Method and apparatus for ai / ML-based mobility measurement and reporting

The UE's AI/ML-based mobility configuration optimizes mobility operations by predicting and reporting measurement results, addressing the lack of efficient methods in current systems and enhancing prediction accuracy and reducing measurement effort.

WO2026239228A1PCT designated stage Publication Date: 2026-11-19FAINITY INNOVATION CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/IB2026/054589
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-15
Filing Date
2026-05-08
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Current wireless communication systems lack efficient configuration and activation/deactivation methods for AI/ML-based mobility, which is essential for enhancing mobility prediction accuracy and reducing measurement effort in next-generation systems like 6G.

Method used

A User Equipment (UE) receives an AI/ML-based Mobility (AMM) configuration including Measurement Identities and Report Configurations, enabling it to perform AI/ML inference for predicted measurement results and transmit them to a Base Station (BS), with options for skipping patterns and activation conditions to optimize mobility operations.

Benefits of technology

This approach enhances mobility prediction accuracy and reduces measurement effort by allowing the UE to predict and report future measurement results, thereby improving handover success rates and conserving power.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IB2026054589_19112026_PF_FP_ABST
    Figure IB2026054589_19112026_PF_FP_ABST
Patent Text Reader

Abstract

Methods and apparatus for Artificial Intelligence / Machine Learning (AI / ML)-based mobility measurement and reporting are provided. The method includes receiving, from a Base Station (BS), an AI / ML-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC), performing measurement based on a first MO of the multiple MOs, performing AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the multiple MOs, and transmitting, to the BS, a measurement report including the predicted measurement result associated with the second MO based on the RC.
Need to check novelty before this filing date? Find Prior Art

Description

METHOD AND APPARATUS FOR AI / ML-B ASED MOBILITY MEASUREMENT AND REPORTINGCROSS-REFERENCE TO RELATED APPLICATION! S)

[0001] The present disclosure claims the benefit of and priority to U.S. Provisional Patent Application Serial No. 63 / 806,055, filed on May 15, 2025, entitled “METHOD AND APPARATUS FOR LIFE CYCLE MANAGEMENT OF AI / ML-BASED MOBILITY,” the content of which is hereby fully incorporated herein by reference into the present disclosure.FIELD

[0002] The present disclosure relates to wireless communication, and in particular, to methods and apparatus for Artificial Intelligence / Machine Learning (AI / ML)-based mobility measurement and reporting.BACKGROUND

[0003] With the continued growth in the number of connected devices, the diversification of communication services, and the rapid increase in user / network traffic volume, various efforts have been made to improve different aspects of wireless communication for nextgeneration wireless communication systems, such as the sixth-generation (6G) system, by improving data rate, latency, reliability, mobility, coverage, energy efficiency, and intelligence.

[0004] The 6G system is designed to provide flexibility, configurability, and native intelligence for optimizing network services and types, and accommodating various use cases, such as immersive communication, hyper-reliable and low-latency communication, massive communication, ubiquitous connectivity, Artificial Intelligence (AI) / Machine Learning (ML)-integrated communication, and integrated sensing and communication.

[0005] However, as the demand for radio access continues to increase and as more diverse services and deployment scenarios emerge, there is a need for further improvements in wireless communication in next-generation wireless communication systems.SUMMARY

[0006] The present disclosure relates to wireless communication, and in particular, to methods and apparatus for AI / ML-based mobility measurement and reporting. The present disclosure provides configuration and reporting mechanisms for artificial intelligence and machine learning based mobility operations, enabling a user equipment to perform AI / MLinference based on measurement results associated with one measurement object to obtain predicted measurement results associated with another measurement object and to report the predicted measurement results based on a report configuration, thereby facilitating enhanced mobility prediction accuracy and reduced measurement effort.

[0007] According to a first aspect of the present disclosure, a User Equipment (UE) is provided. The UE includes at least one processor and at least one non-transitory computer-readable medium coupled to the at least one processor and storing one or more computerexecutable instructions that, when executed by the at least one processor, cause the UE to receive, from a Base Station (BS), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC), perform measurement based on a first MO of the multiple MOs, perform AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the multiple MOs, and transmit, to the BS, a measurement report including the predicted measurement result associated with the second MO based on the RC.

[0008] In some implementations of the first aspect of the present disclosure, the AMM configuration further includes a cell list indicating one or more cells applicable for AMM operation, and the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to perform the AI / ML inference for the one or more cells indicated by the cell list.

[0009] In some implementations of the first aspect of the present disclosure, the AMM configuration further includes a skipping pattern associated with the second MO, and the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to forgo measuring one or more measurement occasions associated with the second MO based on the skipping pattern.

[0010] In some implementations of the first aspect of the present disclosure, the skipping pattern includes a bitmap, each bit of the bitmap indicating to the UE whether to forgo measuring a corresponding measurement occasion.

[0011] In some implementations of the first aspect of the present disclosure, the RC indicates that the UE is to report at least one of the predicted measurement result, or a predicted measurement event.

[0012] In some implementations of the first aspect of the present disclosure, the one or more computer-executable instructions, when executed by the at least one processor, furthercause the UE to transmit, to the BS, capability information indicating a maximum prediction window length supported by the UE for the AI / ML inference.

[0013] In some implementations of the first aspect of the present disclosure, the AMM configuration is associated with an activation condition, and performing the AI / ML inference includes performing the AI / ML inference in response to determining that the activation condition is satisfied.

[0014] In some implementations of the first aspect of the present disclosure, in response to determining that the activation condition is satisfied, the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to report, based on the RC, at least one of the predicted measurement result associated with a future time instance or a predicted measurement event, and forgo measuring one or more measurement occasions on one or more cells associated with the second MO.

[0015] In some implementations of the first aspect of the present disclosure, the second MO is associated with one or more inter-frequency cells, and in response to determining that the activation condition is satisfied, the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to report, based on the RC, the predicted measurement result associated with the one or more inter-frequency cells, and forgo measuring one or more measurement occasions on one or more cells associated with the second MO.

[0016] In some implementations of the first aspect of the present disclosure, the activation condition includes at least one of an identity of a serving cell being in a cell list included in the AMM configuration, a radio link quality of the serving cell satisfying a first threshold, a radio link quality of a neighbor cell satisfying a second threshold, or a speed of the UE satisfying a speed threshold.

[0017] In some implementations of the first aspect of the present disclosure, in response to determining that the activation condition is no longer satisfied, the one or more computerexecutable instructions, when executed by the at least one processor, further cause the UE to comply with a measurement requirement on one or more measurement occasions associated with the second MO.

[0018] According to a second aspect of the present disclosure, a method performed by a User Equipment (UE) is provided. The method includes receiving, from a Base Station (BS), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC), performing measurementbased on a first MO of the multiple MOs, performing AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the multiple MOs, and transmitting, to the BS, a measurement report including the predicted measurement result associated with the second MO based on the RC.

[0019] In some implementations of the second aspect of the present disclosure, the AMM configuration further includes a cell list indicating one or more cells applicable for AMM operation, and performing the AI / ML inference includes performing the AI / ML inference for the one or more cells indicated by the cell list.

[0020] In some implementations of the second aspect of the present disclosure, the AMM configuration further includes a skipping pattern associated with the second MO, and the method further includes forgoing measuring one or more measurement occasions associated with the second MO based on the skipping pattern.

[0021] In some implementations of the second aspect of the present disclosure, the AMM configuration is associated with an activation condition, and performing the AI / ML inference includes performing the AI / ML inference in response to determining that the activation condition is satisfied.

[0022] In some implementations of the second aspect of the present disclosure, the RC indicates that the UE is to report at least one of the predicted measurement result, or a predicted measurement event.

[0023] According to a third aspect of the present disclosure, a Base Station (BS) is provided. The BS includes at least one processor and at least one non -transitory computer-readable medium coupled to the at least one processor and storing one or more computerexecutable instructions that, when executed by the at least one processor, cause the BS to transmit, to a User Equipment (UE), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC), a first MO of the multiple MOs being configured for measurement at the UE, a second MO of the multiple MOs being configured for AI / ML inference at the UE, and receive, from the UE, a measurement report including a predicted measurement result associated with the second MO based on the RC.

[0024] In some implementations of the third aspect of the present disclosure, the AMM configuration further includes a cell list indicating one or more cells applicable for AMM operation.

[0025] In some implementations of the third aspect of the present disclosure, the AMMconfiguration further includes a skipping pattern associated with the second MO.

[0026] In some implementations of the third aspect of the present disclosure, the AMM configuration is associated with an activation condition for the AI / ML inference at the UE.BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Aspects of the present disclosure are best understood from the following detailed disclosure when read with the accompanying drawings. Various features are not drawn to scale. Dimensions of various features may be arbitrarily increased or reduced for clarity of discussion.

[0028] FIG. 1 is a flowchart illustrating a method / process for life cycle management of AI / ML-based mobility, according to an example implementation of the present disclosure.

[0029] FIG. 2 is a flowchart illustrating a method / process for life cycle management of AI / ML-based mobility, according to an example implementation of the present disclosure.

[0030] FIG. 3 is a block diagram illustrating a node for wireless communication, according to an example implementation of the present disclosure.DESCRIPTION

[0031] The following description includes specific information pertaining to example implementations in the present disclosure. The drawings in the present disclosure and the accompanying detailed description are directed to merely example implementations. However, the present disclosure is not limited to merely these example implementations. Other variations and implementations of the present disclosure will occur to those skilled in the art. Unless noted otherwise, like or corresponding elements among the figures may be indicated by like or corresponding reference numerals. Moreover, the drawings and illustrations in the present disclosure are generally not to scale and are not intended to correspond to actual relative dimensions.

[0032] For the purpose of consistency and ease of understanding, like features may be identified (although, in some examples, not shown) by the same numerals in the example figures. However, the features in different implementations may differ in other respects, and thus shall not be narrowly confined to what is shown in the figures.

[0033] The description uses the phrases "in one implementation," or "in some implementations," which may each refer to one or more of the same or different implementations. The term "coupled" is defined as connected, whether directly or indirectly through intervening components, and is not necessarily limited to physical connections. The term "including," when utilized, indicates "including, but not necessarily limited to"; the termspecifically indicates open-ended inclusion or membership in the so-described combination, group, series and the equivalent. The expression "at least one of A, B and C" or "at least one of the following: A, B and C" indicates "only A, or only B, or only C, or any combination of A, B and C."

[0034] Additionally, for the purposes of explanation and non-limitation, specific details, such as functional entities, techniques, protocols, standards, and the like are set forth for providing an understanding of the described technology. In other examples, detailed description of well-known methods, technologies, systems, architectures, and the like are omitted so as not to obscure the description with unnecessary details.

[0035] Persons skilled in the art will immediately recognize that any network function(s) or algorithm(s) described in the present disclosure may be implemented by hardware, software or a combination of software and hardware. Described functions may correspond to modules which may be software, hardware, firmware, or any combination thereof. The software implementation may include computer executable instructions stored on a computer readable medium such as a memory or other type of storage device. For example, one or more microprocessors or general-purpose computers with communication processing capability may be programmed with corresponding executable instructions and carry out the described network function(s) or algorithm(s). The microprocessors or general-purpose computers may be formed of Applications Specific Integrated Circuitry (ASIC), programmable logic arrays, and / or using one or more Digital Signal Processors (DSPs). Although some of the example implementations described in this specification are oriented to software installed and executing on computer hardware, nevertheless, alternative example implementations implemented as firmware or as hardware or a combination of hardware and software are well within the scope of the present disclosure.

[0036] The computer readable medium includes but is not limited to Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), flash memory, Compact Disc Read-Only Memory (CD-ROM), magnetic cassettes, magnetic tape, magnetic disk storage, or any other equivalent medium capable of storing computer-readable instructions.

[0037] A radio communication network architecture (e.g., a Long Term Evolution (LTE) system, an LTE- Advanced (LTE- A) system, an LTE- Advanced Pro system, or a 5GNew Radio (NR) Radio Access Network (RAN)) typically includes at least one Base Station (BS), at least one User Equipment (UE), and one or more optional network elements that provide connectiontowards a network. The UE may communicate with the network (e.g., a Core Network (CN), an Evolved Packet Core (EPC) network, an Evolved Universal Terrestrial Radio Access Network (E-UTRAN), a 5G Core (5GC), or an internet), through a RAN established by one or more BSs.

[0038] In the present disclosure, a UE may include, but is not limited to, a mobile station, a mobile terminal or device, or a user communication radio terminal. For example, a UE may be a portable radio equipment, which includes, but is not limited to, a mobile phone, a tablet, a wearable device, a sensor, a vehicle, or a Personal Digital Assistant (PDA) with wireless communication capability. The UE may be configured to receive and transmit signals over an air interface to one or more cells in a radio access network.

[0039] ABS may be configured to provide communication services according to at least one of the following Radio Access Technologies (RATs): Worldwide Interoperability for Microwave Access (WiMAX), Global System for Mobile communications (GSM, often referred to as 2G), GSM Enhanced Data rates for GSM Evolution (EDGE) Radio Access Network (GERAN), General Packet Radio Service (GPRS), Universal Mobile Telecommunication System (UMTS, often referred to as 3G) based on basic Wideband-Code Division Multiple Access (W-CDMA), High-Speed Packet Access (HSPA), LTE, LTE-A, eLTE (evolved LTE, e.g., LTE connected to 5GC), NR (often referred to as 5G), and / or LTE-A Pro. However, the scope of the present disclosure should not be limited to the above-mentioned protocols.

[0040] ABS may include, but is not limited to, a node B (NB) as in the UMTS, an evolved Node B (eNB) as in the LTE or LTE-A, a Radio Network Controller (RNC) as in the UMTS, a Base Station Controller (BSC) as in the GSM / GERAN, a ng-eNB as in an Evolved Universal Terrestrial Radio Access (E-UTRA) BS in connection with the 5GC, a next generation Node B (gNB) as in the 5G-RAN, and any other apparatus capable of controlling radio communication and managing radio resources within a cell. The BS may serve one or more UEs through a radio interface.

[0041] The BS may be operable to provide radio coverage to a specific geographical area using multiple cells forming the radio access network. The BS may support the operations of the cells. Each cell may be operable to provide services to at least one UE within the radio coverage of the cell. More specifically, each cell (often referred to as a serving cell) may provide services to serve one or more UEs within the radio coverage of the cell (e.g., each cell may schedule the downlink and optionally uplink resources to at least one UE within the radio coverage of the cell for downlink and optionally uplink packet transmissions). The BS maycommunicate with one or more UEs in the radio communication system through multiple cells. A cell may allocate Sidelink (SL) resources for supporting Proximity Service (ProSe) or Vehicle to Everything (V2X) service. Each cell may have overlapped coverage areas with other cells.

[0042] In some communication systems, the frame structure for NR may support flexible configurations for accommodating various next generation (e.g., 5G) communication requirements, such as Enhanced Mobile Broadband (eMBB), Massive Machine Type Communication (mMTC), and Ultra-Reliable and Low-Latency Communication (URLLC), while fulfilling high reliability, high data rate and low latency requirements. The Orthogonal Frequency-Division Multiplexing (OFDM) technology may serve as a baseline for NR waveform. The scalable OFDM numerology, such as the adaptive sub-carrier spacing, the channel bandwidth, and the Cyclic Prefix (CP) may also be used. Additionally, two coding schemes may be considered for NR: Low-Density Parity-Check (LDPC) code and Polar Code. The coding scheme adaptation may be configured based on the channel conditions and / or the service applications.

[0043] Moreover, in a transmission time interval of a single NR frame, a Downlink (DL) transmission data, a guard period, and an Uplink (UL) transmission data may at least be included, where the respective portions of the DL transmission data, the guard period, and the UL transmission data may also be configurable, for example, based on the network dynamics of NR. In addition, SL resources may also be provided in an NR frame to support ProSe services or V2X services.

[0044] In addition, the terms "system" and "network" herein may be used interchangeably. The term "and / or" herein is only an association relationship for describing associated objects, and represents that three relationships may exist. For example, A and / or B may indicate that: A exists alone, A and B exist at the same time, or B exists alone. In addition, the character " / " herein generally represents that the former and latter associated objects are in an "or" relationship. Multiple Public Land Mobile Networks (PLMNs) may operate on the unlicensed spectrum. Multiple PLMNs may share the same unlicensed carrier. The PLMNs may be public or private. Public PLMNs may be (but are not limited to) the operators or virtual operators, which may provide radio services to the public subscribers. Public PLMNs may own the licensed spectrum and may support the radio access technology on the licensed spectrum as well. Private PLMNs may be (but are not limited to) the micro-operators, factories, or enterprises, which may provide radio services to the private users of the private PLMNs (e.g., employees or machines). In some implementations, public PLMNs may support moredeployment scenarios (e.g., carrier aggregation between licensed band NR (PCell) and NR-U (SCell), dual connectivity between licensed band LTE (PCell) and NR-U (PSCell), stand-alone NR-U, an NR cell with DL in unlicensed band and UL in licensed band, dual connectivity between licensed band NR (PCell) and NR-U (PSCell)). In some implementations, private PLMNs may mainly support (but are not limited to) the stand-alone unlicensed radio access technology (e.g., stand-alone NR-U).

