Method and system for handling inference configuration in a wireless communication network

Abstract

Disclosed subject matter relates to method and system for handling inference configuration in a wireless communication network. The present disclosure includes, a DU communicating with UE. The DU configured to generate inference configuration for each of plurality of serving cells and subsets of plurality of serving cells associated with UE or each of a plurality of carriers and subsets of plurality of carriers supported by single serving cell associated with UE. Followed by DU transmitting inference configuration to CU. Further, CU transmits inference configuration to UE. Thereafter, UE performs inference of AI models for each of plurality of serving cells and subsets of plurality of serving cells associated with UE, using inference configuration. The UE further detects occurrence of at least one of plurality of predefined events upon performing inference of AI models. Thereafter, UE releases inference configuration based on detection.

Description

METHOD AND SYSTEM FOR HANDLING INFERENCE CONFIGURATION IN A WIRELESS COMMUNICATION NETWORK

[0001] The present disclosure relates to the field of communication networks. Particularly, but not exclusively, the present disclosure relates to a method and a system for handling inference configuration in a wireless communication network.

[0002] Artificial Intelligence and Machine Learning (AI / ML) are becoming integral to advanced wireless systems such as 5G and 6G, particularly for optimizing physical-layer operations like channel estimation, beamforming, interference management, and modulation recognition. These tasks require precise inference configurations, which define how trained models are deployed and executed in real time. Inference configuration ensures that AI / ML models operate effectively under varying network conditions, but its management becomes complex in scenarios involving Carrier Aggregation (CA) and Dual Connectivity (DC), where multiple serving cells with different frequencies, bandwidths, and subcarrier spacings coexist. Each cell's unique physical-layer characteristics demand granular configuration to maintain performance and avoid interference across heterogeneous environments. Further, in latest wireless technologies like 6G a single cell may support multiple carriers (multiple frequencies or multiple frequencies and subcarrier spacings), unlike earlier generations.

[0003] The challenge is further amplified in split base-station architectures, where Centralized Units (CUs) handle control-plane functions while Distributed Units (DUs) manage physical-layer processing and other lower layer functions. This division raises critical questions about how inference parameters are generated, exchanged between CU and DU, and ultimately delivered to User Equipment (UE) through standardized signalling. While industry discussions acknowledge that inference-related parameters may be conveyed via CSI reporting structures or OtherConfig elements in RRC messages, there is still ambiguity around CU-DU coordination, configuration granularity per serving cell, and lifecycle management at the UE. Events such as RRC state transitions, radio link failures, and recovery attempts introduce additional complexity, as UEs must determine when to retain, release, or deactivate inference configurations to prevent stale data and optimize resource usage.

[0004] Current standardization efforts in 3GPP have recognized these gaps, noting the need for explicit release mechanisms and applicability reporting, but they do not fully address operational details such as handling inference during mobility or failure recovery. Consequently, a comprehensive framework is required to define the context for inference configuration in multi-cell environments, clarify its relationship to physical-layer resources, and establish robust lifecycle handling across diverse network scenarios.

[0005] Fig. 1Ashows an exemplary system for usage of AI / ML in beam prediction, in accordance with the prior art. The UE may indicate its capabilities for AI / ML related functionalities to the network. As seen from Fig. 1A, the UE may be transmitting Radio Resource Control　(RRC) messages such as UECapabilityInformation. In an embodiment, the UE may not be able to perform the AI / ML related functionality even though the functionality is reported as supported. For example, the UE may not be able to perform the inference for the AI / ML related functionality even though it is supported. This may be due to the absence of a relevant model for the functionality or other reasons such as, hardware conditions cannot support the available model, the available model is not applicable for current UE side additional conditions, network side additional conditions and the like. The network may configure the UE for reporting the applicable functionality. The applicable functionality (also referred as applicable functionalities) may be reported using UE Assistance Information (UAI) or any such signalling. The applicable functionality can be the functionality for which the UE to perform inference. The UE may report the applicable functionality in one or more of the following scenarios (also referred to as triggers): (1) upon being configured to provide applicable functionality, (2) upon change of applicable functionality, (3) as response to NW-side additional condition (if available), and (4) upon configuration for inference configuration from the network. Generally the different categories for functionalities (e.g. beam management), whether it is a sub-use case (e.g. beam management Case 1), a group of use cases or other functionalities related to use cases) or may be defined as:

[0006] · Supported functionalities that refer to functionalities that UE can indicate by using UE capability information (via RRC / LPP signalling).

[0007] · Applicable functionalities refer to functionalities that the UE is ready to apply for inference and

[0008] · Activated functionalities refer to functionalities already enabled for performing inference.

[0009] Fig. 1Bshows a sequence flow of inference configuration at a UE, in accordance with prior art. At step 1, the network sends aUECapabilityEnqirymessage to initiate a procedure of the UE reporting its AI / ML supported functionalities.

[0010] At step 2, the UE sends aUECapablityInformationmessage to the network. TheUECapablityInformationmessage comprises supported functionalities at the UE side.

[0011] At step 3, the following configurations are provided from NW to UE:(1) the UE is allowed to do UAI reporting viaOtherConfigand (2) the network may provide NW-side additional condition. FFS on the RRC signalling and whether it is mandatory or optional.

[0012] Between step 3 and step 4, the UE decides the applicable functionalities based on NW-side additional conditions (if provided), UE-side additional conditions (internally known by UE) and model availability in device.

[0013] At step 4, the UE reports applicable functionality in the following scenarios: (1) upon being configured to provide applicable functionality and upon change of applicable functionality via the UAI, (2) as response to NW-side additional condition requesting applicable functionality reporting in step 3, FFS other network configuration (example, inference configuration).

[0014] At step 5, the network configures inference configuration to UE after applicable functionality reporting, if inference configuration based on supported functionality is not provided in step 3 (i.e. inference configuration is provided in step 5). If inference configuration based on supported functionality is provided in step 3, it is up to network implementation whether to provide an updated configuration or not. The applicable functionality may be activated by receiving its inference configuration when it is provided in step 5. In an embodiment, inference configuration may include selection / (de)activation / switching of AI / ML models or AI / ML functionalities, fallback to non-AI / ML operation, and the like. The inference configuration may inform the UE input / output, pre- / post-process. For instance, for the AI / ML for CSI prediction, the inference configuration may include the resources for channel measurements and the resources for channel prediction which serve as inputs for the AI / ML model. It may include the report quantity which can include one of CSI-RS index, SSB index, CSI-RS RSRP, SSB index RSRP etc. which may be the output of the AI / ML model. Similarly the inference configuration for CSI prediction may include post processing information such as how to report the predicted values to the network, for example, whether to use the periodic CSI reporting or semi persistent CSI reporting or aperiodic CSI reporting and various characteristics such as the slot configuration for the reporting, PUCCH resource list if it needs to be used, number of time instances etc. Similarly for the AI / ML for beam management, the inference configuration may include the input / output such as the resources for the beam measurement (input can be set B of beam related parameters when the measurements of set A is derived based on set B and the corresponding set A of beam related parameters). Inference related parameters for beam management case can include the resources for channel measurements and channel predictions.

[0015] Inference configuration may be related to management instructions such as detailed in 3gpp TR 38.843. As described in TR 38.843 v18.0.0 (version at the time of PS filing), the management instruction is the information needed to ensure proper inference operation. This information may include selection / (de)activation / switching of AI / ML models or AI / ML functionalities, fallback to non-AI / ML operation and the like.

[0016] The information disclosed in this background of the disclosure section is only for enhancement of understanding of the general background of the disclosure and should not be taken as an acknowledgement or any form of suggestion that this information forms prior art already known to a person skilled in the art.

[0017] One or more shortcomings of the prior art may be overcome, and additional advantages may be provided through the present disclosure. Additional features and advantages may be realized through the techniques of the present disclosure. Other embodiments and aspects of the disclosure are described in detail herein and are considered a part of the claimed disclosure.

[0018] In a non-limiting embodiment of the disclosure, a DU for handling inference configuration in a wireless communication network is disclosed. The DU comprises a processor and a memory communicatively coupled to the processor, where the memory stores processor executable instructions, which, on execution, may cause the DU to generate an inference configuration for, each of a plurality of serving cells and subsets of the plurality of serving cells or, each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with a User Equipment (UE). Thereafter, the DU may transmit the inference configuration to a Centralized Unit (CU). In an embodiment, the UE receives the inference configuration for performing inference of an Artificial Intelligence (AI) model associated with the UE.

