example(s) herein, wherein the loss is reported by UE or RAN to RAN or CN/OAM, respectively, and the network may update the AI/ML model if loss is higher than certain threshold.
Example B01 includes amethod comprising: triggering an artificial inteligence (Al) and/or machine learning (ML) model update between one or more network nodes in response to one or more trigger conditions.
Example B02 includes the method of example B01 and/or some other example(s) herein, wherein the one or more trigger conditions include expiration of a timer, a accuracy level, a confidence level, performance feedback, and loss of an AI/ML model.
Example B03 includes the method of examples B01-B02 and/or some other example(s) herein, further comprising: sending a message to configure AI/ML model update configuration from a radio access network (RAN) to a user equipment (UE).
Example B04 includes the method of examples B01-B03 and/or some other example(s) herein, further comprising: sending a message to configure AI/ML model update configuration from an operations, administrations, and management function (0AM) and/or core network (CN) to a RAN.
Example B05 includes the method of examples B01-B04 and/or some other example(s) herein, wherein a timer is configured by the one or more network nodes who deploy the AI/ML model to one or more other network nodes.
Example B06 includes the method of example B05 and/or some other example(s) herein, further comprising: as the timer approaches expiration or upon expiration of the timer, sending the AI/ML model update request to a RAN, a CN, or 0AM.
Example B07 includes the method of example B06 and/or some other example(s) herein, further comprising: obtaining an updated AI/ML model from the CN, the 0AM, or the RAN.
Example B08 includes the method of examples B06-B07 and/or some other example(s) herein, wherein the CN, the 0AM, and/or the RAN trigger the AI/ML model update based on an observed accuracy level and/or in response to one or more trigger conditions.
Example B09 includes the method of examples B01-B08 and/or some other example(s) herein, wherein a confidence level is configured by the one or more network nodes who deploy AI/ML model to other network nodes, and a UE or a RAN triggers the AI/ML model update request to the RAN or the UE.
Example BIO includes the method of examples B01-B09 and/or some other example(s) herein, wherein the confidence level is reported by a UE or a RAN to the RAN, the CN, and/or the 0AM, and the method comprises: updating the AI/ML model if the confidence level is lower than a threshold confidence level.
Example Bl l includes the method of examples B01-B10 and/or some other example(s) herein, wherein the UE or the RAN trigger the AI/ML model update request to the RAN or the UE.
Example B12 includes the method of examples B01-B11 and/or some other example(s) herein, wherein a UE or a RAN reports the performance indicator to the RAN, the CN, and/or the 0AM, and the method comprises: indicating performance improvement or down-gradations, and sending or obtaining an updated AI/ML model if the performance is downgraded from the RAN, the CN, and/or the 0AM. Example B13 includes the method of examples B01-B12 and/or some other example(s) herein, wherein the UE or the RAN trigger the AI/ML model update request to the RAN or the UE based on performance observation of its own.
Example B14 includes the method of examples BO 1 -Bl 3 and/or some other example(s) herein, wherein the loss is reported by the UE or the RAN to the RAN, the CN, and/or the 0AM, and the method comprises: updating the AI/ML model if the loss is higher than a threshold amount of loss.
Example B15 includes the method of examples B01-B14 and/or some other example(s) herein, wherein the method is performed by a UE, a RAN, a RAN node, a CN function, and/or an 0 AM.
Example XI includes an apparatus comprising: memory to store user equipment (UE) measurement report information; and processing circuitry, coupled with the memory, to: receive, from a next-generation NodeB (gNB), a request for location information for the UE; retrieve the UE measurement report information from the memory; determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on the UE measurement report information and the request for location information, wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
Example X2 includes the apparatus of example XI or some other example herein, wherein the UE measurement report information includes historical and current information for the UE.
Example X3 includes the apparatus of example X2 or some other example herein, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
Example X4 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
Example X5 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
Example X6 includes the apparatus of example XI or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE. Example X7 includes the apparatus of any of examples XI -X6 or some other example herein, wherein the processing circuitry is to implement a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
Example X8 includes one or more computer-readable media storing instructions that, when executed by one or more processors, cause a network function to: receive, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE); determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on UE measurement report information and the request for location information wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
Example X9 includes the one or more computer-readable media of example X8 or some other example herein, wherein the UE measurement report information includes historical and current information for the UE.
Example XI 0 includes the one or more computer-readable media of example X9 or some other example herein, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
Example XI 1 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
Example XI 2 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
Example XI 3 includes the one or more computer-readable media of example X8 or some other example herein, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE.
Example XI 4 includes the one or more computer-readable media of any of examples X8- XI 3 or some other example herein, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
Example XI 5 includes one or more computer-readable media storing instructions that, when executed by one or more processors, cause a next-generation NodeB (gNB) to: encode, for transmission to a network function, a request for location information for a user equipment (UE); receive, from the network function, the location information for the UE, wherein the location information for the UE includes an indication of a predicted location for the UE; determine configuration information that includes an indication of a combination of candidate handover cells and respective priorities for the candidate handover cells; and encode, for transmission to the UE, a message that includes the configuration information.
Example XI 6 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein determining the configuration information includes determining whether the UE should start a handover.
Example XI 7 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein the configuration information includes an indication of a target cell from the candidate handover cells to perform a handover, and timing information for performing the handover.
Example XI 8 includes the one or more computer-readable media of example XI 5 or some other example herein, wherein the configuration information includes conditional handover (CHO) information.
Example XI 9 includes the one or more computer-readable media of example XI 8 or some other example herein, wherein the CHO information includes an indication of conditional handover execution timing.
Example X20 includes the one or more computer-readable media of example XI 8 or some other example herein, wherein the CHO information includes an indication of timing information for each candidate cell in the combination of candidate handover cells.
Example X21 includes the one or more computer-readable media of examples X15-X20 or some other example herein, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