[0045] Terminology

[0046] The network (NW), cell, camped cell, serving cell, base station, gNB, eNB and ng-eNB may be used interchangeably in the present disclosure. In some implementations, some of these items may refer to the same network entity.

[0047] The RAT may be (but is not limited to) NR, LTE, E-UTRA connected to 5GC, LTE connected to 5GC, E-UTRA connected to EPC, and LTE connected to EPC. The proposed mechanism may be applied for UEs in public networks, or in private networks (e.g., Non-Public Network (NPN), Standalone NPN (SNPN), Public Network Integrated NPN (PNI-NPN)).

[0048] The proposed mechanism may be used for licensed frequency and / or unlicensed frequency. In addition, the proposed mechanism of conditional configuration selection may be applied for the cases that a UE experiences a radio link failure when configured with conditional configurations.

[0049] System information (SI) may refer to one or more of a Master Information Block (MIB), System Information Block 1 (SIB1), and other SI. Minimum SI may include MIB and SIB1. Other SI may refer to one or more of SIB3, SIB4, SIB5, and other SIB(s).

[0050] Dedicated signaling may refer to (but is not limited to) RRC message(s). For example, dedicated signaling may include one or more of an RRC (Connection) Setup Request message, an RRC (Connection) Setup message, an RRC (Connection) Setup Complete message, an RRC (Connection) Reconfiguration message, an RRC Connection Reconfiguration message including the mobility control information, an RRC Connection Reconfiguration message without the mobility control information inside, an RRC Reconfiguration message including the configuration with sync, an RRC Reconfiguration message without the configuration with sync inside, an RRC (Connection) Reconfiguration Complete message, an RRC (Connection) Resume Request message, an RRC (Connection) Resume message, an RRC (Connection) Resume Complete message, an RRC (Connection) Reestablishment Request message, an RRC (Connection) Reestablishment message, an RRC (Connection) Reestablishment Complete message, an RRC (Connection) Reject message, an RRC (Connection) Release message, an RRC System Information Request message, a UEAssistance Information message, a UE Capability Enquiry message, and a UE Capability Information message.

[0051] The RRC CONNECTED UE, RRC INACTIVE UE, and RRC IDLE UE may apply the proposed implementations.

[0052] The source cell may be a suitable cell or an acceptable cell.

[0053] A suitable cell may be a cell on which a UE may camp. The UE may consider a cell as suitable if the following conditions are fulfilled: (1) the cell is part of either the selected PLMN or the registered PLMN or PLMN of the Equivalent PLMN list, and (2) the cell criteria of the cell are fulfilled. Furthermore, according to the latest information provided by NAS, the suitable cell may not be barred. The suitable cell may be part of at least one TA that is not part of the list of "Forbidden Tracking Areas", which may belong to a PLMN that fulfils the condition (1).

[0054] An acceptable cell may be a cell on which the UE may camp to obtain limited service (originate emergency calls and receive ETWS and CMAS notifications). An acceptable cell may fulfil the following requirements, which may be the minimum set of requirements to initiate an emergency call and to receive ETWS and CMAS notification in an NR network: (1) the cell is not barred, and / or (2) the cell selection criteria are fulfilled.

[0055] PCell (Primary Cell) may correspond to the MCG cell, operating on the primary frequency, in which the UE either performs the initial connection establishment procedure or initiates the connection re-establishment procedure.

[0056] PSCell (Primary SCG Cell) may correspond to, for dual connectivity operation, the SCG cell in which the UE performs random access when performing the Reconfiguration with Sync procedure.

[0057] Serving Cell may correspond to, for a UE in RRC CONNECTED not configured with CA / DC, only one serving cell including the primary cell. For a UE in RRC CONNECTED configured with CA / DC the term 'serving cells' may be used to denote the set of cells including the Special Cell(s) and all secondary cells.

[0058] Secondary Cell may correspond to, for a UE configured with CA, a cell providing additional radio resources on top of the Special Cell.

[0059] Special Cell may correspond to, for Dual Connectivity operation, the PCell of the MCG or the PSCell of the SCG; otherwise the term Special Cell may refer to the PCell.

[0060] Master Cell Group may correspond to, in MR-DC, a group of serving cells associated with the Master Node, including the SpCell (PCell) and optionally one or more SCells.

[0061] Master node may correspond to, in MR-DC, the radio access node that provides the control plane connection to the core network. The master node may be a Master eNB (in EN-DC), a Master ng-eNB (in NGEN-DC), or a Master gNB (in NR-DC and NE-DC).

[0062] Secondary Cell Group may correspond to, in MR-DC, a group of serving cells associated with the Secondary Node, including the SpCell (PSCell) and optionally one or more SCells.

[0063] Secondary node may correspond to, in MR-DC, the radio access node, with no control plane connection to the core network, providing additional resources to the UE. The secondary node may be an en-gNB (in EN-DC), a Secondary ng-eNB (in NE-DC), or a Secondary gNB (in NR-DC and NGEN-DC).

[0064] The serving cell in the implementations may be a PCell, an SCell, or a PSCell.

[0065] The term "source node" may refer to the node from which the UE receives a Cell Switch Command (CSC), and the term may be interpreted as a source MN, a source SN, a source PCell, a source PSCell, or a source gNB.

[0066] The term "source cell" may refer to the cell from which the UE receives a CSC, and the term may be interpreted as a source PCell, or a source PSCell.

[0067] The term "candidate node" may refer to the node which is associated with the LTM candidate configuration stored by the UE, and the term may be interpreted as a candidate MN, a candidate SN, a candidate PCell, a candidate PSCell, or a candidate gNB.

[0068] The term "candidate cell" may refer to the cell which is associated with the LTM candidate configuration stored by the UE, and the term may be interpreted as a candidate PCell, or a candidate PSCell.

[0069] The term "target node" may refer to the node which is associated with the LTM candidate configuration ID in the CSC received by the UE, and the term may be interpreted as a target MN, a target SN, a target PCell, a target PSCell, or a target gNB.

[0070] The term "target cell" may refer to the cell which is associated with the LTM candidate configuration ID in the CSC received by the UE, and the term may be interpreted as a target PCell, or a target PSCell.

[0071] The system information may be associated with the serving cell and / or the system information may be associated with the candidate / target cell.

[0072] AI / ML-based Mobility Overview

[0073] Artificial Intelligence / Machine Learning (AI / ML) has become a prevalent prediction method for communication systems. AI / ML may be used for the enhancement of mobility performance in wireless communication systems. Two objectives of AI / ML-basedmobility may be to enhance the Radio Resource Management (RRM) prediction accuracy and to reduce the RRM measurement effort. For example, AI / ML may aide the mobility procedure to achieve these objectives, and as a result, mobility performance such as handover failure rate and ping-pong rate may be improved. In addition, the UE may save power by skipping some measurement occasions which the UE may otherwise have measured to comply with the measurement requirements. Accordingly, the present disclosure addresses AI / ML mobility design, mainly for the life cycle management, aiming to create a flexible, robust, and signaling efficient AI / ML operation in mobility.

[0074] So far, there is no configuration method for the configuration and activation / deactivation of AI / ML-based mobility in NR Release 19. As a result, the present disclosure aims to provide potential solutions for the configuration method as well as the activation / deactivation of the configuration for the management of AI / ML-based mobility. Also, the present disclosure aims at possible next generation design, so the disclosed procedures and signaling may be applicable to the next generation wireless communication (e.g., 6G).

[0075] Scenarios

[0076] In the present disclosure, the focus is on the scenario that the AI / ML models / functionalities are operated (e.g., trained, or implemented, or monitored) at the UE-side (e.g., at the UE, or at a UE-side Over-The-Top (OTT) server). The UE-side OTT server may be located in the RAN nodes and / or in the CN nodes, or may be located in a node independent from RAN and CN nodes. In some other implementations, the AI / ML models / functionalities may be operated at the network side, and the techniques described herein may be adapted accordingly.

[0077] Capability Report

[0078] In some implementations, the UE may receive a first RRC message (e.g., a UECapabilityEnquiry message) from the gNB, where the first RRC message may contain an indication of enquiry of the AI / ML capability for mobility. For example, the first RRC message may request the UE to report the AI / ML-related capabilities of the UE, allowing the network to determine what AI / ML functionalities may be supported by the UE for mobility operations.

[0079] In some implementations, the indication of enquiry of the AI / ML capability for mobility may contain one bit indicating that all the capabilities related to AI / ML for mobility are enquired.

[0080] In some implementations, the indication of enquiry of the AI / ML capability for mobility may contain multiple information elements (IES), where each present IE may indicate one kind of model-type sub-capability associated with AI / ML for mobility.

[0081] In some implementations, the indication of enquiry of the AI / ML capability for mobility may contain multiple IES, where each present IE may indicate one kind of schemetype sub-capability associated with AI / ML for mobility.

[0082] In some implementations, the indication of enquiry of the AI / ML capability for mobility may contain multiple IEs, where each present IE may indicate one kind of model-type sub-capability or scheme-type sub-capability associated with AI / ML for mobility in RRC CONNECTED and / or RRC IDLE and / or RRC INACTIVE state.

[0083] A model-type sub-capability associated with AI / ML for mobility may correspond to a particular type of AI / ML prediction model that the UE may support. The model-type subcapability may take various forms depending on the prediction target, the frequency domain, and the time domain, and several illustrative forms are described below.

[0084] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using AI / ML for the prediction of intra-frequency RRM measurement results in future time instances. For example, the UE may be capable of predicting the RRM measurement results of the intra-frequency measurements for Synchronization Signal Block (SSB) or Channel State Information Reference Signal (CSLRS) in future time instances. This capability may enable the UE to predict measurement results on the same frequency without performing actual measurements, thereby reducing measurement effort and UE power consumption.

[0085] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using AI / ML for the prediction of inter-frequency RRM measurement results in future time instances. For example, the UE may be capable of predicting the RRM measurement results of the inter-frequency measurements for SSB or CSL RS in future time instances. More specifically, the capability may be different based on the length and support of a measurement gap. The enquiry may further request the supporting measurement gap length under this capability query. This capability may enable the UE to reduce or eliminate the need for measurement gaps for inter-frequency measurements in the future, thereby improving resource utilization.

[0086] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using AI / ML for the prediction of inter-frequency RRM measurement results in the current time instance. For example, the UE may be capable of predicting the RRM measurement results of the inter-frequency measurements for SSB or CSL RS in the current time instance. More specifically, the capability may be different based on the length and support of a measurement gap. The enquiry may further request the supportingmeasurement gap length under this capability query. This capability may enable the UE to obtain inter-frequency measurement results without requiring actual measurement gap configurations in the current time instance.

[0087] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using cluster-based AI / ML to predict the RRM measurement results. For example, the UE may be capable of using the measurement results of a first group of cells to predict the measurement results of a second group of cells in the current time instance or in future time instances. More specifically, the cells in the first group of cells may be the same as, or may be totally different from, or may be partially different from, the cells in the second group of cells. More specifically, the cells in the first group of cells may be operated in the same frequency or a different frequency or a partially different frequency and the cells in the second group of cells may be operated in the same frequency or a different frequency or a partially different frequency. This capability may enable the UE to leverage spatial correlation among cells to predict measurement results of cells that the UE may not have directly measured.

[0088] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using AI / ML for the prediction of inter-RAT RRM measurement results in the current and / or future time instances. For example, the UE may be capable of using the measurement results of a first cell in a first RAT to predict the measurement results of a second cell in a second RAT. More specifically, the first RAT may be 6G RAT, NR, LTE, or other radio access technologies. More specifically, the second RAT may be 6G RAT, NR, LTE, or other radio access technologies, and the second RAT may be different from the first RAT. More specifically, the first RAT and the second RAT may be operated in the same frequency or a different frequency.

[0089] In some implementations, a model-type sub -capability associated with AI / ML for mobility may be the capability of using AI / ML for the prediction of Non-Terrestrial Network (NTN)-to-Terrestrial Network (TN) and / or TN-to-NTN RRM measurement results in the current and / or future time instances. For example, the UE may be capable of using the measurement results of a TN cell to predict the measurement results of an NTN cell in the current and / or future time instances. For example, the UE may be capable of using the measurement results of an NTN cell to predict the measurement results of a TN cell in the current and / or future time instances.

[0090] The above-described model-type sub-capabilities are provided as illustrative examples, and the UE may support any one or any combination of the model -type subcapabilities. For example, the UE may simultaneously support the capability of using AI / MLfor intra-frequency prediction and the capability of using cluster-based AI / ML prediction. The specific combination of supported model-type sub-capabilities may depend on the UE implementation and the deployment scenario.

[0091] A scheme-type sub-capability associated with AI / ML for mobility may correspond to a particular type of AI / ML prediction scheme that the UE may support. The scheme-type sub-capability may take various forms depending on the prediction approach and the target event, and several illustrative forms are described below.

[0092] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for the direct prediction of the occurrence of a Measurement Event (ME). For example, the UE may predict the probability of ME occurrence in one or more future time instances. For example, the UE may predict the probability of ME occurrence in a future time period. For example, the UE may predict whether or not the ME will occur in one or more future time instances. For example, the UE may predict whether or not the ME will occur in a future time period.

[0093] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for the indirect prediction of the occurrence of an ME. For example, the UE may predict the time instance / period of an ME occurrence based on predicted RRM measurement results.

[0094] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for the direct prediction of the occurrence of Radio Link Failure (RLF). For example, the UE may predict the probability of an RLF occurrence in one or more future time instances. For example, the UE may predict the probability of RLF occurrence in a future time period. For example, the UE may predict whether or not an RLF will occur in one or more future time instances. For example, the UE may predict whether or not an RLF will occur in a future time period.

[0095] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for the indirect prediction of the occurrence of RLF. For example, the UE may predict the time instance / period of an RLF occurrence based on predicted RRM measurement results.

[0096] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for Conditional Handover (CHO). For example, the UE may execute the CHO by applying a stored RRC configuration according to predicted RRM measurement results. For example, the UE may execute the CHO by applying a stored RRC configuration according to predicted measurement events. For example, the UE mayexecute the CHO by applying a stored RRC configuration according to predicted RLF occurrence. For example, upon the UE predicting an RLF occurrence, the UE may perform a cell selection procedure and select a cell; then the LE may execute the CHO towards the cell if the cell is one of a CHO candidate cell. The LE may perform the cell selection procedure according to the measurement results of the current time instance (e.g., either measured results or predicted results) or according to the measurement results of future time instances (e.g., predicted results). In some implementations, the LE may execute the CHO upon actual RLF. In some implementations, the LE may execute the CHO prior to the actual RLF (e.g., at the predicted RLF timing, or when the LE is considering that RLF has not happened). In some implementations, the LE may execute the CHO at a certain timing before the predicted RLF timing, where the LE may determine the certain timing by the LE implementation, by an IE in the CHO configuration, or by an IE in the system information block of the serving cell. The LE may (1) still report RLF to the network based on the prediction even though CHO is executed before RLF actually happened or (2) not report RLF if the RLF does not actually happen. For example, the UE may suggest an RRC configuration for the CHO according to predicted RRM measurement results and / or predicted measurement events. More specifically, the suggested configuration may include the trigger conditions and candidate cells. With the suggestion, the LE may indicate that the LE has confidence that the CHO may execute within a time period if the suggested RRC configuration is applied. This capability may enable the LE to proactively execute handover procedures based on predicted future conditions, reducing handover failure rates and improving mobility robustness.

[0097] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for Layer 1 / Layer 2-triggered mobility (LTM). For example, the UE may predict the Layer 1 measurement results with AI / ML. In some implementations, the Layer 1 measurement results may be associated with a beam of a cell other than the serving cell. For example, the LE may perform LTM execution by applying a stored RRC configuration according to an indication (e.g., a cell switch command MAC CE) from the network. More specifically, the Layer 1 measurement results may be the Layer 1 measurement results of the source cell or may be the Layer 1 measurement results of one or more of the LTM candidate cells. More specifically, the Layer 1 measurement results may represent the Layer 1 measurement results of one or more SSBs, or may represent the Layer 1 measurement results of one or more CSLRSs. For example, the LE may execute the LTM by applying a stored RRC configuration according to predicted RLF occurrence. For example, upon the LE predicting an RLF occurrence, the UE may perform a cell selection procedure andselect a cell; then the UE may execute the LTM towards the cell if the cell is one of an LTM candidate cell. The UE may perform the cell selection procedure according to the measurement results of the current time instance (e.g., either measured results or predicted results) or according to the measurement results of future time instances (e.g., predicted results). In some implementations, the UE may execute the LTM upon actual RLE In some implementations, the UE may execute the LTM prior to the actual RLF (e.g., at the predicted RLF timing, or when the UE is considering that RLF has not happened). In some implementations, the UE may execute the LTM at a certain timing before the predicted RLF timing, where the UE may determine the certain timing by the UE implementation, by an IE in the LTM configuration, or by an IE in the system information block of the serving cell.