[0019] In a non-limiting embodiment of the disclosure, a UE for handling inference configuration in a wireless communication network is disclosed. The UE comprises a processor and a memory communicatively coupled to the processor, where the memory stores processor executable instructions, which, on execution, may cause the UE to receive an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells or, each of plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE. Followed by performing inference of AI models for each of the plurality of serving cells and subsets of the plurality of serving cells associated with the UE or, each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, using the inference configuration. Further, detecting occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models. Thereafter, releasing the inference configuration based on the detection.

[0020] In a non-limiting embodiment of the disclosure, a method of handling inference configuration in a wireless communication network, for a DU is disclosed. The method comprises generating an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells or, each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with a User Equipment (UE). Thereafter, transmitting the inference configuration to a Centralized Unit (CU). In an embodiment, the UE receives the inference configuration for performing inference of an Artificial Intelligence (AI) model associated with the UE.

[0021] In a non-limiting embodiment of the disclosure, a method of handling inference configuration in a wireless communication network, for a UE is disclosed. The method comprises receive an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells or, each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE. Followed by performing inference of AI models for each of the plurality of serving cells and subsets of the plurality of serving cells associated with the UE, using the inference configuration. Further, detecting occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models. Thereafter, releasing the inference configuration based on the detection.

[0022] The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.

[0023] The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The same numbers are used throughout the figures to reference like features and components. Some embodiments of system and / or methods in accordance with embodiments of the present subject matter are now described, by way of example only, and with reference to the accompanying figures, in which:

[0024] Fig. 1Ashows an exemplary system for usage of AI / ML in beam prediction, in accordance with the prior art;

[0025] Fig. 1Bshows a sequence flow of inference configuration at a UE, in accordance with prior art;

[0026] Fig. 1Cshows an exemplary environment for handling inference configuration in a wireless communication network, in accordance with some embodiments of the present disclosure;

[0027] Fig. 2Ashows a detailed block diagram of a Distributed Unit (DU), in accordance with some embodiments of the present disclosure;

[0028] Fig. 2Bshows a detailed block diagram of a User Equipment (UE), in accordance with some embodiments of the present disclosure;

[0029] Fig. 3Ashows a sequence flow of releasing of inference configuration for AI / ML, in accordance with some embodiments of the present disclosure;

[0030] Fig. 3Bshows a flow chart of inference configuration handling during RRC reestablishment, in accordance with some embodiment of the present disclosure;

[0031] Fig. 3Cshows each of a plurality of serving cells configured with a distinct inference configuration, in accordance with some embodiments of the present disclosure;

[0032] Fig. 4Ashows a flowchart illustrating a method for handling inference configuration in a wireless communication network by a Distributed Unit (DU), in accordance with some embodiments of the present disclosure;

[0033] Fig. 4Bshows a flowchart illustrating a method for handling inference configuration in a wireless communication network by a User Equipment (UE), in accordance with some embodiments of the present disclosure; and

[0034] Fig. 5is a block diagram of an exemplary system for implementing embodiments consistent with the present disclosure.

[0035] It should be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and executed by a computer or processor, whether or not such computer or processor is explicitly shown.

[0036] DETAILED DESCRIPTION

[0037] In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

[0038] While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however, that it is not intended to limit the disclosure to the forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the scope of the disclosure.

[0039] The terms "comprises," "comprising," "includes" or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a setup, device, or method that includes a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus proceeded by "comprises… a" does not, without more constraints, preclude the existence of other elements or additional elements in the system or method.

[0040] The present disclosure relates to User Equipment (UE) and network handling capabilities. More particularly, the present disclosure relates to configuring inference and activation of functionalities during handover.

[0041] Artificial Intelligence (AI) and Machine Learning (ML) is used in a variety of fields. Third Generation Partnership Project (3gpp) is studying the AI / ML on air interface. The AI / ML based algorithms can be used for enhanced performance and / or reduced complexity, overhead and the like. Enhanced performance depends on applications under consideration. The enhanced performance may be improved throughput, robustness, accuracy, reliability, and the like. AI / ML models may reside on a User Equipment (UE) or on network entities such as, base station, Operations, Administration, and Maintenance (OAM) and the like. In an embodiment, the AI / Models are two sided models which reside on both UE and network entities.

[0042] Data collection may be performed for different purposes in Life Cycle Management (LCM), such as, model training, model inference, model monitoring, model selection, model update, and the like. A typical application of AI / ML on air interface is AI / ML for beam management. AI / ML for beam management includes spatial-domain downlink beam prediction for Set A of beams based on measurement results of Set B of beams and temporal Downlink beam prediction for Set A of beams based on historic measurement results of Set B of beams. AI / ML on air interface also may be used for positioning, such as direct AI / ML positioning, AI / ML assisted positioning and the like. The direct AI / ML positioning includes UE-based positioning with UE-side model, UE-assisted / LMF-based positioning with LMF-side model, NG-RAN node assisted positioning with LMF-side model, and the like. The AI / ML assisted positioning includes UE-assisted / LMF-based positioning with UE-side model, NG-RAN node assisted positioning with gNB-side model, and the like. Another application of AI / ML on air interface is for Channel State Information (CSI) feedback enhancements. The CSI feedback enhancements may be used for CSI prediction, CSI compression and the like. AI / ML may be used for other cases such as AI / ML for mobility.

[0043] Training of the AI / ML can be online training or offline training. Core Network (CN) / OAM / Over the Top Server (OTT) collection of UE-sided model training (and even Network sided model training) data may be supported. AI / ML operations typically include multiple steps such as training, inference and performance monitoring.

[0044] However, there still exists need for interference configuration and the activation / deactivation of the functionalities for AI / ML during handover, RRC Reestablishment, transition to RRC_IDLE or transition to RRC_INACTIVE.

[0045] The present disclosure describes how a UE and a network handles inference configuration, activation of the functionalities, deactivation of the functionalities for AI / ML during handover, RRC reestablishment, transition to RRC_IDLE and transition to RRC_INACTIVE but is not limited thereto.

[0046] In an embodiment, the network apparatus configures the UE with a configuration for interference configuration for AI / ML functionalities using a SetupRelease structure. In an embodiment, the network apparatus is a radio access network apparatus such as gNB in fifth Generation (5G) New Radio (NR) or similar nodes in other radio technologies. In an embodiment, the network may configure the UE to add or modify inference configuration by sending SetupRelease as Setup. In an embodiment, the network apparatus may send a list including identifiers for the functionalities or identifies for the inference configuration to add and / or modify the inference configuration. If the identifier for the functionality is provided, UE configures itself to perform inference for the functionality. If the identity of the inference configuration is provided, the network also provides the related functionality also as part of the interference configuration or the RRC configuration. The functionalities in the previous embodiment may include but not limited to AI / ML related to beam management, AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility) and AI / ML related to positioning or AI / ML related to Channel State Information　(CSI)　predictions and so on.

[0047] In an embodiment, the network apparatus configures the UE with configuration for interference configuration for AI / ML functionalities separately for the primary cells and secondary cells. In an embodiment, in dual connectivity, the network apparatus configures the UE with configuration for interference configuration for AI / ML separately for a Primary Cell (PCell) and Primary Secondary Cell Group Cell (PSCell). In an embodiment, the inference configuration from the PSCell or any other cell in the SCG may be provided over Signaling Radio Bearer　(SRB3). In an embodiment, the inference configuration from the PSCell or any other cell in Secondary Node (SN) may be provided over SRB1. In an embodiment, the SN decides the inference configuration from the PSCell or any other cell in the SCG and sends it to the Master Node (MN). The MN sends the inference configuration to the UE over SRB1. The MN may embed the inference configuration for SCG or an RRCReconfiguration including inference configuration for Secondary Cell Group (SCG) in Master Cell Group (MCG) RRCReconfiguration and send it to the UE. The functionalities in the previous embodiment can be the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions, and the like.

