Machine Learning Fallback Model for Wireless Devices

Abstract

According to some embodiments, a method is performed by a wireless device for fallback operation of a machine learning (ML) model. The method includes transmitting, to a network node, a message indicating the wireless device's ability to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality, operating the at least one ML-based feature for the functionality, and operating the at least one preliminary feature for the functionality.

Description

Technical Field

[0001] Embodiments of the present disclosure are directed to wireless communication, and more specifically, to a machine learning fallback model for wireless devices.

Background Art

[0002] Generally, all terms used herein should be interpreted according to their ordinary meanings in the relevant technical field, unless a different meaning is clearly given and / or suggested by the context in which they are used. All references to an element, apparatus, component, means, step, etc. should be construed openly as a reference to at least one instance of those elements, apparatus, components, means, steps, etc., unless otherwise explicitly stated. Any method step disclosed herein need not be performed in the exact order disclosed, unless a step is explicitly described as following or preceding another step and / or it is not implicitly understood that a step must follow or precede another step. Any feature of any embodiment disclosed herein may be applied to any other embodiment, if appropriate. Similarly, any advantage of any embodiment may apply to any other embodiment, and vice versa. Other objects, features, and advantages of the embodiments included will become apparent from the following description.

[0003] Artificial intelligence (AI) and machine learning (ML) are regarded as promising tools for optimizing the design of air interfaces in wireless communication networks in both the academic and industrial communities. Exemplary use cases include the use of autoencoders for channel state information (CSI) to reduce feedback overhead and improve channel prediction accuracy, the use of deep neural networks for classifying line-of-sight (LOS) and non-line-of-sight (NLOS) conditions to improve positioning accuracy, the use of reinforcement learning for beam selection on the network side and / or user equipment (UE) side to reduce signaling overhead and beamforming latency, and the use of deep reinforcement learning to learn optimal precoding policies for complex multiple-input multiple-output (MIMO) precoding problems.

[0004] The standardization work of Release 18 of the 3rd Generation Partnership Project (3GPP) New Radio (NR) includes research items on AI / ML for the NR air interface. The research items will explore the benefits of enhancing the air interface with a set of functions that enable improved support for AI / ML-based algorithms for performance improvement and / or complexity / overhead reduction. Through the study of some selected use cases (CSI feedback, beam management, and positioning), the research items intend to lay the foundation for leveraging AL / ML technologies for future air interface use cases.

[0005] When applying AI / ML to the use cases of the air interface, various levels of cooperation between the network node and the UE can be considered. For one use case, there is no cooperation between the network node and the UE. In this case, a private ML model operating on the existing standard air interface is applied at one end of the communication chain (e.g., on the UE side), and the life cycle management of the model (e.g., model selection / training, model monitoring, model retraining, model updating) is performed at that node without assistance between nodes (e.g., the provision of assistance information by the network node).

[0006] For other use cases, limited cooperation is carried out between the network node and the UE. In this case, an ML model is operating at one end of the communication chain (e.g., on the UE side), but that node receives assistance from the node at the other end of the communication chain (e.g., the next-generation node B (gNB)) for the life cycle management of its AI model (e.g., for the training / retraining of the AI model and model updating).

[0007] The third use case is the combined ML operation between the network node and the UE. In this case, the AI model can be split into one part located on the network side and the other part located on the UE side. Thus, the AI model will include joint training between the network and the UE, and both ends of the communication chain are involved in the life cycle management of the AI model.

[0008] The construction of an AI model or some machine learning model involves several development steps, and the actual training of the AI model is just one step within the training pipeline. An important part in AI development is the life cycle management of the ML model. An example is shown in Figure 1.

[0009] Figure 1 shows the training and inference pipelines, and their interactions in the model lifecycle management procedure. Model lifecycle management typically consists of a training (retraining) pipeline, a deployment stage for incorporating a trained (or retrained) AI model into part of the inference pipeline, the inference pipeline, and a drift detection stage for notifying about drift in model behavior.

[0010] The training (retraining) pipeline may include data ingestion, data preprocessing, model training, model evaluation, and model registration. Data ingestion refers to the collection of raw (training) data from a data storage. After data ingestion, there may be steps to control the validity of the collected data.

[0011] Data preprocessing refers to the engineering features applied to the collected data and may include, for example, normalization and possibly transformation of the data required for input data to an AI model.

[0012] Model training refers to the actual model training stage as outlined above.

[0013] Model evaluation refers to benchmarking the performance against a model baseline. The steps of model training and model evaluation are repeated until an acceptable level of performance (as exemplified above) is achieved.

[0014] Model registration refers to registering the AI model, and possibly the results of the performance of the AI model evaluation, including any corresponding AI metadata that provides information on how the AI model was developed.

[0015] The deployment stage incorporates a trained (or retrained) AI model into part of the inference pipeline.

[0016] The inference pipeline may include data ingestion, data preprocessing, model operation, and data and model monitoring. Data ingestion refers to the collection of raw (inference) data from data storage.

[0017] The data preprocessing stage is typically the same as the corresponding processing that appears in the training pipeline.

[0018] Model operation refers to using the trained and deployed model in the operational mode.

[0019] Data and model monitoring refers to monitoring the model output to detect any performance or operational drift, in addition to verifying that the inference data is from a distribution that matches well with the training data.

[0020] In the drift detection stage, notifications are made about drifts in model operation.

[0021] Currently, there are a number of challenges. For example, in a certain use case category, an ML model is deployed on the UE side and the model output is reported from the UE to the network node. Based on the model output, the network takes actions that affect the current and subsequent wireless communications between the network and the UE.

[0022] The ML model deployed on the UE side is not generalized for some scenarios. Therefore, the output of the ML model (e.g., the estimated Channel Quality Indicator (CQI) value, the Channel State Information (CSI) predicted in one or more sub-bands, the beam measurement results predicted in the time and / or spatial domain, the estimated UE location, etc.) may not be accurate, or the error interval may be higher than an acceptable level, and / or the accuracy (or accuracy interval) may not be acceptable. Since the network performs transmission and reception actions based on the ML model output, inaccurate model output may result in incorrect decisions being made on the network side, thereby affecting the performance of wireless communication.

[0023] For example, based on the incorrect beam measurement prediction reported by the UE, the network may activate the Transmission Configuration Information (TCI) state (and / or trigger beam switching) for the UE that does not correspond to the beam detectable by the UE (or has poor coverage performance). Incorrect decisions may result in beam obstruction, radio link obstruction, poor throughput, and / or excessive signaling due to the subsequent configuration / activation of CSI measurements.

[0024] In other categories of use cases, the ML model is split into one part located on the network side and the other part located on the UE side. One exemplary use case is the Autoencoder (AE)-based CSI feedback / reporting, where the encoder operates at the UE to compress the estimated wireless channel, and the output from the encoder (the estimation result of the compressed wireless channel information) is reported from the UE to the gNB. The gNB uses the decoder to reconstruct the estimated wireless channel information. Therefore, the ML model for this category of use cases requires collaborative operation between the network and the UE. If the part of the ML model located on the UE side does not function well, it will affect the overall performance of the related functionality (e.g., CSI reporting).

[0025] If the performance drift of the ML model is detected, it may be possible to start a new data set to maintain the ML model. However, maintaining such data sets and models can take a long time. Depending on the capabilities of the UE, online model maintenance may not be realistic.

[0026] When an ML model is used for critical functionality, it is important to ensure that the robustness and resilience performance of that functionality are not affected if the ML model does not operate well. If performance issues are detected or predicted in the active ML model(s) associated with that critical functionality, prompt action needs to be taken.

[0027] Regarding the categories of ML use cases mentioned above, the current NR standard does not have a mechanism to guarantee / maintain the robustness and resilience of the functionality when an ML model operating for critical functionality does not operate well. Summary of the Invention

[0028] As described above, there are currently problems with preliminary models of machine learning for wireless devices. Certain aspects of the present disclosure and their embodiments may provide solutions to these or other problems.

[0029] For example, a specific embodiment includes a user equipment (UE) capable of operating at least one machine learning (ML)-based feature for a certain functionality and supporting at least preliminary features for that functionality. The UE indicates to the network its capability to support a combination of at least one ML-based feature and preliminary features for the functionality.

[0030] When a performance problem is detected for at least one ML-based feature, the UE can either be instructed by the network to switch to a preliminary feature for that functionality or autonomously switch to a preliminary feature and indicate the feature switch to the network.

[0031] According to some embodiments, a method in a UE operating with at least one ML-based feature associated with a functionality includes sending a message to a network node indicating its ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality.

[0032] In a specific embodiment, the at least one ML-based feature is based on one or more ML models located in the UE. In a specific embodiment, the at least one ML-based feature is based on one ML model that is divided into two parts, one part is located in the UE, and the other part is located in the network node. In a specific embodiment, the at least one ML-based feature is based on a plurality of ML models, some of those models are located in the UE, and the remainder of those models are located in the network.

[0033] In a specific embodiment, the above preliminary feature is a feature that can satisfy functionality equivalent to the above ML-based feature, but is not preferred compared to the above alternative means by ML. In a specific embodiment, the above preliminary feature is a feature having an ability equal to or lower than that of the above ML-based feature. In a specific embodiment, the above preliminary feature is a feature having a higher ability than the above ML-based feature, provided that the preliminary feature is not preferred for other reasons, and the other reasons include higher complexity, longer processing delay, higher power consumption, excessive consumption of time / frequency resources, etc. The definition of higher ability depends on functionality. For example, for channel state information (CSI), higher ability may refer to more accurate CSI feedback (including sub-band selection, rank indicator (RI), precoding matrix indicator (PMI), modulation and coding scheme (MCS)), for beam management, higher ability may refer to higher accuracy in indicating the best candidate beam, and for positioning, higher ability may refer to more accurate estimation of the UE's position.

[0034] In a specific embodiment, the above preliminary feature is based on a classical non-ML-based algorithm. In a specific embodiment, the above preliminary feature is an ML-based algorithm.

[0035] In a specific embodiment, the message indicates whether the at least one preliminary feature and the ML-based feature can be executed simultaneously. The message indicating its own ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality is a (part of) UE capability parameter associated with the functionality. The message may explicitly indicate that a UE supporting an ML-based feature is assumed to also support a preliminary feature for the associated functionality. The message may point to at least one entry for a combination of composite codebooks in a form where a certain codebook type is associated with the ML-based feature. The message may support different combinations of at least one ML-based feature and at least one preliminary feature between frequency division duplexing (FDD) and time division duplexing (TDD), between FR1 and FR2, and / or between different multiple bands.

[0036] In a specific embodiment, the message indicating its own ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality is a radio resource control (RRC) message, a media access control (MAC) control element (CE), Msg1, MsgA, Msg3, a combination of Msg1 and Msg3, uplink control information (UCI), or scheduling control information (SCI). The message may be transmitted when the UE activates / switches on / registers at least one ML model associated with the ML-based feature.

