WO2025081505A1

WO2025081505A1 - Devices and methods for communication

Info

Publication number: WO2025081505A1
Application number: PCT/CN2023/125795
Authority: WO
Inventors: Peng Guan; Zhen He; Yukai GAO; Gang Wang
Original assignee: NEC Corp
Current assignee: NEC Corp
Priority date: 2023-10-20
Filing date: 2023-10-20
Publication date: 2025-04-24
Anticipated expiration: 2026-04-20

Abstract

Embodiments of the present disclosure provide a solution for a solution for enabling consistency of model-related or functionality-related information. The device determines information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and performs, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage.

Description

DEVICES AND METHODS FOR COMMUNICATION

FIELDS

Example embodiments of the present disclosure generally relate to the field of communication techniques and in particular, to devices and methods for enabling consistency of model-related or functionality-related information.

BACKGROUND

As communication networks and services increase in size, complexity, and number of users, operations in the communication networks may become increasingly more complicated. In order to improve the communication performance, machine learning (ML) /artificial intelligence (AI) technology is proposed to be used in the wireless communication network. For example, the terminal device and the network device may use different ML models to assist communication-related functionalities, such as, beam management (BM) , mobility management and so on.

SUMMARY

In general, embodiments of the present disclosure provide a solution for enabling consistency of model-related or functionality-related information.

In a first aspect, there is provided a device comprising: a processor configured to cause the device to: determine information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and perform, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage.

In a second aspect, there is provided a communication method performed by a device. The method comprises: determining information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and performing, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage.

In a third aspect, there is provided a computer readable medium having instructions stored thereon, the instructions, when executed on at least one processor, causing the at least one processor to carry out the method according to the second aspect.

Other features of the present disclosure will become easily comprehensible through the following description.

BRIEF DESCRIPTION OF THE DRAWINGS

Through the more detailed description of some example embodiments of the present disclosure in the accompanying drawings, the above and other objects, features and advantages of the present disclosure will become more apparent, wherein:

FIG. 1 illustrates an example communication environment in which example embodiments of the present disclosure can be implemented;

FIG. 2A illustrates an example consistency across different ML stages;

FIG. 2B illustrates example sets of beams across different ML stages;

FIG. 2C illustrates example sets of beams across different ML stages;

FIG. 2D illustrates example mappings of sets of beams;

FIG. 3, illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 4 illustrates a 2-D planar antenna structure where each column is a cross-polarized array;

FIG. 5 illustrates example mappings among the resources and the weights;

FIG. 6A illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 6B illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 7A illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 7B illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 8 illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 9 illustrates a signaling flow of communication in accordance with some embodiments of the present disclosure;

FIG. 10 illustrates a flowchart of a method implemented at a device according to some example embodiments of the present disclosure; and

FIG. 11 illustrates a simplified block diagram of an apparatus that is suitable for implementing example embodiments of the present disclosure.

Throughout the drawings, the same or similar reference numerals represent the same or similar element.

DETAILED DESCRIPTION

Principle of the present disclosure will now be described with reference to some example embodiments. It is to be understood that these embodiments are described only for the purpose of illustration and help those skilled in the art to understand and implement the present disclosure, without suggesting any limitation as to the scope of the disclosure. Embodiments described herein can be implemented in various manners other than the ones described below.

In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skills in the art to which this disclosure belongs.

As used herein, the term ‘terminal device’ refers to any device having wireless or wired communication capabilities. Examples of the terminal device include, but not limited to, user equipment (UE) , personal computers, desktops, mobile phones, cellular phones, smart phones, personal digital assistants (PDAs) , portable computers, tablets, wearable devices, internet of things (IoT) devices, Ultra-reliable and Low Latency Communications (URLLC) devices, Internet of Everything (IoE) devices, machine type communication (MTC) devices, devices on vehicle for V2X communication where X means pedestrian, vehicle, or infrastructure/network, devices for Integrated Access and Backhaul (IAB) , Space borne vehicles or Air borne vehicles in Non-terrestrial networks (NTN) including Satellites and High Altitude Platforms (HAPs) encompassing Unmanned Aircraft Systems (UAS) , eXtended Reality (XR) devices including different types of realities such as Augmented Reality (AR) , Mixed Reality (MR) and Virtual Reality (VR) , the unmanned aerial vehicle (UAV) commonly known as a drone which is an aircraft without any human pilot, devices on high speed train (HST) , or image capture devices such as digital cameras, sensors, gaming devices, music storage and playback appliances, or Internet appliances enabling wireless or wired Internet access and browsing and the like. The ‘terminal device’ can further have ‘multicast/broadcast’ feature, to support public safety and mission critical, V2X applications, transparent IPv4/IPv6 multicast delivery, IPTV, smart TV, radio services, software delivery over wireless, group communications and IoT applications. It may also incorporate one or multiple Subscriber Identity Module (SIM) as known as Multi-SIM. The term “terminal device” can be used interchangeably with a UE, a mobile station, a subscriber station, a mobile terminal, a user terminal or a wireless device.