[0048] In an embodiment, the network apparatus may configure the UE to release inference configuration for one or more functionalities. In this embodiment, the inference configurations may be provided by sending SetupRelease as release. The network apparatus may send a release list including identifiers for the one or more functionalities and / or identifiers for the inference configuration. In an embodiment, for SCG, the configuration to release inference configuration for the one or more functionalities may be provided using SRB3 or SRB1. The SN decides to release the inference configuration from the PSCell or any other cell in the SCG and sends the information to release the inference configuration to the MN. The MN sends the configuration to release the inference configuration to the UE over SRB1. The MN may embed the configuration to release inference configuration for SCG or an RRCReconfiguration including the configuration to release inference configuration for SCG in MCG RRCReconfiguration and send it to the UE. The functionalities in the previous embodiment can be the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions, and the like.

[0049] In an embodiment, if the inference configuration for the one or more functionalities for AI / ML is released, the functionality is deactivated by the UE and the network apparatus. The functionality in the previous embodiment can be the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions, and the like. The deactivation may be performed for the corresponding cell or cell group based on the granularity of the inference configuration.

[0050] In an embodiment, upon transitioning to RRC_IDLE, the UE and the network apparatus deactivates AI / ML related functionalities. In an embodiment, upon transitioning to RRC_IDLE, the inference configuration for all the AI / ML related functionalities is released by the UE and by the network apparatus. The functionalities in the previous embodiment can be the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions, and the like, and the like.

[0051] In an embodiment, on transitioning to RRC_IDLE, the inference configuration for the functionalities such as for the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions, and the like are released and deactivated by the UE and by the network apparatus.

[0052] In an embodiment, on transitioning to RRC_INACTIVE, the inference configuration is released, and the AI / ML related functionalities are deactivated.

[0053] In an embodiment, on transitioning to RRC_INACTIVE, the inference configuration for one or more of the functionalities such as for AI / ML related to beam management, AI / ML related to mobility (AI / ML for L3 mobility or Lower layer triggered mobility), AI / ML related to positioning and AI / ML related to CSI predictions are released by the UE and the network apparatus. The UE may also deactivate the related functionality.

[0054] In an embodiment, on transitioning to RRC_INACTIVE, the UE keeps the inference configuration for one or more functionalities such as for AI / ML related to mobility (AI / ML for L3 mobility or Lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions but may deactivate the related functionality. The UE may further release the inference configuration during RRC resume procedure. For example, on initiating RRC resume procedure, the UE releases the inference configuration stored while moving to RRC_INACTIVE.

[0055] In an embodiment, on transitioning to RRC_INACTIVE, the inference configuration for the AI / ML related to beam management is released by the UE and the network apparatus. The UE and the network apparatus may deactivate the AI / ML functionalities related to beam management.

[0056] In an embodiment, the inference configuration for AI / ML including one or more inference configuration for one or more functionalities such as for AI / ML related to beam management, AI / ML related to mobility (AI / ML for L3 mobility or Lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions, and the like is decided by gNB Control Unit (CU) and gNB CU informs the gNB Distributed Unit (DU) the inference configuration using F1AP interface. In an embodiment, if the gNB CU decides to activate inference configuration as described previously, the gNB CU informs the gNB DU through a F1Application Protocol (AP) message.

[0057] In an embodiment, the inference configuration for the AI / ML related to beam management is decided by the gNB DU. The gNB DU informs the gNB CU of the inference configuration using F1AP interface. In an embodiment, when the gNB DU decides to activate inference configuration related to beam management, the gNB DU informs the gNB CU through a F1AP message.

[0058] In an embodiment, on performing handover for Intra-RAT handover (such as from a source NR cell to a target NR cell), the inference configuration for AI / ML related to beam management is released. In an embodiment, upon performing handover from a source NR cell to a target NR cell, the UE deactivates the functionality for AI / ML related to beam management.

[0059] In an embodiment, on performing PSCell change, the inference configuration for AI / ML related to beam management for SCG is released by the UE and the network apparatus. The UE and the network apparatus also may deactivate the functionality for AI / ML related to beam management for SCG.

[0060] In an embodiment, on performing handover for inter-RAT handover (such as from a source NR cell to a target E-UTRA Cell), the inference configuration for AI / ML related to beam management is released by the UE. In an embodiment, upon performing handover inter-RAT handover (such as from a source NR cell to a target E-UTRA Cell), UE deactivates the functionality for AI / ML related to beam management.

[0061] In an embodiment, on performing handover for Intra-RAT handover (such as from a source NR cell to a target NR cell), the inference configuration for one or more of AI / ML functionalities such as AI / ML related to mobility (AI / ML for Level-3 (L3) mobility or lower layer triggered mobility), the AI / ML related to positioning, the AI / ML related to CSI predictions, and the like is kept by the UE. The UE may also keep the activation status of the functionality as in the source cell unless explicitly changed by the network. In an embodiment, if the mobility is for inter-RAT handover (such as from a source NR cell to a target E-UTRA cell), the inference configuration for one or more of AI / ML functionalities such as AI / ML related to mobility (AI / ML for L3 mobility or Lower layer triggered mobility), the AI / ML related to positioning and the AI / ML related to CSI predictions is released, and the UE deactivates the corresponding functionality for AI / ML.

[0062] In an embodiment, upon performing serving cell addition, the serving cell configuration (comprising the inference configuration for a functionality), the AI / ML related functionality is activated for the corresponding SCell by the UE and the network apparatus. In an embodiment, upon performing serving cell modification and the serving cell configuration releases the inference configuration for a functionality, the AI / ML related functionality is deactivated for the corresponding SCell by the UE and the network apparatus. The functionalities in the previous embodiment can be AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or Lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions and the like.

[0063] In an embodiment, on performing RRC reestablishment, the inference configuration for one or more of AI / ML functionalities such as one or more the AI / ML related to beam management, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions, and the like is released by the UE. The UE also may deactivate the corresponding functionality for AI / ML.

[0064] In an embodiment, the UE may also deactivate the corresponding functionality. In an embodiment, according to TS 38.331,

[0065]

[0066]

[0067] In an embodiment, if the UE is not able to perform inference for an AI / ML related functionality, the UE may deactivate the functionality. On deactivating the functionality, the UE sends UAI to inform the network apparatus that the corresponding functionality is deactivated. i.e. the UE deactivates functionality without waiting for the response (such as RRCReconfiguration) from the network.

[0068] In an embodiment, on receiving a UAI that the applicable functionality has changed, the network apparatus sends a configuration to switch from AI / ML based functionality to non-AI / ML based functionality. This may be provided in RRC Reconfiguration message. Alternatively, this may be provided in L1 or L2 signalling. The CU may transfer the received UAI to the DU. The DU may send L1 or L2 signalling to deactivate the functionality.

[0069] In an embodiment, the AI / ML configuration / activation for a functionality also means the AI / ML configuration / activation for the sub-applications for the functionality. For example, the AI / ML configuration / activation of the beam management also encompasses Spatial-domain Downlink beam prediction, temporal downlink beam prediction and the like.

[0070] Disclosed herein is a method and a system for handling inference configuration in a wireless communication network. The existing approach lacks clear mechanisms for managing inference configurations in multi-cell, multi carrier single cell and split architecture scenarios. Current standards do not specify which network node generates the configuration, how CU-DU coordination should occur, or how UEs should handle these configurations during events like RRC state transitions and radio link failures. This results in ambiguity, potential inefficiencies, and inconsistent behavior across devices and networks.

[0071] To solve the above problem, the present disclosure discloses a method and a system for handling inference configuration in a wireless communication network. In the present disclosure, a structured flow is disclosed that begins with the Distributed Unit (DU) generating inference configurations for each serving cell, considering their unique physical-layer characteristics in scenarios like carrier aggregation and dual connectivity. When a single cell supports multiple carriers, the DU may generate inference configuration for each carrier separately. These configurations are then communicated to the Centralized Unit (CU) over the F1 interface, ensuring proper coordination in split base-station architectures. The CU delivers the configurations to the User Equipment (UE) through standardized RRC signalling, using either CSI-ReportConfig or OtherConfig information elements. When CSI-ReportConfig is used, the CU may just relay what is received from the DU. When OtherConfig is used, DU may construct a new structure using the inference configuration received from CU. It is possible that the network uses both options at different instances for the same UE. For instance, CSI-ReportConfig may be used for full inference configuration and OtherConfig may use for partial inference configuration. Once received, the UE performs AI / ML inference based on these configurations and manages them dynamically, releasing them upon explicit network instruction, RRC state transitions, or Radio Link Failures (RLFs) without successful recovery. This approach overcomes existing drawbacks by defining clear roles for DU and CU, enabling granular per-cell configuration, and establishing robust lifecycle handling at the UE, thereby eliminating ambiguity and inefficiencies present in existing methods.