Example Z01 may include an apparatus comprising means to perform one or more elements of a method described in or related to any of examples 1-X21, or any other method or process described herein.
Example Z02 may include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of a method described in or related to any of examples 1- X21, or any other method or process described herein. Example Z03 may include an apparatus comprising logic, modules, or circuitry to perform one or more elements of a method described in or related to any of examples 1- X21, or any other method or process described herein.
Example Z04 may include a method, technique, or process as described in or related to any of examples 1- X21, or portions or parts thereof.
Example Z05 may include an apparatus comprising: one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
Example Z06 may include a signal as described in or related to any of examples 1- X21, or portions or parts thereof.
Example Z07 may include a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
Example Z08 may include a signal encoded with data as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
Example Z09 may include a signal encoded with a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1- X21, or portions or parts thereof, or otherwise described in the present disclosure.
Example Z10 may include an electromagnetic signal carrying computer-readable instructions, wherein execution of the computer-readable instructions by one or more processors is to cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
Example Zll may include a computer program comprising instructions, wherein execution of the program by a processing element is to cause the processing element to carry out the method, techniques, or process as described in or related to any of examples 1- X21, or portions thereof.
Example Z12 may include a signal in a wireless network as shown and described herein.
Example Z13 may include a method of communicating in a wireless network as shown and described herein.
Example Z14 may include a system for providing wireless communication as shown and described herein.
Example Z15 may include a device for providing wireless communication as shown and described herein. Any of the above-described examples may be combined with any other example (or combination of examples), unless explicitly stated otherwise. The foregoing description of one or more implementations provides illustration and description, but is not intended to be exhaustive or to limit the scope of embodiments to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of various embodiments.
Abbreviations
Unless used differently herein, terms, definitions, and abbreviations may be consistent with terms, definitions, and abbreviations defined in 3GPP TR 21.905 V16.0.0 (2019-06). For the purposes of the present document, the following abbreviations may apply to the examples and embodiments discussed herein.
3GPP Third Generation AP Application BRAS Broadband Partnership 35 Protocol, Antenna Remote Access
Project Port, Access Point Server
4G Fourth API Application 70 BSS Business
Generation Programming Interface Support System
5G Fifth Generation APN Access Point BS Base Station 5GC 5G Core network 40 Name BSR Buffer Status AC ARP Allocation and Report
Application Retention Priority 75 BW Bandwidth Client ARQ Automatic BWP Bandwidth Part
ACR Application Repeat Request C-RNTI Cell Context Relocation 45 AS Access Stratum Radio Network ACK ASP Temporary
Acknowledgeme Application Service 80 Identity nt Provider CA Carrier
ACID Aggregation,
Application 50 ASN.1 Abstract Syntax Certification
Client Identification Notation One Authority
AF Application AUSF Authentication 85 CAPEX CAPital
Function Server Function Expenditure
AM Acknowledged AWGN Additive CBRA Contention Mode 55 White Gaussian Based Random
AMBRAggregate Noise Access Maximum Bit Rate BAP Backhaul 90 CC Component AMF Access and Adaptation Protocol Carrier, Country
Mobility BCH Broadcast Code, Cryptographic
Management 60 Channel Checksum Function BER Bit Error Ratio CCA Clear Channel
AN Access Network BFD Beam 95 Assessment
ANR Automatic Failure Detection CCE Control Channel
Neighbour Relation BLER Block Error Rate Element AOA Angle of 65 BPSK Binary Phase CCCH Common Arrival Shift Keying Control Channel
100 CE Coverage Enhancement CDM Content Delivery CoMP Coordinated Resource Network Multi-Point Indicator
CDMA Code- CORESET Control C-RNTI Cell Division Multiple Resource Set RNTI
Access 40 COTS Commercial Off- 75 CS Circuit Switched
CDR Charging Data The-Shelf CSCF call Request CP Control Plane, session control function
CDR Charging Data Cyclic Prefix, CSAR Cloud Service Response Connection Archive
CFRA Contention Free 45 Point 80 CSI Channel-State Random Access CPD Connection Information CG Cell Group Point Descriptor CSI-IM CSI CGF Charging CPE Customer Interference
Gateway Function Premise Measurement CHF Charging 50 Equipment 85 CSI-RS CSI
Function CPICHCommon Pilot Reference Signal
CI Cell Identity Channel CSI-RSRP CSI CID Cell-ID (e g., CQI Channel Quality reference signal positioning method) Indicator received power CIM Common 55 CPU CSI processing 90 CSI-RSRQ CSI Information Model unit, Central reference signal CIR Carrier to Processing Unit received quality Interference Ratio C/R CSI-SINR CSI CK Cipher Key C ommand/Respo signal-to-noise and CM Connection 60 nse field bit 95 interference ratio Management, CRAN Cloud Radio CSMA Carrier Sense
Conditional Access Network, Multiple Access Mandatory Cloud RAN CSMA/CA CSMA CM AS Commercial CRB Common with collision Mobile Alert Service 65 Resource Block 100 avoidance CMD Command CRC Cyclic CSS Common Search CMS Cloud Redundancy Check Space, Cell- specific Management System CRI Channel-State Search Space CO Conditional Information Resource CTF Charging Optional 70 Indicator, CSI-RS 105 Trigger Function CTS Clear-to-Send DSL Domain Specific 70 ECSP Edge
CW Codeword Language. Digital Computing Service
CWS Contention Subscriber Line Provider
Window Size DSLAM DSL EDN Edge
D2D Device-to- 40 Access Multiplexer Data Network
Device DwPTS 75 EEC Edge
DC Dual Downlink Pilot Enabler Client
Connectivity, Direct Time Slot EECID Edge
Current E-LAN Ethernet Enabler Client
DCI Downlink 45 Local Area Network Identification
Control E2E End-to-End 80 EES Edge
Information EAS Edge Enabler Server
DF Deployment Application Server EESID Edge
Flavour ECCA extended clear Enabler Server
DL Downlink 50 channel Identification
DMTF Distributed assessment, 85 EHE Edge
Management Task extended CCA Hosting Environment
Force ECCE Enhanced EGMF Exposure
DPDK Data Plane Control Channel Governance
Development Kit 55 Element, Management
DM-RS, DMRS Enhanced CCE 90 Function
Demodulation ED Energy EGPRS Enhanced
Reference Signal Detection GPRS
DN Data network EDGE Enhanced EIR Equipment
DNN Data Network 60 Datarates for GSM Identity Register
Name Evolution (GSM 95 eLAA enhanced
DNAI Data Network Evolution) Licensed Assisted
Access Identifier EAS Edge Access,
Application Server enhanced LAA
DRB Data Radio 65 EASID Edge EM Element
Bearer Application Server 100 Manager
DRS Discovery Identification eMBB Enhanced
Reference Signal ECS Edge Mobile
DRX Discontinuous Configuration Server Broadband
Reception EMS Element E-UTRAN Evolved FDM Frequency
Management System UTRAN Division Multiplex eNB evolved NodeB, EV2X Enhanced V2X FDMA F requency E-UTRAN Node B F1AP Fl Application Division Multiple EN-DC E- 40 Protocol 75 Access UTRA-NR Dual Fl-C Fl Control plane FE Front End
Connectivity interface FEC Forward Error
EPC Evolved Packet Fl-U Fl User plane Correction Core interface FFS For Further
EPDCCH enhanced 45 FACCH Fast 80 Study
PDCCH, enhanced Associated Control FFT Fast Fourier
Physical CHannel Transformation
Downlink Control FACCH/F Fast feLAA further enhanced
Cannel Associated Control Licensed Assisted
EPRE Energy per 50 Channel/Full 85 Access, further resource element rate enhanced LAA EPS Evolved Packet FACCH/H Fast FN Frame Number System Associated Control FPGA Field-
EREG enhanced REG, Channel/Half Programmable Gate enhanced resource 55 rate 90 Array element groups FACH Forward Access FR Frequency ETSI European Channel Range