The term “network device” refers to a device which is capable of providing or hosting a cell or coverage where terminal devices can communicate. Examples of a network device include, but not limited to, a Node B (NodeB or NB) , an evolved NodeB (eNodeB or eNB) , a next generation NodeB (gNB) , a transmission reception point (TRP) , a remote radio unit (RRU) , a radio head (RH) , a remote radio head (RRH) , an IAB node, a low power node such as a femto node, a pico node, a reconfigurable intelligent surface (RIS) , and the like.

The terminal device or the network device may have Artificial intelligence (AI) or Machine learning capability. It generally includes a model which has been trained from numerous collected data for a specific function, and can be used to predict some information.

The terminal or the network device may work on several frequency ranges, e.g., FR1 (e.g., 450 MHz to 6000 MHz) , FR2 (e.g., 24.25GHz to 52.6GHz) , frequency band larger than 100 GHz as well as Tera Hertz (THz) . It can further work on licensed/unlicensed/shared spectrum. The terminal device may have more than one connection with the network devices under Multi-Radio Dual Connectivity (MR-DC) application scenario. The terminal device or the network device can work on full duplex, flexible duplex and cross division duplex modes.

The embodiments of the present disclosure may be performed in test equipment, e.g., signal generator, signal analyzer, spectrum analyzer, network analyzer, test terminal device, test network device, channel emulator. In some embodiments, the terminal device may be connected with a first network device and a second network device. One of the first network device and the second network device may be a master node and the other one may be a secondary node. The first network device and the second network device may use different radio access technologies (RATs) . In some embodiments, the first network device may be a first RAT device and the second network device may be a second RAT device. In some embodiments, the first RAT device is eNB and the second RAT device is gNB. Information related with different RATs may be transmitted to the terminal device from at least one of the first network device or the second network device. In some embodiments, first information may be transmitted to the terminal device from the first network device and second information may be transmitted to the terminal device from the second network device directly or via the first network device. In some embodiments, information related with configuration for the terminal device configured by the second network device may be transmitted from the second network device via the first network device. Information related with reconfiguration for the terminal device configured by the second network device may be transmitted to the terminal device from the second network device directly or via the first network device.

As used herein, the singular forms ‘a’ , ‘an’ and ‘the’ are intended to include the plural forms as well, unless the context clearly indicates otherwise. The term ‘includes’ and its variants are to be read as open terms that mean ‘includes, but is not limited to. ’ The term ‘based on’ is to be read as ‘at least in part based on. ’ The term ‘one embodiment’ and ‘an embodiment’ are to be read as ‘at least one embodiment. ’ The term ‘another embodiment’ is to be read as ‘at least one other embodiment. ’ The terms ‘first, ’ ‘second, ’ and the like may refer to different or same objects. Other definitions, explicit and implicit, may be included below.

In some examples, values, procedures, or apparatus are referred to as ‘best, ’ ‘lowest, ’ ‘highest, ’ ‘minimum, ’ ‘maximum, ’ or the like. It will be appreciated that such descriptions are intended to indicate that a selection among many used functional alternatives can be made, and such selections need not be better, smaller, higher, or otherwise preferable to other selections.

As used herein, the term “resource, ” “transmission resource, ” “uplink resource, ” or “downlink resource” may refer to any resource for performing a communication, such as a resource in time domain, a resource in frequency domain, a resource in space domain, a resource in code domain, or any other resource enabling a communication, and the like. In the following, unless explicitly stated, a resource in both frequency domain and time domain will be used as an example of a transmission resource for describing some example embodiments of the present disclosure. It is noted that example embodiments of the present disclosure are equally applicable to other resources in other domains.