[0037] In a specific embodiment, the method further includes the UE receiving from the network node a first configuration message that configures the UE to simultaneously execute / operate at least the ML-based feature and at least the preliminary feature for the associated functionality.

[0038] In a specific embodiment, the method further includes the UE receiving from the network node a second configuration message that configures the UE to deactivate / stop / switch off at least one ML-based feature for the associated functionality and to activate / switch on an associated fallback feature. The network may send the second configuration message when a performance impairment of at least one ML-based feature for the associated functionality is detected / predicted. The method further includes the UE deactivating / stopping the ML-based feature and activating / switching on the fallback feature according to the information contained in the second configuration message in response to receiving the second configuration message from the network node.

[0039] In a specific embodiment, the method further includes the UE monitoring the ML model performance of the one or more ML-based features. The UE detects or predicts a performance impairment of at least one ML-based feature for the associated functionality, autonomously deactivates / stops at least the detected ML-based feature, and activates / switches on at least the associated fallback feature. The method further includes the UE indicating to the network node the switching information of the features (e.g., the ML-based feature that is deactivated / stopped / switched off).

[0040] In a specific embodiment, when there are multiple fallback features supported by the UE for the associated functionality, the order / sequence for the UE to perform function switching (e.g., first switch to fallback feature 1, and if there is a problem, then switch to fallback feature 2, etc.) is preconfigured by the network node or predefined in the standard specification.

[0041] In a specific embodiment, examples of functionality include CSI reporting, time domain beam prediction or beam selection, spatial domain beam prediction or beam selection, beam obstruction prediction, wireless link obstruction prediction, mobility management (e.g., handover decision), location estimation, link adaptation (e.g., MCS selection).

[0042] In a specific embodiment, the functionality is CSI reporting, the at least one ML-based feature is an ML-based CSI report, and the at least one preliminary feature is a legacy CSI reporting type (e.g., type 2 codebook-based CSI reporting, e type 2 codebook-based CSI reporting, or type 1 single panel-based CSI reporting).

[0043] According to some embodiments, a method in a network node includes receiving, from the UE, a message indicating the UE's ability to support a combination of at least one ML-based feature and at least one preliminary feature for a certain functionality.

[0044] In a specific embodiment, the method further includes the network node transmitting a first configuration message to the UE instructing the UE to simultaneously execute / operate at least the ML-based feature and at least the preliminary feature for the associated functionality in response to receiving the UE's capability information.

[0045] In a specific embodiment, the message further includes transmitting a second configuration message to the UE instructing the UE to deactivate / stop / switch off at least the detected / predicted ML-based feature having a performance issue and activate / switch on the associated preliminary feature in response to detecting / predicting a performance impairment of at least one ML-based feature for the associated functionality.

[0046] In a specific embodiment, the method further includes receiving, from the UE, an indication about the switching information of the features (e.g., deactivated / stopped / switched-off ML-based features and activated / switched-on backup features) for the associated functionality of the UE.

[0047] In a specific embodiment, the method further includes, for a case where the ML-based feature is split into two parts in a form that one part is located in the UE and the other part is located in the network, or for a case where the ML-based feature is based on multiple ML models in a form that a part of the model is located in the UE and the rest of the model is located in the network, deactivating / stopping / switching off the associated ML model on the network side for at least the deactivated ML-based feature.

[0048] In a specific embodiment, the method further includes the network node accordingly sending an adjusted configuration and / or a scheduling message to the UE. The adjustment by the adjusted configuration and / or the scheduling message may include an updated reference signal resource configuration for UE measurement and / or an updated CSI reporting configuration for the UE to report CSI using the backup feature.

[0049] According to some embodiments, a method is performed by a wireless device for fallback operation of an ML model. The method includes sending, to a network node, a message indicating the ability of the wireless device to support a combination of at least one ML-based feature for a functionality and at least one backup feature for the functionality, operating the at least one ML-based feature for the functionality, and operating the at least one backup feature for the functionality.

[0050] In a specific embodiment, the at least one ML-based feature is based on one ML model that is divided into two parts, one part being located in the wireless device and the other part being located in the network node.

[0051] In a specific embodiment, the at least one preliminary feature is a feature that satisfies functionality equivalent to that of the ML-based feature, but is not preferred as compared to the ML-based feature. The at least one preliminary feature may be a feature having a higher ability than the ML-based feature, but the preliminary feature is not preferred. The at least one preliminary feature may be based on a non-ML-based algorithm, or may be another ML-based algorithm (for example, a more general ML-based algorithm).

[0052] In a specific embodiment, the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously (for example, for comparison of performance between the two).

[0053] In a specific embodiment, the method further includes receiving a first configuration message that configures the wireless device to operate the at least one ML-based feature.

[0054] In a specific embodiment, the method further includes receiving a first configuration message that configures the wireless device to operate the at least one ML-based feature and the at least one preliminary feature simultaneously.

[0055] In a specific embodiment, the method further includes receiving a second configuration message that configures the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature.

[0056] In a specific embodiment, the method further includes autonomously determining that the at least one ML-based feature should be deactivated and the at least one preliminary feature should be activated.

[0057] According to some embodiments, a wireless device comprises processing circuitry operable to execute any of the methods of the wireless device described above.

[0058] Also disclosed is a computer program product including a non-transitory computer-readable medium storing computer-readable program code, the computer-readable program code being operable, when executed by processing circuitry, to perform any of the methods executed by the wireless device described above.

[0059] According to some embodiments, a method is performed by a network node to configure a wireless device for fallback operation of an ML model. The method includes receiving, from the wireless device, a message indicating the wireless device's ability to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality, determining that the at least one preliminary feature should be activated, and sending a configuration message to the wireless device to configure the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature.

[0060] In a specific embodiment, the at least one ML-based feature is based on one ML model that is divided into two parts, one part located at the wireless device and the other part located at the network node.

[0061] In a specific embodiment, the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously.

[0062] In a specific embodiment, the method further includes transmitting, to the wireless device, a configuration message that configures the wireless device to operate the at least one ML-based feature.

[0063] In a specific embodiment, the method further includes transmitting a configuration message that configures the wireless device to simultaneously operate the at least one ML-based feature and the at least one preliminary feature.

[0064] Other computer program products include a non-transitory computer-readable medium storing computer-readable program code that, when executed by a processing circuit, is operable to perform any of the methods performed by the network node described above.

[0065] Certain embodiments may provide one or more of the following technical advantages. For example, a specific embodiment ensures that a UE supporting an ML-based feature for a certain critical functionality also supports a preliminary feature for that functionality. By sharing the capability information of such a UE with a network node, the UE (and the network node) can switch to a preliminary feature when a performance problem is detected / predicted for the ML-based feature. Thus, a specific embodiment guarantees / maintains the robustness and resilience of a functionality when the ML model operating for that critical functionality is not performing well.

Brief Description of the Drawings

[0066] For a more complete understanding of the disclosed embodiments and their features and advantages, reference is now made to the following description taken in conjunction with the accompanying drawings:

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

DETAILED DESCRIPTION OF THE INVENTION

[0067] As described above, there are currently certain problems with the preliminary models of machine learning for wireless devices. Certain aspects of the present disclosure and their embodiments may provide solutions to these or other problems.

[0068] For example, a specific embodiment includes a user equipment (UE) that can operate at least one machine learning (ML)-based feature for a certain functionality and also supports at least preliminary features for that functionality. The UE indicates to the network its own capability to support a combination of at least one ML-based feature and preliminary features for the above functionality.

[0069] Specific embodiments are described more fully with reference to the accompanying drawings. However, other embodiments are also included within the scope of the subject matter disclosed herein, and the disclosed subject matter should not be construed as being limited only to the embodiments described herein. Rather, those embodiments are provided as examples to convey the scope of the subject matter to those skilled in the art.

[0070] As used herein, the terms "ML model", "AI-based feature", and "ML-based feature" are interchangeable. An AI / ML model may be defined as a functionality or a part of a functionality deployed / implemented within a first node. The first node may receive, from a second node, a message indicating that the functionality is not operating correctly, for example, that the prediction error is higher than a predefined value, the error interval is not within an acceptable level, or the prediction accuracy is lower than a predefined value.

[0071] Furthermore, an AI / ML model may be defined as a functionality or a part of a functionality implemented / supported within a first node. The first node may indicate the version of the functionality to the second node. When the above ML model is updated, the version of the functionality may be changed by the first node.

[0072] The ML model may correspond to a function that receives one or more inputs (e.g., measurement results) and provides one or more prediction / estimation results of a certain type as outputs. In one example, the ML model may correspond to a function that receives, as an input, the measurement result of a reference signal (e.g., transmitted in beam X) at time t0 and provides, as an output, a prediction of the reference signal at time t0+T. In another example, the ML model may correspond to a function that receives, as an input, the measurement result of a reference signal X (such as a synchronization signal block (SSB) with index 'x', e.g., transmitted in beam X) and provides, as an output, a prediction of another reference signal transmitted in a different beam (such as reference signal Y, e.g., transmitted in beam X) such as an SSB with index 'x'.

[0073] Another example is an ML model that assists in the estimation of channel state information (CSI). In such a setup, the ML models are an ML model specific to the UE and an ML model on the network side. Both ML models cooperate to provide combined network functionality. The function of the ML model in the UE is to compress the channel input, and the function of the ML model on the network side is to decompress the output received from the UE.

[0074] Furthermore, it is possible to apply the ML model to positioning. In that case, the input may be a channel impulse related to a certain temporal reference point (typically, a transmission point (TP)). The objective on the network side is to detect different peaks within the impulse response that reflect the multipaths experienced by the radio signal reaching the UE side. Another positioning method is to input multiple sets of measurement results into the ML network and derive the estimated position of the UE based on that.

[0075] Other ML models are ML models that assist in channel estimation or interference estimation for channel estimation in the UE. The channel estimation may be, for example, for the physical downlink shared channel (PDSCH) and may be associated with a specific set of reference signal patterns transmitted from the network to the UE. The ML model is part of the receiver chain within the UE and may not be directly visible within the reference signal patterns configured / scheduled for use between the network and the UE. Other examples of ML models for CSI estimation are those that predict appropriate CQI, PMI, RI, CRI (CSI-RS resource indicator) or similar values towards the future. The future may be a number of slots after the UE has made its last measurement or may target a specific slot at a future point in time.

[0076] The network node may be one of a general network node (gNB) base station unit within a base station handling at least some of the above functionality, a relay node, a core network node, a core network node handling at least some of the above functionality, a device supporting device-to-device (D2D) communication, a location management function (LMF), or another type of location server.