[0098] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for conditional LTM (C-LTM). For example, the UE may predict the Layer 1 measurement event with AI / ML (or predict the Layer 1 measurement event with predicted Layer 1 measurement results by AI / ML). For example, the UE may perform early Transmission Configuration Indicator (TCI) state activation for an LTM candidate cell according to AI / ML inference results. For example, the UE may perform early uplink synchronization for an LTM candidate cell according to AI / ML inference results. For example, the UE may perform C-LTM execution by applying a stored RRC configuration according to AI / ML inference results. More specifically, the AI / ML inference results may be predicted Layer 1 measurement results or Layer 3 measurement results. The Layer 3 measurement results may be directly or indirectly predicted. More specifically, the Layer 1 measurement results may be the Layer 1 measurement results of the source cell or may be the Layer 1 measurement results of one or more of the LTM candidate cells. More specifically, the Layer 1 measurement results may represent the Layer 1 measurement results of one or more SSBs, or may represent the Layer 1 measurement results of one or more CSLRSs. For example, the UE may execute the C-LTM by applying a stored RRC configuration according to predicted RLF occurrence. For example, upon the UE predicting an RLF occurrence, the UE may perform a cell selection procedure and select a cell; then the UE may execute the C-LTM towards the cell if the cell is one of a C-LTM candidate cell. The UE may perform the cell selection procedure according to the measurement results of the current time instance (e.g., either measured results or predicted results) or according to the measurement results of future time instances (e.g., predicted results). In some implementations, the UE may execute the C-LTM upon actual RLF. In some implementations, the UE may execute the C-LTM prior to the actual RLF (e.g., at the predicted RLF timing, or when the UE is considering that RLF has nothappened). In some implementations, the UE may execute the C-LTM at a certain timing before the predicted RLF timing, where the UE may determine the certain timing by the UE implementation, by an IE in the C-LTM configuration, or by an IE in the system information block of the serving cell.

[0099] In some implementations, a scheme-type sub-capability associated with AI / ML for mobility may be the capability of using AI / ML for Conditional PSCell Addition and / or Change (CPAC) and subsequent CPAC (S-CPAC). For example, the UE may execute the CPAC or S-CPAC by applying a stored RRC configuration according to AI / ML inference results. More specifically, the AI / ML inference results may be predicted Layer 1 measurement results or Layer 3 measurement results.

[0100] The above-described scheme-type sub-capabilities are provided as illustrative examples, and the UE may support any one or any combination of the scheme-type subcapabilities. For example, the UE may simultaneously support the capability of using AI / ML for direct ME prediction and the capability of using AI / ML for CHO. The specific combination of supported scheme-type sub-capabilities may depend on the UE implementation and the deployment scenario. Furthermore, any of the model-type sub-capabilities described above may be used in combination with any of the scheme-type sub-capabilities described herein. For example, the capability of using AI / ML for inter-frequency RRM prediction may be combined with the capability of using AI / ML for CHO to enable the UE to predict inter-frequency measurement results and use the predicted results for conditional handover decisions.

[0101] In some implementations, in response to the reception of the first RRC message (e.g., if the first RRC message includes the indication of requesting the capability of AI / ML for mobility), the UE may transmit a second RRC message (e.g., a UECapability Information message) to the gNB to indicate the capability of the UE for AI / ML for mobility.

[0102] In some implementations, the UE may indicate that the UE supports the AI / ML for mobility in the second RRC message.

[0103] In some implementations, the UE may indicate that the UE supports one or more model-type sub-capabilities associated with AI / ML for mobility in the second RRC message. More specifically, the UE may do so if the network enquires the one or more model-type subcapabilities associated with AI / ML for mobility in the first RRC message.

[0104] In some implementations, the UE may indicate that the UE supports one or more scheme-type sub-capabilities associated with AI / ML for mobility in the second RRC message. More specifically, the UE may do so if the network enquires the one or more scheme-type subcapabilities associated with AI / ML for mobility in the first RRC message.

[0105] In some implementations, the UE may indicate model-related parameters of the UE in the second RRC message. For example, the UE may indicate the information of frequency bands associated with the model. For example, the UE may indicate the information of supported RATs associated with the model. For example, the UE may indicate the clusters of cells associated with the model. For example, the UE may indicate the acceptable / preferred skipping patterns of measurements. For example, the UE may indicate the maximum Measurement Reduction Rate in Time (MRRT). For example, the UE may indicate the maximum Measurement Reduction Rate in Space (MRRS). For example, the UE may indicate the maximum prediction window length. For example, the UE may indicate the maximum observation window length. For example, the UE may indicate the maximum ratio of prediction window length to observation window length. These model-related parameters may enable the network to configure the AI / ML-based mobility operation in a manner that is compatible with the capabilities and limitations of the AI / ML model at the UE.

[0106] In some implementations, if the UE indicates the capability of using AEML for mobility and the UE additionally indicates the capability of applying CHO and / or LTM and / or C-LTM and / or CPAC, the combined indication may imply that the UE may use AI / ML for CHO and / or LTM and / or C-LTM and / or CPAC if both indications show the support. In some implementations, even if the UE indicates the support of using AI / ML for mobility and the UE additionally indicates the support of applying CHO and / or LTM and / or C-LTM and / or CPAC, the combined indication may not imply that the UE may use AI / ML for CHO and / or LTM and / or C-LTM and / or CPAC, and a separate indication may be needed. In some other implementations, the capability of using AI / ML for mobility may be signaled via a MAC CE or a DCI, or a combination thereof.

[0107] Configuration of AI / ML-based Mobility

[0108] The following section describes the configuration of AI / ML-based Mobility (AMM). The AMM configuration may enable the network to configure the UE with AI / ML-related measurement and reporting parameters for mobility operations. The implementations described in this section are provided as illustrative examples, and variations, modifications, and alternative implementations are possible without departing from the scope of the present disclosure.

[0109] In some implementations, the UE may receive the configuration related to AI / ML-based mobility (AMM) via RRC signaling. The AMM configuration may include, but is not limited to, a Measurement Object (MO) information element (IE), and / or a Report Configuration (RC) IE, and / or a Measurement Identity (MI) IE. An MI may associate an MOwith an RC; the UE may consider that an MO and an RC are associated if the MI contains the ID of the MO and the ID of the RC. For example, the AMM configuration may enable the network to instruct the UE regarding which measurements to perform, which AI / ML inferences to execute, and how to report the corresponding results. In some other implementations, the AMM configuration may be provided via a MAC CE, a DCI, or a combination thereof.

[0110] The AMM configuration may include various types of information that collectively define the AI / ML-based mobility operation for the UE. The AMM configuration may include one or more of an associated ID, a skipping pattern, a value for maximum MRRT, a value for maximum MRRS, an indication of AMM Case A or Case B, cluster information, a reporting configuration for future time instances, and an activation condition. Each of these parameters is described in detail below.

[0111] Associated ID

[0112] Regarding the associated ID, in some implementations, the associated ID may indicate the functionality or the model corresponding to the AMM configuration. In some implementations, the associated ID may be used to ensure that the wireless communication environment for data collection (for AMM model / functionality training) is consistent with the wireless communication environment for AMM inference. For example, the associated ID may enable the network and the UE to maintain consistency between the training environment and the inference environment, thereby improving the reliability of AI / ML predictions.

[0113] Skipping Pattern

[0114] Regarding the skipping pattern, in some implementations, the skipping pattern may indicate that the UE may skip the measurement on the associated measurement object (e.g., the UE may ignore the measurement requirement). In some implementations, the UE may use the predicted measurement result for the skipped time instance. For example, the skipping pattern may enable the UE to reduce measurement effort by relying on AI / ML -predicted results for certain measurement occasions, thereby reducing UE power consumption.

[0115] The skipping pattern may take various forms depending on the configuration approach, and several illustrative forms are described below.

[0116] In some implementations, the skipping pattern may be a bitmap, where each bit may indicate whether the measurement of the corresponding time instance can be skipped. For example, a bit with value 'O' may indicate that the measurement of the corresponding time instance can be skipped, and a bit with value T may indicate that the measurement of the corresponding time instance cannot be skipped. For another example, a bit with value 'O' may indicate that the measurement of the corresponding time instance cannot be skipped, and a bitwith value '1' may indicate that the measurement of the corresponding time instance can be skipped.

[0117] In some implementations, the skipping pattern may be a pre-configured and / or a pre-defined pattern and may be associated with a skipping pattern ID, and the network may use the skipping pattern ID in the AMM configuration to indicate to the UE the skipping pattern to be used for the MO.

[0118] In some implementations, the skipping pattern may be applied in a time duration. In some implementations, the time duration may be defined as a starting time to the starting time plus a period.

[0119] In some implementations, the starting time may be the time when the UE performs a performance monitoring of the AI / ML.

[0120] In some implementations, the starting time may be the time when the UE receives a performance monitoring configuration of the AI / ML.

[0121] In some implementations, the starting time may be the time after an offset when the UE performs a performance monitoring of the AI / ML. In some implementations, the offset may be configured by the network or may be a predefined value.

[0122] In some implementations, the period may be configured by the network. In some implementations, the period may be configured as an ENUMERATED value indicating a length of time. In some implementations, the period may be configured as an ENUMERATED value indicating a portion of a monitoring period.

[0123] In some implementations, the skipping pattern may be based on a (pre-)configured and / or (pre-)defined rule. For example, the AMM configuration may include a field / IE / parameter to indicate the rule corresponding to the skipping pattern. The field / IE may use an integer number to indicate the rule. If the field / IE / parameter indicates 'O', the indication may mean that the UE does not skip any measurements on the associated MO. If the field / IE / parameter indicates '2', the indication may mean that the UE skips the measurements on the associated MO every predefined number of time instances from the first time instance and / or a reference time. For example, the predefined number may be 2, though other values are also possible. The first time instance may be the time that the UE begins to perform the measurement on the associated MO.

[0124] In some implementations, the granularity unit of the time instance may be per frame, per slot, per DRX cycle, or per reporting periodicity (if configured).

[0125] In some implementations, the reference time of the skipping pattern may be defined. For example, the start time of the skipping pattern (e.g., the reference time) may be the absoluteSFN / frame / slot while receiving the pattern configuration.

[0126] In some implementations, the skipping pattern may be activated / deactivated via another message or another MAC CE or DCI.

[0127] The above-described forms of the skipping pattern are provided as illustrative examples, and the skipping pattern may take any one or any combination of these forms or other forms not explicitly described herein.

[0128] Maximum MRRT

[0129] Regarding the value for maximum MRRT, in some implementations, the maximum MRRT may indicate the maximum ratio of the measurements that the UE may skip in the total measurements that the UE should have measured in the time domain. In some implementations, the UE may use the predicted measurement result for the skipped time instance. In some implementations, the maximum MRRT may be indicated by an ENUMERATED value. For example, the value 'oneThird' may indicate that the UE may skip at most one measurement occasion in every predefined number of measurement occasions. For example, the predefined number of measurement occasions may be three, though other ratios are also possible. This parameter may enable the network to control the degree of measurement relaxation, balancing between measurement effort reduction and prediction reliability.

[0130] Maximum MRRS

[0131] Regarding the value for maximum MRRS, in some implementations, the maximum MRRS may indicate the maximum ratio of the measurements that the UE may skip in the total measurements that the UE should have measured in the space domain (e.g., the beam domain). In some implementations, the UE may use the predicted measurement result for the skipped measurement in the space domain (e.g., for the skipped measurement of a beam). In some implementations, the maximum MRRS may be indicated by an ENUMERATED value. For example, the value 'oneThird' may indicate that the UE may skip at most one measurement on a beam in every predefined number of measurements on beams. For example, the predefined number may be three, though other ratios are also possible. In some implementations, the MRRS may be different in different frequency ranges (e.g., FR1 / FR2 / FR3) such that respective MRRS values may be provided. This parameter may enable the network to control measurement effort reduction in the spatial domain while maintaining sufficient beam-level measurement accuracy.

[0132] Indication of AMM Case A or Case B

[0133] Regarding the indication of AMM Case A or Case B, in some implementations, AMM Case A may represent the operation where the UE is required to measure themeasurement occasion which has been predicted in advance, while AMM Case B may represent the operation where the UE is not required to measure the measurement occasion (or, may skip the measurement occasion) which has been predicted in advance.

[0134] In some implementations, for AMM Case B, the UE may use the predicted measurement result for the input of AI / ML to predict the measurement result in a future time instance. In some implementations, for AMM Case A, the UE may use the predicted measurement result and / or the measured measurement result for the input of AI / ML to predict the measurement result in a future time instance.

[0135] In some implementations, the indication may be an ENUMERATED value, where a first value (e.g., 'CaseA') may indicate that the AI / ML is operated in AMM Case A and a second value (e.g., 'CaseB') may indicate that the AI / ML is operated in AMM Case B. In some implementations, if the indication is absent, the UE may assume a default case value, where the default case value may be 'CaseA' or 'CaseB'.

[0136] In some implementations, the indication may be an ENUMERATED value 'true', where the AI / ML may be operated in AMM Case A if the indication is present and the AI / ML may be operated in AMM Case B if the indication is absent.

[0137] In some implementations, the indication may be an ENUMERATED value 'true', where the AI / ML may be operated in AMM Case B if the indication is present and the AI / ML may be operated in AMM Case A if the indication is absent.

[0138] The AMM Case A and AMM Case B are provided as illustrative examples of the AI / ML operation mode, and the distinction between requiring or not requiring measurement of predicted occasions may enable the network to flexibly control the trade-off between prediction accuracy and measurement effort.

[0139] Cluster Information

[0140] Regarding the cluster information, the cluster information may indicate entities applicable for cluster-based AI / ML operation. The cluster information may take various forms depending on the granularity of the cluster definition, and several illustrative forms are described below.

[0141] In some implementations, the cluster information may indicate one or more cells, one or more frequency bands, or one or more areas.

[0142] In some implementations, the cluster information may be a list of cell IDs (e.g., the global cell ID or the physical cell ID) indicating that the cells can be operated in cluster-based AI / ML. For example, the UE may take the measurement results of a first subset of the indicated cells as the input of AI / ML and obtain the predicted measurement results of a second subset ofthe indicated cells, in time and / or in space (e.g., beam) domain.

[0143] In some implementations, the cluster information may be a list of frequency bands (e.g., ARFCN value) indicating that the cells associated with the frequency bands can be operated in cluster-based AI / ML. For example, the UE may take the measurement results of a first subset of the cells associated with the indicated frequency bands as the input of AI / ML and obtain the predicted measurement results of a second subset of the cells associated with the indicated frequency bands, in time and / or in space (e.g., beam) domain.

[0144] In some implementations, the cluster information may be a list of area IDs (e.g., tracking area ID, or TCE ID) indicating that the cells associated with the area can be operated in cluster-based AI / ML. For example, the UE may take the measurement results of a first subset of the cells associated with the area as the input of AI / ML and obtain the predicted measurement results of a second subset of the cells associated with the area, in time and / or in space (e.g., beam) domain.

[0145] In some implementations, the cluster information may be a list of PLMN IDs indicating that the cells associated with the PLMN can be operated in cluster-based AI / ML.

[0146] The above-described forms of the cluster information are provided as illustrative examples, and the cluster information may take any one or any combination of these forms or other forms not explicitly described herein. The cluster-based AI / ML operation may enable the UE to leverage measurement results from one set of cells to predict measurement results for another set of cells, thereby further reducing measurement effort.

[0147] Reporting Configuration for Future Time Instances

[0148] Regarding the reporting configuration for future time instances, in some implementations, the reporting configuration may indicate that the UE reports the predicted measurement results in specific time instances and / or specific time periods. In some implementations, the reporting configuration may indicate the UE to report periodically and / or by event-triggered manner. In some implementations, the reporting configuration may indicate that the UE reports the predicted measurement results when the UE predicts one or more measurement events in a given future time period and / or a future time instance. In some implementations, the reporting configuration may indicate that the UE performs CHO when the UE predicts one or more measurement events in a given future time period and / or a future time instance. In some implementations, when a UE is indicated / triggered to perform AI / ML based measurement for an MO, MI and / or RC, the reporting configuration may indicate that the UE reports both of measurement results (e.g., non-predicted measurements) and predicted measurement results in specific time instances and / or specific time periods. This parametermay enable the network to control the timing and triggering conditions for the UE to submit predicted measurement reports.