[0072] In the following detailed description of the embodiments of the disclosure, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be taken in a limiting sense.

[0073] Fig. 1Cillustrates an environment 100 for managing inference configurations in a wireless communication network. As shown, the environment 100 includes a Core Network 102, a plurality of Distributed Units (DUs) 104 (including a DU 104₁and a DU 104₂, collectively referred to as the plurality of DUs 104), a Centralized Unit (CU) 106, and a User Equipment (UE) 108. For illustration purposes, only two DUs are shown in Fig. 1C, although in practice there may be more.

[0074] The Core Network 102 refers to the central part of the mobile network responsible for mobility management, session management, and overall control of network resources. The Distributed Unit (DU) is a component of a gNB (next-generation NodeB) that primarily handles lower-layer functionalities such as the physical layer (PHY) and parts of the MAC layer. The Centralized Unit (CU) is another gNB component that manages higher-layer functions, including Radio Resource Control (RRC) and PDCP. The User Equipment (UE) 108 represents a mobile device or terminal that communicates with the network for data and control signaling. In an embodiment, it may be possible that the CU 106 and the DU 1041may be implemented in the same network apparatus. In such scenarios, any embodiment related to either CU or DU is applicable for network apparatus in general.

[0075] Any embodiment related to gNB is equally applicable to any base station which performs functionalities similar to the gNB, for instance for a 6G base station (gNB). The gNB can mean any network apparatus, which performs Radio Access Network (RAN) functions.

[0076] As depicted, the CU 106 is connected to the Core Network 102 and interfaces with the plurality of DUs 104 over the F1 interface, enabling coordination between control-plane and user-plane functions. The UE 108 communicates with the CU 106 for RRC signaling and with the DUs 104 for physical-layer or RLC / MAC operations. This environment provides the architectural basis for implementing AI / ML inference configuration management across split base-station components and the UE 108.

[0077] Fig. 2Ashows a detailed block diagram of a Distributed Unit (DU), in accordance with some embodiments of the present disclosure.

[0078] Fig. 2Ashows internal architecture of the DU in accordance with some embodiments of the present disclosure. The DU 1041may include at least one Central Processing Unit ("CPU" or "processor") 206 and a memory 204storing instructions executable by the at least one processor 206. The processor 206 may comprise at least one data processor for executing program components for executing user or system-generated requests. The memory 204 is communicatively coupled to the processor 206. The DU 1041further comprises an Input / Output (I / O) interface 202. The I / O interface 202 is coupled with the processor 206through which an input signal or / and an output signal is communicated.

[0079] In some implementations, the DU 1041may include data 208 and modules 210. As an example, the data 208 may be stored within the memory 204 associated with the DU 1041. In some embodiments, data 208 may include, for example, inference data 212 and other data 214. In some embodiments, the data 208 may be stored in the memory 204 in form of various data structures.

[0080] The inference data 212 comprises inference configuration corresponding to a set of parameters and settings that define how AI / ML models will perform inference for physical-layer functionalities in the wireless communication network.

[0081] The other data 212 may be stored data, including temporary data and temporary files, generated by the modules 202 for performing the various functions of the DU 1041.

[0082] In an embodiment, the data 208 in the memory 204 are processed by the one or more modules 210 present within the memory 204 of the DU 1041.

[0083] One or more modules 210 along with the data 208, functions to handle inference configuration in a wireless communication network. In one implementation, the one or more modules 210 may include, but are not limited to, a generating module 216, a transmitting module 218, and one or more other modules 220.

[0084] In an embodiment, the one or more modules 210 may be implemented as dedicated units. As used herein, the term module refers to an application specific integrated circuit (ASIC), an electronic circuit, a field-programmable gate arrays (FPGA), Programmable System-on-Chip　 (PSoC), a combinational logic circuit, and / or other suitable components that provide the described functionality. In some implementations, the one or more modules 210 may be communicatively coupled to the processor 206 for performing one or more functions of the DU 1041. The said modules 210 when configured with the functionality defined in the present disclosure will result in a novel hardware.

[0085] In an embodiment, the generating module 216 may be configured to generate an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells associated with the UE 108. In NR (New Radio) architecture, the plurality of serving cells correspond to Primary Cells (PCell) and Secondary Cells (SCell). Particularly the generating module 216 generates the inference configuration separately for the PCell and the SCell. The generating module 216 may be configured to generate the inference configuration when the UE 108 is configured with at least one of, a Carrier Aggregation (CA) mode or Dual Connectivity (DC). Typically, in the CA mode, the UE may have one PCell and multiple SCells within same cell group, while in DC, there are two cell groups corresponding to a Master Cell Group (MCG) with its PCell and associated SCells, and a Secondary Cell Group (SCG) with its Primary Secondary Cell Group Cell (PSCell) and associated SCells. All these cells collectively form the plurality of serving cells for the UE 108, and the inference configurations are generated by the generating module 216 for per serving cell to account for their distinct physical layer characteristics. Therefore, the generating module 216 may be configured to generate the inference configuration based on the physical layer characteristics associated with the corresponding serving cell of the plurality of serving cells. In an embodiment, the generated inference configuration is stored in the inference data 212. The inference data 212 comprising the inference configuration generated for each of the plurality of serving cells and the subsets of the plurality of serving cells, is further sent to the transmitting module 218.

[0086] In an embodiment, the generating module 216 may be configured to generate an inference configuration for each of a plurality of carriers in a serving cell and subsets of the plurality of carriers in a serving cell associated with the UE 108. Plurality of carriers in a serving cell may correspond to multiple carriers, such as multiple frequencies supported by a single serving cell. Particularly the generating module 216 generates the inference configuration separately for the each of the carriers in a serving cell. The generating module 216 may be configured to generate the inference configuration when the UE 108 is configured with multi carrier single cell operation in the same serving cell. The inference configurations are generated by the generating module 216 for per carrier to account for their distinct physical layer characteristics. Therefore, the generating module 216 may be configured to generate the inference configuration based on the physical layer characteristics associated with the corresponding carrier of the plurality of carriers. In an embodiment, the generated inference configuration is stored in the inference data 212. The inference data 212 comprising the inference configuration generated for each of the plurality of carriers in a serving cell. There may be also a plurality of serving cells where each serving cell may have a plurality of carriers associated to it and the inference generation may be generated for the plurality of carriers for each of the plurality of serving cells.

[0087] In an embodiment, the transmitting module 218 may be configured to receive the generated data 212 from the generating module 216. The transmitting module 218 may further be configured to transmit the inference configuration to the CU 106. In an embodiment, the inference configuration may be transmitted by the transmitting module 218 to the CU 106 over an F1AP interface. For example, when the DU 1041decides to activate the inference configuration related to beam management, the transmitting module 218 is configured to inform the CU 106 through the F1AP interface. In an embodiment, the transmitting module 218 may transmit the inference configuration in one of, a CSI-ReportConfig information element or an extendedConfig information element, configured in a RRCReconfiguration message. In an embodiment, the CSI-ReportConfig information element may be part of CSI-MeasConfig provided per serving cell of the plurality of serving cells. Particularly, the CSI-MeasConfig is used to configure CSI-RS (reference signals) belonging to the serving cell in which the CSI-MeasConfig is included. In an embodiment, the extendedConfig information element corresponds to an extension beyond existing CSI configurations. A person skilled in the art may appreciate that any element defined in future, which may beyond the existing CSI configurations, may correspond to the extendedConfig information element. In an embodiment, the DU 1041may send the inference configuration during F1AP UEContext Setup procedure. In an embodiment, the DU 1041may send the inference configuration during F1AP UEContext Modification procedure.

[0088] Fig. 2Bshows a detailed block diagram of a User Equipment (UE), in accordance with some embodiments of the present disclosure.

[0089] Fig. 2Bshows internal architecture of the UE in accordance with some embodiments of the present disclosure. The UE 108 may include at least one Central Processing Unit ("CPU" or "processor") 228 and a memory 226storing instructions executable by the at least one processor 228. The processor 228 may comprise at least one data processor for executing program components for executing user or system-generated requests. The memory 226 is communicatively coupled to the processor 228. The UE 108 further comprises an Input / Output (I / O) interface 224. The I / O interface 224 is coupled with the processor 228 through which an input signal or / and an output signal is communicated.