[0077] In the use cases described herein, the ML-based features are at least partially in the UE. In certain types of use cases, the ML-based features may be based on multiple ML models deployed on the UE side (e.g., for RX beam prediction, the ML model is located on the UE side). In other types of use cases, the ML-based features are based on one ML model that is split into two parts, one part located in the UE and the other part located in the network node (e.g., AE-based CSI feedback / reporting). In yet another type of use case, the ML-based features are based on multiple ML models, some of those models located in the UE and the remainder of those models located in the network (e.g., in ML-based beam pair prediction between a network node and a UE, an ML model is located in the network node for its TX beam prediction and another ML model is located in the UE for its RX beam prediction).

[0078] If a performance drift of the ML model is detected, one option is to start a new data set to maintain the ML model. However, maintaining such a data set and model may take a long time. Depending on the UE's capabilities, online model maintenance may not be realistic.

[0079] When an ML model is used for critical functionality, it is important to ensure that the robustness and recovery performance of that functionality are not affected if the ML model does not operate well. If performance issues with the active ML model associated with that critical functionality are detected or predicted, rapid action needs to be taken.

[0080] The specific embodiments described herein enable a UE operating with at least one ML-based feature for critical functionality to quickly switch to a backup feature for that functionality when performance issues with the active ML model associated with that critical functionality are detected or predicted.

[0081] A preliminary feature may be a feature having an ability equal to or lower than that of an ML-based feature. The preliminary feature may be based on a classical non-ML-based algorithm. For example, considering the use case of AE-based CSI feedback / reporting, the functionality is CSI feedback / reporting, and one ML-based feature for this functionality may be AE-based CSI feedback / reporting (dual-side ML algorithm), and one preliminary feature may be a legacy CSI reporting type (e.g., type 2 codebook-based CSI reporting, e type 2 codebook-based CSI reporting, or type 1 single-panel-based CSI reporting).

[0082] In some embodiments, the preliminary feature may have a higher ability than the ML-based feature, provided that the preliminary feature is not preferred for other reasons, where the other reasons include higher complexity, longer processing delay, higher power consumption, excessive consumption of time / frequency resources, etc. Generally, the preliminary feature satisfies the functionality equivalent to that of the ML-based feature, but is not preferred compared to the ML-based alternative.

[0083] What is regarded as higher ability may depend on the function. For example, for CSI, higher ability may refer to more accurate CSI feedback (including sub-band selection, RI, PMI, MCS). For beam management, higher ability may refer to higher accuracy in identifying the best candidate beam. For positioning, higher ability may refer to more accurate estimation of the UE's position.

[0084] Although specific examples focus on embodiments where the preliminary feature is a classical non-ML-based algorithm, in some embodiments, the preliminary feature may also be ML-based.

[0085] In one example, the UE is capable of supporting at least two ML models. The first ML model is a generalized model that can be used in a wide range of deployments (e.g., indoor and outdoor, dense urban and rural, high mobility and low mobility), while the second ML model is a specialized model that is trained to perform best for a specific deployment (e.g., an indoor factory). In this case, when the UE is deployed in the trained environment, the first ML model can be used as a preliminary feature, while the second model is activated as the preferred feature. Generally, one ML model is more basic and can be used as a preliminary feature, while the other ML model is more refined and can be used as the preferred model unless it is considered inappropriate, e.g., if excessive errors are detected during the monitoring period.

[0086] In other examples, the at least two ML models may be different versions or models / algorithms for the same functionality, considering them independently from the case where one is more generalized than the other. The two ML models are then identified by a model ID or model version in that case. They may, for example, both support CSI reporting, but the resolution or detail of the reporting may still differ. Similar aspects may be generalized such that the two ML models only support a subset of features or a lower resolution than the other.

[0087] In other examples, the UE is capable of supporting at least two ML models, and these at least two ML models use different inputs and / or different outputs. In some embodiments, the first ML model is equivalent to a classical non-ML-based algorithm in terms of the input and output of the algorithm, that is, the first ML model does not require any interface changes in terms of signaling, configuration, measurement, reporting, etc. Therefore, a functional "black box" can be equally realized by either the classical algorithm or the first ML model, and the UE does not need to notify (and the network node does not need to recognize) whether the UE is running a classical algorithm or the first ML model. This first ML model can be used as a preliminary feature. The second ML model requires explicit cooperation to fulfill a more advanced algorithm between the network node and the UE, and the explicit cooperation is reflected in the changes of the Uu interface compared to the classical algorithm, including different configurations of RS, different measurement modes, different reports from the UE, etc.

[0088] Some embodiments include UE Capability Indication. To support the UE's switching from a certain ML-based feature to a preliminary feature for critical functionality, the UE is capable of supporting a combination of the ML-based feature and the preliminary feature for that functionality.

[0089] In some embodiments, a UE capable of operating at least one ML-based feature for a certain functionality is required to also support at least a preliminary feature for that functionality.

[0090] The requirement may be explicitly defined as part of the UE capability parameters in the specification. For example, in the UE capability parameters in the specification, it can be explicitly described that a UE supporting ML-based feature X is also considered to support preliminary feature Y for the functionality associated with it. The UE can indicate this capability information to the network node by various methods described below.

[0091] In some embodiments, the UE indicates, in the report of its capabilities, its ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality.

[0092] In some embodiments, a UE capable of operating at least one ML-based feature associated with a certain functionality sends a message to the network node indicating its ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality.

[0093] In some embodiments, the above message indicating the UE's ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality is (part of) the UE capability parameter associated with the above functionality.

[0094] In some embodiments, the above message explicitly indicates that a UE supporting a certain ML-based feature is also considered to support a preliminary feature for the functionality associated with it.

[0095] In some embodiments, the UE indicates having its own capabilities that support at least one combination of ML-based features in the reporting of the UE's capabilities, and support for at least one preliminary feature regarding the functionality associated therewith, a feature that is not explicitly declared as a capability. The network can require the preliminary functionality for many UEs, regardless of their capabilities. For example, the UE may declare its capabilities for ML-based demodulation reference signal (DMRS) channel estimation. This may be in the form of supporting multiple different DMRS patterns or multiple different receiver requirements for a particular DMRS pattern. Also, all UEs may implement legacy non-ML channel estimation, and its use may be required by the network.

[0096] For example, for the functionality of codebook-based CSI feedback / reporting, the UE capability parameter codebookComboParametersAddition-r16 has been incorporated into 3GPP TS 38.306 v17.0.0 of Release 16 of NR, and as shown in the following table, this indicates that the UE supports a combined codebook combination.

Table 1

[0097] Regarding the use cases of CSI feedback / reporting, a message indicating the UE's ability to support a combination of an ML-based feature (e.g., ML-based CSI feedback / reporting) and at least one preliminary feature (e.g., Type 1 single panel) may be defined by a similar UE capability parameter, e.g., codebookComboParametersAddition-r18 with a new entry {Type 1 Single Panel, Type 3}, where Type 3 means the function of ML-based CSI feedback. For more advanced UEs, new entries representing combinations of more than two codebook types may be added, e.g., {Type 1 Single Panel, Type 2, Type 3}.

[0098] In some embodiments, the above message indicating the UE's ability to support a combination of at least one ML-based feature and at least one preliminary feature regarding the functionality of CSI reporting is part of the UE capability parameter. The UE capability parameter includes at least one entry regarding a combination of composite codebooks in a form where a certain codebook type is associated with an ML-based feature.

[0099] In some embodiments, the above message may support different combinations of at least one ML-based feature and at least one preliminary feature between FDD and TDD, and / or between FR1 and FR2 (or other multiple spectrum ranges), multiple feature sets, and / or different multiple bands.

[0100] In some embodiments, when the UE activates / switches on / registers at least one ML model for a certain ML-based feature, the UE may send information regarding that ML-based feature (e.g., ML model ID) to the network node, and in addition, may also send the preliminary feature associated with that ML-based feature.

[0101] A message containing information about ML-based features and a message containing preliminary features associated with the ML-based features may be the same message or different messages.

[0102] The UE may activate / switch on / register an ML model, for example, by initiating a random access procedure or by sending an RRC message or a MAC CE message, and indicate such an action to the network node. When the payload size of the information is small, the above indication may be performed using the transmission of uplink control information (UCI) (for example, encoding such an action as part of the UCI and the UE sending it to the gNB), or it may be performed using the transmission of sidelink control information (SCI) (for example, encoding such an action as part of the SCI and the UE sending it to another UE).

[0103] In some embodiments, the above message is sent when the UE activates / switch on / registers at least one ML model associated with the ML-based features.

[0104] In some embodiments, the above message indicating its own ability to support a combination of at least one ML-based feature and at least one preliminary feature for the associated functionality is an RRC message, a MAC CE Msg1, MsgA, Msg3, a combination of Msg1 and Msg3, UCI, or SCI.

[0105] In some embodiments, the UE indication of its own ability to support a combination of at least one ML-based feature and at least one preliminary feature for the same functionality includes an indication that those ML and preliminary features can be executed simultaneously. Alternatively, the UE may indicate that they can be executed one by one instead of simultaneously.

[0106] Some embodiments include fallback of ML-based features assisted by a network node. For example, a UE may be instructed by the network to perform fallback / switching of ML-based features.

[0107] In some embodiments, when the UE reports its concurrent execution capability, in response to receiving the UE's capability information, the network node may send a first configuration message instructing the UE to concurrently execute / operate the ML-based feature and its associated preliminary feature. The outputs of both features (the ML-based feature and the preliminary feature) may be used by the network node to monitor or predict the performance of the ML-based feature using instantaneous or short-term comparison.

[0108] To provide the network node with concurrent feedback regarding the model and preliminary output, the UE may be configured with a special reporting mode and reserved reporting resources. For example, some related reporting procedure may be repeated once or twice for the features of the ML model and the preliminary feature. Alternatively, the reporting may be configured such that the ML-based and preliminary-based outputs are signaled to the network according to a predetermined pattern, such as an alternating pattern. The parallel execution of the preliminary operation may be initiated with a low duty cycle, such as 5% of the total operation time, for example, during which the ML feature may be initiated alone for most of the time.

[0109] In some embodiments, if the UE indicates a lack of ability for simultaneous execution, the UE may be configured to operate, for example, using ML and fallback features alternately, and performance monitoring may be performed by comparing the long-term performance of the two operating modes. The time durations of the ML-based and fallback activities may be asymmetric / unequal in such a way that the ML mode operates for most of the time if no performance issues are detected. In some embodiments, the network node may perform performance monitoring of the ML features based on a comparison with the reference performance of high-level KPIs (key performance indicators) such as typical TP detection, SINR, serving beam selection, etc.

[0110] In response to the detection / prediction of a performance impairment of at least one ML-based feature for the associated functionality, the network node transmits a second configuration message to the UE instructing the UE to deactivate / stop / switch off the ML-based feature for the associated functionality and to activate / switch on the fallback feature.