[0149] Activation Condition

[0150] Regarding the activation condition, the activation condition may specify criteria for determining when the UE activates the MO / MI / RC for AI / ML operation. The activation condition may be based on various types of information available to the UE, such as the identity of cells, the radio link quality, or the mobility state of the UE, and the activation condition may take various forms depending on the deployment scenario and network configuration. Several illustrative forms of the activation condition are described below.

[0151] In some implementations, the activation condition may be a list of cell IDs, where the UE may activate the MO / MI / RC when the ID of the serving cell is in the list; the UE may deactivate the MO / MI / RC when the ID of the serving cell is not in the list. For example, the UE may forgo operating AI / ML for the MO / MI / RC when the serving cell ID is not in the cell ID list. For example, the UE may transmit a notification to the network indicating that the activation condition is no longer satisfied.

[0152] In some implementations, the activation condition may be based on radio link quality and may be one of the following conditions: the serving cell quality becomes higher than a threshold, the serving cell quality becomes lower than a threshold, a neighbor cell quality becomes higher than a threshold, a neighbor cell quality becomes lower than a threshold, the serving cell differential quality (e.g., the quality measured at a first time instance minus the quality measured at a second time instance) becomes higher than a threshold, the serving cell differential quality becomes lower than a threshold, the number of qualified beams (e.g., a beam of which the measured quality is higher than a threshold) of the serving cell is higher than a threshold, or the number of qualified beams of the serving cell is lower than a threshold.

[0153] In some implementations, the activation condition may be based on UE mobility speed and / or location. For example, if LowMobilityEvaluation and / or CellEdgeEvaluation is configured and the condition is met, the UE may activate the MO / MI / RC.

[0154] The serving cell quality may be determined based on the measured quality of one or more SSBs or one or more CSLRSs of the serving cell. The serving cell quality may be based on Reference Signal Received Power (RSRP), Reference Signal Received Quality (RSRQ), or Signal to Interference plus Noise Ratio (SINR).

[0155] The above-described forms of the activation condition are provided as illustrative examples, and the activation condition may take any one or any combination of these forms or other forms not explicitly described herein. For example, the activation condition mayadditionally or alternatively be based on a timer expiration, a network-signaled indication, or a UE-autonomous determination.

[0156] Any of the AMM configuration parameters described above (e.g., the associated ID, the skipping pattern, the maximum MRRT, the maximum MRRS, the indication of AMM Case A or Case B, the cluster information, the reporting configuration for future time instances, and the activation condition) may be used in combination with any other AMM configuration parameter. For example, the skipping pattern may be combined with the activation condition such that the UE applies the skipping pattern only when the activation condition is satisfied. The specific combination of AMM configuration parameters may depend on the deployment scenario, the network configuration, and the UE capability.

[0157] AMM Configuration Placement in MO, RC, and MI

[0158] The AMM configuration may be placed in different information elements (e.g., the MO, the RC, or the MI), and the placement may determine how the UE interprets whether the associated measurement and reporting operations are to be performed in AI / ML mode. Several placement approaches are described below.

[0159] In some implementations, the UE may receive the AMM configuration in an MO. Upon receiving the AMM configuration, the UE may consider that the MO is operated in AI / ML mode.

[0160] For example, for an MO without AMM configuration, the UE may consider that the MO configures the measurement without AI / ML; the UE may perform measurements according to the MO without AI / ML operation.

[0161] For example, for an MO with AMM configuration, the UE may consider that the MO configures the measurement with AI / ML; the UE may perform measurements and / or skip measurements according to the MO according to AI / ML inference results.

[0162] In some implementations, an MO with AMM and an MO without AMM may be associated with the same frequency band (e.g., the same ARFCN value may be contained in the MO with AMM and the MO without AMM), in case the network wants to operate both AL based and non-ALbased measurement procedures. This approach may enable the network to maintain a fallback measurement procedure alongside the AI / ML-based measurement procedure on the same frequency.

[0163] In some implementations, the UE may receive the AMM configuration in an MO. Upon receiving the AMM configuration, the UE may consider that all the Mis associated with the MO are operated in AI / ML mode.

[0164] For example, for an MO without AMM configuration, the UE may consider thatthe Mis associated with the MO are operated in non-AI mode; the UE may report and / or execute conditional configurations without AI / ML operation.

[0165] For example, for an MO with AMM configuration, the UE may consider that the Mis associated with the MO are operated in AI / ML mode; the UE may report and / or execute conditional configurations with AI / ML operation.

[0166] In some implementations, an MO with AMM and an MO without AMM may be associated with the same frequency band (e.g., the same ARFCN value may be contained in the MO with AMM and the MO without AMM), in case the network wants to operate both AI / ML-based and non-AI / ML-based measurement procedures.

[0167] In some implementations, the UE may receive the AMM configuration in an RC. Upon receiving the AMM configuration, the UE may consider that all the Mis associated with the RC are operated in AI / ML mode.

[0168] For example, for an RC without AMM configuration, the UE may consider that the Mis associated with the RC are operated in non-AI / ML mode; the UE may report and / or execute conditional configurations without AI / ML operation.

[0169] For example, for an RC with AMM configuration, the UE may consider that the Mis associated with the RC are operated in AI / ML mode; the UE may report and / or execute conditional configurations with AI / ML operation.

[0170] In some implementations, an RC with AMM and an RC without AMM may be associated with the same MO (e.g., the same MO ID may be contained in a first MI including a first RC ID and a second MI including a second RC ID, where the first RC ID is associated with a non-AI-based RC and the second RC ID is associated with an Al-based RC), in case the network wants to operate both Al-based and non-AI-based measurement reporting.

[0171] In some implementations, an MI may associate the RC with a first MO and a second MO by including the RC ID, the first MO ID, and the second MO ID. In some implementations, the UE may consider that the measurement results from the first MO may be used for the input of the AI / ML and the UE may consider that the UE may report the predicted measurement results of the second MO to the network. The first MO ID and the second MO ID may be the same or different. This approach may enable the UE to use measurement results from one MO as the AI / ML inference input and to report predicted results for another MO, thereby supporting flexible AI / ML-based mobility operations across different measurement objects.

[0172] In some implementations, for cluster-based AMM operation, an MI may associate the RC with a first list of MO IDs and a second list of MO IDs. In some implementations, the UE may consider that the measurement results from the first list of MOs may be used for theinput of the AI / ML and the UE may consider that the UE may report the predicted measurement results of the second list of MOs to the network. The first list of MO IDs and the second list of MO IDs may be totally the same, or may be partially the same, or may be mutually exclusive. The UE may use all or partial measurement results from the first list of MOs for the input of AI / ML.

[0173] In some implementations, the UE may receive AMM configuration in an MI. Upon receiving the AMM configuration, the UE may consider that the MI is operated in AI / ML mode.

[0174] For example, for an MI without AMM configuration, the UE may consider that the MI is operated in non-AI / ML mode; the UE may report and / or execute conditional configurations without AI / ML operation.

[0175] For example, for an MI with AMM configuration, the UE may consider that the MI is operated in AI / ML mode; the UE may report and / or execute conditional configurations with AI / ML operation.

[0176] In some implementations, a first MI may associate a first MO with a first RC and a second MI may also associate the first MO with the first RC, while the first MI includes AMM configuration and the second MI does not include AMM configuration, in case the network wants to operate both Al-based and non-AI-based measurement for reporting and / or CHO execution.

[0177] In some implementations, a first MI may associate with two MOs (a first MO and a second MO) and may further associate with a first RC and an associated ID. The UE may consider the MI is operated in AI / ML mode where the UE may measure the first MO and report the predicted measurement results of the second MO to the network followed by the RC. If the associated ID changes upon camped cells, the UE may treat the MI as inapplicable. For example, the UE may release or suspend the AI / ML operation for the MI when the associated ID no longer corresponds to the current serving cell environment.

[0178] AMM-MO and AMM-RC Configuration

[0179] In some implementations, the UE may receive AMM configuration specific to MO (AMM-MO configuration) in an MO and may receive AMM configuration specific to RC (AMM-RC configuration) in an RC and may receive an MI associating the MO and the RC.

[0180] In some implementations, the AMM-MO configuration may include AMM configuration related to measurement. For example, the AMM-MO configuration may include one or more of an associated ID, a skipping pattern indication / maximum MRRT value / maximum MRRS value, an indication of operation in AMM Case A or AMM Case B, and cluster-based information.

[0181] In some implementations, the AMM-RC configuration may include AMM configuration related to reporting. For example, the AMM-RC configuration may include one or more of an associated ID, cluster-based information, and a reporting configuration for future time instances.

[0182] Determining AI / ML Mode of the MI

[0183] In some implementations, the MI may additionally include information related to the operation and / or activation of AI / ML. Several types of indications that may be included in the MI are described below.

[0184] In some implementations, the MI may include an indication whether the MI is operated in AI / ML mode. In some implementations, the indication may take ENUMERATED values, and a first value (e.g., 'Al') may indicate that the MI is operated in AI / ML mode, and a second value (e.g., 'non- Al') may indicate that the MI is operated in non- AI / ML mode. In some implementations, the indication may take ENUMERATED value 'true', and the UE may infer that the MI is operated in AI / ML mode if the indication is present, and the UE may infer that the MI is operated in non-AI / ML mode if the indication is absent.

[0185] In some implementations, the MI may include an indication whether the MI can be operated in AI / ML mode. In some implementations, the indication may take ENUMERATED value 'enable', and the UE may infer that the MI can be operated in AI / ML mode. The UE may perform AI / ML for the MI if the condition(s) of AI / ML operation activation are satisfied. If the indication takes a different ENUMERATED value and / or the indication is absent, the UE may infer that the MI cannot be operated in AI / ML mode.

[0186] In some implementations, the MI may include an indication whether the AI / ML is initially activated for the MI. In some implementations, the indication may take ENUMERATED values, and a first value (e.g., 'activated') may indicate that the MI is initially activated for AI / ML, and a second value (e.g., 'deactivated') may indicate that the MI is initially deactivated for AI / ML. In some implementations, the indication may take ENUMERATED value 'true', and the UE may infer that the MI is initially activated for AI / ML if the indication is present, and the UE may infer that the MI is initially deactivated for AI / ML if the indication is absent.

[0187] In some implementations, the UE may consider that the MI is operated in AI / ML mode if the MI is associated with an MO including AMM-MO configuration and an RC including AMM-RC configuration. In some implementations, the UE may consider that the MI is operated in AI / ML mode if the MI further contains an indication indicating that the MI is operated in AI / ML mode. In some implementations, the UE may consider that the MI is ableto be operated in AI / ML mode if the MI further contains an indication indicating that the MI can be operated in AI / ML mode.

[0188] In some implementations, the UE may consider that the MI is operated in non-AI / ML mode if the MI is associated with an MO including AMM-MO configuration and an RC not including AMM-RC configuration. In some implementations, the UE may consider that the MI is operated in non- AI / ML mode if the MI is associated with an MO not including AMM-MO configuration and an RC including AMM-RC configuration. In some implementations, the UE may consider that the MI is operated in non-AI / ML mode if the MI is associated with an MO not including AMM-MO configuration and an RC not including AMM-RC configuration. In some implementations, the UE may consider that the MI is operated in non-AI / ML mode if the MI is associated with an MO including AMM-MO configuration, an RC including AMM-RC configuration, and the condition(s) of activating AI / ML operation are not satisfied. In some implementations, the UE may consider that the MI is operated in non-AI / ML mode if the MI contains an indication indicating that the MI is operating in non-AI / ML mode. In some implementations, the UE may consider that the MI is operated in non-AI / ML mode if the MI contains an indication indicating that the MI cannot operate in AI / ML mode.

[0189] For an MI operated in AI / ML mode, the UE may consider that the associated MO and / or RC is operated in AI / ML mode. When an MI / MO / RC is operated in AI / ML mode, the UE may operate AI / ML for the MI / MO / RC when the MI / MO / RC is activated for AI / ML; otherwise (e.g., when the MI / MO / RC is deactivated for AI / ML), the UE may not operate AI / ML for the MI / MO / RC. For example, the UE may forgo performing AI / ML inference for the MI / MO / RC when the MI / MO / RC is deactivated. In some implementations, the network may activate and / or deactivate the MI / MO / RC for AI / ML via control signaling (e.g., by RRC signaling, MAC CE, DCI, etc.). In some implementations, the UE may activate and / or deactivate the MI / MO / RC for AI / ML by the UE (e.g., autonomous (de)activation, event-triggered (de)activation).

[0190] When an MI / MO / RC is operated in non-AI / ML mode, the UE may perform non-Al-based operation for the MI / MO / RC. The MI / MO / RC may not be activated and / or deactivated for AI / ML.

[0191] AMM Configuration for RRC IDLE and / or RRC INACTIVE State

[0192] Beyond the RRC CONNECTED state described above, the AMM configuration may also be applicable in RRC IDLE and / or RRC INACTIVE state.

[0193] In some implementations, the UE may receive the AMM configuration from the network in RRC IDLE and / or RRC INACTIVE state via system information and / or a releasingmessage that switches the UE from RRC CONNECTED state to RRC IDLE / RRC INACTIVE state.

[0194] In some implementations, the AMM configuration for RRC IDLE and / or INACTIVE state may include an ID to ensure that the communication environment for training data collection is consistent with the communication environment for AI / ML inference.

[0195] In some implementations, the ID may be an associated ID. The associated ID for AMM in the RRC CONNECTED case, the associated ID for AMM in the RRC IDLE / INACTIVE case, and the associated ID for AI / ML for air interface cases may or may not share the same value. For example, a CONNECTED AMM configuration may have the same associated ID as an IDLE / INACTIVE AMM configuration. In some implementations, the same associated ID may represent the same communication environment. In some implementations, the same associated ID for different AI / ML cases (e.g., CONNECTED AMM, IDLE / INACTIVE AMM, and AI / ML for air interface) may represent different communication environments.

[0196] In some implementations, the ID may be a cell ID (e.g., a cell global ID, a physical cell ID), where the ID may be specific to a cell. In some implementations, the UE may use the cell ID to determine whether to use an AI / ML functionality and / or which AI / ML functionality is used.

[0197] In some implementations, the ID may be an area ID (e.g., a tracking area ID), where the ID may be common to a group of cells. In some implementations, the UE may use the area ID to determine whether to use an AI / ML functionality and / or which AI / ML functionality is used. In some implementations, a cell may have multiple IDs such that the cell may be grouped into different area IDs. If the UE recognizes one of the IDs matched the stored associated ID of the UE, the UE may treat the camped cell as applicable for AI / ML mode.

[0198] In some implementations, the AMM configuration for RRC IDLE and / or INACTIVE state may include information related to relaxed measurement period. In some implementations, the information may indicate the skipping pattern to the UE. In some implementations, the information may indicate the MRRT to the UE. In some implementations, the information may indicate the MRRS to the UE.

[0199] The implementations described in this section are provided as illustrative examples, and other implementations achieving similar technical effects may be employed. The abovedescribed AMM configuration placement approaches (e.g., placement in MO, RC, or MI) may be used individually or in any combination, and other placement approaches not explicitly described herein are also within the scope of the present disclosure.

[0200] Activation / D eactivation of AI / ML-based Mobility

[0201] The following section describes the activation and deactivation mechanisms for AI / ML-based mobility. The activation and deactivation of the MO, RC, and MI may be triggered by various methods, including RRC signaling, MAC CE signaling, and conditionbased (e.g., autonomous) activation / deactivation. The implementations described in this section are provided as illustrative examples, and variations, modifications, and alternative implementations are possible without departing from the scope of the present disclosure.

[0202] Any of the activation / deactivation mechanisms described in this section may be used in combination with any of the AMM configuration parameters described in the preceding section (e.g., the skipping pattern, the activation condition, the associated ID, the reporting configuration for future time instances). For example, the MAC CE-based activation / deactivation mechanism may be combined with the activation condition to provide a two-layer activation control, where the MAC CE provides network-initiated activation control and the activation condition provides event-triggered activation control. The specific combination of activation / deactivation mechanisms and AMM configuration parameters may depend on the deployment scenario and the network configuration.

[0203] RRC-based Activation of MO

[0204] In some implementations, the UE may consider that an MO is activated upon receiving the RRC signaling containing the MO associated with AMM configuration.

[0205] In some implementations, if the received MO contains AMM configuration or AMM-MO configuration, the UE may consider that the MO is activated.