[0090] In some implementations, the UE 108 may include data 230 and modules 232. As an example, the data 230 may be stored within the memory 226 associated with the UE 108. In some embodiments, data 230 may include, for example, inference data 234, detected data 236 and other data 238. In some embodiments, the data 230 may be stored in the memory 226 in form of various data structures.

[0091] The inference data 234 comprises inference configuration corresponding to the set of parameters and settings that define how AI / ML models will perform inference for physical-layer functionalities in the wireless communication network.

[0092] The detected data 236 comprises a plurality of predefined events further comprising receiving an explicit instruction to release the inference configuration, transitioning of the UE 108 to an RRC_IDLE state or an RRC_INACTIVE state and detecting a Radio Link Failure (RLF).

[0093] The other data 238 may be stored data, including temporary data and temporary files, generated by the modules 232 for performing the various functions of the UE 108.

[0094] In an embodiment, the data 208 in the memory 226 are processed by the one or more modules 232 present within the memory 226 of the UE 108.

[0095] One or more modules 232 along with the data 230, functions to handle inference configuration in a wireless communication network. In one implementation, the one or more modules 232 may include, but are not limited to, a receiving module 240, an inference module 242, detecting module 244, releasing module 246 and one or more other modules 248.

[0096] In an embodiment, the one or more modules 232 may be implemented as dedicated units. As used herein, the term module refers to an application specific integrated circuit (ASIC), an electronic circuit, a field-programmable gate arrays (FPGA), Programmable System-on-Chip　 (PSoC), a combinational logic circuit, and / or other suitable components that provide the described functionality. In some implementations, the one or more modules 232 may be communicatively coupled to the processor 228 for performing one or more functions of the UE 108. The said modules 232 when configured with the functionality defined in the present disclosure will result in a novel hardware.

[0097] In an embodiment, the receiving module 240 may be configured to receive the inference configuration for each of the plurality of serving cells associated with the UE 108 and subsets of the plurality of serving cells. In an embodiment, the receiving module 240 may receive the inference configuration from the CU 106. In one embodiment, the inference configuration may be received as a encoded information via a CSI-ReportConfig information element, over which the DU 1041transmits the inference configuration to the CU 106. In another embodiment, the inference configuration may be received as a restructured information via an OtherConfig information element, which is restructured by the CU 106. In yet another embodiment, the receiving module 240 may receive the inference configuration from any other network entity. The UE 108 may receive the RRCReconfiguration message including one set or multiple sets of inference related parameters via the OtherConfig information element.. In an embodiment, the inference configuration may be stored in the inference data 234 and sent to the inference module 242 for performing further functionalities of the UE 108.

[0098] In an embodiment, the inference module 242 may be configured to receive the inference data 234 from the receiving module 240. Upon receiving the inference data 234, the inference module 242 may be configured to perform inference of AI models for each of the plurality of serving cells and subsets of the plurality of serving cells associated with the UE 108. The inference module 242 may be configured to perform inference of AI models for each of the plurality of carriers in the same serving cell. The inference module 242 may perform the inference using the inference configuration in the inference data 234.

[0099] In an embodiment, the detecting module 244 may be configured to detect occurrence of at least one of the plurality of predefined events upon performing the inference of the AI models. As mentioned previously, the plurality of predefined events comprise of, receiving the explicit instruction to release the inference configuration, transitioning of the UE 108 to the RRC_IDLE state or the RRC_INACTIVE state and detecting the RLF. Particularly, the detecting module 244 may detect a SetupRelease element when the explicit instruction to release the inference configuration is received. When the SetupRelease element is detected, a release list including identifiers of the one or more physical-layer functionalities and / or one or more identifiers of the inference configuration are further detected. In an embodiment, for SCG, the information to release inference configuration for the one or more physical-layer functionalities may be provided using SRB3 or SRB1. In an embodiment, the one or more physical-layer functionalities may be the AI / ML functionalities related to beam management, mobility (AI / ML for L3 mobility or lower layer triggered mobility), positioning or CSI feedback enhancements, and the like. The explicit instruction is generated by the Secondary Node (SN) that hosts the SCG which includes the PSCell. Particularly the SN may decide to release the inference configuration from the PSCell or any other cell in the SCG and sends the information to release the inference configuration to the Master Node (MN), which hosts the MCG and its PCell. The MN may then send the information to release the inference configuration to the UE 108 over SRB1. The MN may embed the information to release the inference configuration for the SCG in MCG RRCReconfiguration message and send it to the UE 108. The detecting module 244 may detect the MCG RRCReconfiguration message and determines the occurrence of the explicit instruction. Upon detecting at least one of the plurality of predefined events, the detecting module 244 stores the detected information in the detected data 236 and sends to the releasing module 246. Releasing the inference configuration based on explicit instruction allows the network fine control of the inference configuration. This allows the network to switch to non AI methods if it finds that the AI / ML is not working properly in some cases.

[0100] In an embodiment, the releasing module 246 may be configured to receive the detected data 236 from the detecting module 244. When the detected data 236 comprises the explicit instruction to release the inference configuration or the UE 108 transitioned into the RRC_IDLE state or the RRC_INACTIVE state, then the releasing module 246 may be configured to perform the release and deactivation of the configuration information for one or more functionalities.Fig. 3Ashows a sequence flow of releasing of inference configuration for AI / ML, in accordance with some embodiments of the present disclosure. As seen from the figure, if the inference configuration for the one or more functionalities for AI / ML is released, the functionality is deactivated by the UE 108 and the network apparatus. Releasing the inference configuration while transitioning to RRC_IDLE / RRC_INACTIVE saves the UE memory and the network memory. It also avoids the overhead of the signalling between old and new network nodes when the UE 108 moves to RRC_CONNECTED again. This also helps to make the inference more robust as the inference will not be based on old configuration, i.e. before the transition to RRC_IDLE or RRC_INACTIVE.

[0101] In an embodiment, when the detected data 236 comprises of the RLF, the releasing module 246 in conjunction to one or more other modules 248 may be configured to initiate an RRC reestablishment procedure. Particularly, the one or more other modules 248 may be configured to determine whether the UE 108 is configured for mobility recovery procedures. In an embodiment, the mobility recovery procedures may be defined as mechanisms that enable UE 108 to restore connectivity after detecting the RLF without performing a full initial access procedure. These mobility recovery procedures may be designed to minimize service interruption and signalling overhead during mobility events. These mobility recovery procedures may include:

[0102] · Conditional Handover (CHO): A pre-configured handover process where the UE 108 switches to a target cell upon meeting specific conditions, allowing rapid recovery from RLF.

[0103] · Lower-layer Triggered Mobility (LTM): A recovery method initiated based on the configuration for Lower Layer Triggered Mobility candidates (mobility which is executed based on the triggers from lower protocol layers (such as PHY or MAC)), including Conditional Lower Layer Triggered Mobility, enabling the UE 108 to move to a suitable cell quickly when link quality deteriorates.

[0104] Fig. 3Bshows a flow chart of inference configuration handling during RRC reestablishment, in accordance with some embodiment of the present disclosure. As seen from the figure,upon initiation of RRC re-establishment procedure, the UE 108 checks if it is configured for attempting conditional reconfiguration for Radio Link Failure　(RLF) recovery (for example, NR RRC IE attemptCondReconfig is configured) or if it is configured for attempting Lower-layer Triggered Mobility (LTM) for RLF recovery (for example, NR RRC IE attemptLTM-Switch is configured) and if any of these configurations are available (i.e. if any of these flags are configured). The UE 108 keeps the inference configuration for one or more of AI / ML functionalities such as the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions and the AI / ML related to beam management. If none of these configurations for the recovery are available (i.e. none of these flags such as attemptCondReconfig or attemptLTM-Switch are configured), the UE 108 releases the inference configuration for one or more AI / ML functionalities such as, the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions and the AI / ML related to beam management. The UE 108 may also deactivate the corresponding functionality. If the UE 108 is not able to perform Conditional Handover (CHO) based recovery or LTM based recovery while the timer for cell selection in the RRC Reestablishment procedure (such as T311 timer in NR is running), UE 108 releases theinference configuration for one or more of AI / ML functionalities such as the AI / ML related to mobility (AI / ML for L3 mobility or lower layer triggered mobility), the AI / ML related to positioning or AI / ML related to CSI predictions and the AI / ML related to beam management following cell selection while T311 is running.