[0111] FIG. 2 is a flowchart showing an example of fallback of an ML-based feature assisted by a network node. A specific embodiment may include at least a portion of the following set of steps. The order of some steps may be interchanged and some steps may be optional.

[0112] Step 1: The UE (e.g., UE200, described in more detail below with respect to FIG. 4) transmits a message to the network node (e.g., network node 300, described in more detail below with respect to FIG. 4) indicating the UE's ability to support a combination of at least one ML-based feature and at least one fallback feature for a certain functionality. The network node receives from the UE a message indicating the UE's ability to support a combination of at least one ML-based feature and at least one fallback feature for a certain functionality.

[0113] Step 2 [Optional]: In response to receiving the UE's capability information, the network node transmits a first configuration message instructing the UE to simultaneously execute / operate at least the ML-based feature and its associated preliminary features for the associated functionality. In response to receiving the message, the UE simultaneously executes / operates the configured ML-based feature and preliminary feature for the associated functionality.

[0114] Step 3 [Optional]: The network node uses the output of the ML-based feature and the preliminary feature to monitor or predict the performance of the ML-based feature.

[0115] Step 4: The network node detects or predicts a performance impairment of at least one ML-based feature for the associated functionality.

[0116] Step 5: The network node transmits a second configuration message instructing the UE to deactivate / stop / switch off the ML-based feature with performance issues and activate / switch on the associated preliminary feature. In response to the second message, the UE deactivates / stop / switch off the indicated ML-based feature with performance issues and activates / switch on the associated preliminary feature.

[0117] Step 5 [Optional]: For cases where the ML-based feature is split into two parts, for example, in a form where one part is located in the UE and the other part is located in the network, or for cases based on multiple ML models in a form where part of the model is located in the UE and the rest of the model is located in the network, the network node deactivates / stop / switch off the associated ML model on the network side for at least the ML-based feature to be deactivated.

[0118] Step 6 [Optional]: The network node transmits to the UE an adjusted configuration and / or a scheduling message. For example, the adjusted configuration and / or the scheduling message may include an updated reference signal resource configuration for UE measurements and / or an updated CSI reporting configuration for the UE to report CSI using a preliminary feature.

[0119] Some embodiments include a fallback of UE-autonomous ML-based features, and the UE reports its actions to the network node. For some use cases involving more advanced UEs, the UE may monitor the ML model performance of one or more ML-based features and detect / predict issues with the ML-based features on its own. Some embodiments are such that the UE autonomously performs a fallback / switch of the ML-based features and indicates that information to the network node.

[0120] Figure 3 is a flowchart showing an example of a fallback of UE-autonomous ML-based features and reporting of its actions to the network node. Specific embodiments may include at least a portion of the following group of steps. The order of some steps may be interchanged, and some steps may be optional.

[0121] Step 1: The UE (e.g., UE 200, described in more detail below with respect to FIG. 4) transmits to the network node (e.g., network node 300, described in more detail below with respect to FIG. 4) a message indicating the UE's capabilities to support a combination of at least one ML-based feature and at least one preliminary feature for a certain functionality. The network node receives from the UE a message indicating the UE's capabilities to support a combination of at least one ML-based feature and at least one preliminary feature for a certain functionality.

[0122] Step 2: The network configures at least one ML-based feature.

[0123] Step 3: The UE monitors the ML model performance of one or more ML-based features.

[0124] Step 4: The UE detects or predicts a performance impairment of at least one ML-based feature with respect to the associated functionality. This is similar to the network-side performance monitoring approach. In some embodiments, when the UE has the ability to execute simultaneously, it may execute / operate the ML-based feature and its associated preliminary features simultaneously. The UE may use the outputs of both features (the ML-based feature and the preliminary feature) for monitoring or predicting the performance of the ML-based feature using instantaneous or short-term comparisons. The parallel operation may be initiated with a low duty cycle, such as 5% of the total operation time, for example. Simultaneous execution does not necessarily mean that it is carried out exactly simultaneously. Rather, it means that the same or very similar input data is used for both functions so that the performance can be compared between the ML-based feature and the preliminary feature. Thus, the original data may need to be acquired simultaneously or at very close time positions.

[0125] For example, consider a use case where the functionality is related to the construction of CSI reports. The same CSI-RS resource set / index may act as source data for both the ML-based and preliminary features, although the actual data processing processes for both features do not need to occur simultaneously. Rather, they can be spread out in time for subsequent comparison of the results or reporting to the gNB.

[0126] In some embodiments, if the UE lacks the ability to execute simultaneously, the UE may operate using ML and the fallback features alternately, and the performance monitoring may be performed by comparing the long-term periods of the two operating modes. The time durations of the ML-based and fallback activities may be asymmetric / unequal in such a way that the ML mode is operated for most of the time if no performance issues are detected. In some embodiments, the UE may perform performance monitoring of the ML features based on a comparison with the reference performance of high-level KPIs such as, for example, typical TP detection, SINR, serving beam quality, etc.

[0127] Step 5: The UE autonomously deactivates / halts at least the detected ML-based features having performance issues and activates / switches on the associated fallback features.

[0128] Step 6: The UE indicates to the network node about the fallback / switching information of the above features (e.g., the deactivated / stopped / switched-off ML-based features).

[0129] Step 7 [Optional]: The network node deactivates / halts / switches off the associated ML model on the network side for at least the deactivated ML-based features, for example, in cases based on an ML model that is split into two parts such that one part is located at the UE and the other part is located at the network, or in cases based on multiple ML models such that a part of the model is located at the UE and the rest of the model is located at the network.

[0130] Step 8 [Optional]: The network node transmits an adjusted configuration and / or a scheduling message to the UE. For example, the adjusted configuration and / or the scheduling message may include an updated reference signal resource configuration for the UE's measurement and / or an updated CSI reporting configuration for the UE to report CSI using the fallback features.

[0131] Although the examples described herein mainly focus on the reporting of UE capabilities for the Uu interface, the same methodology may be applied to support fallback of ML-based features using signaling on the PC5 interface between different UEs.

[0132] FIG. 4 shows an example of a communication system 100 according to some embodiments. In this example, the communication system 100 includes a telecommunication network 102 that includes an access network 104, such as a radio access network (RAN), and a core network 106 that includes one or more core network nodes 108. The access network 104 includes one or more access network nodes, such as network nodes 110a and 110b (one or more of which may generally be referred to as network node 110), or some other similar 3rd Generation Partnership Project (3GPP) access node or non-3GPP access point. The network node 110 facilitates direct or indirect connection of user equipment (UE), such as by connecting UEs 112a, 112b, 112c, and 112d (one or more of which may generally be referred to as UE 112) to the core network 106 over one or more wireless connections.

[0133] Exemplary wireless communication over a wireless connection includes transmitting and / or receiving wireless signals using electromagnetic waves, radio waves, infrared waves, and / or other types of signals suitable for carrying information without using wires, cables, or other physical conductors. Moreover, in various embodiments, the communication system 100 may include any number of wired or wireless networks, network nodes, UEs, and / or any other components or systems that can facilitate or participate in the communication of data and / or signals, whether over a wired or wireless connection. The communication system 100 may include any type of communication, telecommunications, data, cellular, wireless network, and / or other similar types of systems, and / or interface with them.

[0134] UE112 may be any of a wide variety of communication devices, including wireless devices arranged, configured, and / or operable to communicate wirelessly with network node 110 and other communication devices. Similarly, network node 110 is arranged, enabled, and / or operable to communicate directly or indirectly with UE112 and / or other network nodes or devices within the telecommunications network 102 to enable and / or provide network access, such as wireless network access, and / or to perform other functions, such as management within the telecommunications network 102.

[0135] In the illustrated example, the core network 106 connects network nodes 110 to one or more hosts such as host 116. Those connections may be direct or may be indirect via one or more intermediate networks or devices. In other examples, network nodes may be directly connected to hosts. The core network 106 includes one or more core network nodes (e.g., core network node 108) structured with hardware and software components. The functions of those components may be substantially similar to those described with respect to the UE, network nodes, and / or hosts, and thus those descriptions are generally applicable to the corresponding components of core network node 108. Exemplary core network nodes include one or more of a mobile switching center (MSC), a mobility management entity (MME), a home subscriber server (HSS), an access and mobility management function (AMF), a session management function (SMF), an authentication server function (AUSF), a subscription identifier deconcealment function (SIDF), a unified data management (UDM), a security edge protection proxy (SEPP), a network exposure function (NEF), and / or a user plane function (UPF).

[0136] Host 116 may be under the ownership or control of a service provider other than the operator, or of the provider of access network 104 and / or telecommunications network 102, and may be operated by or for that service provider. Host 116 may host various applications to provide one or more services. Examples of such applications include live and pre-recorded audio / video content, data collection services such as the acquisition and editing of data regarding various ambient conditions sensed by multiple UEs, analytical functionality, social media, functions for the control or otherwise interaction with remote devices, functions for alarm and monitoring centers, or any other such functions performed by a server.

[0137] Overall, the communication system 100 of FIG. 4 enables connectivity between UEs, network nodes, and hosts. In that sense, the communication system may be configured to operate according to predefined rules or procedures, such as certain standards including, but not limited to: GSM (Global System for Mobile Communications), UMTS (Universal Mobile Telecommunications System), Long-Term Evolution (LTE), and / or other suitable 2G, 3G, 4G, 5G standards, or any applicable future generation standards (e.g., 6G), WLAN (wireless local area network) standards such as IEEE (Institute of Electrical and Electronics Engineers) 802.11 standards (WiFi), and / or any other suitable wireless communication standards such as WiMax (Worldwide Interoperability for Microwave Access), Bluetooth, Z-Wave, NFC (Near Field Communication) ZigBee, LiFi, and / or any LPWAN (low-power wide-area network) standards such as LoRa and Sigfox.

[0138] In some examples, the telecommunications network 102 is a cellular network implementing functions standardized by 3GPP. Thus, the telecommunications network 102 may support network slicing to provide various logical networks to the various devices connected to the telecommunications network 102. For example, the telecommunications network 102 may provide ultra-reliable low-latency communication (URLLC) services to some UEs while providing enhanced mobile broadband (eMBB) services to other UEs, and may further provide massive machine type communication (mMTC) / massive IoT services to additional UEs.

[0139] In some examples, UE 112 is configured to send and / or receive information without direct human interaction. For example, the UE may be designed to send information to access network 104 at a predetermined schedule, when triggered by an internal or external event, or in response to a request from access network 104. Additionally, the UE may be configured to operate in single or multi-RAT, or in multi-standard mode. For example, the UE may be configured and operate in any one or combination of Wi-Fi, NR (New Radio), and LTE, i.e., for multi-radio dual connectivity (MR-DC) such as E-UTRAN (Evolved-UMTS Terrestrial Radio Access Network) New Radio - Dual Connectivity (EN-DC).