Telecommunicat FAUSCH Fast FQDN Fully Qualified ions Standards Uplink Signalling Domain Name Institute 60 Channel 95 G-RNTI GERAN
ETWS Earthquake and FB Functional Block Radio Network Tsunami Warning FBI Feedback Temporary System Information Identity eUICC embedded FCC Federal GERAN UICC, embedded 65 Communications 100 GSM EDGE
Universal Commission RAN, GSM EDGE
Integrated Circuit FCCH Frequency Radio Access Card Correction CHannel Network
E-UTRA Evolved FDD Frequency GGSN Gateway GPRS
UTRA 70 Division Duplex 105 Support Node GLONASS 35 GTP-UGPRS 70 HSUPA High
GLObal'naya Tunnelling Protocol Speed Uplink Packet
NAvigatsionnay for User Plane Access a Sputnikovaya GTS Go To Sleep HTTP Hyper Text Sistema (Engl.: Signal (related to Transfer Protocol Global Navigation 40 WUS) 75 HTTPS Hyper
Satellite System) GUMMEI Globally Text Transfer Protocol gNB Next Generation Unique MME Identifier Secure (https is NodeB GUTI Globally Unique http/ 1.1 over gNB-CU gNB- Temporary UE SSL, i.e. port 443) centralized unit, Next 45 Identity 80 I-Block
Generation HARQ Hybrid ARQ, Information
NodeB Hybrid Block centralized unit Automatic ICCID Integrated gNB-DU gNB- Repeat Request Circuit Card distributed unit, Next 50 HANDO Handover 85 Identification
Generation HFN HyperFrame IAB Integrated
NodeB Number Access and Backhaul distributed unit HHO Hard Handover ICIC Inter-Cell GNSS Global HLR Home Location Interference Navigation Satellite 55 Register 90 Coordination
System HN Home Network ID Identity,
GPRS General Packet HO Handover identifier Radio Service HPLMN Home IDFT Inverse Discrete
GPSI Generic Public Land Mobile Fourier
Public Subscription 60 Network 95 Transform
Identifier HSDPA High IE Information GSM Global System Speed Downlink element for Mobile Packet Access IBE In-Band
Communications HSN Hopping Emission , Groupe Special 65 Sequence Number 100 IEEE Institute of Mobile HSPA High Speed Electrical and GTP GPRS Tunneling Packet Access Electronics Protocol HSS Home Engineers
Subscriber Server IEI Information Ipsec IP Security, 70 kB Kilobyte (1000
Element Identifier Internet Protocol bytes)
IEIDL Information Security kbps kilo-bits per
Element Identifier IP-CAN IP- second
Data Length 40 Connectivity Access Kc Ciphering key
IETF Internet Network 75 Ki Individual
Engineering Task IP-M IP Multicast subscriber
Force IPv4 Internet Protocol authentication
IF Infrastructure Version 4 key
IIOT Industrial 45 IPv6 Internet Protocol KPI Key
Internet of Things Version 6 80 Performance Indicator
IM Interference IR Infrared KQI Key Quality
Measurement, IS In Sync Indicator
Intermodulation, IRP Integration KSI Key Set
IP Multimedia 50 Reference Point Identifier
IMC IMS Credentials ISDN Integrated 85 ksps kilo-symbols per
IMEI International Services Digital second
Mobile Network KVM Kernel Virtual
Equipment ISIM IM Services Machine
Identity 55 Identity Module LI Layer 1
IMGI International ISO International 90 (physical layer) mobile group identity Organisation for Ll-RSRP Layer 1 IMPI IP Multimedia Standardisation reference signal
Private Identity ISP Internet Service received power
IMPU IP Multimedia 60 Provider L2 Layer 2 (data
PUblic identity IWF Interworking- 95 link layer)
IMS IP Multimedia Function L3 Layer 3 (network
Subsystem I-WLAN layer)
IMSI International Interworking LAA Licensed
Mobile 65 WLAN Assisted Access
Subscriber Constraint length 100 LAN Local Area
Identity of the convolutional Network loT Internet of code, USIM LADN Local
Things Individual key Area Data Network
IP Internet Protocol LBT Listen Before MAC Medium Access 70 MCOT Maximum
Talk Control (protocol Channel
LCM LifeCycle layering context) Occupancy Time
Management MAC Message MCS Modulation and
LCR Low Chip Rate 40 authentication code coding scheme
LCS Location (security/encryption 75 MD AF Management
Services context) Data Analytics
LCID Logical MAC-A MAC Function
Channel ID used for MDAS Management
LI Layer Indicator 45 authentication Data Analytics
LLC Logical Link and key 80 Service
Control, Low Layer agreement (TSG MDT Minimization of
Compatibility T WG3 context) Drive Tests
LMF Location MAC -IMAC used for ME Mobile
Management Function 50 data integrity of Equipment
LOS Line of signalling messages 85 MeNB master eNB
Sight (TSG T WG3 context) MER Message Error
LPLMN Local MANO Ratio
PLMN Management and MGL Measurement
LPP LTE Positioning 55 Orchestration Gap Length
Protocol MBMS 90 MGRP Measurement
LSB Least Significant Multimedia Gap Repetition
Bit Broadcast and Multicast Period
LTE Long Term Service MIB Master
Evolution 60 MBSFN Information Block,
LWA LTE-WLAN Multimedia 95 Management aggregation Broadcast multicast Information Base
LWIP LTE/WLAN service Single MIMO Multiple Input
Radio Level Frequency Multiple Output
Integration with 65 Network MLC Mobile Location
IPsec Tunnel MCC Mobile Country 100 Centre
LTE Long Term Code MM Mobility
Evolution MCG Master Cell Management
M2M Machine-to- Group MME Mobility
Machine Management Entity MN Master Node MSIN Mobile Station NE-DC NR-E- MNO Mobile Identification 70 UTRA Dual Network Operator Number Connectivity MO Measurement MSISDN Mobile NEF Network Object, Mobile 40 Subscriber ISDN Exposure Function
Originated Number NF Network
MPBCH MTC MT Mobile 75 Function
Physical Broadcast Terminated, Mobile NFP Network
CHannel Termination Forwarding Path
MPDCCH MTC 45 MTC Machine-Type NFPD Network Physical Downlink Communications Forwarding Path
Control CHannel mMTCmassive MTC, 80 Descriptor
MPDSCH MTC massive Machine- NFV Network Physical Downlink Type Communications Functions
Shared CHannel 50 MU-MIMO Multi Virtualization
MPRACH MTC User MIMO NFVI NFV Physical Random MWUS MTC 85 Infrastructure
Access CHannel wake-up signal, MTC NFVO NFV
MPUSCH MTC wus Orchestrator Physical Uplink Shared 55 NACKNegative NG Next Generation,
Channel Acknowledgement Next Gen
MPLS MultiProtocol NAI Network Access 90 NGEN-DC NG-RAN
Label Switching Identifier E-UTRA-NR Dual
MS Mobile Station NAS Non-Access Connectivity MSB Most Significant 60 Stratum, Non- Access NM Network Bit Stratum layer Manager
MSC Mobile NCT Network 95 NMS Network Switching Centre Connectivity Topology Management System MSI Minimum NC-JT Non- N-PoP Network Point of
System 65 Coherent Joint Presence
Information, Transmission NMIB, N-MIB MCH Scheduling NEC Network 100 Narrowband MIB Information Capability Exposure NPBCH
MSID Mobile Station Narrowband Identifier Physical Broadcast NSA Non-Standalone 70 OSI Other System
CHannel operation mode Information
NPDCCH NSD Network Service OSS Operations
Narrowband Descriptor Support System
Physical 40 NSR Network Service OTA over-the-air
Downlink Record 75 PAPR Peak-to-Average
Control CHannel NSSAINetwork Slice Power Ratio
NPDSCH Selection PAR Peak to Average
Narrowband Assistance Ratio
Physical 45 Information PBCH Physical
Downlink S-NNSAI Single- 80 Broadcast Channel
Shared CHannel NSSAI PC Power Control,
NPRACH NSSF Network Slice Personal
Narrowband Selection Function Computer
Physical Random 50 NW Network PCC Primary
Access CHannel NWUSNarrowband 85 Component Carrier,
NPUSCH wake-up signal, Primary CC
Narrowband Narrowband WUS P-CSCF Proxy
Physical Uplink NZP Non-Zero Power CSCF
Shared CHannel 55 O&M Operation and PCell Primary Cell
NPSS Narrowband Maintenance 90 PCI Physical Cell ID,
Primary ODU2 Optical channel Physical Cell
Synchronization Data Unit - type 2 Identity
Signal OFDM Orthogonal PCEF Policy and
NSSS Narrowband 60 Frequency Division Charging
Secondary Multiplexing 95 Enforcement
Synchronization OFDMA Function
Signal Orthogonal PCF Policy Control
NR New Radio, Frequency Division Function
Neighbour Relation 65 Multiple Access PCRF Policy Control
NRF NF Repository OOB Out-of-band 100 and Charging Rules
Function OOS Out of Sync Function
NRS Narrowband OPEX OPerating PDCP Packet Data
Reference Signal EXpense Convergence Protocol,
NS Network Service Packet Data Convergence PNFD Physical 70 PSCCH Physical Protocol layer Network Function Sidelink Control PDCCH Physical Descriptor Channel Downlink Control PNFR Physical PSSCH Physical Channel 40 Network Function Sidelink Shared PDCP Packet Data Record 75 Channel Convergence Protocol POC PTT over PSCell Primary SCell PDN Packet Data Cellular PSS Primary Network, Public PP, PTP Point-to- Synchronization