[0105] Upon performing the determination, the one or more other modules 248 may be configured to release the inference configuration during the RRC reestablishment procedure when the UE 108 is not configured for the mobility recovery procedures. For example, if none of the configurations for the recovery are available (i.e. none of the flags such as attemptCondReconfig or attemptLTM-Switch in NR are configured), the releasing module 246 may release the inference configuration for the one or more AI / ML functionalities. In another embodiment, the one or more other modules 248 may be configured to perform the mobility recovery procedures when the UE 108 is configured for the mobility recovery procedures. In an embodiment, when the mobility recovery procedures are successful, the one or more other modules 248 may further be configured to maintain the inference configuration during the RRC reestablishment procedure. For example, if the UE 108 is configured for attempting the LTM for RLF recovery (for example, NR RRC IE attemptLTM-Switch is configured) and if any of these configurations are available (i.e. if any of these flags are configured), then the UE 108 may maintain the inference configuration for the one or more of AI / ML functionalities such as the AI / ML functionalities. In another embodiment, when the mobility recovery procedures are not successful, the releasing module 246 may be configured to release the inference configuration during the RRC reestablishment procedure. For example, if the UE 108 is not able to perform the CHO based recovery or the LTM based recovery while the timer for cell selection in the RRC Reestablishment procedure (such as T311 timer in NR is running), then the releasing module 246 may release the inference configuration for the one or more of AI / ML functionalities. This allows the UE 108 and the network to be in synchronisation with respect to the inference configuration when the recovery is successful and also to ensure that the inference configuration is not kept when it is not possible to recover the radio link there by avoiding potentially wrong UE inference after Reestablishment.

[0106] Fig. 3Cshows each of a plurality of serving cells configured with a distinct inference configuration, in accordance with some embodiments of the present disclosure. As shown, the figure includes a plurality of serving cells, such as a PCell, a PSCell, an MCG Secondary Cell, and a SCG Secondary Cell. Each serving cell is associated with a distinct inference configuration, depicted as conf1 for the PCell, conf2 for the PSCell, conf3 for the MCG Secondary Cell, and conf4 for the SCG Secondary Cell. In the context of carrier aggregation and dual connectivity, the PCell refers to the primary cell of the Master Cell Group, while the PSCell refers to the primary cell of the Secondary Cell Group. The MCG and SCG may include additional secondary cells beyond those shown for illustration purposes. Generating inference configurations per serving cell ensures that AI / ML models used for physical-layer functionalities, such as beam management or channel state prediction, are tailored to the specific characteristics of each cell, including frequency, bandwidth, and subcarrier spacing. This approach provides fine-grained control and enables optimized inference performance across heterogeneous radio conditions. In an embodiment, the one or more of these serving cell may be configured with multiple carriers (i.e the cell may support multiple frequencies) and in such a case inference configuration may be provided separately by the network apparatus to the UE for one or more of these carriers.

[0107] Fig. 4Ais a flowchart illustrating a method for a DU for handling inference configuration in a wireless communication network, in accordance with some embodiments of the present disclosure.

[0108] As illustrated inFig. 4A, the method 400A comprises one or more blocks illustrating a method for a DU for handling inference configuration in a wireless communication network, in accordance with some embodiments of the present disclosure. The method 400A may be described in the general context of computer-executable instructions. Generally, computer-executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform functions or implement abstract data types.

[0109] The order in which the method 400A is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method 400A. Additionally, individual blocks may be deleted from the methods without departing from scope of the subject matter described herein. Furthermore, the method 400A can be implemented in any suitable hardware, software, firmware, or combination thereof.

[0110] At block 402, the method 400A may include generating, by the DU 1041, the inference configuration for each of the plurality of serving cells and the subsets of the plurality of serving cells or the plurality of carriers and subsets of the plurality of carriers supported by the single serving cell, associated with the UE 108. In an embodiment, the inference configuration is generated for each of the plurality of serving cells and the subsets of the plurality of serving cells, when the UE 108 is configured with at least one of, the carrier aggregation mode or the dual connectivity. In an embodiment, the inference configuration is generated based on the characteristics of the physical layer associated with the corresponding serving cell of the plurality of serving cells. In an embodiment, the inference configuration is generated for each of the plurality of carriers and the subsets of the plurality of carriers supported by the single serving cell when the UE 108 is configured with multi carrier single cell configuration.

[0111] At block 404, the method 400A may include transmitting, by the DU 1041, the inference configuration to the CU 106. In an embodiment, the UE 108 receives the inference configuration for performing inference of the AI model associated with the UE 108. In an embodiment, the inference configuration is transmitted to the CU 106 over the F1AP interface. In an embodiment, the inference configuration is transmitted in one of, the CSI-ReportConfig information element or the extendedConfig information element, configured in the RRCReconfiguration message.

[0112] Fig. 4Bis a flowchart illustrating a method for a UE for handling inference configuration in a wireless communication network, in accordance with some embodiments of the present disclosure.

[0113] As illustrated inFig. 4B, the method 400B comprises one or more blocks illustrating a method for a DU for handling inference configuration in a wireless communication network, in accordance with some embodiments of the present disclosure. The method 400B may be described in the general context of computer-executable instructions. Generally, computer-executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform functions or implement abstract data types.

[0114] The order in which the method 400B is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method 400B. Additionally, individual blocks may be deleted from the methods without departing from scope of the subject matter described herein. Furthermore, the method 400B can be implemented in any suitable hardware, software, firmware, or combination thereof.

[0115] At block 406, the method 400B may include receiving, by the UE 108, the inference configuration for each of the plurality of serving cells and subsets of the plurality of serving cells or a plurality of carriers and subsets of the plurality of carriers supported by the single serving cell, associated with the UE 108. In an embodiment, the inference configuration is received via one of, the CSI-ReportConfig information element or the OtherConfig information element, configured in the RRCReconfiguration message. In an embodiment, the inference configuration is received over the F1AP interface.

[0116] At block 408, the method 400B may include performing, by the UE 108, inference of the AI models for each of the plurality of serving cells and subsets of the plurality of serving cells associated with the UE 108, using the inference configuration.

[0117] At block 410, the method 400B may include detecting, by the UE 108, the occurrence of the at least one of the plurality of predefined events upon performing the inference of the AI models. In an embodiment, the plurality of predefined events comprises receiving the explicit instruction to release the inference configuration, transitioning of the UE 108 to the RRC_IDLE state or the RRC_INACTIVE state, or detecting the RLF. In an embodiment, when the radio link failure is detected, the method comprises initiating the RRC reestablishment procedure. The RRC reestablishment procedure comprises determining whether the UE 108 is configured for mobility recovery procedures.

[0118] At block 412, the method 400B may include releasing, by the UE 108, the inference configuration based on the detection. In an embodiment, the inference configuration is released during the RRC reestablishment procedure when the UE 108 is not configured for the mobility recovery procedures. In another embodiment, the mobility recovery procedures are performed when the UE 108 is configured for mobility recovery procedures. Further, performing the mobility recovery procedures comprises performing one of, maintaining the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are successful or releasing the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are not successful.

[0119] Fig. 5is a block diagram of an exemplary computer system for implementing embodiments consistent with the present disclosure.

[0120] In some embodiments, FIG. 5 illustrates a block diagram of an exemplary computer system 500 for implementing embodiments consistent with the present disclosure. In some embodiments, the computer system 500 may be at least one of, the DU 1041or the UE 108, that comprise a processor (also referred as a processor 502 in this FIG. 5) that is used for handling inference configuration in a wireless communication network. The processor 502 may include at least one data processor for executing program components for executing user or system-generated business processes. The processor 502 may include specialized processing units such as integrated system (bus) controllers, memory management control units, floating point units, graphics processing units, digital signal processing units, etc.

[0121] The processor 502 may be disposed in communication with input devices 510 and output devices 511 via I / O interface 501. The I / O interface 501 may employ communication protocols / methods such as, without limitation, audio, analog, digital, stereo, IEEE-1394, serial bus, Universal Serial Bus (USB), infrared, PS / 2, BNC, coaxial, component, composite, Digital Visual Interface (DVI), high-definition multimedia interface (HDMI), Radio Frequency (RF) antennas, S-Video, Video Graphics Array (VGA), IEEE 802.n / b / g / n / x, Bluetooth, cellular (e.g., Code-Division Multiple Access (CDMA), High-Speed Packet Access (HSPA+), Global System For Mobile Communications (GSM), Long-Term Evolution (LTE), WiMax, or the like), etc.