[0140] In the above example, the hub 114 communicates with the access network 104 to facilitate indirect communication between one or more UEs (e.g., UEs 112c and / or 112d) and a network node (e.g., network node 110b). In some examples, the hub 114 may be any of a controller, a router, a content source and analytics, or other communication devices described herein with respect to the UE. For example, the hub 114 may be a broadband router that enables access to the core network 106 for the UE. As another example, the hub 114 may be a controller that sends commands or instructions to one or more actuators within the UE. The commands or instructions may be received from the UE or the network node 110, or may be received by executable code, scripts, processes, or other instructions within the hub 114. As another example, the hub 114 may be a data controller that operates as temporary storage for the data of the UE, and in some embodiments, may perform analysis or other processing of that data. As another example, the hub 114 may be a content source. For example, for a UE that is a VR headset, a display, a loudspeaker, or other media delivery device, the hub 114 may obtain media or data related to VR assets, video, audio, or other sensory information via the network node, and in that case, the hub 114 provides it to the UE either directly, after local processing is performed, and / or after additional local content is added. In yet another example, the hub 114 operates as a proxy server or orchestrator for the UE, especially if one or more of the UEs are low-energy IoT devices.

[0141] The hub 114 may have a steady / permanent or intermittent connection to the network node 110b. Also, the hub 114 may enable different communication methods and / or schedules between the hub 114 and the UEs (UE112c and / or 112d), as well as between the hub 114 and the core network 106. In other examples, the hub 114 is connected to the core network 106 and / or one or more UEs via a wired connection. Moreover, the hub 114 may be configured to be connected to an M2M service provider on the access network 104 and / or to other UEs on a direct connection. In some scenarios, the UEs may establish a wireless connection with the network node 110 while still being connected via the hub 114 via a wired or wireless connection. In some embodiments, the hub 114 may be a dedicated hub, i.e., a hub whose main function is to route communications between the UEs and the network node 110b. In other embodiments, the hub 114 may be a non-dedicated hub, i.e., a device that is operable to route communications between the UEs and the network node 110b, but in addition is operable as a starting point and / or an end point of communication for some data channel.

[0142] FIG. 5 shows UE 200 according to some embodiments. As used herein, a UE is a device that can communicate wirelessly with a network node and / or another UE, and is configured, arranged, and / or operable to do so. Examples of UEs include, but are not limited to, smartphones, mobile phones, cell phones, VoIP (Voice over IP) phones, wireless local loop phones, desktop computers, personal digital assistants (PDAs), wireless cameras, game consoles or devices, music storage devices, playback appliances, wearable terminal devices, wireless endpoints, mobile stations, tablets, laptops, laptop embedded equipment (LEE), laptop-mounted equipment (LME), smart devices, wireless customer premise equipment (CPE), vehicle-mounted or vehicle-embedded / integrated wireless devices, etc. Other examples include any UE identified by the Third Generation Partnership Project (3GPP), including narrowband Internet of Things (NB-IoT) UEs, machine type communication (MTC) UEs, and / or enhanced MTC (eMTC) UEs.

[0143] The UE may support device-to-device (D2D) communication, for example, by implementing 3GPP standards for sidelink communication, dedicated short range communication (DSRC), vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), or vehicle-to-everything (V2E). In other examples, the UE may not necessarily have a user in the sense of a human user who owns and / or operates the associated device. Instead, the UE may represent a device (e.g., a smart sprinkler controller) that is intended for sale to or operation by a human user but is not initially associated with a particular human user. Alternatively, the UE may represent a device (e.g., a smart power meter) that is not intended for sale to or operation by an end user and may be associated with or operated for the benefit of a user.

[0144] UE 200 includes a processing circuit 202, a power supply 208, a memory 210, a communication interface 212, and / or any other component, or any combination thereof, operably coupled to an input / output interface 206 via a bus 204. A UE may utilize all or a subset of the components shown in FIG. 5. The level of integration between components may vary between one UE and another. Further, a UE may include multiple instances of components, such as multiple processors, memories, transceivers, transmitters, receivers, and the like.

[0145] The processing circuit 202 is configured to process instruction sets and data and may be configured to implement some sequential state machine operable to execute instruction sets stored as a machine-readable computer program within the memory 210. The processing circuit 202 may be implemented as one or more hardware-implemented state machines (e.g., in discrete logic, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), etc.), programmable logic with appropriate firmware, one or more stored computer programs, a general-purpose processor such as a microprocessor or a digital signal processor (DSP) with appropriate software, or any combination of the above. For example, the processing circuit 202 may include multiple central processing units (CPUs).

[0146] In the above example, the input / output interface 206 may be configured to provide an interface to an input device, an output device, or one or more input / output devices. Examples of output devices include speakers, sound cards, video cards, displays, monitors, printers, actuators, emitters, smart cards, other output devices, or any combination thereof. The input device may enable a user to capture information for the UE 200. Examples of input devices include touch-sensitive or presence-sensitive displays, cameras (e.g., digital cameras, digital video cameras, webcams, etc.), microphones, sensors, mice, trackballs, directional pads, trackpads, scroll wheels, and smart cards. A presence-sensitive display may include a capacitive or resistive touch sensor for sensing input from a user. The sensor may be, for example, an accelerometer, gyroscope, tilt sensor, force sensor, magnetometer, light sensor, proximity sensor, biometric sensor, etc., or any combination thereof. The output device may use the same type of interface port as the input device. For example, a Universal Serial Bus (USB) port may be used to provide the input device and the output device.

[0147] In some embodiments, the power supply 208 is structured as a battery or battery pack. Other types of power supplies such as an external power supply (e.g., an electrical outlet), a solar power generation device, or a battery may also be used. The power supply 208 may further include a power circuit for transmitting power from the power supply 208 itself and / or an external power supply to various parts of the UE 200 via an interface such as an input circuit or a power cable. The transmission of power may be, for example, for charging the power supply 208. The power circuit may perform some shaping, conversion, or other modification to the power from the power supply 208 to make it suitable for each component of the UE 200 that is the power recipient.

[0148] The memory 210 may be, for example, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic disk, an optical disk, a hard disk, a removable cartridge, and a flash drive, etc., or may be configured to include such memories. In one example, the memory 210 includes one or more application programs 214, such as an operating system, a web browser application, a widget, a gadget engine or other applications, and corresponding data 216. The memory 210 may store any of a wide variety of operating systems or combinations of multiple operating systems for use by the UE 200.

[0149] The memory 210 may be configured to include one or more physical drive units such as RAID (Redundant Array of Independent Disks), flash memory, USB flash drives, external hard disk drives, thumb drives, pen drives, key drives, HD-DVD (High-Density Digital Versatile Disc), optical disc drives, internal hard disk drives, Blu-Ray optical disc drives, HDDS (Holographic Digital Data Storage) optical disc drives, external mini DIMM (Dual In-Line Memory Module), SDRAM (Synchronous Dynamic Random Access Memory), external micro DIMM SDRAM, a UICC (universal integrated circuit card) in the form of a smart card memory including one or more SIMs (subscriber Identity Module) such as USIM and / or ISIM, other memories, or any combination thereof. The UICC may be, for example, an embedded UICC (eUICC), an integrated UICC (iUICC), or a removable UICC commonly known as a "SIM card". The memory 210 may enable the UE200 to access instruction groups and application programs stored in a temporary or non-temporary storage medium to offload or upload data. Items of a product such as those using a communication system may be embodied tangibly as or within the memory 210 which may be or include a device-readable storage medium.

[0150] The processing circuit 202 may be configured to communicate with an access network or other network using the communication interface 212. The communication interface 212 may include one or more communication subsystems, may include the antenna 222, or may be communicatively coupled to the antenna 222. The communication interface 212 may include one or more transceivers used to perform communication, such as by communicating with one or more remote transceivers of other wirelessly communicable devices (e.g., other UEs or network nodes within the access network). Each transceiver may include a transmitter 218 and / or a receiver 220 appropriate for providing network communication (e.g., optical, electrical, frequency allocation, etc.). Moreover, the transmitter 218 and the receiver 220 may be coupled to one or more antennas (e.g., antenna 222), they may share circuit components, software, or firmware, or alternatively may be implemented separately.

[0151] In the illustrated embodiment, the communication functions of the communication interface 212 may include cellular communication, Wi-Fi communication, LPWAN communication, data communication, voice communication, multimedia communication, short-range communication such as Bluetooth, near-field communication, location-based communication such as the use of GPS (Global Positioning System) for location determination, other similar communication functions, or any combination thereof. The communication may be implemented according to one or more communication protocols and / or standards such as IEEE802.11, code division multiple access (CDMA), wideband code division multiple access (WCDMA), GSM, LTE, new radio (NR), UMTS, WiMax, Ethernet, TCP / IP (transmission control protocol / internet protocol), SONET (synchronous optical networking), ATM (Asynchronous Transfer Mode), QUIC, HTTP (Hypertext Transfer Protocol), etc.

[0152] Regardless of the type of sensor, the UE may provide an output of data captured by its sensors to a network node via a wireless connection through its communication interface 212. The data captured by the UE's sensors may be communicated to the network node via the wireless connection through other UEs. The output may be periodic (e.g., every 15 minutes when reporting the measured temperature), random (e.g., to balance the load from reports from multiple sensors), made in response to a trigger event (e.g., an alert is sent when moisture is detected), made in response to a request (e.g., a request initiated by the user), or a continuous stream (e.g., a live video feed of a patient).

[0153] As another example, the UE includes an actuator, a motor, or a switch associated with a communication interface configured to receive a wireless input from a network node via a wireless connection. The state of the actuator, motor, or switch may change in response to the received wireless input. For example, the UE may include a motor that adjusts the control surfaces or rotors of a drone in flight according to the received input, or a robotic arm that performs a medical procedure according to the received input.

[0154] When the UE is in the form of an Internet of Things (IoT) device, it may be a device for use in one or more application domains, which may include, without limitation, wearable technology in the street, extended industrial applications, and healthcare. Non-limiting examples of such IoT devices include connected refrigerators or freezers, TVs, connected lighting devices, electricity meters, robotic vacuum cleaners, voice-controlled smart speakers, home security cameras, motion detectors, thermostats, smoke detectors, door / window sensors, flood / moisture sensors, electric door locks, connected doorbells, air conditioning systems such as heat pumps, autonomous vehicles, monitoring systems, climate monitoring devices, parking monitoring devices, vehicle charging stations, smartwatches, fitness trackers, head-mounted displays for augmented reality (AR) or virtual reality (VR), wearables for tactile or sensory augmentation, water sprinklers, animal or item tracking devices, sensors for monitoring plants or animals, industrial robots, unmanned aerial vehicles (UAVs), and any type of medical device such as a heart rate monitor or a remotely controlled surgical robot, or a device incorporated therein. The UE in the form of an IoT device includes other components as described in connection with UE200 shown in FIG. 5, in addition to circuits and / or software that depend on the application for which the IoT device is intended.