As used herein, the terms “UE expects” , “UE does not expect, “terminal device expects” , “terminal device does not expect” may imply restrictions on a configuration of a network device (also referred to as NW configuration) . The terms “UE is not expected to” and “terminal device is not expected to” may imply a terminal implementation, also referred to as UE implementation. In some embodiments, the terms “UE does not expect” and “UE is not expected to” may be used equally.

As discussed above, the terminal device and the network device may use different ML models to assist communication-related functionalities, such as, beam management (BM) , mobility management and so on.

By far, for AI/ML-based beam management, it has been agreed to support below BM-Case1 and BM-Case2:

BM-Case1: Spatial-domain downlink (or uplink) beam prediction for Set A of beams based on measurement results of Set B of beams.

BM-Case2: Temporal downlink (or uplink) beam prediction for Set A of beams based on the historic measurement results of Set B of beams.

Further, for BM-Case1 and BM-Case2, beams in above Set A and Set B may be in the same frequency range (FR) .

In case of BM-Case1 and BM-Case2, the following alternatives for the predicted beams may be supported: downlink transmitting (TX) beam prediction, downlink receiving (RX) beam prediction, beam pair prediction (abeam pair consists of a downlink TX beam and a corresponding downlink RX beam) .

Regarding the sub-use cases of BM-Case1 and BM-Case2, the following alternatives for AI/ML output may be supported: TX and/or RX Beam identity (ies) ID (s) and/or the predicted layer 1 (L1) -reference signal receiving power (RSRP) of the N predicted downlink TX and/or RX beams, e.g., N predicted beams can be the top-N predicted beams; TX and/or RX beam ID (s) of the N predicted downlink TX and/or TX beams and other information (e.g., probability for the beam to be the best beam, the associated confidence, beam application time/dwelling time, predicted beam failure) , where N predicted beams may be the top-N predicted beams; TX and/or RX Beam angle (s) and/or the predicted L1-RSRP of the N predicted DL TX and/or RX beams, where N predicted beams can be the top-N predicted beams.

For BM-Case1 and BM-Case2 with a UE-side AI/ML model, it is expected to study the necessity and potential BM-specific conditions/additional conditions for functionality (ies) and/or model (s) at least from the following aspects: information regarding model inference; Set A/Set B configuration; performance monitoring; data collection; and assistance information.

For BM-Case1 and BM-Case2 with a UE-side AI/ML model, below indication may be considered: an indication of the associated Set A from network to UE, e.g., association/mapping of beams within Set A and beams within Set B if applicable; a beam indication from network for UE reception.

Regarding the data collection for AI/ML model training at UE side, configurations (e.g., configuration related to set A and/or Set B, information on association/mapping of Set A and Set B) would be provided, and assistance information may be provided by the network to UE.

In the following, “reference pattern information” / “reference pattern” will be used as an example of information used for enabling the consistency for describing some specific example embodiments of the present disclosure. It is noted that example embodiments described with regard to the “reference pattern information” / “reference pattern” are equally applicable to the information used for enabling the consistency.

For better descriptions, some terms used herein are listed as below:

● AI/ML model: refers to a data driven algorithm that applies AI/ML techniques to generate a set of outputs based on a set of inputs;

● AI/ML model delivery: refers to a generic term referring to delivery of an AI/ML model from one entity to another entity in any manner. Note: An entity could mean a network node/function (e.g., gNB, LMF, etc. ) , UE, proprietary server, and so on;

● Functionality: refers to an AI/ML-enabled feature/feature group (FG) enabled by configuration (s) , where configuration (s) is (are) supported based on conditions indicated by UE capability;

● Functionality-based LCM: operates based on, at least, one configuration of AI/ML-enabled feature/FG or specific configurations of an AI/ML-enabled feature/FG;

● Model-ID-based LCM: operates based on identified models, where a model may be associated with specific configurations/conditions associated with UE capability of an AI/ML-enabled Feature/FG and additional conditions (e.g., scenarios, sites, and datasets) as determined/identified between UE-side and NW-sider;

● AI/ML model inference: refers to a process of using a trained AI/ML model to produce a set of outputs based on a set of inputs;

● AI/ML model testing: refers to a subprocess of training, to evaluate the performance of a final AI/ML model using a dataset different from one used for model training and validation. Differently from AI/ML model validation, testing does not assume subsequent tuning of the model;