Data Network 45 Point Signal PDSCH Physical PPP Point-to-Point 80 PSTN Public Switched
Downlink Shared Protocol Telephone Network Channel PRACH Physical PT-RS Phase-tracking PDU Protocol Data RACH reference signal Unit 50 PRB Physical PTT Push-to-Talk PEI Permanent resource block 85 PUCCH Physical Equipment PRG Physical Uplink Control
Identifiers resource block Channel PFD Packet Flow group PUSCH Physical Description 55 ProSe Proximity Uplink Shared P-GW PDN Gateway Services, 90 Channel PHICH Physical Proximity-Based QAM Quadrature hybrid-ARQ indicator Service Amplitude channel PRS Positioning Modulation PHY Physical layer 60 Reference Signal QCI QoS class of PLMN Public Land PRR Packet 95 identifier Mobile Network Reception Radio QCL Quasi coPIN Personal PS Packet Services location Identification Number PSBCH Physical QFI QoS Flow ID, PM Performance 65 Sidelink Broadcast QoS Flow Identifier Measurement Channel 100 QoS Quality of PMI Precoding PSDCH Physical Service Matrix Indicator Sidelink Downlink QPSK Quadrature PNF Physical Channel (Quaternary) Phase Network Function Shift Keying QZSS Quasi-Zenith RL Radio Link 70 RRC Radio Resource
Satellite System RLC Radio Link Control, Radio
RA-RNTI Random Control, Radio Resource Control
Access RNTI Link Control layer
RAB Radio Access 40 layer RRM Radio Resource
Bearer, Random RLC AM RLC 75 Management
Access Burst Acknowledged Mode RS Reference Signal
RACH Random Access RLC UM RLC RSRP Reference Signal
Channel Unacknowledged Mode Received Power
RADIUS Remote 45 RLF Radio Link RSRQ Reference Signal
Authentication Dial In Failure 80 Received Quality
User Service RLM Radio Link RS SI Received Signal
RAN Radio Access Monitoring Strength Indicator
Network RLM-RS RSU Road Side Unit
RANDRANDom 50 Reference Signal RSTD Reference Signal number (used for for RLM 85 Time difference authentication) RM Registration RTP Real Time
RAR Random Access Management Protocol
Response RMC Reference RTS Ready-To-Send
RAT Radio Access 55 Measurement Channel RTT Round Trip
Technology RMSI Remaining MSI, 90 Time
RAU Routing Area Remaining Rx Reception,
Update Minimum Receiving, Receiver
RB Resource block, System S1AP SI Application
Radio Bearer 60 Information Protocol
RBG Resource block RN Relay Node 95 SI -MME SI for the group RNC Radio Network control plane
REG Resource Controller Sl-U SI for the user
Element Group RNL Radio Network plane
Rel Release 65 Layer S-CSCF serving
REQ REQuest RNTI Radio Network 100 CSCF
RF Radio Frequency Temporary Identifier S-GW Serving Gateway
RI Rank Indicator ROHC RObust Header S-RNTI SRNC
RIV Resource Compression Radio Network indicator value Temporary SDAP Service Data SI System Identity Adaptation Protocol, Information S-TMSI SAE Service Data SI-RNTI System Temporary Mobile Adaptation Information RNTI
Station Identifier 40 Protocol layer 75 SIB System
SA Standalone SDL Supplementary Information Block operation mode Downlink SIM Subscriber SAE System SDNF Structured Data Identity Module Architecture Evolution Storage Network SIP Session Initiated SAP Service Access 45 Function 80 Protocol Point SDP Session SiP System in
SAPD Service Access Description Protocol Package Point Descriptor SDSF Structured Data SL Sidelink SAPI Service Access Storage Function SLA Service Level Point Identifier 50 SDT Small Data 85 Agreement SCC Secondary Transmission SM Session Component Carrier, SDU Service Data Management Secondary CC Unit SMF Session
SCell Secondary Cell SEAF Security Anchor Management Function SCEF Service 55 Function 90 SMS Short Message Capability Exposure SeNB secondary eNB Service Function SEPP Security Edge SMSF SMS Function
SC-FDMA Single Protection Proxy SMTC SSB-based Carrier Frequency SFI Slot format Measurement Timing Division 60 indication 95 Configuration
Multiple Access SFTD Space-Frequency SN Secondary Node, SCG Secondary Cell Time Diversity, SFN Sequence Number Group and frame timing SoC System on Chip
SCM Security Context difference SON Self-Organizing Management 65 SFN System Frame 100 Network
SCS Subcarrier Number SpCell Special Cell Spacing SgNB Secondary gNB SP-CSI-RNTISemi-
SCTP Stream Control SGSN Serving GPRS Persistent CSI RNTI Transmission Support Node SPS Semi-Persistent
Protocol 70 S-GW Serving Gateway 105 Scheduling SQN Sequence Signal based Signal to TCP Transmission number Noise and Interference 70 Communication
SR Scheduling Ratio Protocol
Request SSS Secondary TDD Time Division
SRB Signalling Radio 40 Synchronization Duplex
Bearer Signal TDM Time Division
SRS Sounding SSSG Search Space Set 75 Multiplexing
Reference Signal Group TDMATime Division
SS Synchronization SSSIF Search Space Set Multiple Access
Signal 45 Indicator TE Terminal
SSB Synchronization SST Slice/Service Equipment
Signal Block Types 80 TEID Tunnel End
SSID Service Set SU-MIMO Single Point Identifier
Identifier User MIMO TFT Traffic Flow
SS/PBCH Block 50 SUL Supplementary Template
SSBRI SS/PBCH Block Uplink TMSI Temporary
Resource Indicator, TA Timing 85 Mobile
Synchronization Advance, Tracking Subscriber
Signal Block Area Identity
Resource Indicator 55 TAC Tracking Area TNL Transport
SSC Session and Code Network Layer
Service TAG Timing Advance 90 TPC Transmit Power
Continuity Group Control
SS-RSRP TAI Tracking TPMI Transmitted
Synchronization 60 Area Identity Precoding Matrix
Signal based TAU Tracking Area Indicator
Reference Signal Update 95 TR Technical Report
Received Power TB Transport Block TRP, TRxP
SS-RSRQ TBS Transport Block Transmission
Synchronization 65 Size Reception Point
Signal based TBD To Be Defined TRS Tracking
Reference Signal TCI Transmission 100 Reference Signal
Received Quality Configuration Indicator TRx Transceiver
SS-SINR TS Technical
Synchronization Specifications, Technical 35 UMTS Universal 70 V2X Vehicle-to-
Standard Mobile every thing
TTI Transmission Telecommunicat VIM Virtualized
Time Interval ions System Infrastructure Manager
Tx Transmission, UP User Plane VL Virtual Link,
Transmitting, 40 UPF User Plane 75 VLAN Virtual LAN,
Transmitter Function Virtual Local Area
U-RNTI UTRAN URI Uniform Network
Radio Network Resource Identifier VM Virtual Machine
Temporary URL Uniform VNF Virtualized
Identity 45 Resource Locator 80 Network Function
UART Universal URLLC UltraVNFFG VNF
Asynchronous Reliable and Low Forwarding Graph
Receiver and Latency VNFFGD VNF
Transmitter USB Universal Serial Forwarding Graph
UCI Uplink Control 50 Bus 85 Descriptor Information USIM Universal VNFMVNF Manager
UE User Equipment Subscriber Identity VoIP Voice-over-IP,
UDM Unified Data Module Voice-over- Internet
Management USS UE-specific Protocol
UDP User Datagram 55 search space 90 VPLMN Visited Protocol UTRA UMTS Public Land Mobile
UDSF Unstructured Terrestrial Radio Network
Data Storage Network Access VPN Virtual Private
Function UTRAN Universal Network
UICC Universal 60 Terrestrial Radio 95 VRB Virtual Resource
Integrated Circuit Access Network Block
Card UwPTS Uplink WiMAX
UL Uplink Pilot Time Slot Worldwide
UM V2I Vehicle-to- Interoperability
Unacknowledge 65 Infrastruction 100 for Microwave d Mode V2P Vehicle-to- Access
UML Unified Pedestrian WLANWireless Local Modelling Language V2V Vehicle-to- Area Network
Vehicle WMAN Wireless
Metropolitan Area
Network
WPANWireless Personal Area Network
X2-C X2-Control plane
X2-U X2-User plane
XML extensible Markup Language
XRES EXpected user
RESponse
XOR exclusive OR
ZC Zadoff-Chu ZP Zero Power
Terminology
For the purposes of the present document, the following terms and definitions are applicable to the examples and embodiments discussed herein.
The term “circuitry” as used herein refers to, is part of, or includes hardware components such as an electronic circuit, a logic circuit, a processor (shared, dedicated, or group) and/or memory (shared, dedicated, or group), an Application Specific Integrated Circuit (ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-capacity PLD (HCPLD), a structured ASIC, or a programmable SoC), digital signal processors (DSPs), etc., that are configured to provide the described functionality. In some embodiments, the circuitry may execute one or more software or firmware programs to provide at least some of the described functionality. The term “circuitry” may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) with the program code used to carry out the functionality of that program code. In these embodiments, the combination of hardware elements and program code may be referred to as a particular type of circuitry.