[0155] As yet another specific example, in an IoT scenario, the UE may represent a machine or other device that performs monitoring and / or measurement and transmits the results of such monitoring and / or measurement to other UEs and / or network nodes. The UE may be an M2M device in this case and may be referred to as an MTC device in the context of 3GPP. As one specific example, the UE may implement the 3GPP NB-IoT standard. In other scenarios, the UE may represent a passenger vehicle, bus, truck, ship or aircraft, or other equipment that is capable of monitoring and / or reporting on its operating status or other functions associated with its operation.

[0156] In fact, any number of UEs may be used together for a single use case. For example, a first UE may be a drone or integrated into a drone and provide speed information of the drone (obtained through a speed sensor) to a second UE that is a remote controller for operating the drone. When the user makes a change from the remote controller, the first UE may adjust the throttle of the drone (e.g., by controlling an actuator) to increase or decrease the speed of the drone. The first and / or second UE may also include more than one of the above-described functionality. For example, the UE may include sensors and actuators and handle communication of data for both the speed sensor and the actuator.

[0157] FIG. 6 shows a network node 300 according to some embodiments. As used herein, a network node refers to a device that is capable of communicating directly or indirectly with a UE and / or other network nodes or devices within a telecommunication network and is configured, arranged, and / or operable as such. Examples of network nodes include, but are not limited to, access points (APs) (e.g., wireless access points) and base stations (BSs) (e.g., wireless base stations, Node B, evolved Node B (eNB), and NR Node B (gNB)).

[0158] Base stations may be categorized based on the amount of coverage they provide (or, put another way, their transmit power levels), and thus may be referred to as femto base stations, pico base stations, micro base stations, or macro base stations, depending on the amount of coverage provided. A base station may also be a relay donor node that controls relay nodes or repeaters. A network node may include one or more (or all) parts of a distributed radio base station, such as a centralized digital unit and / or a remote radio unit (RRU), which may also be referred to as a remote radio head (RRH). Such remote radio units may or may not be integrated with an antenna, such as an antenna-integrated radio. A part of a distributed radio base station may also be referred to as a node within a distributed antenna system (DAS).

[0159] Other examples of network nodes include multi-transmission point (multi-TRP) 5G access nodes, multi-standard radio (MSR) devices such as MSR BS, network controllers such as radio network controllers (RNC) or base station controllers (BSC), base transceiver stations (BTS), transmission points, transmission nodes, multi-cell / multicast coordination entities (MCE), operation and maintenance (O&M) nodes, operation support system (OSS) nodes, self-organizing network (SON) nodes, positioning nodes (e.g., including evolved serving mobile location center (E-SMLC) and / or drive test minimization (MDT)).

[0160] The network node 300 includes a processing circuit 302, a memory 304, a communication interface 306, and a power supply 308. The network node 300 may be composed of a plurality of physically distinct components (e.g., a Node B component and an RNC component, or a BTS component and a BSC component, etc.) each of which may have its own respective components. In a scenario where the network node 300 comprises a plurality of distinct components (e.g., BTS and BSC components), one or more of those distinct components may be shared among several network nodes. For example, a single RNC may control a plurality of Node Bs. In such scenarios, each unique pair of Node B and RNC may, in some instances, be regarded as a single distinct network node. In some embodiments, the network node 300 may be configured to support a plurality of radio access technologies (RATs). In such embodiments, some components may be made redundant (e.g., separate memories 304 for different RATs), and some components may be reused (e.g., the same antenna 310 may be shared by a plurality of different RATs). Also, the network node 300 may include a plurality of sets of various exemplary components for various wireless technologies integrated into the network node 300, such as, for example, GSM, WCDMA, LTE, NR, WiFi, Zigbee, Z-wave, LoRaWAN, RFID (Radio Frequency Identification), or Bluetooth wireless technologies. Those wireless technologies may be integrated into the same or different chips or sets of chips and other components within the network node 300.

[0161] The processing circuit 302 can include one or more combinations of a microprocessor, a controller, a microcontroller, a central processing unit, a digital signal processor, an application specific integrated circuit, a field programmable gate array, or other suitable computing device, resource, or hardware, software, and / or encoded logic, which can operate alone or in cooperation with components of other network nodes 300 such as the memory 304 to provide the functionality of the network node 300.

[0162] In some embodiments, the processing circuit 302 includes a system on chip (SOC). In some embodiments, the processing circuit 302 includes one or more of a radio frequency (RF) transceiver circuit 312 and a baseband processing circuit 314. In some embodiments, the radio frequency (RF) transceiver circuit 312 and the baseband processing circuit 314 may be on separate chips (or a set of chips), substrates, or units, such as a radio unit and a digital unit. In alternative embodiments, some or all of the RF transceiver circuit 312 and the baseband processing circuit 314 may be on the same chip or a set of chips, substrate, or unit.

[0163] Memory 304 includes, without limitation, any form of volatile or non-volatile computer-readable memory, such as persistent storage, solid-state memory, remotely mounted memory, magnetic media, optical media, random access memory (RAM), read-only memory (ROM), mass storage media (e.g., hard disk), removable storage media (e.g., flash drive, compact disc (CD) or digital video disc (DVD)), and / or any other suitable volatile or non-volatile non-transitory device-readable and / or computer-executable memory device that can store information, data, and / or instructions that can be used by processing circuit 302. Memory 304 can store a computer program, software, logic, rules, code, table, or other suitable instructions, data, or information, including one or more of the above, that is executable by processing circuit 302 and available for use by network node 300. Memory 304 may be used to store any calculation results generated by processing circuit 302 and / or any data received via interface 306. In some embodiments, processing circuit 302 and memory 304 are integrated.

[0164] The communication interface 306 is used for wired or wireless communication of signaling and / or data between network nodes, access networks, and / or UEs. As illustrated, the communication interface 306 includes, for example, ports / terminals 316 for transmitting and receiving data to and from a network over a wired connection. The communication interface 306 also includes a radio front-end circuit 318 that may be connected to the antenna 310 or in some embodiments be part of the antenna 310. The radio front-end circuit 318 includes a filter 320 and an amplifier 322. The radio front-end circuit 318 may be connected to the antenna 310 and the processing circuit 302. The radio front-end circuit may be configured to condition signals communicated between the antenna 310 and the processing circuit 302. The radio front-end circuit 318 may receive digital data to be transmitted to other network nodes or UEs via a wireless connection. The radio front-end circuit 318 may convert the digital data into a radio signal having appropriate channel and bandwidth parameters using a combination of the filter 320 and / or the amplifier 322. The radio signal may then be transmitted via the antenna 310. Similarly, when data is received, the antenna 310 may collect the radio signal, which may then be converted into digital data by the radio front-end circuit 318. The digital data may be passed to the processing circuit 302. In other embodiments, the communication interface may include different components and / or different combinations of components.

[0165] In some alternative embodiments, network node 300 may not include a separate radio front-end circuit 318. Instead, processing circuit 302 may include a radio front-end circuit and may be connected to antenna 310. Similarly, in some embodiments, all or some of RF transceiver circuit 312 is part of communication interface 306. In yet another embodiment, communication interface 306 includes one or more ports or terminals 316, radio front-end circuit 318, and RF transceiver circuit 312 as part of a wireless unit (not shown), and communication interface 306 communicates with baseband processing circuit 314, which is part of a digital unit (not shown).

[0166] Antenna 310 may include one or more antennas or antenna arrays configured to transmit and / or receive wireless signals. Antenna 310 may be coupled to radio front-end circuit 318 and may be any type of antenna capable of wirelessly transmitting and receiving data and / or signals. In some embodiments, antenna 310 is separate from network node 300 and can be connected to network node 300 through an interface or port.

[0167] Antenna 310, communication interface 306, and / or processing circuit 302 may be configured to perform any of the receiving operations and / or certain acquisition operations described herein as being performed by a network node. Any information, data, and / or signals may be received from a UE, other network nodes, and / or any other network device. Similarly, antenna 310, communication interface 306, and / or processing circuit 302 may be configured to perform any of the transmitting operations described herein as being performed by a network node. Any information, data, and / or signals may be transmitted to a UE, other network nodes, and / or any other network device.

[0168] Power supply 308 provides power to the various components of network node 300 in a form suitable for each component (e.g., at the voltage and current levels required for each respective component). The power supply 308 may include, or be coupled to, a power management circuit for supplying power to the components of network node 300 for performing the functionality described herein. For example, network node 300 may be connectable to an external power source (e.g., a power grid, an electrical outlet) via an input circuit or interface such as an electrical cable, whereby the external power source supplies power to the power circuit of power supply 308. As a further example, power supply 308 may include a source of power in the form of a battery or battery pack that is connected to or integrated into the power circuit. The battery may provide backup power in the event of a failure of the external power source.

[0169] Embodiments of network node 300 may include additional components other than those shown in FIG. 6 to provide a functional perspective of the network node that includes any of the functionality described herein and / or any functionality necessary to support the subject matter described herein. For example, network node 300 may include a user interface device that enables input of information to network node 300 and output of information from network node 300. This may enable a user to perform diagnostic, maintenance, repair, and other administrative functions with respect to network node 300.

[0170] FIG. 7 is a block diagram of a host 400 that can be an embodiment of the host 116 of FIG. 4 according to the various aspects described herein. As used herein, the host 400 can be hardware and / or software in various combinations, including a stand-alone server, a blade server, a cloud-implemented server, a distributed server, a virtual machine, a container, or processing resources within a server farm, or can include the same. The host 400 can provide one or more services to one or more UEs.

[0171] The host 400 includes a processing circuit 402, a network interface 408, a power supply 410, and a memory 412 that are operably coupled via a bus 404 to an input / output interface 406. In other embodiments, other components may be included. The functions of those components may be substantially similar to those described with respect to the devices in the previous figures such as FIGS. 3 and 4, and thus the descriptions thereof are generally applicable to the corresponding components of the host 400.

[0172] Memory 412 may include one or more computer programs including one or more host application programs 414, and data 416 that may include user data such as data generated by the UE for the host 400 or data generated by the host 400 for the UE. Embodiments of the host 400 may utilize only a subset or all of the illustrated components. The host application program 414 may be implemented in a container-based architecture and may provide support for video codecs (VVC (Versatile Video Coding), HEVC (High Efficiency Video Coding), AVC (Advanced Video Coding), MPEG, VP9) and audio codecs (e.g., FLAC, AAC (Advanced Audio Coding), MPEG, G.711) including transcoding for different classes, types or implementations of multiple UEs (e.g., handsets, desktop computers, wearable display systems, head-up display systems). Also, the host application program 414 may provide user authentication and license checking and may periodically report health, route and content availability to a central node such as a device within or at the edge of the core network. Thus, the host 400 may select and / or indicate different hosts for over-the-top services for the UE. The host application program 414 may support various protocols such as the HLS (HTTP Live Streaming) protocol, RTMP (Real-Time Messaging Protocol), RTSP (Real-Time Streaming Protocol), MPEG-DASH (Dynamic Adaptive Streaming over HTTP).