[0206] In some implementations, if the received MO contains AMM configuration / AMM-MO configuration, the MI associated with the MO contains an indication that the MI is able to be operated in AI / ML mode, and the condition(s) of activating AI / ML mode are satisfied, the UE may consider that the MO is activated.

[0207] In some implementations, if the received MO contains AMM-MO configuration, and an RC associated with the MO contains AMM-RC configuration, the UE may consider that the MO is activated.

[0208] In some implementations, if the received MO contains AMM-MO configuration, and an MI associated with the MO contains an indication that the MI is operated in AI / ML mode, the UE may consider that the MO is activated.

[0209] In some implementations, if the received MO contains AMM-MO configuration, and an MI associated with the MO contains an indication that the MI is operated in AI / ML mode, and the MI also contains an indication that the AI / ML state for the MI is activated, theUE may consider that the MO is activated.

[0210] In some implementations, if the received MO is associated with an MI, and the UE considers that the MI is activated, the UE may consider that the MO is activated.

[0211] The above-described approaches for RRC -based activation of MO are provided as illustrative examples, and other approaches for determining MO activation based on received RRC signaling are also within the scope of the present disclosure.

[0212] RRC-based Activation of RC

[0213] In some implementations, the UE may consider that an RC is activated upon receiving the RRC signaling containing the RC associated with AMM configuration.

[0214] In some implementations, if the received RC contains AMM configuration or AMM-RC configuration, the UE may consider that the RC is activated.

[0215] In some implementations, if the received RC contains AMM configuration / AMM-RC configuration, the MI associated with the RC contains an indication that the MI is able to be operated in AI / ML mode, and the condition(s) of activating AI / ML mode are satisfied, the UE may consider that the RC is activated.

[0216] In some implementations, if the received RC contains AMM-RC configuration, and an MO associated with the RC contains AMM-MO configuration, the UE may consider that the RC is activated.

[0217] In some implementations, if the received RC contains AMM-RC configuration, and an MI associated with the RC contains an indication that the MI is operated in AI / ML mode, the UE may consider that the RC is activated.

[0218] In some implementations, if the received RC contains AMM-RC configuration, and an MI associated with the RC contains an indication that the MI is operated in AI / ML mode, and the MI also contains an indication that the AI / ML state for the MI is activated, the UE may consider that the RC is activated.

[0219] In some implementations, if the received RC is associated with an MI, and the UE considers that the MI is activated, the UE may consider that the RC is activated.

[0220] In some implementations, if the received RC is associated with a specific reporting type (e.g., periodic reporting), the UE may consider that the RC is activated immediately; otherwise the UE may consider that the RC is configured and may await another activation signaling from the network / gNB.

[0221] RRC-based Activation of MI

[0222] In some implementations, the UE may consider that an MI is activated upon receiving the RRC signaling containing the MI associated with AMM configuration.

[0223] In some implementations, if the received MI contains AMM configuration, the UE may consider that the MI is activated.

[0224] In some implementations, if the received MI associates an MO including AMM-MO configuration with an RC including AMM-RC configuration, the UE may consider that the MI is activated.

[0225] In some implementations, if the received MI includes an indication that the MI is activated, the UE may consider that the MI is activated.

[0226] In some implementations, if the received MI includes an indication that the MI is able to be operated in AI / ML mode and the condition(s) of activating AI / ML mode are satisfied, the UE may consider that the MI is activated.

[0227] MAC CE-based Activation / Deactivation of MO

[0228] In some implementations, the UE may consider that an MO is activated or deactivated upon receiving a signaling from the network. More specifically, the signaling may be a MAC CE. In some implementations, the UE may activate the MO after a first offset after receiving the signaling. In some implementations, the UE may deactivate the MO after a second offset after receiving the signaling. In some implementations, the first offset and the second offset may be equal or may be different. This approach may enable the network to dynamically control the AI / ML operation state of the MO without requiring full RRC reconfiguration, thereby reducing signaling overhead and latency.

[0229] In some implementations, the signaling may include a field indicating an MO configuration to be activated or deactivated. In some implementations, the field may include bits corresponding to the MO ID.

[0230] In some implementations, the UE may consider that the indicated MO is activated upon receiving the signaling.

[0231] In some implementations, the UE may consider that the indicated MO is deactivated upon receiving the signaling.

[0232] In some implementations, the UE may consider that the indicated MO is activated if the MO is originally deactivated.

[0233] In some implementations, the UE may consider that the indicated MO is deactivated if the MO is originally activated.

[0234] In some implementations, the signaling may include a first field indicating an MO configuration and a second field indicating the activation or deactivation of the MO.

[0235] In some implementations, the UE may consider that the indicated MO is activated if the second field indicates the activation.

[0236] In some implementations, the UE may consider that the indicated MO is deactivated if the second field indicates the deactivation.

[0237] In some implementations, the signaling may include multiple first fields indicating multiple MOs.

[0238] In some implementations, the UE may consider that the indicated MOs are activated upon receiving the signaling.

[0239] In some implementations, the UE may consider that the indicated MOs are deactivated upon receiving the signaling.

[0240] In some implementations, the UE may consider that the indicated MOs which are originally activated are deactivated and the indicated MOs which are originally deactivated are activated.

[0241] In some implementations, the signaling may include multiple first fields indicating multiple MOs and a second field indicating the activation or deactivation.

[0242] In some implementations, the UE may consider that the indicated MOs are activated if the second field indicates the activation.

[0243] In some implementations, the UE may consider that the indicated MOs are deactivated if the second field indicates the deactivation.

[0244] In some implementations, the signaling may include multiple first fields indicating multiple MOs and multiple second fields indicating the activation or deactivation. In some implementations, the number of the first fields may be equal to the number of the second fields. In some implementations, each first field may correspond to a second field.

[0245] In some implementations, for a first field, the UE may consider that the MO indicated by the first field is activated if the second field corresponding to the first field indicates the activation.

[0246] In some implementations, for a first field, the UE may consider that the MO indicated by the first field is deactivated if the second field corresponding to the first field indicates the deactivation.

[0247] In some implementations, the field indicating an MO configuration may include bits corresponding to the MO ID.

[0248] In some implementations, the field indicating activation or deactivation may include a predefined number of bits. For example, the predefined number may be 1, though other bit widths are also possible. A first value (e.g., 0) may indicate the deactivation and a second value (e.g., 1) may indicate the activation.

[0249] In some implementations, the signaling may further include one or more fieldsindicating a subset of AMM-related configuration which is included in the AMM-MO configuration. In some implementations, the UE may apply the subset of AMM-related configuration indicated by the field for the AMM operation.

[0250] In some implementations, the activation / deactivation of the MO may be based on the activation / deactivation of SSB and / or the activation / deactivation of SCells. For example, if the UE receives the deactivation of SSB subject to respective cells, the UE may also deactivate the MO under AI / ML mode.

[0251] MAC CE-based Activation / Deactivation of RC

[0252] In some implementations, the UE may consider that an RC is activated or deactivated upon receiving a signaling from the network. More specifically, the signaling may be a MAC CE. In some implementations, the UE may activate the RC after a first offset after receiving the signaling. In some implementations, the UE may deactivate the RC after a second offset after receiving the signaling. In some implementations, the first offset and the second offset may be equal or may be different.

[0253] In some implementations, the signaling may include a field indicating an RC configuration to be activated or deactivated. In some implementations, the field may include bits corresponding to the RC ID.

[0254] In some implementations, the UE may consider that the indicated RC is activated upon receiving the signaling.

[0255] In some implementations, the UE may consider that the indicated RC is deactivated upon receiving the signaling.

[0256] In some implementations, the UE may consider that the indicated RC is activated if the RC is originally deactivated.

[0257] In some implementations, the UE may consider that the indicated RC is deactivated if the RC is originally activated.

[0258] In some implementations, the signaling may include a first field indicating an RC configuration and a second field indicating the activation or deactivation of the RC.

[0259] In some implementations, the UE may consider that the indicated RC is activated if the second field indicates the activation.

[0260] In some implementations, the UE may consider that the indicated RC is deactivated if the second field indicates the deactivation.

[0261] In some implementations, the signaling may include multiple first fields indicating multiple RCs.

[0262] In some implementations, the UE may consider that the indicated RCs are activatedupon receiving the signaling.

[0263] In some implementations, the UE may consider that the indicated RCs are deactivated upon receiving the signaling.

[0264] In some implementations, the UE may consider that the indicated RCs which are originally activated are deactivated and the indicated RCs which are originally deactivated are activated.

[0265] In some implementations, the signaling may include multiple first fields indicating multiple RCs and a second field indicating the activation or deactivation.

[0266] In some implementations, the UE may consider that the indicated RCs are activated if the second field indicates the activation.

[0267] In some implementations, the UE may consider that the indicated RCs are deactivated if the second field indicates the deactivation.

[0268] In some implementations, the signaling may include multiple first fields indicating multiple RCs and multiple second fields indicating the activation or deactivation. In some implementations, the number of the first fields may be equal to the number of the second fields. In some implementations, each first field may correspond to a second field.

[0269] In some implementations, for a first field, the UE may consider that the RC indicated by the first field is activated if the second field corresponding to the first field indicates the activation.

[0270] In some implementations, for a first field, the UE may consider that the RC indicated by the first field is deactivated if the second field corresponding to the first field indicates the deactivation.

[0271] In some implementations, the field indicating an RC configuration may include bits corresponding to the RC ID.

[0272] In some implementations, the field indicating activation or deactivation may include a predefined number of bits. For example, the predefined number may be 1, though other bit widths are also possible. A first value (e.g., 0) may indicate the deactivation and a second value (e.g., 1) may indicate the activation.

[0273] In some implementations, the signaling may further include one or more fields indicating a set of parameters (e.g., the starting time, the interval, the number of reported time instances, etc.) which is included in the AMM-RC configuration. In some implementations, the UE may apply the set of parameters indicated by the field for the AMM operation. For example, the AMM-RC configuration may contain a first set of parameters (e.g., indicating a first starting time, a first interval and a first number of reported time instances) and a second set ofparameters (e.g., indicating a second starting time, a second interval and a second number of reported time instances); if the field of the signaling indicates that the first set of parameters is to be applied, the UE may consider using the first set of parameters for the reporting procedure (e.g., reporting results of the time instances corresponding to the arithmetic sequence, where the first starting time is the initial term, the first interval is the common difference, and the first number of reported time instances is the number of terms). The first set of parameters and the second set of parameters may be configured via a full configuration and a partial configuration accordingly.

[0274] MAC CE-based Activation / Deactivation of MI

[0275] In some implementations, the UE may consider that an MI is activated or deactivated upon receiving a signaling from the network. More specifically, the signaling may be a MAC CE. In some implementations, the UE may activate the MI after a first offset after receiving the signaling. In some implementations, the UE may deactivate the MI after a second offset after receiving the signaling. In some implementations, the first offset and the second offset may be equal or may be different.

[0276] In some implementations, the signaling may include a field indicating an MI configuration to be activated or deactivated. In some implementations, the field may include bits corresponding to the MI ID.

[0277] In some implementations, the UE may consider that the indicated MI is activated upon receiving the signaling.

[0278] In some implementations, the UE may consider that the indicated MI is deactivated upon receiving the signaling.

[0279] In some implementations, the UE may consider that the indicated MI is activated if the MI is originally deactivated.

[0280] In some implementations, the UE may consider that the indicated MI is deactivated if the MI is originally activated.

[0281] In some implementations, the signaling may include a first field indicating an MI configuration and a second field indicating the activation or deactivation of the MI.

[0282] In some implementations, the UE may consider that the indicated MI is activated if the second field indicates the activation.

[0283] In some implementations, the UE may consider that the indicated MI is deactivated if the second field indicates the deactivation.

[0284] In some implementations, the signaling may include multiple first fields indicating multiple Mis.

[0285] In some implementations, the UE may consider that the indicated Mis are activated upon receiving the signaling.

[0286] In some implementations, the UE may consider that the indicated Mis are deactivated upon receiving the signaling.

[0287] In some implementations, the UE may consider that the indicated Mis which are originally activated are deactivated and the indicated Mis which are originally deactivated are activated.

[0288] In some implementations, the signaling may include multiple first fields indicating multiple Mis and a second field indicating the activation or deactivation.

[0289] In some implementations, the UE may consider that the indicated Mis are activated if the second field indicates the activation.

[0290] In some implementations, the UE may consider that the indicated Mis are deactivated if the second field indicates the deactivation.

[0291] In some implementations, the signaling may include multiple first fields indicating multiple Mis and multiple second fields indicating the activation or deactivation. In some implementations, the number of the first fields may be equal to the number of the second fields. In some implementations, each first field may correspond to a second field.

[0292] In some implementations, for a first field, the UE may consider that the MI indicated by the first field is activated if the second field corresponding to the first field indicates the activation.

[0293] In some implementations, for a first field, the UE may consider that the MI indicated by the first field is deactivated if the second field corresponding to the first field indicates the deactivation.

[0294] In some implementations, the field indicating an MI configuration may include bits corresponding to the MI ID.

[0295] In some implementations, the field indicating activation or deactivation may include a predefined number of bits. For example, the predefined number may be 1, though other bit widths are also possible. A first value (e.g., 0) may indicate the deactivation and a second value (e.g., 1) may indicate the activation.

[0296] In some implementations, the signaling may further include one or more fields indicating a subset of AMM-related configuration which is included in the AMM-RC configuration of the RC associated with the MI. In some implementations, the UE may apply the subset of AMM-related configuration indicated by the field for the AMM operation.

[0297] In some implementations, the signaling may further include one or more fieldsindicating a subset of AMM-related configuration which is included in the AMM-MO configuration of the MO associated with the MI. In some implementations, the UE may apply the subset of AMM-related configuration indicated by the field for the AMM operation.

[0298] In some other implementations, the activation / deactivation of the MO, the RC, or the MI may be signaled via a DCI, or via a combination of MAC CE and DCI, or via RRC signaling with a shorter processing delay than full RRC reconfiguration.

[0299] Condition-based Activation / Deactivation

[0300] In some implementations, the UE may consider that an MO / MI / RC is activated or deactivated if the UE considers that one or more conditions associated with the MO / MI / RC are satisfied.

[0301] In some implementations, the conditions may be configured by the network. More specifically, the conditions may be one of the conditions described in measurement events in a wireless communication system.

[0302] In some implementations, the satisfaction of the conditions may be determined according to the radio link quality (e.g., RSRP, RSRQ, SINR) of the SSBs and / or CSI-RSs in the cells associated with the MI / MO / RC.

[0303] In some implementations, a UE may be configured with an RRC counter and the initial value of the counter may be 0. If the UE considers an MO / MI / RC is activated since the conditions for activating the MO / MI / RC are satisfied, when the conditions for activating the MO / MI / RC become unsatisfied and / or the conditions for deactivating the MO / MI / RC are satisfied, the RRC counter may be increased by 1. Once the number of the RRC counter reaches a configured number (which may be configured to the UE via RRC signaling), the UE may consider that the MO / MI / RC is deactivated. Furthermore, the UE may consider that the MO / MI / RC falls back to non-AI / ML mode. The UE may reset the RRC counter when the conditions for activating the MO / MI / RC become satisfied. This approach may enable the network to control the persistence of deactivation decisions, preventing premature deactivation due to transient fluctuations in conditions.

[0304] In some implementations, when the conditions for deactivating the MO / MI / RC are satisfied, the UE may consider that the MO / MI / RC is deactivated. Furthermore, the UE may consider that the MO / MI / RC falls back to non-AI / ML mode.

[0305] In some implementations, the satisfaction of the conditions may be determined according to a configured time. For example, a timer may start / restart while receiving the RRC configuration. If the timer expires, the MI / MO / RC may be deactivated or activated.

[0306] In some implementations, the satisfaction of the conditions may be determinedaccording to the area information (e.g., cell ID, tracking area ID) associated with the MI / MO / RC. In some implementations, a cell ID and / or tracking area ID may be included in AMM-related MI / MO / RC. For example, when the UE identifies that the serving cell has a cell ID and has a tracking area ID, the UE may consider that the MO / MI / RC which are associated with the cell ID and / or the tracking area ID are activated; the UE may also consider that the MO / MI / RC which are not associated with the cell ID and / or the tracking area ID are deactivated.

[0307] In some implementations, the satisfaction of the conditions may be determined according to the location information (e.g., the location of the UE) associated with the MI / MO / RC. In some implementations, area information may be included in the AMM-related MO / MI / RC; the area information may be characterized by a reference location and a distance, and the UE may consider that the area is a circle with the reference location as the center and the distance as the radius. For example, when the UE identifies that the UE is located within the area, the UE may consider that the MO / MI / RC are activated; when the UE identifies that the UE is located outside the area, the UE may consider that the MO / MI / RC are deactivated.

[0308] In some implementations, the satisfaction of the conditions may be determined according to the speed of the UE.