● AI/ML model training: refers to a process to train an AI/ML Model [by learning the input/output relationship] in a data driven manner and obtain the trained AI/ML Model for inference;

● AI/ML model transfer: refers to delivery of an AI/ML model over the air interface in a manner that is not transparent to 3GPP signalling, either parameters of a model structure known at the receiving end or a new model with parameters. Delivery may contain a full model or a partial model;

● AI/ML model validation: refers to a subprocess of training, to evaluate the quality of an AI/ML model using a dataset different from one used for model training, that helps selecting model parameters that generalize beyond the dataset used for model training;

● Data collection: refers to a process of collecting data by the network nodes, management entity, or UE for the purpose of AI/ML model training, data analytics and inference;

● Federated learning/federated training: refers to a machine learning technique that trains an AI/ML model across multiple decentralized edge nodes (e.g., UEs, gNBs) each performing local model training using local data samples. The technique requires multiple interactions of the model, but no exchange of local data samples;

● Functionality identification: refers to a process/method of identifying an AI/ML functionality for the common understanding between the network and the UE. Note: Information regarding the AI/ML functionality may be shared during functionality identification. Where AI/ML functionality resides depends on the specific use cases and sub use cases;

● Model activation: refers to enable an AI/ML model for a specific AI/ML-enabled feature;

● Model deactivation: refers to disable an AI/ML model for a specific AI/ML-enabled feature;

● Model download: refers to transfer a Model from the network to UE;

● Model identification: refers to a process/method of identifying an AI/ML model for the common understanding between the network and the UE. Note: The process/method of model identification may or may not be applicable;

● Information regarding the AI/ML model may be shared during model identification;

● Model monitoring: refers to a procedure that monitors the inference performance of the AI/ML model;

● Model parameter update: refers to a process of updating the model parameters of a model;

● Model selection: refers to a process of selecting an AI/ML model for activation among multiple models for the same AI/ML enabled feature. Note: Model selection may or may not be carried out simultaneously with model activation;

● Model switching: refers to deactivating a currently active AI/ML model and activating a different AI/ML model for a specific AI/ML-enabled feature;

● Model update: refers to a process of updating the model parameters and/or model structure of a model;

● Model upload: refers to transfer a Model from UE to the network;

● Network-side (AI/ML) model: refers to an AI/ML Model whose inference is performed entirely at the network;

● Offline field data: refers to the data collected from field and used for offline training of the AI/ML model;

● Offline training: refers to an AI/ML training process where the model is trained based on collected dataset, and where the trained model is later used or delivered for inference. Note: This definition only serves as a guidance. There may be cases that may not exactly conform to this definition but could still be categorized as offline training by commonly accepted conventions;

● Online field data: refers to the data collected from field and used for online training of the AI/ML model;

● Online training: refers to an AI/ML training process where the model being used for inference) is (typically continuously) trained in (near) real-time with the arrival of new training samples. Note: the notion of (near) real-time and non real-time are context-dependent and is relative to the inference time-scale. Note: This definition only serves as a guidance. There may be cases that may not exactly conform to this definition but could still be categorized as online training by commonly accepted conventions. Note: Fine-tuning/re-training may be done via online or offline training. (This note could be removed when we define the term fine-tuning) ;

● Reinforcement Learning (RL) : refers to a process of training an AI/ML model from input (a.k.a. state) and a feedback signal (a. k. a. reward) resulting from the model’s output (a.k.a. action) in an environment the model is interacting with;

● Semi-supervised learning A process of training a model with a mix of labelled data and unlabeled data;

● Supervised learning: refers to a process of training a model from input and its corresponding labels;

● Two-sided (AI/ML) model: refers to a paired AI/ML Model (s) over which joint inference is performed, where joint inference comprises AI/ML Inference whose inference is performed jointly across the UE and the network, i.e., the first part of inference is firstly performed by UE and then the remaining part is performed by gNB, or vice versa;

● UE-side (AI/ML) model: refers to an AI/ML Model whose inference is performed entirely at the UE;

● Unsupervised learning: refers to a process of training a model without labelled data;

● Proprietary-format models ML models of vendor-/device-specific proprietary format, from 3GPP perspective. They are not mutually recognizable across vendors and hide model design information from other vendors when shared. Note: An example is a device-specific binary executable format;