[0122] Using the I / O interface 501, computer system 500 may communicate with input devices 510 and output devices 511.

[0123] In some embodiments, the processor 502 may be disposed in communication with a communication network 509 via a network interface 503. The network interface 503 may communicate with the communication network 509. The network interface 503 may employ connection protocols including, without limitation, direct connect, Ethernet (e.g., twisted pair 10 / 100 / 1000 Base T), Transmission Control Protocol / Internet Protocol (TCP / IP), token ring, IEEE 802.11a / b / g / n / x, etc. Using the network interface 503 and the communication network 509, the computer system 500 may communicate with the CU 106.

[0124] The communication network 509 can be implemented as one of the different types of networks, such as intranet or Local Area Network (LAN) and such within the organization. The communication network 509 may either be a dedicated network or a shared network, which represents an association of the different types of networks that use a variety of protocols, for example, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol / Internet Protocol (TCP / IP), Wireless Application Protocol (WAP), etc., to communicate with each other.

[0125] Further, the communication network 509 may include a variety of network devices, including routers, bridges, servers, computing devices, storage devices, etc. In some embodiments, the processor 502 may be disposed in communication with a memory 505 (e.g., RAM, ROM, etc. not shown in FIG. 5) via a storage interface 504. The storage interface 504 may connect to memory 505 including, without limitation, memory drives, removable disc drives, etc., employing connection protocols such as Serial Advanced Technology Attachment (SATA), Integrated Drive Electronics (IDE), IEEE-1394, Universal Serial Bus (USB), fibre channel, Small Computer Systems Interface (SCSI), etc. The memory drives may further include a drum, magnetic disc drive, magneto-optical drive, optical drive, Redundant Array of Independent Discs (RAID), solid-state memory devices, solid-state drives, etc.

[0126] The memory 505 may store a collection of program or database components, including, without limitation, a user interface 506, an operating system 507, a web browser 508 etc. In some embodiments, the computer system　500 may store user / application data, such as the data, variables, records, etc. as described in this invention. Such databases may be implemented as fault-tolerant, relational, scalable, secure databases such as Oracle or Sybase.

[0127] Operating system　507　may facilitate resource management and operation of computer system　500. Examples of operating systems include, without limitation, APPLE®MACINTOSH® OS X®, UNIX®, UNIX-like system distributions (E.G., BERKELEY SOFTWARE DISTRIBUTION® (BSD), FREEBSD®, NETBSD®, OPENBSD, etc.), LINUX® DISTRIBUTIONS (E.G., RED HAT®, UBUNTU®, KUBUNTU®, etc.), IBM®OS / 2®, MICROSOFT® WINDOWS® (XP®, VISTA® / 7 / 8, 10 etc.), APPLE® IOS®, GOOGLE™ ANDROID™, BLACKBERRY® OS, or the like. User interface　506　may facilitate display, execution, interaction, manipulation, or operation of program components through textual or graphical facilities. For example, user interfaces may provide computer interaction interface elements on a display system operatively connected to computer system　500, such as cursors, icons, check boxes, menus, scrollers, windows, widgets, etc. Graphical User Interfaces (GUIs) may be employed, including, without limitation, Apple® Macintosh® operating systems' Aqua®, IBM® OS / 2®, Microsoft® Windows® (e.g., Aero, Metro, etc.), web interface libraries (e.g., ActiveX®, Java®,　Javascript®, AJAX, HTML, Adobe® Flash®, etc.), or the like.

[0128] The computer system　500　may implement web browser　508　stored program components. Web browser　508　may be a hypertext viewing application, such as MICROSOFT® INTERNET EXPLORER®, GOOGLE™ CHROME™, MOZILLA® FIREFOX®, APPLE® SAFARI®, etc. Secure web browsing may be provided using Secure Hypertext Transport Protocol (HTTPS), Secure Sockets Layer (SSL), Transport Layer Security (TLS), etc. Web browsers 508 may utilize facilities such as AJAX, DHTML, ADOBE® FLASH®, JAVASCRIPT®, JAVA®, Application Programming Interfaces (APIs), etc. The computer system 500　may implement a mail server stored program component. The mail server may be an Internet mail server such as Microsoft Exchange, or the like. The mail server may utilize facilities such as ASP, ACTIVEX®, ANSI® C++ / C#, MICROSOFT®,. NET, CGI SCRIPTS, JAVA®, JAVASCRIPT®, PERL®, PHP, PYTHON®, WEBOBJECTS®, etc. The mail server may utilize communication protocols such as Internet Message Access Protocol (IMAP), Messaging Application Programming Interface (MAPI), MICROSOFT® exchange, Post Office Protocol (POP), Simple Mail Transfer Protocol (SMTP), or the like. In some embodiments, the computer system 500 may implement a mail client stored program component. The mail client may be a mail viewing application, such as APPLE® MAIL, MICROSOFT® ENTOURAGE®, MICROSOFT® OUTLOOK®, MOZILLA® THUNDERBIRD®, etc.

[0129] Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present invention. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term "computer-readable medium" should be understood to include tangible items and exclude carrier waves and transient signals, i.e., non-transitory. Examples include Random Access Memory (RAM), Read-Only Memory (ROM), volatile memory, non-volatile memory, hard drives, Compact Disc (CD) ROMs, Digital Video Disc (DVDs), flash drives, disks, and any other known physical storage media.

[0130] An embodiment of the present disclosure the present disclosure provides a method and a system for handling inference configuration in a wireless communication network. The present disclosure facilitates generation of the inference configurations per serving cell ensures that AI / ML models used for physical-layer functionalities, such as beam management or channel state prediction, are tailored to the specific characteristics of each cell, including frequency, bandwidth, and subcarrier spacing. This approach provides fine-grained control and enables optimized inference performance across heterogeneous radio conditions.

[0131] According to embodiments of the disclosure a method of handling inference configuration in a wireless communication network is provided. The method comprises generating, by a Distributed Unit (DU), an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells associated with a User Equipment (UE), or each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with the UE; and transmitting, by the DU, the inference configuration to a Centralized Unit (CU). The inference configuration is used for performing inference of an artificial intelligence (AI) model associated with the UE. The UE receives the inference configuration for performing inference of an Artificial Intelligence (AI) model associated with the UE.

[0132] For example, the inference configuration is generated for each of the plurality of serving cells and the subsets of the plurality of serving cells, when the UE is configured with at least one of, a carrier aggregation mode or dual connectivity.

[0133] For example, the inference configuration is generated based on characteristics of a physical layer associated with one of the corresponding serving cell of the plurality of serving cells, the corresponding subset of the plurality of serving cells, the corresponding carrier of the plurality of carriers, or the corresponding subset of the plurality of carriers.

[0134] For example, the inference configuration is transmitted to the CU over an F1AP interface.

[0135] For example, the inference configuration is transmitted in one of, a CSI-ReportConfig information element or an extendedConfig information element, configured in a RRCReconfiguration message.

[0136] According to embodiments of the disclosure, a method of handling inference configuration in a wireless communication network is provided. The method comprises receiving, by a User Equipment (UE), an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells associated with the UE, or each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE; performing, by the UE, inference of AI models for each of the plurality of serving cells and the subsets of the plurality of serving cells associated with the UE, or each of the plurality of carriers and the subsets of the plurality of carriers, using the inference configuration; detecting, by the UE, occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models; and releasing, by the UE, the inference configuration based on the detection.

[0137] For example, the inference configuration is received via one of, a CSI-ReportConfig information element or an OtherConfig information element, configured in a RRCReconfiguration message.

[0138] For example, the inference configuration is received over an F1AP interface.

[0139] For example, the plurality of predefined events comprises receiving an explicit instruction to release the inference configuration; transitioning of the UE to an RRC_IDLE state or an RRC_INACTIVE state; or detecting a radio link failure.

[0140] For example, when the predefined event corresponds to the detecting of the radio link failure, the method comprises initiating an RRC reestablishment procedure, wherein the RRC reestablishment procedure comprises determining whether the UE is configured for mobility recovery procedures; and performing one of releasing the inference configuration during the RRC reestablishment procedure when the UE is not configured for the mobility recovery procedures, or performing the mobility recovery procedures when the UE is configured for mobility recovery procedures.

[0141] For example, performing the mobility recovery procedures comprises performing one of maintaining the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are successful; or releasing the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are not successful.