The term “processor circuitry” as used herein refers to, is part of, or includes circuitry capable of sequentially and automatically carrying out a sequence of arithmetic or logical operations, or recording, storing, and/or transferring digital data. Processing circuitry may include one or more processing cores to execute instructions and one or more memory structures to store program and data information. The term “processor circuitry” may refer to one or more application processors, one or more baseband processors, a physical central processing unit (CPU), a single-core processor, a dual-core processor, a triple-core processor, a quad-core processor, and/or any other device capable of executing or otherwise operating computerexecutable instructions, such as program code, software modules, and/or functional processes. Processing circuitry may include more hardware accelerators, which may be microprocessors, programmable processing devices, or the like. The one or more hardware accelerators may include, for example, computer vision (CV) and/or deep learning (DL) accelerators. The terms “application circuitry” and/or “baseband circuitry” may be considered synonymous to, and may be referred to as, “processor circuitry.”
The term “interface circuitry” as used herein refers to, is part of, or includes circuitry that enables the exchange of information between two or more components or devices. The term “interface circuitry” may refer to one or more hardware interfaces, for example, buses, I/O interfaces, peripheral component interfaces, network interface cards, and/or the like. The term “user equipment” or “UE” as used herein refers to a device with radio communication capabilities and may describe a remote user of network resources in a communications network. The term “user equipment” or “UE” may be considered synonymous to, 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 term “user equipment” or “UE” may include any type of wireless/wired device or any computing device including a wireless communications interface.
The term “network element” as used herein 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 to and/or referred to as a networked computer, networking hardware, network equipment, network node, router, switch, hub, bridge, radio network controller, RAN device, RAN node, gateway, server, virtualized VNF, NFVI, and/or the like.
The term “computer system” as used herein refers to any type interconnected electronic devices, computer devices, or components thereof. Additionally, the term “computer system” and/or “system” may refer to various components of a computer that are communicatively coupled with one another. Furthermore, the term “computer system” and/or “system” may refer to multiple computer devices and/or multiple computing systems that are communicatively coupled with one another and configured to share computing and/or networking resources.
The term “appliance,” “computer appliance,” or the like, as used herein refers to a computer device or computer system with program code (e.g., software or firmware) that is specifically designed to provide a specific computing resource. A “virtual appliance” is a virtual machine image to be implemented by a hypervisor-equipped device that virtualizes or emulates a computer appliance or otherwise is dedicated to provide a specific computing resource.
The term “resource” as used herein refers to a physical or virtual device, a physical or virtual component within a computing environment, and/or a physical or virtual component within a particular device, such as computer devices, mechanical devices, memory space, processor/CPU time, processor/CPU usage, processor and accelerator loads, hardware time or usage, electrical power, input/output operations, ports or network sockets, channel/link allocation, throughput, memory usage, storage, network, database and applications, workload units, and/or the like. A “hardware resource” may refer to compute, storage, and/or network resources provided by physical hardware element(s). A “virtualized resource” may refer to compute, storage, and/or network resources provided by virtualization infrastructure to an application, device, system, etc. The term “network resource” or “communication resource” may refer to resources that are accessible by computer devices/systems via a communications network. The term “system resources” may refer to any kind of shared entities to provide services, and may include computing and/or network resources. System resources may be considered as a set of coherent functions, network data objects or services, accessible through a server where such system resources reside on a single host or multiple hosts and are clearly identifiable.
The term “channel” as used herein refers to any transmission medium, either tangible or intangible, which is used to communicate data or a data stream. The term “channel” may be synonymous with and/or equivalent to “communications channel,” “data communications channel,” “transmission channel,” “data transmission channel,” “access channel,” “data access channel,” “link,” “data link,” “carrier,” “radiofrequency carrier,” and/or any other like term denoting a pathway or medium through which data is communicated. Additionally, the term “link” as used herein refers to a connection between two devices through a RAT for the purpose of transmitting and receiving information.
The terms “instantiate,” “instantiation,” and the like as used herein refers to the creation of an instance. An “instance” also refers to a concrete occurrence of an object, which may occur, for example, during execution of program code.
The terms “coupled,” “communicatively coupled,” along with derivatives thereof are used herein. The term “coupled” may mean two or more elements are in direct physical or electrical contact with one another, may mean that two or more elements indirectly contact each other but still cooperate or interact with each other, and/or may mean that one or more other elements are coupled or connected between the elements that are said to be coupled with each other. The term “directly coupled” may mean that two or more elements are in direct contact with one another. The term “communicatively coupled” may mean that two or more elements may be in contact with one another by a means of communication including through a wire or other interconnect connection, through a wireless communication channel or link, and/or the like.
The term “information element” refers to a structural element containing one or more fields. The term “field” refers to individual contents of an information element, or a data element that contains content.
The term “SMTC” refers to an SSB-based measurement timing configuration configured by SSB-MeasurementTimingConflguration.
The term “SSB” refers to an SS/PBCH block. The term “a “Primary Cell” refers to the MCG cell, operating on the primary frequency, in which the UE either performs the initial connection establishment procedure or initiates the connection re-establishment procedure.
The term “Primary SCG Cell” refers to the SCG cell in which the UE performs random access when performing the Reconfiguration with Sync procedure for DC operation.
The term “Secondary Cell” refers to a cell providing additional radio resources on top of a Special Cell for a UE configured with CA.
The term “Secondary Cell Group” refers to the subset of serving cells comprising the PSCell and zero or more secondary cells for a UE configured with DC. The term “Serving Cell” refers to the primary cell for a UE in RRC CONNECTED not configured with CA/DC there is only one serving cell comprising of the primary cell.
The term “serving cell” or “serving cells” refers to the set of cells comprising the Special Cell(s) and all secondary cells for a UE in RRC_CONNECTED configured with CA/.
The term “Special Cell” refers to the PCell of the MCG or the PSCell of the SCG for DC operation; otherwise, the term “Special Cell” refers to the Pcell.