[0309] In some implementations, the condition may be characterized by a threshold. In some implementations, the UE may obtain the speed by the implementation of the UE. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the MO / MI / RC is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the MO / MI / RC is deactivated. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the MO / MI / RC is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the MO / MI / RC is activated. More specifically, the network may configure whether the MO / MI / RC is activated while the speed "exceeds" or is "below" the threshold and whether the MO / MI / RC is deactivated while the speed "exceeds" or is "below" the threshold.

[0310] In some implementations, the condition may be characterized by a threshold and an indication of higher or lower for triggering the condition. For example, if the indication indicates 'higher' for triggering the condition, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the MO / MI / RC is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the MO / MI / RC is deactivated. For example, if the indication indicates 'lower' for triggering thecondition, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the MO / MI / RC is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the MO / MI / RC is activated. More specifically, the network may configure whether the MO / MI / RC is activated while the speed "exceeds" or is "below" the threshold and whether the MO / MI / RC is deactivated while the speed "exceeds" or is "below" the threshold.

[0311] In some implementations, a time-to-trigger method may apply to determine the satisfaction of the conditions described above. For example, when the UE considers that a condition transits from unmet to met, the UE may reset and / or start / restart a timer for the condition; if the timer is running and the corresponding condition becomes unmet, the UE may stop the timer; if the timer expires, the UE may consider that the corresponding condition is satisfied. This approach may prevent premature activation or deactivation due to transient fluctuations in the monitored parameter.

[0312] The above-described condition-based activation / deactivation mechanisms (e.g., radio link quality-based, counter-based, timer-based, area-based, location-based, speed-based, and time-to-trigger-based mechanisms) are provided as illustrative examples, and the UE may use any one or any combination of these mechanisms for determining the activation or deactivation of the MO / MI / RC. Other condition-based mechanisms not explicitly described herein are also within the scope of the present disclosure.

[0313] Non-Connected-Mode AMM Configuration

[0314] Beyond the RRC CONNECTED state described above, the AMM configuration may also be applicable in non-connected modes. The following section describes the Non-Connected-Mode AMM (NCM-AMM) configuration, which may be used for AI / ML-based mobility operations when the UE is in RRC IDLE and / or RRC INACTIVE state. The implementations described in this section are provided as illustrative examples, and variations, modifications, and alternative implementations are possible without departing from the scope of the present disclosure.

[0315] In some implementations, the UE may receive a non-connected-mode AMM configuration (NCM-AMM configuration) from the network.

[0316] In some implementations, the UE may receive the NCM-AMM configuration via an RRC message. In some implementations, the RRC message may be an RRC Release message.

[0317] In some implementations, the UE may receive the NCM-AMM configuration via a system information.

[0318] In some implementations, the UE may store the NCM-AMM configuration upon receiving the RRC message and / or the system information.

[0319] In some implementations, the UE may store the NCM-AMM configuration if the UE considers that the NCM-AMM configuration is applicable (e.g., the configuration corresponds to an applicable AMM functionality).

[0320] In some implementations, the UE may not store the NCM-AMM configuration (e.g., the UE may ignore the NCM-AMM configuration) if the UE considers that the NCM-AMM configuration is non-applicable (e.g., the configuration corresponds to a non-applicable AMM functionality). For example, the UE may forgo storing the NCM-AMM configuration when the configuration does not match any AI / ML functionality supported by the UE.

[0321] In some implementations, the NCM-AMM configuration may include an MO to be measured and the AMM-MO configuration indicating the AMM-related operation to the MO. In some implementations, the NCM-AMM configuration may include the information related to AI / ML-based cell reselection such as one or more of the radio link quality (e.g., RSRP, RSRQ, SINR) thresholds for AI / ML-based cell reselection, and / or the frequency band priority for AI / ML-based cell reselection, and / or the indication of cells supporting (or not supporting) the operation of AMM in RRC CONNECTED state, and / or the indication of cells supporting (or not supporting) the operation of AMM in RRC IDLE / INACTIVE state.

[0322] For example, the UE may store a first cell reselection configuration associated with AI / ML and a second cell reselection configuration not associated with AI / ML (e.g., not used for AI / ML operation), and the UE may apply the first cell reselection configuration for a cell reselection procedure when the NCM-AMM configuration is activated; the UE may apply the second cell reselection configuration for a cell reselection procedure when the NCM-AMM configuration is deactivated.

[0323] For example, the UE may store a first frequency band priority configuration associated with AI / ML and a second frequency band priority configuration not associated with AI / ML (e.g., not used for AI / ML operation), and the UE may apply the first frequency band priority configuration for a cell reselection procedure when the NCM-AMM configuration is activated; the UE may apply the second frequency band priority configuration for a cell reselection procedure when the NCM-AMM configuration is deactivated.

[0324] For example, the UE may prioritize a first cell for cell reselection if the first cell is indicated to support AI / ML operation and may deprioritize a second cell for cell reselection if the second cell is indicated to not support AI / ML operation; in some implementations, the UE may do so if the UE supports AI / ML operation.

[0325] For example, the UE may deprioritize a first cell for cell reselection if the first cell is indicated to support AI / ML operation and may prioritize a second cell for cell reselection if the second cell is indicated to not support AI / ML operation; in some implementations, the UE may do so if the UE does not support AI / ML operation.

[0326] In some implementations, the NCM-AMM configuration may contain an AI / ML part configuration and a non-AI / ML part configuration, and the UE may apply one of the configurations in the NCM-AMM configuration. In some implementations, an AI / ML part configuration may include an association ID. This approach may enable the network to provide the UE with both AI / ML-based and non-AI / ML-based configurations for non-connected mode, allowing the UE to switch between the two modes based on activation conditions.

[0327] Activation of NCM-AMM Configuration

[0328] In some implementations, the UE in RRC IDLE / INACTIVE state may activate or may be activated an NCM-AMM configuration.

[0329] In some implementations, the UE may consider that the NCM-AMM configuration is activated upon receiving the RRC message and / or the system information including the NCM-AMM configuration.

[0330] In some implementations, the UE may reset and start / restart a timer upon activating the NCM-AMM configuration; the UE may stop and / or release the timer upon deactivating and / or releasing the NCM-AMM configuration; the UE may deactivate and / or release the NCM-AMM configuration upon the timer expiration (e.g., the timer exceeds a timer length); the UE may reset and start / restart the timer upon the UE updating the NCM-AMM configuration; the UE may update the NCM-AMM configuration upon the timer expiration or at any time by the implementation of the UE. In some implementations, the timer length may be configured by the network (e.g., included in the RRC message and / or system information including the NCM-AMM configuration). In some implementations, the UE updating the NCM-AMM configuration may include the UE receiving an RRC message and / or a system information including an NCM-AMM configuration from the network. In some implementations, if the received RRC message and / or the system information does not include an NCM-AMM configuration having the same ID as the stored NCM-AMM configuration, the UE may release the stored NCM-AMM configuration. In some implementations, if the received RRC message and / or the system information includes an NCM-AMM configuration having the same ID as the stored NCM-AMM configuration, the UE may release the stored NCM-AMM configuration and apply and / or store the NCM-AMM configuration in the received RRC message and / or the system information.

[0331] In some implementations, the UE may set a counter to be 0 upon activating the NCM-AMM configuration; the UE may stop and / or release the counter upon deactivating and / or releasing the NCM-AMM configuration; the UE may deactivate and / or release the NCM-AMM configuration upon the counter reaching a configured number (which may be configured to the UE by the network); the UE may reset and start / restart the counter upon the UE updating the NCM-AMM configuration; the UE may update the NCM-AMM configuration upon the counter reaching a configured number (which may be configured to the UE by the network) or at any time by the implementation of the UE. In some implementations, the maximum number of the counter (e.g., the configured number) may be configured by the network (e.g., included in the RRC message and / or system information including the NCM-AMM configuration). In some implementations, the UE updating the NCM-AMM configuration may include the UE receiving an RRC message and / or a system information including an NCM-AMM configuration from the network. In some implementations, if the received RRC message and / or the system information does not include an NCM-AMM configuration having the same ID as the stored NCM-AMM configuration, the UE may release the stored NCM-AMM configuration. In some implementations, if the received RRC message and / or the system information includes an NCM-AMM configuration having the same ID as the stored NCM-AMM configuration, the UE may release the stored NCM-AMM configuration and apply and / or store the NCM-AMM configuration in the received RRC message and / or the system information.

[0332] In some implementations, the UE may activate the NCM-AMM configuration according to the location information.

[0333] In some implementations, the location information may be one or more cell IDs. In some implementations, the UE may activate the NCM-AMM configuration if the ID of the camping cell is in the one or more cell IDs.

[0334] In some implementations, the location information may be one or more area IDs (e.g., a tracking area ID). In some implementations, the UE may activate the NCM-AMM configuration if the area ID associated with the camping cell is in the one or more area IDs.

[0335] In some implementations, the NCM-AMM configuration may be activated if the UE recognizes the change of Access and Mobility Management Function (AMF) or the change of Central Unit (CU).

[0336] In some implementations, the UE may activate the NCM-AMM configuration according to the speed of the UE.

[0337] In some implementations, the NCM-AMM configuration may be associated with athreshold (e.g., the NCM-AMM configuration may include the threshold). In some implementations, the UE may obtain the speed by the implementation of the UE. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is deactivated. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is activated.

[0338] In some implementations, the NCM-AMM configuration may be associated with a threshold and an indication of higher or lower for triggering the activation. For example, if the indication indicates 'higher' for triggering the activation, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is deactivated. For example, if the indication indicates 'lower' for triggering the condition, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is activated.

[0339] In some implementations, the network may activate NCM-AMM configurations for a UE via a paging message where a short message may indicate that there is activation / deactivation of an AI / ML model and then the UE may decode the corresponding paging message to understand which functionality (by checking the MI or associated ID) is activated or deactivated. This approach may enable the network to control the NCM-AMM configuration of the UE in RRC IDLE / INACTIVE state without requiring the UE to establish an RRC connection, thereby reducing signaling overhead.

[0340] In some implementations, when the UE activates the NCM-AMM configuration, the UE may perform a measurement procedure, and / or a cell reselection procedure according to the NCM-AMM configuration.

[0341] In some implementations, when the UE activates the NCM-AMM configuration, the UE may perform a measurement procedure, and / or a cell reselection procedure according to the AI / ML part of the NCM-AMM configuration.

[0342] Deactivation of NCM-AMM Configuration

[0343] In some implementations, the UE in RRC IDLE / INACTIVE state may deactivatean NCM-AMM configuration.

[0344] In some implementations, the UE may deactivate the NCM-AMM configuration upon receiving the RRC message and / or the system information including the NCM-AMM configuration.

[0345] In some implementations, the UE may deactivate the NCM-AMM configuration upon a timer expiration.

[0346] In some implementations, the UE may deactivate the NCM-AMM configuration upon a counter reaching a configured number.

[0347] In some implementations, the UE may deactivate the NCM-AMM configuration when at least one of the conditions for deactivating the NCM-AMM configuration is satisfied.

[0348] In some implementations, the UE may deactivate the NCM-AMM configuration upon RRC connection establishment or upon RRC resume procedures.

[0349] In some implementations, the UE may deactivate the NCM-AMM configuration upon releasing the NCM-AMM configuration.

[0350] In some implementations, the UE may deactivate the NCM-AMM configuration when the UE identifies that the ID of the camping cell is not in the one or more cell IDs associated with the NCM-AMM configuration.

[0351] In some implementations, the UE may deactivate the NCM-AMM configuration when the UE identifies that the area ID associated with the camping cell is not in the one or more area IDs associated with the NCM-AMM configuration.

[0352] In some implementations, the UE may deactivate the NCM-AMM configuration according to the speed of the UE.

[0353] In some implementations, the NCM-AMM configuration may be associated with a threshold (e.g., the NCM-AMM configuration may include the threshold). In some implementations, the UE may obtain the speed by the implementation of the UE. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is deactivated. For example, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is activated.

[0354] In some implementations, the NCM-AMM configuration may be associated with a threshold and an indication of higher or lower for triggering the activation. For example, if theindication indicates 'higher' for triggering the activation, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is activated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is deactivated. For example, if the indication indicates 'lower' for triggering the condition, when the UE identifies that the speed of the UE is higher than the threshold, the UE may consider that the NCM-AMM configuration is deactivated; when the UE identifies that the speed of the UE is lower than the threshold, the UE may consider that the NCM-AMM configuration is activated.

[0355] In some implementations, when the UE deactivates the NCM-AMM configuration, the UE may ignore the NCM-AMM configuration (e.g., the UE may not perform a measurement procedure and / or a cell reselection procedure according to the NCM-AMM configuration). For example, the UE may forgo applying the AI / ML-based measurement and cell reselection parameters in the NCM-AMM configuration.

[0356] In some implementations, when the UE deactivates the NCM-AMM configuration, the UE may perform a measurement procedure, and / or a cell reselection procedure according to the non-AI / ML part of the NCM-AMM configuration.

[0357] The above-described NCM-AMM activation and deactivation mechanisms (e.g., timer-based, counter-based, location-based, speed-based, paging-based, and RRC connection establishment-based mechanisms) are provided as illustrative examples, and the UE may use any one or any combination of these mechanisms. Other NCM-AMM activation / deactivation mechanisms not explicitly described herein are also within the scope of the present disclosure.

[0358] UE Behavior upon MO Activation

[0359] In some implementations, when the UE considers an MO to be activated, the UE may perform one or more of the following operations.

[0360] In some implementations, the UE may skip measurements associated with the MO (e.g., the UE may ignore the requirement for RRM measurement).

[0361] More specifically, the UE may skip the measurement in the time domain. In some implementations, the UE may skip the measurement at one or more measurement occasions, and the ratio of skipped measurement occasions may be less than a value configured by the network. In some implementations, the UE may measure a first number of consecutive measurement occasions and skip a second number of consecutive measurement occasions, where the first number may be larger than a first value and the second number may be less than a second value, and the first value and the second value may be configured by the network. In some implementations, the UE may start a first timer upon a measurement, and may keepcomplying with the measurement requirement when the timer is running and is not exceeding a first timer length; the UE may stop the first timer and start a second timer upon the first timer exceeding the first timer length; the UE may have a relaxed measurement (e.g., the UE may not comply with the measurement requirement) when the second timer is running and is not exceeding a second timer length; the first timer length and the second timer length may be configured by the network.

[0362] More specifically, the UE may skip the measurement in the space domain. For example, the UE may skip the measurement of one or more beams in a cell associated with the MO. In some implementations, the beams which are skipped for measurement may be configured by the network (e.g., if the network indicates the beams to be skipped for measurement, the UE may skip the measurement on the indicated beams). In some implementations, the UE may select one or more beams to be skipped for measurement, and the number of selected beams may be less than a value configured by the network.

[0363] In some implementations, the UE may use the measurement results in an observation window to predict the measurement results in a prediction window for the MO.

[0364] In some implementations, one or more of the measurement results in the observation window may be measured based on the SSBs and / or CSI-RSs of the cell associated with the MO.

[0365] In some implementations, one or more of the measurement results in the observation window may be predicted measurement results from the inference at one or more former time instances. More specifically, the UE may do so if there is an indication in the MO and / or MI / RC associated with the MO that contains an indication allowing the UE to use predicted results for AI / ML inference input.

[0366] In some implementations, one or more of the measurement results in the observation window may be filtered by LI or L3 filter and predicted measurement results from the inference at one or more former time instances may be filtered by LI or L3 filter, where the network may configure which kind of filter is to be used for measurement and inference.

[0367] In some implementations, the measurement results in the observation window and / or in the prediction window may be associated with the MO.

[0368] In some implementations, the measurement results in the observation window and / or in the prediction window may be associated with one or more other MOs. More specifically, if there is an indication configuring cluster-based AMM operation for the MO, the UE may use the measurement results from other MOs as the AI / ML inference input, and the UE may predict the measurement results for other MOs. More specifically, if there is anindication of the MOs applicable to cluster-based AMM operation, the UE may use the measurement results from the indicated MOs as the input for cluster-based AI / ML inference, and the UE may predict the measurement results for the indicated MOs.

[0369] These operations upon MO activation may enable the UE to reduce measurement effort while maintaining prediction accuracy through AI / ML inference, thereby reducing UE power consumption and improving the efficiency of the measurement process.

[0370] UE Behavior upon MO Deactivation

[0371] In some implementations, when the UE considers an MO to be deactivated, the UE may perform one or more of the following operations.

[0372] In some implementations, the UE may ignore the AMM related configuration associated with the MO.

[0373] In some implementations, the UE may perform measurement of the cells associated with the MO to comply with the measurement requirement. In some implementations, the UE may do so even if the MO includes AMM-related RRM relaxation configurations (e.g., maximum MRRT, maximum MRRS, skipping pattern).