[0142] According to embodiments of the disclosure, a Distributed Unit (DU) for handling inference configuration in a wireless communication network is provided. The DU comprises a processor; and a memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to generate an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells associated with a User Equipment (UE), or each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with the UE; and transmit the inference configuration to a Centralized Unit (CU). The inference configuration is used for performing inference of an artificial intelligence (AI) model associated with the UE. The UE receives the inference configuration for performing inference of an Artificial Intelligence (AI) model associated with the UE.

[0143] For example, the processor is configured to generate the inference configuration based on characteristics of a physical layer associated with one of the corresponding serving cell of the plurality of serving cells, the corresponding subset of the plurality of serving cells, the corresponding carrier of the plurality of carriers, or the corresponding subset of the plurality of carriers.

[0144] For example, the processor is configured to transmit the inference configuration to the CU over an F1AP interface.

[0145] For example, the processor is configured to transmit the inference configuration in one of, a CSI-ReportConfig information element or an extendedConfig information element, configured in a RRCReconfiguration message.

[0146] According to embodiments of the disclosure, a User Equipment (UE) for handling inference configuration in a wireless communication network is provided. The UE comprises a processor; and a memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to receive an inference configuration for each of a plurality of serving cells and subsets of the plurality of serving cells associated with the UE, or each of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE; perform inference of AI models for each of the plurality of serving cells and the subsets of the plurality of serving cells associated with the UE, or each of the plurality of carriers and the subsets of the plurality of carriers, using the inference configuration; detect occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models; and release the inference configuration based on the detection.

[0147] For example, the processor is configured to receive the inference configuration via one of, a CSI-ReportConfig information element or an OtherConfig information element, configured in a RRCReconfiguration message.

[0148] For example, the processor is configured to receive the inference configuration over an F1AP interface.

[0149] For example, the plurality of predefined events comprises receiving an explicit instruction to release the inference configuration; transitioning of the UE to an RRC_IDLE state or an RRC_INACTIVE state; or detecting a radio link failure.

[0150] For example, when the predefined event corresponds to the detecting of the radio link failure, the processor is configured to initiate an RRC reestablishment procedure, wherein the RRC reestablishment procedure comprises determining whether the UE is configured for mobility recovery procedures; and performing one of releasing the inference configuration during the RRC reestablishment procedure when the UE is not configured for the mobility recovery procedures, or performing the mobility recovery procedures when the UE is configured for mobility recovery procedures.

[0151] For example, the processor is configured to perform the mobility recovery procedures by performing one of:

[0152] maintaining the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are successful; or

[0153] releasing the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are not successful.

[0154] A description of an embodiment with several components in communication with each other does not imply that all such components are required. On the contrary a variety of optional components are described to illustrate the wide variety of possible embodiments of the invention. When a single device or article is described herein, it will be apparent that more than one device / article (whether or not they cooperate) may be used in place of a single device / article. Similarly, where more than one device or article is described herein (whether or not they cooperate), it will be apparent that a single device / article may be used in place of the more than one device or article, or a different number of devices / articles　may be used instead of the shown number of devices or programs. The functionality and / or the features of a device may be alternatively embodied by one or more other devices which are not explicitly described as having such functionality / features. Thus, other embodiments of the invention need not include the device itself.

[0155] The specification has described a system and a method for performing edge federation in the federated network. The illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that on-going technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope and spirit of the disclosed embodiments. Also, the words "comprising," "having," "containing," and "including," and other similar forms are intended to be equivalent in meaning and be open-ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise.

[0156] Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based here on. Accordingly, the embodiments of the present invention are intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

[0157] Referral numeralsReference NumberDescription100Environment102Core network104Plurality of DUs106CU108UE202, 224I / O Interface204, 226Memory206, 228Processor208, 230Data210, 232Modules212, 234Inference data214, 238Other data216Generating module218Transmitting module220, 248Other modules236Detected data240Receiving module242Inference module244Detecting module246Releasing module500Exemplary computer system501I / O Interface of the exemplary computer system502Processor of the exemplary computer system503Network interface504Storage interface505Memory of the exemplary computer system506User interface507Operating system508Web browser509Communication network510Input devices511Output devices

Claims

1.A method of handling inference configuration in a wireless communication network, the method comprising:generating, by a distributed unit (DU), an inference configuration for:each of a plurality of serving cells and subsets of the plurality of serving cells associated with a user equipment (UE), oreach of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with the UE; andtransmitting, by the DU, the inference configuration to a centralized unit (CU), wherein the inference configuration is used for performing inference of an artificial intelligence (AI) model associated with the UE.2.The method of claim 1, wherein the inference configuration is generated for each of the plurality of serving cells and the subsets of the plurality of serving cells, when the UE is configured with at least one of, a carrier aggregation mode or dual connectivity.3.The method of claim 1, wherein the inference configuration is generated based on characteristics of a physical layer associated with one of,the corresponding serving cell of the plurality of serving cells,the corresponding subset of the plurality of serving cells,the corresponding carrier of the plurality of carriers, orthe corresponding subset of the plurality of carriers.4.The method of claim 1, wherein the inference configuration is transmitted to the CU over an F1AP interface.5.The method of claim 1, wherein the inference configuration is transmitted in one of, a CSI-ReportConfig information element or an extendedConfig information element, configured in a RRCReconfiguration message.6.A method of handling inference configuration in a wireless communication network, the method comprising:receiving, by a user equipment (UE), an inference configuration for:each of a plurality of serving cells and subsets of the plurality of serving cells associated with the UE, oreach of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE;performing, by the UE, inference of AI models for each of the plurality of serving cells and the subsets of the plurality of serving cells associated with the UE, or each of the plurality of carriers and the subsets of the plurality of carriers, using the inference configuration;detecting, by the UE, occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models; andreleasing, by the UE, the inference configuration based on the detection.7.The method of claim 6, wherein the inference configuration is received via one of, a CSI-ReportConfig information element or an OtherConfig information element, configured in a RRCReconfiguration message.8.The method of claim 6, wherein the inference configuration is received over an F1AP interface.9.The method of claim 6, wherein the plurality of predefined events comprises:receiving an explicit instruction to release the inference configuration;transitioning of the UE to an RRC_IDLE state or an RRC_INACTIVE state; ordetecting a radio link failure.10.The method of claim 9, wherein when the predefined event corresponds to the detecting of the radio link failure, the method comprises initiating an RRC reestablishment procedure, wherein the RRC reestablishment procedure comprises:determining whether the UE is configured for mobility recovery procedures; andperforming one of:releasing the inference configuration during the RRC reestablishment procedure when the UE is not configured for the mobility recovery procedures, orperforming the mobility recovery procedures when the UE is configured for mobility recovery procedures.11.The method of claim 10, wherein performing the mobility recovery procedures comprises performing one of:maintaining the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are successful; orreleasing the inference configuration during the RRC reestablishment procedure when the mobility recovery procedures are not successful.12.A distributed unit (DU) for handling inference configuration in a wireless communication network, comprising:a processor; anda memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to:generate an inference configuration for:each of a plurality of serving cells and subsets of the plurality of serving cells associated with a user equipment (UE), oreach of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell associated with the UE; andtransmit the inference configuration to a centralized unit (CU),wherein the inference configuration is used for performing inference of an artificial intelligence (AI) model associated with the UE.13.The DU of claim 12, wherein the processor is configured to generate the inference configuration based on characteristics of a physical layer associated with one ofthe corresponding serving cell of the plurality of serving cells,the corresponding subset of the plurality of serving cells,the corresponding carrier of the plurality of carriers, orthe corresponding subset of the plurality of carriers.14.The DU of claim 12, wherein the processor is configured to transmit the inference configuration to the CU over an F1AP interface.15.A user equipment (UE) for handling inference configuration in a wireless communication network, comprising:a processor; anda memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to:receive an inference configuration for:each of a plurality of serving cells and subsets of the plurality of serving cells associated with the UE, oreach of a plurality of carriers and subsets of the plurality of carriers supported by a single serving cell, associated with the UE;perform inference of AI models for each of the plurality of serving cells and the subsets of the plurality of serving cells associated with the UE, or each of the plurality of carriers and the subsets of the plurality of carriers, using the inference configuration;detect occurrence of at least one of a plurality of predefined events upon performing the inference of the AI models; andrelease the inference configuration based on the detection.

Images