Claims

CLAIMS What is claimed is:
1. An apparatus comprising: memory to store user equipment (UE) measurement report information; and processing circuitry, coupled with the memory, to: receive, from a next-generation NodeB (gNB), a request for location information for the UE; retrieve the UE measurement report information from the memory; determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on the UE measurement report information and the request for location information, wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
2. The apparatus of claim 1, wherein the UE measurement report information includes historical and current information for the UE.
3. The apparatus of claim 2, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
4. The apparatus of claim 1, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
5. The apparatus of claim 1, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
6. The apparatus of claim 1, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE.
7. The apparatus of any of claims 1-6, wherein the processing circuitry is to implement a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
8. One or more computer-readable media storing instructions that, when executed by one or more processors, cause a network function to: receive, from a next-generation NodeB (gNB), a request for location information for a user equipment (UE); determine, using a positioning artificial intelligence (Al) or machine learning (ML) model, the location information for the UE based on UE measurement report information and the request for location information wherein the location information for the UE includes an indication of a predicted location for the UE; and encode a message for transmission to the gNB that includes an indication of the determined location information for the UE.
9. The one or more computer-readable media of claim 8, wherein the UE measurement report information includes historical and current information for the UE.
10. The one or more computer-readable media of claim 9, wherein the historical and current information for the UE includes: location information, speed information, or trajectory information.
11. The one or more computer-readable media of claim 8, wherein the determined location information for the UE includes an indication of a predicted load of the UE.
12. The one or more computer-readable media of claim 8, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a current cell of the UE.
13. The one or more computer-readable media of claim 8, wherein the determined location information for the UE includes an indication of a current or predicted resource status for a target cell of the UE.
14. The one or more computer-readable media of any of claims 8-13, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
15. One or more computer-readable media storing instructions that, when executed by one or more processors, cause a next-generation NodeB (gNB) to: encode, for transmission to a network function, a request for location information for a user equipment (UE); receive, from the network function, the location information for the UE, wherein the location information for the UE includes an indication of a predicted location for the UE; determine configuration information that includes an indication of a combination of candidate handover cells and respective priorities for the candidate handover cells; and encode, for transmission to the UE, a message that includes the configuration information.
16. The one or more computer-readable media of claim 15, wherein determining the configuration information includes determining whether the UE should start a handover.
17. The one or more computer-readable media of claim 15, wherein the configuration information includes an indication of a target cell from the candidate handover cells to perform a handover, and timing information for performing the handover.
18. The one or more computer-readable media of claim 15, wherein the configuration information includes conditional handover (CHO) information.
19. The one or more computer-readable media of claim 18, wherein the CHO information includes an indication of conditional handover execution timing.
20. The one or more computer-readable media of claim 18, wherein the CHO information includes an indication of timing information for each candidate cell in the combination of candidate handover cells.
21. The one or more computer-readable media of any of claims 15-20, wherein the network function includes: a network data analytics function (NWDAF), operation administration and maintenance (0AM) function, or location management function (LMF).
PCT/US2022/039441 2021-08-05 2022-08-04 User equipment trajectory-assisted handover Ceased WO2023014896A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US18/556,421 US20240205781A1 (en) 2021-08-05 2022-08-04 User equipment trajectory-assisted handover
EP22853917.7A EP4381807A4 (en) 2021-08-05 2022-08-04 User equipment trajectory-assisted handover

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
CNPCT/CN2021/110778 2021-08-05
CN2021110778 2021-08-05
US202163236456P 2021-08-24 2021-08-24
US63/236,456 2021-08-24
US202163270445P 2021-10-21 2021-10-21
US63/270,445 2021-10-21

Publications (1)

Publication Number Publication Date
WO2023014896A1 true WO2023014896A1 (en) 2023-02-09

Family

ID=85156400

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2022/039441 Ceased WO2023014896A1 (en) 2021-08-05 2022-08-04 User equipment trajectory-assisted handover

Country Status (3)

Country Link
US (1) US20240205781A1 (en)
EP (1) EP4381807A4 (en)
WO (1) WO2023014896A1 (en)