[0374] In some implementations, if the MO includes non -AMM-related RRM relaxation configurations (e.g., a relaxation feature) and AMM-related RRM relaxation configurations, the UE may select one based on the least measurement samples.

[0375] In some implementations, the UE may not perform measurement of the cells not associated with the MO even if the MO includes cluster-based AMM configuration. For example, the UE may forgo measuring cells that are not associated with the MO, even when the MO includes cluster-based AMM configuration indicating other cells.

[0376] UE Behavior upon RC Activation

[0377] In some implementations, when the UE considers an RC to be activated, the UE may perform one or more of the following operations.

[0378] In some implementations, the UE may include the measured measurement results and corresponding timestamps in the measurement report and may transmit the measurement report to the network.

[0379] In some implementations, the UE may include the predicted measurement results and corresponding timestamps in the measurement report and may transmit the measurement report to the network. The measurement report for predicted measurement results may be transmitted in a different message from the measurement report for measured measurement results and these two messages may be transmitted based on different trigger conditions.

[0380] In some implementations, the UE may indicate the timestamps by integers, each ofwhich may be interpreted as the starting time plus the corresponding integer multiplied by the period, where the starting time may be the time when the UE generates the measurement report, and the period may be a configured value by the network.

[0381] In some implementations, the UE may trigger the measurement report procedure periodically: the UE may start or restart a first timer upon receiving an MI associated with the RC and / or upon transmitting a measurement report; the UE may stop the first timer and generate the measurement report when the first timer expires (e.g., the first timer exceeds a first timer length); the UE may stop or release the first timer when the RC is released.

[0382] In some implementations, the UE may trigger the measurement report procedure conditionally: the UE may generate a measurement report and transmit the measurement report if a condition is satisfied; the condition may be configured by the network and may be one of the measurement events in a wireless communication system. The UE may be configured with a first RC without AMM-related configuration and with a second RC with AMM-related configuration; if the UE considers that the second RC is activated, the UE may use a non-AI-based method to determine whether the event configured in the first RC is satisfied according to measured radio link quality; the UE may also use an Al-based method to determine whether the event in the second RC is satisfied according to the inference results.

[0383] In some implementations, the UE may trigger the CHO execution: the UE may apply a stored RRC configuration associated with a candidate PCell if a condition is satisfied; the condition may be configured by the network and may be one of the measurement events in a wireless communication system. The UE may be configured with a first RC without AMM-related configuration and with a second RC with AMM-related configuration; if the UE considers that the second RC is activated, the UE may use a non-AI-based method to determine whether the event configured in the first RC is satisfied according to measured radio link quality; the UE may also use an Al-based method to determine whether the event in the second RC is satisfied according to the inference results. In some implementations, the UE may trigger the CHO execution by one or more conditions configured by the network, based on the predicted measurement results of the associated MO. In some implementations, the network may include an indication in the RC, indicating that the UE may determine whether to trigger the CHO execution by the implementation of the UE; the UE may determine to trigger the CHO execution by the implementation of the UE (e.g., without a network-configured condition) if the indication is configured.

[0384] In some implementations, the UE may trigger the CPAC execution: the UE may apply a stored RRC configuration associated with a candidate PSCell if a condition is satisfied;the condition may be configured by the network and may be one of the measurement events in a wireless communication system. The UE may be configured with a first RC without AMM-related configuration and with a second RC with AMM-related configuration; if the UE considers that the second RC is activated, the UE may use a non-AI-based method to determine whether the event configured in the first RC is satisfied according to measured radio link quality; the UE may also use an Al-based method to determine whether the event in the second RC is satisfied according to the inference results. In some implementations, the UE may trigger the CPAC execution by one or more conditions configured by the network, based on the predicted measurement results of the associated MO. In some implementations, the network may include an indication in the RC, indicating that the UE may determine whether to trigger the CPAC execution by the implementation of the UE; the UE may determine to trigger the CPAC execution by the implementation of the UE (e.g., without a network-configured condition) if the indication is configured.

[0385] In some implementations, the UE may trigger the C-LTM execution: the UE may apply a stored RRC configuration associated with a candidate PCell or PSCell if a condition is satisfied; the condition may be configured by the network and may be one of the measurement events in a wireless communication system. The UE may be configured with a first RC without AMM-related configuration and with a second RC with AMM-related configuration; if the UE considers that the second RC is activated, the UE may use a non-AI-based method to determine whether the event configured in the first RC is satisfied according to measured radio link quality; the UE may also use an Al-based method to determine whether the event in the second RC is satisfied according to the inference results. The condition may be based on predicted / measured Layer 1 measurements and / or predicted / measured Layer 3 measurements. The condition may be based on predicted / measured beam-level qualities and / or predicted / measured cell-level qualities. In some implementations, the UE may trigger the C-LTM execution by one or more conditions configured by the network, based on the predicted measurement results of the associated MO. In some implementations, the network may include an indication in the RC, indicating that the UE may determine whether to trigger the C-LTM execution by the implementation of the UE; the UE may determine to trigger the C-LTM execution by the implementation of the UE (e.g., without a network-configured condition) if the indication is configured.

[0386] UE Behavior upon RC Deactivation

[0387] In some implementations, when the UE considers an RC to be deactivated, the UE may perform one or more of the following operations.

[0388] In some implementations, the UE may report the measurement results according to the part of non- AMM-r elated report configuration. For example, the RC may include an AMM-RC configuration and a non-AMM-RC configuration; the UE may report the measurement results according to the part of non-AMM-RC configuration and may ignore the part of AMM-RC configuration.

[0389] In some implementations, the UE may ignore the RC. For example, the UE may not perform measurement report according to the RC. For example, the UE may forgo reporting measurement results according to the RC. For example, the UE may not perform CHO / CPAC / C-LTM execution according to the RC. For example, the UE may forgo executing CHO / CPAC / C-LTM according to the RC.

[0390] UE Behavior upon MI Activation

[0391] In some implementations, when the UE considers an MI to be activated, the UE may perform one or more of the following operations.

[0392] In some implementations, the UE may perform measurement report and / or CHO / CPAC / C-LTM execution according to the MO and the RC associated with the MI.

[0393] In some implementations, the UE may perform measurement procedures for the cells associated with the MO associated with the MI.

[0394] In some implementations, the UE may apply the AMM-MO configuration in the MO associated with the MI and the AMM-RC configuration in the RC associated with the MI for the AMM operation.

[0395] UE Behavior upon MI Deactivation

[0396] In some implementations, when the UE considers an MI to be deactivated, the UE may perform one or more of the following operations.

[0397] In some implementations, the UE may ignore the MO and RC associated with the MI.

[0398] In some implementations, the UE may apply the non-AMM part of the MO associated with the MI and the non-AMM part of the RC associated with the MI, and the UE may perform the measurement and report procedures accordingly.

[0399] In some implementations, the AMM configuration may further include a cell list indicating one or more cells applicable for AMM operation. For example, the cell list may indicate one or more cells for which the UE may perform AI / ML inference. In some implementations, the UE may perform the AI / ML inference for the one or more cells indicated by the cell list. For example, the cell list may include cell identifiers (e.g., physical cell IDs or global cell IDs) of cells whose measurement results may be predicted by the AI / ML model.The cell list may enable the network to specify the scope of the AI / ML-based mobility operation by explicitly identifying the cells that are subject to AI / ML inference at the UE.

[0400] In some implementations, the AMM configuration may further include a skipping pattern associated with the second MO. For example, the skipping pattern may indicate one or more measurement occasions on which the UE may forgo measuring the second MO. In some implementations, the UE may forgo measuring one or more measurement occasions associated with the second MO based on the skipping pattern. For example, the UE may rely on the predicted measurement result obtained via AI / ML inference for the measurement occasions that are skipped on the second MO, thereby reducing the measurement effort on the second MO while maintaining the availability of measurement results through AI / ML prediction.

[0401] In some implementations, the skipping pattern may include a bitmap, and each bit of the bitmap may indicate to the UE whether to forgo measuring a corresponding measurement occasion. For example, a bit with a first value may indicate that the UE may forgo measuring the corresponding measurement occasion, and a bit with a second value may indicate that the UE may not forgo measuring the corresponding measurement occasion (e.g., the UE may perform the measurement on the corresponding measurement occasion). This bitmap-based skipping pattern may provide fine-grained control over which measurement occasions the UE may skip.

[0402] In some implementations, the AMM configuration may be associated with an activation condition, and the UE may perform the AI / ML inference in response to determining that the activation condition is satisfied. For example, when the UE determines that the activation condition associated with the AMM configuration is satisfied, the UE may activate the AI / ML-based mobility operation and may begin performing AI / ML inference based on the measurement results of the first MO to obtain predicted measurement results associated with the second MO. When the UE determines that the activation condition is not satisfied, the UE may not perform the AI / ML inference and may instead comply with the measurement requirement on the measurement occasions associated with the second MO. For example, the UE may forgo performing AI / ML inference when the activation condition is not satisfied.

[0403] In some implementations, in response to determining that the activation condition is satisfied, the UE may report, based on the RC, at least one of the predicted measurement result associated with a future time instance or a predicted measurement event, and may forgo measuring one or more measurement occasions on one or more cells associated with the second MO. For example, when the activation condition is satisfied, the UE may rely on AI / ML inference to predict the measurement results for the cells associated with the second MO atfuture time instances, and may skip the actual measurement on these cells, thereby reducing UE power consumption while still providing the network with measurement information through the predicted results.

[0404] In some implementations, the second MO may be associated with one or more inter-frequency cells. For example, the second MO may correspond to a measurement object configured for a frequency different from the frequency of the first MO. In response to determining that the activation condition is satisfied, the UE may report, based on the RC, the predicted measurement result associated with the one or more inter-frequency cells, and may forgo measuring one or more measurement occasions on one or more cells associated with the second MO. For example, the UE may use the measurement results of the first MO (e.g., intrafrequency measurement results) as the AI / ML inference input to predict the measurement results of the inter-frequency cells associated with the second MO, thereby potentially eliminating the need for measurement gaps for inter-frequency measurements.

[0405] In some implementations, in response to determining that the activation condition is no longer satisfied, the UE may comply with a measurement requirement on one or more measurement occasions associated with the second MO. For example, when the activation condition transitions from satisfied to unsatisfied, the UE may resume measuring the measurement occasions associated with the second MO that the UE previously forewent during the period when the activation condition was satisfied. For example, the UE may resume complying with the measurement requirement for the second MO by performing actual measurements rather than relying on AI / ML-predicted results.

[0406] FIG. 1 is a flowchart illustrating a method / process 100 for life cycle management of AI / ML-based mobility, according to an example implementation of the present disclosure. Although actions 102, 104, 106, and 108 are illustrated as separate actions, represented as independent blocks in FIG. 1, these separately illustrated actions should not be construed as necessarily order-dependent. The order in which the actions are performed in FIG. 1 is not intended to be construed as a limitation, and any number of the disclosed blocks may be combined in any order to implement the method, or an alternative method. Each of actions 102, 104, 106, and 108 may be performed independent of the other actions, and may be omitted in some implementations of the present disclosure. Moreover, the method / process 100 may be combined with other procedures / methods described in the present disclosure. The process 100 may be performed by a UE, with each action of the process 100 corresponding to an operation executed by the UE.

[0407] In action 102, the UE may receive, from a Base Station (BS), an ArtificialIntelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC). For example, the AMM configuration may be received via RRC signaling from the BS. The MI may associate the multiple MOs with the RC such that the UE may determine which MOs are to be used for actual measurement and which MOs are to be used for AI / ML inference. For example, the AMM configuration may include various parameters such as one or more of an associated ID, a skipping pattern, a maximum MRRT value, a maximum MRRS value, an indication of AMM Case A or Case B, cluster information, a reporting configuration for future time instances, and an activation condition. The AMM configuration may enable the network to instruct the UE regarding AI / ML-based measurement and reporting operations. In some other implementations, the AMM configuration may be received via a MAC CE, a DCI, or a combination thereof.

[0408] In action 104, the UE may perform measurement based on a first MO of the multiple MOs. For example, the UE may measure signal strength, signal quality, or other radio parameters of one or more cells associated with the first MO. The measurement may be based on SSBs and / or CSLRSs of the cells associated with the first MO. The measurement results may be collected within an observation window and may serve as input for AI / ML inference. In some implementations, the first MO may correspond to an intra-frequency measurement object, though the first MO may also correspond to an inter-frequency or inter-RAT measurement object.

[0409] In action 106, the UE may perform AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the multiple MOs. For example, the UE may input the measurement results of the first MO into an AI / ML model and may obtain predicted measurement results for one or more cells associated with the second MO. The predicted measurement results may correspond to future time instances and / or a current time instance. In some implementations, the second MO may be associated with one or more inter-frequency cells, and the UE may predict the measurement results of the inter-frequency cells based on the measurement results of the first MO without performing actual inter-frequency measurements. In some implementations, the UE may use cluster-based AI / ML inference, where the measurement results of multiple cells associated with the first MO may be used to predict the measurement results of multiple cells associated with the second MO.

[0410] In action 108, the UE may transmit, to the BS, a measurement report including the predicted measurement result associated with the second MO based on the RC. For example,the measurement report may include the predicted measurement results and corresponding timestamps. The RC may indicate reporting conditions such as periodic reporting, event-triggered reporting, or both. In some implementations, the measurement report may also include measured (non-predicted) measurement results in addition to the predicted measurement results. The measurement report may enable the network to make mobility management decisions (e.g., handover, conditional handover, LTM, or CPAC decisions) based on the predicted measurement results reported by the UE.

[0411] The process 100 may enable efficient AI / ML-based mobility management, allowing the UE to predict measurement results for cells that the UE may not have directly measured, thereby reducing measurement effort and UE power consumption while providing the network with timely measurement information for mobility decisions. The process 100 may also enable the UE to skip measurement occasions on the second MO and to rely on AI / ML-predicted results, which may reduce the need for measurement gaps and improve resource utilization.

[0412] FIG. 2 is a flowchart illustrating a method / process 200 for life cycle management of AI / ML-based mobility, according to an example implementation of the present disclosure. Although actions 202 and 204 are illustrated as separate actions, represented as independent blocks in FIG. 2, these separately illustrated actions should not be construed as necessarily order-dependent. The order in which the actions are performed in FIG. 2 is not intended to be construed as a limitation, and any number of the disclosed blocks may be combined in any order to implement the method, or an alternative method. Each of actions 202 and 204 may be performed independent of the other actions, and may be omitted in some implementations of the present disclosure. Moreover, the method / process 200 may be combined with other procedures / methods described in the present disclosure. The process 200 may be performed by a BS, with each action of the process 200 corresponding to an operation executed by the BS.

[0413] In action 202, the BS may transmit, to a User Equipment (UE), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration including a Measurement Identity (MI) associated with multiple Measurement Objects (MOs) and a Report Configuration (RC), a first MO of the multiple MOs being configured for measurement at the UE, a second MO of the multiple MOs being configured for AI / ML inference at the UE. For example, the BS may transmit the AMM configuration via RRC signaling. The BS may configure the first MO such that the UE performs actual measurement on cells associated with the first MO, and may configure the second MO such that the UE performs AI / ML inference to predict measurement results for cells associated withthe second MO. The AMM configuration may include various parameters such as one or more of an associated ID, a skipping pattern, a cell list, cluster information, a reporting configuration for future time instances, and an activation condition. In some other implementations, the AMM configuration may be transmitted via a MAC CE, a DCI, or a combination thereof.

[0414] In action 204, the BS may receive, from the UE, a measurement report including a predicted measurement result associated with the second MO based on the RC. For example, the measurement report may include predicted measurement results and corresponding timestamps. The BS may use the predicted measurement results for mobility management decisions, such as handover triggering, conditional handover configuration, LTM execution, or CPAC execution. In some implementations, the BS may also receive measured (non-predicted) measurement results from the UE in the same or a different measurement report.

[0415] The process 200 may enable the BS to configure AI / ML-based mobility operations at the UE and to receive predicted measurement results from the UE, thereby facilitating proactive mobility management based on AI / ML predictions while reducing measurement overhead at the UE.

[0416] In some implementations, the AMM configuration may further include a cell list indicating one or more cells applicable for AMM operation, and the UE may perform the AI / ML inference for the one or more cells indicated by the cell list. For example, the cell list may include cell identifiers (e.g., physical cell IDs or global cell IDs) that specify the cells for which the UE may perform AI / ML-based measurement prediction. The cell list may enable the network to limit the scope of AI / ML inference to specific cells, which may improve prediction accuracy by focusing the AI / ML model on a defined set of cells. This feature may also enable the network to dynamically adjust the set of cells subject to AI / ML inference based on the mobility scenario.