[0173] FIG. 8 is a block diagram showing a virtualization environment 500 in which functions implemented according to some embodiments can be virtualized. In the present context, the virtualization means for generating a virtual version of a device or apparatus may include a virtualization hardware platform, a storage device, and networking resources. As used herein, virtualization can be applied to any device or their components described herein, and is related to an implementation example in which at least a part of its functionality is implemented as one or more virtual components. Some or all of the functions described herein are implemented as virtual components executed by one or more virtual machines (VMs) implemented within one or more virtual environments 500 hosted by one or more hardware nodes such as a hardware computing device operating as a network node, a UE, a core network node, or a host. Further, in embodiments where the virtual node does not require wireless connectivity (e.g., a core network node or a host), the node may be virtualized as a whole.

[0174] An application 502 (alternatively, may be referred to as a software instance, a virtual appliance, a network function, a virtual node, a virtual network function, etc.) operates in a virtualization environment 500 for implementing some of the features, functions, and / or benefits of some of the embodiments disclosed herein.

[0175] Hardware 504 includes a processing circuit, a memory storing software and / or a set of instructions executable by the processing circuit which is hardware, and / or hardware devices as described herein such as a network interface and an input / output interface. The software is executed by the processing circuit to instantiate one or more virtualization layers 506 (also referred to as a hypervisor or a virtual machine monitor (VMM)), provide VMs 508a and VM508b (one or more of which may be collectively referred to as VM508), and / or execute any of the functions, features, and / or benefits described in relation to several embodiments described herein. The virtualization layer 506 may present a virtual operating platform that appears to the virtual machines 508 as networking hardware.

[0176] VMs 508 include virtual processing, virtual memory, virtual networking or interfaces, and virtual storage, and may be executed by the corresponding virtualization layer 506. Various embodiments of instances of the virtual appliance 502 may be implemented in one or more of the VMs 508, and the implementation may be made in various ways. Virtualization of hardware is referred to in some contexts as network function virtualization (NFV). NFV can be used to consolidate many types of network devices into industry-standard high-volume server hardware, physical switches, and physical storage that can be located within data centers and customer premise equipment.

[0177] In the context of NFV, VM508 may be a software implementation of a physical machine that runs a program as if it were running on a physical, non-virtualized machine. Each of the VM508, and the portion of the hardware 504 that executes the VM, forms a separate virtual network element, whether it is hardware dedicated to the VM and / or hardware shared with other VMs by the VM. Also in the context of NFV, the virtual network function is responsible for handling the native network functions running in one or more VMs508 at the top level of the hardware 504 and corresponds to the application 502.

[0178] The hardware 504 may be implemented in a stand-alone network node with general or proprietary components. The hardware 504 may implement some functions via virtualization. Alternatively, the hardware 504 may be part of a larger class of hardware (such as those within a data center or CPE) where multiple hardware nodes cooperate and are managed via management and orchestration 510, which oversees, among other things, the lifecycle management of the application 502. In some embodiments, the hardware 504 is coupled to one or more radio units, each including one or more transmitters and one or more receivers, which may be coupled to one or more antennas. The radio unit may communicate directly with other hardware nodes via one or more appropriate network interfaces and may be used in combination with virtual components to provide radio capabilities to virtual nodes, such as radio access nodes or base stations. In some embodiments, some signaling can be provided in conjunction with the use of the control system 512, which may alternatively be used for communication between the hardware node and the radio unit.

[0179] Figure 9 shows a communication diagram of host computer 602 communicating with UE 606 via network node 604 over a partially wireless connection according to some embodiments. Exemplary implementations according to various embodiments of the UE (UE 112a of FIG. 4 and / or UE 200 of FIG. 5), network node (network node 110a of FIG. 4 and / or network node 300 of FIG. 6), and host (host 116 of FIG. 4 and / or host 400 of FIG. 7) discussed in the paragraphs thus far will be described with reference to FIG. 9 hereinafter.

[0180] Similar to host 400, embodiments of host 602 include hardware such as a communication interface, processing circuitry, and memory. Host 602 further includes software stored within host 602 or accessible by host 602, the software being executable by the processing circuitry. The software may include a host application operable to provide services to remote users such as UE 606 connected via an over-the-top (OTT) connection 650 extending between UE 606 and host computer 602. During provision of services to a remote user, the host application may provide user data transmitted using OTT connection 650.

[0181] Network node 604 includes hardware that enables communication with host 602 and UE 606. Connection 660 is direct or may pass through one or more other intermediate networks such as a core network (such as core network 106 of FIG. 4) and / or one or more public, private, or hosted networks. For example, the intermediate network may be a backbone network or the Internet.

[0182] UE606 includes software stored within or accessible by UE606, including such software executable by the processing circuitry of the UE. The software can include client applications, such as a web browser or a carrier-specific "app", that are operable to provide services to human or non-human users via UE606, with the support of host 602. At host 602, the host application to be executed can communicate with the client application to be executed via OTT connection 650 that terminates at UE606 and host 602. During service provision to the user, the client application of the UE can receive request data from the host application of the host and provide user data as a response to the request data. OTT connection 650 can transfer both request data and user data. The client application of the UE can interact with the user to generate user data that the client application provides to the host application through OTT connection 650.

[0183] OTT connection 650 extends via connection 660 between host 602 and network node 604 and via wireless connection 670 between network node 604 and UE606 to provide a connection between host 602 and UE606. Connection 660 and wireless connection 670 through which OTT connection 650 can be provided are abstractly depicted to illustrate communication via network node 604 between host 602 and UE606 without any explicit reference to any intermediate devices and the exact routing of messages through those devices.

[0184] As an example of transmitting data via the OTT connection 650, at step 608, the host 602 provides user data, which can be done by executing a host application. In some embodiments, the user data is associated with a particular human user interacting with the UE 606. In other embodiments, the user data is associated with the UE 606 sharing data with the host 602 without explicit human interaction. At step 610, the host 602 begins transmitting the user data to the UE 606 that will carry it. The host 602 may begin the transmission in response to a request transmitted by the UE 606. The request may be caused by a human interaction with the UE 606 or by the operation of a client application running on the UE 606. The transmission may pass through the network node 604 in accordance with the teachings of the embodiments described throughout this disclosure. Accordingly, at step 612, the network node 604 transmits the user data carried in the transmission initiated by the host 602 to the UE 606 in accordance with the teachings of the embodiments described throughout this disclosure. At step 614, the UE 606 receives the user data carried in the transmission, which can be done by a client application running on the UE 606 that is associated with a host application executed by the host 602.

[0185] In some examples, UE 606 executes a client application, thereby providing user data to host 602. The user data may be provided in reaction or response to receiving data from host 602. Accordingly, at step 616, UE 606 may provide user data, which may be done by executing a client application. During the provision of the user data, the client application may further consider user input received from the user via the input / output interface of UE 606. Regardless of the specific manner in which the user data is provided, at step 618, UE 606 initiates transmission of the user data to host 602 via network node 604. At step 620, in accordance with the teachings of the embodiments described throughout this disclosure, network node 604 receives the user data from UE 606 and initiates transmission of the received user data to host 602. At step 622, host 602 receives the user data carried in the above transmission initiated by UE 606.

[0186] One or more of the various embodiments improve the performance of the OTT service provided to UE 606 using OTT connection 650, and the wireless connection 670 forms its last segment. More precisely, the teachings of these embodiments may provide benefits such as reducing the latency of directly activating the SCell by RRC and the power consumption of the user equipment, thereby reducing the user's waiting time and increasing the battery life.

[0187] In an exemplary scenario, the host 602 may collect and analyze the status information of a factory. As another example, the host 602 may process audio and video data obtained from the UE for use in generating a map. As another example, the host 602 may collect and analyze real-time data to assist in controlling vehicle congestion (e.g., traffic signal control). As another example, the host 602 may store surveillance videos uploaded by the UE. As another example, the host 602 may perform storage or access control for media content such as video, audio, VR, or AR that can be broadcast, multicast, or unicast to the UE. As another example, the host 602 may be used for energy pricing, remote control of non-time-critical power loads for balancing power generation needs, location services, presentation services (such as editing of diagrams from data collected from remote devices), or any other function that collects, obtains, stores, analyzes, and / or transmits data.

[0188] In some examples, a measurement procedure may be provided for the purpose of monitoring data rate, latency, and other factors that are improved by one or more embodiments. There may further be network functionality as an option to reconfigure the OTT connection 650 between the host 602 and the UE 606 in response to variations in the measurement results. The measurement procedure and / or the network functionality for reconfiguring the OTT connection may be implemented in the software and hardware of the host 602 and / or the UE 606. In some embodiments, sensors (not shown) through which the OTT connection 650 passes may be deployed within or associated with other devices, and those sensors may participate in the measurement procedure by supplying the quantitative values of the monitoring results exemplified above or by supplying the values of other physical quantities, from which the quantity to be monitored may be calculated or estimated by software. The reconfiguration of the OTT connection 650 may include a message format, retransmission settings, a suitable routing, etc., and it is not necessary for the reconfiguration to directly change the operation of the network node 604. Such procedures and functionality may be known or in use in the art. In one embodiment, the measurement may include unique UE signaling that facilitates measurements such as throughput, propagation time, and latency by the host 602. The measurement may be implemented in such a way that the software transmits a message that is specifically empty or a "dummy" message using the OTT connection 650 while monitoring the propagation time, errors, etc.

[0189] FIG. 10 is a flowchart showing an exemplary method in a wireless device according to some embodiments. In a specific embodiment, one or more steps of FIG. 10 may be performed by the UE 200 described with respect to FIG. 5. The wireless device is capable of fallback operation of the ML model.

[0190] The above method starts at step 1012, and the wireless device (e.g., UE200) transmits a message to the network node indicating the wireless device's ability to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality.

[0191] In a specific embodiment, the at least one ML-based feature is based on one ML model that is divided into two parts, one part being located at the wireless device and the other part being located at the network node.

[0192] In a specific embodiment, the at least one preliminary feature is a feature that satisfies functionality equivalent to the ML-based feature, but is not preferred compared to the ML-based feature. The at least one preliminary feature may be a feature having a higher ability than the ML-based feature, but the preliminary feature is not preferred. The at least one preliminary feature may be based on a non-ML-based algorithm, or may be another ML-based algorithm (e.g., a more general ML-based algorithm). Other examples of preliminary features are described in the embodiments and examples of the UE.

[0193] In a specific embodiment, the above message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously (e.g., for comparison of performance between the two).