● Open-format models: refers to ML models of specified format that are mutually recognizable across vendors and allow interoperability, from 3GPP perspective. They are mutually recognizable between vendors and do not hide model design information from other vendors when shared;

● Measurement result (s) may include but be not limited to, (L1/L3) -reference signal received power (RSRP) , L1/L3) -SINR, (L1/L3) -received signal strengthen indicator (RSSI) , or (L1/L3) -reference signal received quality (RSRQ) ;

● “Conditions” : configurations supported indicated via UE capability reporting related to model training, model inference, Performance monitoring, validation procedure, fallback, of an AI/ML model/functionality or a group of models/functionalities;

● “Additional conditions” : e.g., application conditions, scenarios, datasets, cell ID, timestamp and SNR, and so on; For an AI/ML-enabled feature/FG, additional conditions refer to any aspects that are assumed for the training of the model but are not a part of UE capability for the AI/ML-enabled feature/FG; Additional conditions can be divided into two categories: NW-side additional conditions and UE-side additional conditions. For an AI/ML-enabled feature/feature group, additional conditions refer to any aspects that are assumed for the training of the model but are not a part of UE capability for the AI/ML-enabled feature/feature group. Additional conditions can be divided into two categories: NW-side additional conditions and UE-side additional conditions;

● “UE internal conditions” : e.g., memory, battery, computation resource, overheating and other hardware limitations;

● NW/network device: may be an access network device, or a core network device, such as, “Operation Administration and Maintenance (OAM) ” , “server” , “Access and Mobility Management Function (AMF) /Location Management Function (LMF) ” .

In the present disclosure,

terms of “information used for enabling the consistency” “calibration information” and “consistency information” may be used interchangeably;

terms of “model” “functionality” and “model/functionality” may be used interchangeably;

terms of “ID” , “index” , “indicator” and “identifier” may be used interchangeably;

terms “model” , “model group” , “group of model” , “model set” , “a set of group” may be used interchangeably;

terms “functionality” , “functionality group” , “group of functionalities” , “functionality set” and “set of functionalities” may be used interchangeably.

In the present disclosure, a beam may correspond to a channel state information-reference signal (CSI-RS) , a synchronization signal and physical broadcast channel (PBCH) block (SSB) , a CSI-RS resource, or an SSB resource. Accordingly, a beam identity (ID) may be a CSI-RS resource indicator (CRI) , an SSB resource indicator (SSBRI) , or a RS ID. It also should be understood that in fact, a beam refers to a resource that enables a spatially directional communication, and thus may be identified by other suitable parameter in other embodiments. In present disclosure is not limited in this regard.

As used herein, the term “ML stage” may be replaced by “ML phase” , “ML procedure” , “LCM stage” , “LCM phase” , “LCM procedure” , including but not limited to, model delivery procedure/phase/stage, model inference procedure/phase/stage, model testing procedure/phase/stage, model training procedure/phase/stage, model monitoring procedure/phase/stage, model transfer procedure/phase/stage, model validation procedure/phase/stage, data collection procedure/phase/stage, model learning procedure/phase/stage and so on.

As used herein, a physical model ID may refer to a real AI/ML model, a real implementation; a logical model ID may be associated with one or a group of physical models for the same purpose. Further, the global model ID and the local model ID also may be used to identity a model.

It is noted that when the term “a set of” is used, it may mean one or more elements/items, which may be replaced by terms of “at least one” , “a group of” or “a list of” . For example, “a set of X” means “at least one X” or “one or more Xs” .

As used herein, a model may be equivalent to at least one of the following: an AI/ML model, a ML model, an AI model, a data-driven, a data processing model, an algorithm, a functionality, a procedure, a process, an entity, a function, a feature, a feature group, a model identifier (ID) , an ID, a functionality ID, a configuration ID, a scenario ID, a site ID, or a dataset ID. As a result, the above terms may be used interchangeably.

In some embodiments, the model may comprise a set of weights values that may be learned during training, for example for a specific architecture or configuration, where a set of weights values may also be called a parameter set.

In some embodiments, the model may be used to predict a target cell, or measurements of a set of beams of a set of candidate cells in future based on at least historical measurements (e.g., L1-RSRP, L1-SINR) of a set of beams of a set of candidate cells.

In some embodiments, an input of the ML model (i.e., AI input) may refer to the input of a model and indicate data inputted into the model, which may be equivalent to data.