[0417] In some implementations, the AMM configuration may further include a skipping pattern associated with the second MO, and the UE may forgo measuring one or more measurement occasions associated with the second MO based on the skipping pattern. For example, the skipping pattern may specify which measurement occasions on the second MO the UE may skip, and the UE may rely on the AI / ML-predicted results for the skipped measurement occasions. The skipping pattern may enable the UE to reduce measurement effort on the second MO, thereby reducing UE power consumption while maintaining the availability of measurement results through AI / ML prediction.

[0418] In some implementations, the skipping pattern may include a bitmap, and each bit of the bitmap may indicate to the UE whether to forgo measuring a corresponding measurementoccasion. For example, a bit with a first value may indicate that the UE may forgo measuring the corresponding measurement occasion, and a bit with a second value may indicate that the UE may perform the measurement on the corresponding measurement occasion. The bitmapbased skipping pattern may provide fine-grained, per-measurement-occasion control over which measurement occasions the UE may skip on the second MO.

[0419] In some implementations, the RC may indicate that the UE is to report at least one of the predicted measurement result or a predicted measurement event. For example, the RC may indicate that the UE reports the predicted measurement results (e.g., predicted RSRP, predicted RSRQ, or predicted SINR values) for cells associated with the second MO. For example, the RC may indicate that the UE reports a predicted measurement event (e.g., an event indicating that a predicted measurement result satisfies a configured threshold at a future time instance). The RC may enable the network to flexibly control the type of prediction information reported by the UE, which may support different mobility management strategies (e.g., event-triggered handover based on predicted events vs. periodic reporting of predicted measurement results).

[0420] In some implementations, the UE may transmit, to the BS, capability information indicating a maximum prediction window length supported by the UE for the AI / ML inference. For example, the maximum prediction window length may indicate the furthest future time instance for which the UE may provide a reliable predicted measurement result. The capability information may be transmitted via an RRC message (e.g., a UECapability Information message). The maximum prediction window length may enable the network to configure the reporting configuration for future time instances in a manner that is compatible with the AI / ML model capabilities of the UE.

[0421] In some implementations, the AMM configuration may be associated with an activation condition, and performing the AI / ML inference may include performing the AI / ML inference in response to determining that the activation condition is satisfied. For example, when the activation condition is satisfied, the UE may activate the AEML-based mobility operation and may begin performing AI / ML inference. When the activation condition is not satisfied, the UE may not perform the AI / ML inference and may instead perform measurements without AI / ML operation. The activation condition may enable the network to control when the UE activates the AI / ML-based mobility operation, ensuring that AI / ML inference is performed only under conditions where the AI / ML model is expected to provide reliable predictions.

[0422] In some implementations, in response to determining that the activation conditionis satisfied, the UE may report, based on the RC, at least one of the predicted measurement result associated with a future time instance or a predicted measurement event, and may forgo measuring one or more measurement occasions on one or more cells associated with the second MO. For example, when the activation condition is satisfied, the UE may rely on AI / ML inference to predict the measurement results for the cells associated with the second MO at future time instances, and may skip the actual measurement on these cells. This approach may enable the UE to reduce power consumption by avoiding unnecessary measurements while still providing the network with measurement information through AI / ML-predicted results.

[0423] In some implementations, the second MO may be associated with one or more inter-frequency cells, and in response to determining that the activation condition is satisfied, the UE may report, based on the RC, the predicted measurement result associated with the one or more inter-frequency cells, and may forgo measuring one or more measurement occasions on one or more cells associated with the second MO. For example, the UE may use the measurement results of the first MO (e.g., intra-frequency measurement results) as the AI / ML inference input to predict the measurement results of the inter-frequency cells associated with the second MO. This approach may enable the UE to obtain inter-frequency measurement results without requiring measurement gaps, thereby improving resource utilization and reducing interruption to data transmission.

[0424] In some implementations, the activation condition may include at least one of an identity of a serving cell being in a cell list included in the AMM configuration, a radio link quality of the serving cell satisfying a first threshold, a radio link quality of a neighbor cell satisfying a second threshold, or a speed of the UE satisfying a speed threshold. For example, the UE may activate the AI / ML inference when the serving cell ID matches one of the cell IDs in the cell list, or when the radio link quality (e.g., RSRP, RSRQ, or SINR) of the serving cell satisfies the first threshold, or when the radio link quality of a neighbor cell satisfies the second threshold, or when the speed of the UE satisfies the speed threshold. The multiple types of activation conditions may provide the network with flexibility to tailor the AI / ML activation criteria to different deployment scenarios and mobility patterns. In some implementations, in response to determining that the activation condition is no longer satisfied, the UE may comply with a measurement requirement on one or more measurement occasions associated with the second MO. For example, when the activation condition transitions from satisfied to unsatisfied, the UE may resume measuring the measurement occasions associated with the second MO that the UE previously forewent during the period when the activation condition was satisfied. This approach may enable the UE to fall back to non-AI / ML-based measurement when theconditions for reliable AI / ML prediction are no longer met, thereby maintaining measurement accuracy under varying conditions.

[0425] In some implementations, the AMM configuration may further include a cell list indicating one or more cells applicable for AMM operation, and performing the AI / ML inference may include performing the AI / ML inference for the one or more cells indicated by the cell list. The cell list may enable the network to scope the AI / ML inference to a defined set of cells, improving prediction accuracy and reducing computational overhead at the UE.

[0426] In some implementations, the AMM configuration may further include a skipping pattern associated with the second MO, and the method may further include forgoing measuring one or more measurement occasions associated with the second MO based on the skipping pattern. The skipping pattern may enable the UE to reduce measurement effort on the second MO while relying on AI / ML-predicted results for the skipped occasions.

[0427] In some implementations, the AMM configuration may be associated with an activation condition, and performing the AI / ML inference may include performing the AI / ML inference in response to determining that the activation condition is satisfied. The activation condition may enable the network to control when the AI / ML inference is performed at the UE, ensuring that inference is performed under conditions favorable for accurate prediction.

[0428] In some implementations, the RC may indicate that the UE is to report at least one of the predicted measurement result or a predicted measurement event. The RC may enable the network to control whether the UE reports predicted measurement values, predicted measurement events, or both, providing flexibility for different mobility management strategies.

[0429] In some implementations, the AMM configuration may further include a cell list indicating one or more cells applicable for AMM operation. The cell list may enable the BS to specify the scope of the AI / ML-based mobility operation at the UE by identifying the cells for which the UE may perform AI / ML inference. This feature may enable the BS to align the cells subj ect to AI / ML inference with the cells for which the B S needs predicted measurement results for mobility management purposes.

[0430] In some implementations, the AMM configuration may further include a skipping pattern associated with the second MO. The skipping pattern may enable the BS to control the degree of measurement relaxation at the UE on the second MO, balancing between measurement effort reduction and the availability of actual measurement results from the UE.

[0431] In some implementations, the AMM configuration may be associated with an activation condition for the AI / ML inference at the UE. The activation condition may enable the BS to control under what conditions the UE activates the AI / ML-based mobility operation,ensuring that AI / ML inference at the UE is performed under conditions where the AI / ML model is expected to provide reliable predictions.

[0432] The network device, such as the BS, may perform methods / actions corresponding to those performed by the UE. For example, the receiving actions performed by the UE may correspond to the transmitting / configuring actions of the network device; the transmitting actions performed by the UE may correspond to the receiving actions of the network device. That is, the network device and the UE may have reciprocally aligned roles in transmission and reception. For example, when the UE receives the AMM configuration from the BS (action 102 in FIG. 1), the BS may transmit the AMM configuration to the UE (action 202 in FIG. 2). When the UE transmits the measurement report including the predicted measurement result to the BS (action 108 in FIG. 1), the BS may receive the measurement report from the UE (action 204 in FIG. 2). The BS may use the predicted measurement results received in the measurement report for mobility management decisions, such as handover triggering, conditional handover configuration, or resource allocation.

[0433] The UE may perform methods / actions corresponding to those performed by the BS. For example, the receiving actions performed by the BS may correspond to the transmitting actions of the UE; the transmitting actions performed by the BS may correspond to the receiving actions of the UE. That is, the UE and the BS may have reciprocally aligned roles in transmission and reception. For example, when the BS transmits the AMM configuration to the UE (action 202 in FIG. 2), the UE may receive the AMM configuration from the BS (action 102 in FIG. 1). When the BS receives the measurement report from the UE (action 204 in FIG.2), the UE may transmit the measurement report to the BS (action 108 in FIG. 1). The UE may perform measurement, AI / ML inference, and measurement reporting in accordance with the AMM configuration received from the BS, thereby enabling the BS to obtain predicted measurement results without requiring the UE to perform actual measurements on all configured measurement objects.

[0434] FIG. 3 is a block diagram illustrating a node for wireless communication, according to an example implementation of the present disclosure. As shown in FIG. 3, a node 300 may include a transceiver 320, a processor 328, a memory 334, one or more presentation components 338, and at least one antenna 336. The node 300 may also include an RF spectrum band module, a base station communications module, a network communications module, and a system communications management module, Input / Output (I / O) ports, I / O components, and a power supply (not explicitly shown in FIG. 3). Each of these components may be in communication with each other, directly or indirectly, over one or more buses 340. In oneimplementation, the node 300 may be a UE or a base station that performs various functions described herein.

[0435] The transceiver 320 having a transmitter 322 (e.g., transmitting / transmission circuitry) and a receiver 324 (e.g., receiving / reception circuitry) may be configured to transmit and / or receive time and / or frequency resource partitioning information. In some implementations, the transceiver 320 may be configured to transmit in different types of subframes and slots including, but not limited to, usable, non-usable and flexibly usable subframes and slot formats. The transceiver 320 may be configured to receive data and control channels.

[0436] The node 300 may include a variety of computer-readable media. Computer-readable media can be any available media that can be accessed by the node 300 and may include both volatile and non-volatile media, removable and non-removable media. By way of example, and not limitation, computer-readable media may include computer storage media and communication media. Computer storage media may include both volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, Artificial Intelligence (Al) / Machine Learning (ML) module(s) or data.

[0437] Computer storage media may include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices. Computer storage media does not include a propagated data signal. Communication media typically embodies computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and may include any information delivery media. The term "modulated data signal" may correspond to a signal that has one or more of the characteristics of the signal set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media may include wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media. Combinations of any of the above should also be included within the scope of computer-readable media.

[0438] The memory 334 may include computer- storage media in the form of volatile and / or non-volatile memory. The memory 334 may be removable, non-removable, or a combination thereof. Exemplary memory includes solid-state memory, hard drives, optical-disc drives, etc. As illustrated in FIG. 3, the memory 334 may store computer-readable,computer-executable instructions 332 (e.g., software codes) that are configured to, when executed, cause the processor 328 to perform various functions described herein. The instructions 332 may not be directly executable by the processor 328 but may be configured to cause the node 300 (e.g., when compiled and executed) to perform various functions described herein. The memory 334 may also store data 330.

[0439] The processor 328 (e.g., having processing circuitry) may include an intelligent hardware device, e.g., a Central Processing Unit (CPU), a microcontroller, an ASIC, etc. The processor 328 may include memory. The processor 328 may process the data 330 and the instructions 332 received from the memory 334, and information through the transceiver 320, the base band communications module, and / or the network communications module. The processor 328 may also process information to be sent to the transceiver 320 for transmission through the antenna 336, to the network communications module for transmission to a core network.

[0440] One or more presentation components 338 may present data indications to a person or other device. Exemplary presentation components 338 may include a display device, speaker, printing component, vibrating component, etc.

[0441] From the above description, various techniques may be used for implementing the concepts described in the present disclosure without departing from the scope of those concepts. Moreover, while the concepts have been described with specific reference to certain implementations, a person of ordinary skill in the art would recognize that changes may be made in form and detail without departing from the scope of those concepts. As such, the described implementations are to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A User Equipment (UE), the UE comprising:at least one processor; andat least one non-transitory computer-readable medium coupled to the at least one processor and storing one or more computer-executable instructions that, when executed by the at least one processor, cause the UE to:receive, from a Base Station (BS), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration comprising a Measurement Identity (MI) associated with a plurality of Measurement Objects (MOs) and a Report Configuration (RC);perform measurement based on a first MO of the plurality of MOs;perform AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the plurality of MOs; and transmit, to the BS, a measurement report comprising the predicted measurement result associated with the second MO based on the RC.

2. The UE of claim 1, wherein the AMM configuration further comprises a cell list indicating one or more cells applicable for AMM operation, and wherein the one or more computerexecutable instructions, when executed by the at least one processor, further cause the UE to perform the AI / ML inference for the one or more cells indicated by the cell list.

3. The UE of claim 1, wherein the AMM configuration further comprises a skipping pattern associated with the second MO, and wherein the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to forgo measuring one or more measurement occasions associated with the second MO based on the skipping pattern.

4. The UE of claim 3, wherein the skipping pattern comprises a bitmap, each bit of the bitmap indicating to the UE whether to forgo measuring a corresponding measurement occasion.

5. The UE of claim 1, wherein the RC indicates that the UE is to report at least one of:the predicted measurement result, ora predicted measurement event.

6. The UE of claim 1, wherein the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to transmit, to the BS, capability information indicating a maximum prediction window length supported by the UE for the AI / ML inference.

7. The UE of claim 1, wherein the AMM configuration is associated with an activation condition, and wherein performing the AI / ML inference comprises performing the AI / ML inference in response to determining that the activation condition is satisfied.

8. The UE of claim 7, wherein, in response to determining that the activation condition is satisfied, the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to:report, based on the RC, at least one of the predicted measurement result associated with a future time instance or a predicted measurement event; andforgo measuring one or more measurement occasions on one or more cells associated with the second MO.

9. The UE of claim 7, wherein the second MO is associated with one or more inter-frequency cells, and wherein, in response to determining that the activation condition is satisfied, the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to:report, based on the RC, the predicted measurement result associated with the one or more inter-frequency cells; andforgo measuring one or more measurement occasions on one or more cells associated with the second MO.

10. The UE of claim 7, wherein the activation condition comprises at least one of:an identity of a serving cell being in a cell list comprised in the AMM configuration, a radio link quality of the serving cell satisfying a first threshold,a radio link quality of a neighbor cell satisfying a second threshold, ora speed of the UE satisfying a speed threshold.

11. The UE of claim 7, wherein, in response to determining that the activation condition is no longer satisfied, the one or more computer-executable instructions, when executed by the at least one processor, further cause the UE to comply with a measurement requirement on one or more measurement occasions associated with the second MO.

12. A method performed by a User Equipment (UE), the method comprising:receiving, from a Base Station (BS), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration comprising a Measurement Identity (MI) associated with a plurality of Measurement Objects (MOs) and a Report Configuration (RC);performing measurement based on a first MO of the plurality of MOs;performing AI / ML inference based on a measurement result of the first MO to obtain a predicted measurement result associated with a second MO of the plurality of MOs; and transmitting, to the BS, a measurement report comprising the predicted measurement result associated with the second MO based on the RC.

13. The method of claim 12, wherein:the AMM configuration further comprises a cell list indicating one or more cells applicable for AMM operation, andperforming the AI / ML inference comprises performing the AI / ML inference for the one or more cells indicated by the cell list.

14. The method of claim 12, wherein the AMM configuration further comprises a skipping pattern associated with the second MO, and the method further comprises:forgoing measuring one or more measurement occasions associated with the second MO based on the skipping pattern.

15. The method of claim 12, wherein the AMM configuration is associated with an activation condition, and performing the AI / ML inference comprises performing the AI / ML inference in response to determining that the activation condition is satisfied.

16. The method of claim 12, wherein the RC indicates that the UE is to report at least one of: the predicted measurement result, ora predicted measurement event.

17. A Base Station (BS), the BS comprising:at least one processor; andat least one non-transitory computer-readable medium coupled to the at least one processor and storing one or more computer-executable instructions that, when executed by the at least one processor, cause the BS to:transmit, to a User Equipment (UE), an Artificial Intelligence / Machine Learning (AI / ML)-based Mobility (AMM) configuration, the AMM configuration comprising a Measurement Identity (MI) associated with a plurality of Measurement Objects (MOs) and a Report Configuration (RC), a first MO of the plurality of MOs being configured for measurement at the UE, a second MO of the plurality of MOs being configured for AI / ML inference at the UE; andreceive, from the UE, a measurement report comprising a predicted measurement result associated with the second MO based on the RC.

18. The BS of claim 17, wherein the AMM configuration further comprises a cell list indicating one or more cells applicable for AMM operation.

19. The BS of claim 17, wherein the AMM configuration further comprises a skipping pattern associated with the second MO.

20. The BS of claim 17, wherein the AMM configuration is associated with an activation condition for the AI / ML inference at the UE.