[0194] In step 1014, the wireless device may receive a first configuration message that configures the wireless device to operate at least one ML-based feature. In a specific embodiment, the method may further include receiving a first configuration message that configures the wireless device to simultaneously operate at least one ML-based feature and at least one preliminary feature. In other embodiments, the wireless device may autonomously determine whether to operate and simultaneously operate at least one ML-based feature and / or at least one preliminary feature.

[0195] In step 1016, the wireless device operates at least one ML-based feature for the functionality.

[0196] In step 1018, the wireless device may receive a second configuration message that configures the wireless device to deactivate at least one ML-based feature and activate at least one preliminary feature.

[0197] In other embodiments, at step 10120, the wireless device autonomously determines that it should deactivate at least one ML-based feature and activate at least one preliminary feature.

[0198] In step 1022, the wireless device operates at least one preliminary feature for the functionality.

[0199] Modifications, additions, or omissions may be made to method 1000 of FIG. 10. Additionally, one or more steps in the method of FIG. 10 may be executed in parallel or in any suitable order.

[0200] FIG. 11 is a flowchart showing an exemplary method in a network node according to an embodiment. In a specific embodiment, one or more steps of FIG. 11 may be performed by the network node 300 described with respect to FIG. 6. The network node can configure a wireless device for fallback operation of the ML model.

[0201] The method starts at step 1112, where the network node (e.g., network node 300) receives from the wireless device a message indicating the ability of the wireless device to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality.

[0202] In a specific embodiment, the at least one ML-based feature is based on one ML model that is divided into two parts, one part located in the wireless device and the other part located in the network node.

[0203] In a specific embodiment, the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously.

[0204] In step 1114, the network node may send a configuration message to the wireless device to configure the wireless device to operate at least one ML-based feature. In a specific embodiment, the method further includes sending (1114) a configuration message to configure the wireless device to operate at least one ML-based feature and at least one preliminary feature simultaneously.

[0205] In step 1116, the network node determines that at least one preliminary feature should be activated.

[0206] In step 1118, the network node transmits a configuration message to the wireless device to configure the wireless device to deactivate at least one ML-based feature and activate at least one preliminary feature.

[0207] In step 1120, the network node may deactivate the portion of at least one ML-based feature that operates at the network node.

[0208] Modifications, additions, or omissions may be made to method 1100 of FIG. 11. Additionally, one or more steps in the method of FIG. 11 may be executed in parallel or in any suitable order.

[0209] Modifications, additions, or omissions may be made to the methods disclosed herein without departing from the scope of the present invention. The methods may include more, fewer, or other steps. Additionally, the steps may be executed in any suitable order.

[0210] The foregoing description has set forth numerous specific details. However, it is understood that embodiments may be practiced without these specific details. In other instances, well-known circuits, structures, and techniques have not been shown in detail so as not to obscure the understanding of the description. Those skilled in the art will be able to implement appropriate functionality without undue experimentation based on the description provided.

[0211] References to "one embodiment", "an embodiment", "exemplary embodiments", etc. in this specification indicate that while the described embodiments may include certain features, structures, or characteristics, not every embodiment necessarily includes the particular features, structures, or characteristics. Moreover, such phrases do not necessarily refer to the same embodiment. Further, when a particular feature, structure, or characteristic is described in relation to one embodiment, it is contemplated that implementing such feature, structure, or characteristic in relation to other embodiments, whether or not explicitly described, is within the knowledge of those skilled in the art.

[0212] Although the disclosure has been described from the perspective of certain embodiments, modifications and substitutions of those embodiments will be apparent to those skilled in the art. Accordingly, the foregoing description of those embodiments does not limit the disclosure. Other changes, substitutions, and modifications are possible without departing from the scope of the disclosure as defined by the following claims.

Claims

1. A method performed by a wireless device for preliminary operation of a machine learning (ML) model, comprising: sending (1012) to a network node a message indicating the ability of the wireless device to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality; operating (1016) the at least one ML-based feature for the functionality; operating (1022) the at least one preliminary feature for the functionality.

2. The method according to claim 1, wherein the at least one ML-based feature is based on one ML model divided into two parts, one part being located at the wireless device and the other part being located at the network node.

3. The method according to any one of claims 1 to 2, wherein the at least one preliminary feature is a feature that satisfies functionality equivalent to that of the ML-based feature, but is not preferred as compared to the ML-based feature.

4. The method according to any one of claims 1 to 3, wherein the at least one preliminary feature is a feature having a higher ability than the ML-based feature, but the preliminary feature is not preferred.

5. The method according to any one of claims 1 to 4, wherein the at least one preliminary feature is based on a non-ML-based algorithm.

6. The method according to any one of claims 1 to 4, wherein the at least one preliminary feature is an ML-based algorithm.

7. The method according to any one of claims 1 to 6, wherein the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously.

8. The method according to any one of claims 1 to 7, further comprising receiving (1014) a first configuration message for configuring the wireless device to operate the at least one ML-based feature.

9. The method according to any one of claims 1 to 7, further comprising receiving (1014) a first configuration message that configures the wireless device to operate the at least one ML-based feature and the at least one preliminary feature simultaneously.

10. The method according to any one of claims 1 to 9, further comprising receiving (1018) a second configuration message that configures the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature.

11. The method according to any one of claims 1 to 9, further comprising autonomously determining (1020) that the at least one ML-based feature should be deactivated and the at least one preliminary feature should be activated.

12. A wireless device (200) capable of preliminary operation of a machine learning (ML) model, the wireless device comprising a processing circuit (202), the processing circuit sending a message to a network node indicating the wireless device's ability to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality; operating the at least one ML-based feature for the functionality; operating the at least one preliminary feature for the functionality; A wireless device operable to perform.

13. The wireless device according to claim 12, wherein the at least one ML-based feature is based on one ML model that is divided into two parts, one part being located in the wireless device and the other part being located in the network node.

14. The wireless device according to any one of claims 12 to 13, wherein the at least one preliminary feature is a feature that satisfies functionality equivalent to the ML-based feature, but is not preferred compared to the ML-based feature.

15. A wireless device according to any one of claims 12 to 14, wherein the at least one preliminary feature is a feature having a higher level of ability than the ML-based feature, provided that the preliminary feature is not preferred.

16. A wireless device according to any one of claims 12 to 15, wherein the at least one preliminary feature is based on a non-ML-based algorithm.

17. A wireless device according to any one of claims 12 to 15, wherein the at least one preliminary feature is an ML-based algorithm.

18. A wireless device according to any one of claims 12 to 17, wherein the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously.

19. A wireless device according to any one of claims 12 to 18, wherein the processing circuit is further operable to receive a first configuration message that configures the wireless device to operate the at least one ML-based feature.

20. A wireless device according to any one of claims 12 to 18, wherein the processing circuit is further operable to receive a first configuration message that configures the wireless device to operate the at least one ML-based feature and the at least one preliminary feature simultaneously.

21. A wireless device according to any one of claims 12 to 20, wherein the processing circuit is further operable to receive a second configuration message that configures the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature.

22. A wireless device according to any one of claims 12 to 20, further comprising autonomously determining (1020) that the at least one ML-based feature should be deactivated and the at least one preliminary feature should be activated.

23. A method performed by a network node for configuring a wireless device for preliminary operation of a machine learning (ML) model, comprising: Receiving (1112) from the wireless device a message indicating the wireless device's ability to support a combination of at least one ML-based feature for functionality and at least one preliminary feature for the functionality; Determining (1116) that the at least one preliminary feature should be activated; Sending (1118) to the wireless device a configuration message for configuring the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature. **Claim 24** The method according to claim 23, wherein the at least one ML-based feature is based on one ML model that is divided into two parts, one part being located at the wireless device and the other part being located at the network node. **Claim 25** The method according to any one of claims 23 to 24, wherein the at least one preliminary feature is a feature that satisfies functionality equivalent to the ML-based feature, but is not preferred as compared to the ML-based feature. **Claim 26** The method according to any one of claims 23 to 25, wherein the at least one preliminary feature is a feature having a higher level of ability than the ML-based feature, but the preliminary feature is not preferred. **Claim 27** The method according to any one of claims 23 to 26, wherein the at least one preliminary feature is based on a non-ML-based algorithm. **Claim 28** The method according to any one of claims 23 to 26, wherein the at least one preliminary feature is an ML-based algorithm. **Claim 29** The method according to any one of claims 23 to 28, wherein the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously. **Claim 30** The method according to any one of claims 23 to 29, further comprising transmitting (1114) a configuration message to the wireless device to configure the wireless device to operate the at least one ML-based feature.

31. The method according to any one of claims 23 to 29, further comprising transmitting (1114) a configuration message to the wireless device to configure the wireless device to simultaneously operate the at least one ML-based feature and at least one preliminary feature.

32. The method according to any one of claims 23 to 31, further comprising deactivating (1120) a portion of the at least one ML-based feature that operates at the network node.

34. A network node (300) capable of configuring a wireless device (200) for parallel operation of a plurality of machine learning (ML) models, the network node comprising a processing circuit (302), the processing circuit receiving, from the wireless device, a message indicating the capabilities of the wireless device regarding supporting a combination of at least one ML-based feature regarding functionality and at least one preliminary feature regarding the functionality; detecting a performance degradation of the at least one ML-based feature; transmitting, to the wireless device, a configuration message to configure the wireless device to deactivate the at least one ML-based feature and activate the at least one preliminary feature; A network node operable to perform the above.

35. The network node according to claim 34, wherein the at least one ML-based feature is based on one ML model that is divided into two parts, one part being located at the wireless device and the other part being located at the network node.

36. The network node according to any one of claims 34 to 35, wherein the at least one preliminary feature is a feature that satisfies functionality equivalent to that of the ML-based feature, but is not preferred as compared to the ML-based feature.

37. A network node according to any one of claims 34 to 36, wherein the at least one preliminary feature is a feature having a higher level of ability than the ML-based feature, provided that the preliminary feature is not preferred, the network node.

38. A network node according to any one of claims 34 to 37, wherein the at least one preliminary feature is based on a non-ML-based algorithm, the network node.

39. A network node according to any one of claims 34 to 37, wherein the at least one preliminary feature is an ML-based algorithm, the network node.

40. A network node according to any one of claims 34 to 39, wherein the message indicates whether the at least one preliminary feature and the at least one ML-based feature can be executed simultaneously, the network node.

41. A network node according to any one of claims 34 to 40, wherein the processing circuit is further operable to transmit a configuration message to the wireless device to configure the wireless device to operate the at least one ML-based feature, the network node.

42. A network node according to any one of claims 34 to 40, wherein the processing circuit is further operable to transmit a configuration message to configure the wireless device to operate the at least one ML-based feature and at least one preliminary feature simultaneously, the network node.

43. A network node according to any one of claims 34 to 42, wherein the processing circuit is further operable to deactivate a portion of the at least one ML-based feature that operates in the network node, the network node.

Images