Cited By (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116346198A (en) * 2023-03-01 2023-06-27 西安电子科技大学 A Mobility Management Method for Satellite Networks Based on Pre-interaction
WO2024156136A1 (en) * 2023-03-24 2024-08-02 Zte Corporation Intelligent optimization for dual connectivity in wireless communication systems
WO2024178346A1 (en) * 2023-02-24 2024-08-29 Interdigital Patent Holdings, Inc Determination of future candidate cells based on predictive model
WO2024176074A1 (en) * 2023-02-23 2024-08-29 Nokia Technologies Oy Control plane initiated delivery of machine learning model via user plane for wireless networks
WO2024179933A1 (en) * 2023-02-27 2024-09-06 Telefonaktiebolaget Lm Ericsson (Publ) Methods for flexible prediction timing configuration
GB2628603A (en) * 2023-03-30 2024-10-02 Nokia Technologies Oy Model execution in a wireless telecommunications network
WO2024205714A1 (en) * 2023-03-30 2024-10-03 Qualcomm Incorporated Efficient initial acquisition with gain state prediction using machine learning
WO2024197735A1 (en) * 2023-03-30 2024-10-03 Mediatek Singapore Pte. Ltd. Model performance monitor mechanism for ai/ml assisted positioning
WO2024207292A1 (en) * 2023-04-06 2024-10-10 Mediatek Singapore Pte. Ltd. Model performance monitor mechanism for direct ai/ml positioning based on soft information
WO2024207241A1 (en) * 2023-04-04 2024-10-10 Nokia Shanghai Bell Co., Ltd. Wake-up signal based measurement
WO2024223570A1 (en) * 2023-04-24 2024-10-31 Continental Automotive Technologies GmbH Method of data-driven model signaling for multi-usim
WO2024231675A1 (en) * 2023-05-09 2024-11-14 Vodafone Group Services Limtied Configuration of mobility parameters in a cellular network
WO2024237511A1 (en) * 2023-05-12 2024-11-21 Samsung Electronics Co., Ltd. Method and apparatus for managing ue privacy in wireless communication system
WO2025032399A1 (en) * 2023-08-08 2025-02-13 Nokia Technologies Oy Model selection and testing for functionality/model change
WO2025035367A1 (en) * 2023-08-14 2025-02-20 Shenzhen Tcl New Technology Co., Ltd. Ue mobility management method, user equipment, and base station
WO2025047814A1 (en) * 2023-08-30 2025-03-06 京セラ株式会社 Communication control method, base station, and user equipment
WO2025108574A1 (en) * 2023-11-22 2025-05-30 Telefonaktiebolaget Lm Ericsson (Publ) Dynamic configuration of uplink configured grants
WO2025111557A1 (en) * 2023-11-22 2025-05-30 Interdigital Patent Holdings, Inc. Methods and apparatus for measurement reporting based on prediction of time of stay in a cell
WO2025124715A1 (en) * 2023-12-14 2025-06-19 Telefonaktiebolaget Lm Ericsson (Publ) Machine-learning based cell management
WO2025124135A1 (en) * 2023-12-12 2025-06-19 华为技术有限公司 Communication method, and apparatus
WO2025137641A1 (en) * 2023-12-23 2025-06-26 Cisco Technology, Inc. Inverse basic service set transition management for temporal roaming avoidance
EP4598131A1 (en) * 2024-02-01 2025-08-06 Nokia Solutions and Networks Oy Mobile terminal handover decision optimizations
EP4598122A1 (en) * 2024-02-02 2025-08-06 Telia Company AB Management of handover
EP4601350A1 (en) * 2024-02-06 2025-08-13 Nokia Technologies Oy Computational mode switch
GB2638266A (en) * 2024-02-16 2025-08-20 Sony Group Corp Methods, communications devices, infrastructure equipment, and information processing servers
WO2025176427A1 (en) * 2024-02-19 2025-08-28 Continental Automotive Technologies GmbH System and apparatus for determining a model accuracy and a method in association thereto
WO2025209068A1 (en) * 2024-04-03 2025-10-09 大唐移动通信设备有限公司 Information processing method and apparatus, terminal, and network device
WO2025212285A1 (en) * 2024-04-01 2025-10-09 Interdigital Patent Holdings, Inc. Methods for user equipment trajectory prediction and reporting
EP4646894A4 (en) * 2023-02-10 2026-01-21 Zte Corp TRANSFER OF ARTIFICIAL INTELLIGENCE NETWORK MANAGEMENT MODELS TO WIRELESS COMMUNICATION SYSTEMS
EP4637205A4 (en) * 2023-02-16 2026-04-01 Huawei Tech Co Ltd COMMUNICATION METHOD AND DEVICE

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11716308B2 (en) * 2020-03-05 2023-08-01 Telefonaktiebolaget Lm Ericsson (Publ) Application triggered setup of distributed anchor for edge computing
WO2022156910A1 (en) * 2021-01-25 2022-07-28 Nokia Technologies Oy Enablement of federated machine learning for terminals to improve their machine learning capabilities
KR20240073862A (en) * 2021-09-28 2024-05-27 미쓰비시덴키 가부시키가이샤 communication system
CN116918418A (en) * 2021-11-23 2023-10-20 中兴通讯股份有限公司 Systems and methods for model management
US20230199591A1 (en) * 2021-12-17 2023-06-22 Intel Corporation Handover processes
US12507125B2 (en) * 2022-03-29 2025-12-23 Samsung Electronics Co., Ltd. Monitoring for application AI/ML-based services and operations
US12406232B2 (en) 2023-04-12 2025-09-02 Tyco Fire & Security Gmbh Building management system with generative AI-based automated flexible customer report generation and standardization
US12242937B1 (en) 2023-04-12 2025-03-04 Tyco Fire & Security Gmbh Building management system with generative AI-based root cause prediction
US12455896B2 (en) 2023-04-12 2025-10-28 Tyco Fire & Security Gmbh Building management system with generative AI-based interactive service tool
US20250113282A1 (en) * 2023-09-29 2025-04-03 Qualcomm Incorporated Enhanced AI/ML Based Mobility Optimization
EP4568218A1 (en) * 2023-12-05 2025-06-11 Nokia Solutions and Networks Oy Methods and apparatus for quantifying probability of expected network performance
WO2026019240A1 (en) * 2024-07-16 2026-01-22 주식회사 케이티 Method and device for performing communication using artificial intelligence and machine learning
WO2026008187A1 (en) * 2025-03-20 2026-01-08 Lenovo International Coöperatief U.A. Supporting creation and/or training of a learning model for a user equipment
CN119919125B (en) * 2025-04-01 2025-07-25 比塔（上海）数据科技有限公司 AI-based predictive maintenance system and method for distributed energy storage equipment
CN121164969B (en) * 2025-11-21 2026-02-27 科大智能科技股份有限公司 System and method for improving operation reliability of backup battery of electric energy meter

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100069070A1 (en) 2008-09-17 2010-03-18 Qualcomm Incorporated Location-assisted network entry, scan and handover
US20190223073A1 (en) * 2018-01-12 2019-07-18 FG Innovation Company Limited Conditional handover procedures
WO2020168321A1 (en) 2019-02-15 2020-08-20 Apple Inc. Location services (lcs) client in a next generation (ng) radio access network (ran)
WO2021032280A1 (en) 2019-08-19 2021-02-25 Nokia Technologies Oy Ue location tracking in an inactive state

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10512047B2 (en) * 2016-03-30 2019-12-17 Intel IP Corporation Access point (AP), station (STA) and method of multi-user (MU) location measurment
US20240007988A1 (en) * 2021-01-05 2024-01-04 Qualcomm Incorporated Multi-measurement reporting per reference signal

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100069070A1 (en) 2008-09-17 2010-03-18 Qualcomm Incorporated Location-assisted network entry, scan and handover
US20190223073A1 (en) * 2018-01-12 2019-07-18 FG Innovation Company Limited Conditional handover procedures
WO2020168321A1 (en) 2019-02-15 2020-08-20 Apple Inc. Location services (lcs) client in a next generation (ng) radio access network (ran)
WO2021032280A1 (en) 2019-08-19 2021-02-25 Nokia Technologies Oy Ue location tracking in an inactive state

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
CMCC: "Remaining Issues of LTE Conditional Handover", 3GPP DRAFT; R2-1905948, vol. RAN WG2, no. Reno, 13 May 2019 (2019-05-13)
HUAWEI, HISILICON: "Discussion on remaining issues for CHO in NTN", 3GPP DRAFT; R2-2105613, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. RAN WG2, no. Electronic Meeting; 20210519 - 20210527, 11 May 2021 (2021-05-11), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP052007184 *
HUAWEI: "Further discussions on AI/ML assisted load balancing", 3GPP DRAFT; R3-212524, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. RAN WG3, no. E-meeting; 20210517 - 20210527, 7 May 2021 (2021-05-07), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP052002571 *
See also references of EP4381807A4

Cited By (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4646894A4 (en) * 2023-02-10 2026-01-21 Zte Corp TRANSFER OF ARTIFICIAL INTELLIGENCE NETWORK MANAGEMENT MODELS TO WIRELESS COMMUNICATION SYSTEMS
EP4637205A4 (en) * 2023-02-16 2026-04-01 Huawei Tech Co Ltd COMMUNICATION METHOD AND DEVICE
WO2024176074A1 (en) * 2023-02-23 2024-08-29 Nokia Technologies Oy Control plane initiated delivery of machine learning model via user plane for wireless networks
WO2024178346A1 (en) * 2023-02-24 2024-08-29 Interdigital Patent Holdings, Inc Determination of future candidate cells based on predictive model
WO2024179933A1 (en) * 2023-02-27 2024-09-06 Telefonaktiebolaget Lm Ericsson (Publ) Methods for flexible prediction timing configuration
CN116346198A (en) * 2023-03-01 2023-06-27 西安电子科技大学 A Mobility Management Method for Satellite Networks Based on Pre-interaction
WO2024156136A1 (en) * 2023-03-24 2024-08-02 Zte Corporation Intelligent optimization for dual connectivity in wireless communication systems
EP4684556A4 (en) * 2023-03-24 2026-04-01 Zte Corp INTELLIGENT OPTIMIZATION FOR DUAL CONNECTIVITY IN WIRELESS COMMUNICATION SYSTEMS
WO2024205714A1 (en) * 2023-03-30 2024-10-03 Qualcomm Incorporated Efficient initial acquisition with gain state prediction using machine learning
WO2024197735A1 (en) * 2023-03-30 2024-10-03 Mediatek Singapore Pte. Ltd. Model performance monitor mechanism for ai/ml assisted positioning
US12316398B2 (en) 2023-03-30 2025-05-27 Qualcomm Incorporated Efficient initial acquisition with gain state prediction using machine learning
GB2628603A (en) * 2023-03-30 2024-10-02 Nokia Technologies Oy Model execution in a wireless telecommunications network
WO2024207241A1 (en) * 2023-04-04 2024-10-10 Nokia Shanghai Bell Co., Ltd. Wake-up signal based measurement
WO2024207292A1 (en) * 2023-04-06 2024-10-10 Mediatek Singapore Pte. Ltd. Model performance monitor mechanism for direct ai/ml positioning based on soft information
WO2024223570A1 (en) * 2023-04-24 2024-10-31 Continental Automotive Technologies GmbH Method of data-driven model signaling for multi-usim
WO2024231675A1 (en) * 2023-05-09 2024-11-14 Vodafone Group Services Limtied Configuration of mobility parameters in a cellular network
WO2024237511A1 (en) * 2023-05-12 2024-11-21 Samsung Electronics Co., Ltd. Method and apparatus for managing ue privacy in wireless communication system
WO2025032399A1 (en) * 2023-08-08 2025-02-13 Nokia Technologies Oy Model selection and testing for functionality/model change
WO2025035367A1 (en) * 2023-08-14 2025-02-20 Shenzhen Tcl New Technology Co., Ltd. Ue mobility management method, user equipment, and base station
WO2025047814A1 (en) * 2023-08-30 2025-03-06 京セラ株式会社 Communication control method, base station, and user equipment
WO2025108574A1 (en) * 2023-11-22 2025-05-30 Telefonaktiebolaget Lm Ericsson (Publ) Dynamic configuration of uplink configured grants
WO2025111557A1 (en) * 2023-11-22 2025-05-30 Interdigital Patent Holdings, Inc. Methods and apparatus for measurement reporting based on prediction of time of stay in a cell
WO2025124135A1 (en) * 2023-12-12 2025-06-19 华为技术有限公司 Communication method, and apparatus
WO2025124715A1 (en) * 2023-12-14 2025-06-19 Telefonaktiebolaget Lm Ericsson (Publ) Machine-learning based cell management
WO2025137641A1 (en) * 2023-12-23 2025-06-26 Cisco Technology, Inc. Inverse basic service set transition management for temporal roaming avoidance
US12389302B1 (en) 2024-02-01 2025-08-12 Nokia Solutions And Networks Oy Mobile terminal handover decision optimizations
EP4598131A1 (en) * 2024-02-01 2025-08-06 Nokia Solutions and Networks Oy Mobile terminal handover decision optimizations
EP4598122A1 (en) * 2024-02-02 2025-08-06 Telia Company AB Management of handover
WO2025168233A1 (en) * 2024-02-06 2025-08-14 Nokia Technologies Oy Apparatus, methods and computer programs
EP4601350A1 (en) * 2024-02-06 2025-08-13 Nokia Technologies Oy Computational mode switch
GB2638266A (en) * 2024-02-16 2025-08-20 Sony Group Corp Methods, communications devices, infrastructure equipment, and information processing servers
WO2025172711A1 (en) * 2024-02-16 2025-08-21 Sony Group Corporation Transmitting a request for updating an ai/ml model based on the fulfilment of certain configured triggered conditions
WO2025176427A1 (en) * 2024-02-19 2025-08-28 Continental Automotive Technologies GmbH System and apparatus for determining a model accuracy and a method in association thereto
WO2025212285A1 (en) * 2024-04-01 2025-10-09 Interdigital Patent Holdings, Inc. Methods for user equipment trajectory prediction and reporting
WO2025209068A1 (en) * 2024-04-03 2025-10-09 大唐移动通信设备有限公司 Information processing method and apparatus, terminal, and network device

Also Published As

Publication number Publication date
EP4381807A1 (en) 2024-06-12
US20240205781A1 (en) 2024-06-20
EP4381807A4 (en) 2025-06-25

Similar Documents

Publication Publication Date Title
US20240205781A1 (en) User equipment trajectory-assisted handover
US12413475B2 (en) A1 policy functions for open radio access network (O-RAN) systems
US20240349082A1 (en) Enhanced collaboration between user equpiment and network to facilitate machine learning
US20230199868A1 (en) Policy enhancement to support group application function (af) session from artificial intelligence/machine learning (aiml) provider af with required quality of service (qos)
US20240162955A1 (en) Beamforming for multiple-input multiple-output (mimo) modes in open radio access network (o-ran) systems
US20230171592A1 (en) Enhancing ran ue id based ue identification in o-ran
US20230189058A1 (en) Ran-aware traffic distribution rules and ran measurements for enhanced access traffic steering switching and splitting
US20230164598A1 (en) Self-organizing network coordination and energy saving assisted by management data analytics
US20220408445A1 (en) Link adaptation for 5g systems
US20240259277A1 (en) Edge computing network deployment for fifth-generation (5g) systems
WO2023014910A1 (en) Collision handling for sounding reference signal (srs) transmission
US20240243936A1 (en) Charging for edge enabling infrastructure resources
WO2022174070A1 (en) Performance measurements for edge computing applications
US12593301B2 (en) A1 enrichment information for user equipment (UE) physical positioning information
US20230422038A1 (en) Cyber attack detection function
WO2023018779A1 (en) Remote direct memory access (rdma) support in cellular networks
WO2022094039A1 (en) Computing offloading for next generation cellular networks
EP4646821A1 (en) Retrieval of user equipment (ue) identity in a mobile system in presence of network address translation
WO2024172887A1 (en) Resource allocation of sidelink positioning reference signal in a resource pool
WO2024167590A1 (en) Flight path update triggering
WO2022178127A1 (en) Performance measurements for data management and background data transfer policy control for next-generation systems
US20250393067A1 (en) Transmit (tx) carrier selection for new radio (nr) sidelink operation
EP4633288A1 (en) Network sharing at the distributed unit in the next generation cellular networks
US20240275552A1 (en) Positioning bandwidth aggregation of positioning reference signal (prs) and sounding reference signal (srs)
WO2024167569A1 (en) Management data analytics (mda) capability for network function (nf) resource utilization analysis

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22853917

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18556421

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2022853917

Country of ref document: EP

Effective date: 20240305

WWW Wipo information: withdrawn in national office

Ref document number: 2022853917

Country of ref document: EP