In some embodiments, an output of ML model (i.e., AI output) may refers to the output of a model and indicate result (s) outputted by the model, which is equivalent to label/data.

Principles and implementations of the present disclosure will be described in detail below with reference to the figures.

Example environment

FIG. 1 illustrates a schematic diagram of an example communication environment 100 in which example embodiments of the present disclosure can be implemented. In the communication environment 100, a plurality of communication devices, including a device 110 and a further device 120, can communicate with each other.

Further, multiple input multiple output (MIMO) is supported in the communication environment 100, such that the further device 120 and the device 110 may communicate with each other via different beams to enable a directional communication.

In the example of FIG. 1, in some embodiments, the device 110 may a terminal device or a network device (including an access network device or a core network device) , and the further device 120 may be a terminal device or a network device (including an access network device or a core network device) .

Further, in some embodiments, one of the device 110 and the further device 120 may comprise a terminal device, which may be referred to as a first device, and the other one of the device 110 and the further device 120 may comprise a network device, which may be referred to as a second device. In this specific example embodiment, a link from the first device to the second device is referred to as uplink, while a link from the second device to the first device is referred to as a downlink.

In downlink, the second device is a transmitting (TX) device (or a transmitter) and the first device is a receiving (RX) device (or a receiver) . Correspondingly, in uplink, the second device is an RX device (or a receiver) and the first device is a TX device (or a transmitter) .

In FIG. 1, the device 110 and the further device 120 may communicate with each other via one or more beams. As illustrated in FIG. 1, the device 110 may communicate with the further device 120 via the beams 130-1 to 130-3. For purpose of discussion, the beams 130-1 to 130-3 are collectively or individually referred to as beam 130. As illustrated in FIG. 1, the further device 120 may communicate with the device 110 via the one or more of beams 140-1, 140-2 and 140-3. For purpose of discussion, the beams 140-1 to 140-3 are collectively or individually referred to as beam 140.

In some embodiments, one or more models may be deployed at the further device 120 and/or the device 110. As illustrated in FIG. 1, the model 115 is deployed at the device 110 and/or the model 125 is deployed at the further device 120. Further, in the example of FIG. 1, the model 115 and/or the model 125 may assist such as BM, i.e., obtain input data and derive related output.

It is to be understood that the number of devices and their connections shown in FIG. 1 are only for the purpose of illustration without suggesting any limitation. The communication environment 100 may include any suitable number of devices configured to implementing example embodiments of the present disclosure.

In some embodiments, the device 110 and the further device 120 may communicate with each other via a channel such as a wireless communication channel on an air interface (e.g., Uu interface) . The wireless communication channel may comprise a physical uplink control channel (PUCCH) , a physical uplink shared channel (PUSCH) , a physical random-access channel (PRACH) , a physical downlink control channel (PDCCH) , a physical downlink shared channel (PDSCH) and a physical broadcast channel (PBCH) . Of course, any other suitable channels are also feasible.

The communications in the communication environment 100 may conform to any suitable standards including, but not limited to, Global System for Mobile Communications (GSM) , Long Term Evolution (LTE) , LTE-Evolution, LTE-Advanced (LTE-A) , New Radio (NR) , Wideband Code Division Multiple Access (WCDMA) , Code Division Multiple Access (CDMA) , GSM EDGE Radio Access Network (GERAN) , Machine Type Communication (MTC) and the like. The embodiments of the present disclosure may be performed according to any generation communication protocols either currently known or to be developed in the future. Examples of the communication protocols include, but not limited to, the first generation (1G) , the second generation (2G) , 2.5G, 2.75G, the third generation (3G) , the fourth generation (4G) , 4.5G, the fifth generation (5G) communication protocols, 5.5G, 5G-Advanced networks, or the sixth generation (6G) networks.

Example processes

Generally speaking, for BM-Case1 and BM-Case2 with a UE-side AI/ML model, consistency/association of Set B beams and Set A beams across training and inference is beneficial from performance perspective. Thus, for BM-Case1 and BM-Case2 with a UE-side AI/ML model, regarding aspects related to association/mapping of beams within Set A and beams within Set B, mechanisms to ensure consistency of Set B beams and Set A beams across at least training and inference is needed (such as order/indexing consistently) .

Reference is now made to FIG. 2A, which illustrates an example consistency 200A across different ML stages. In some embodiments, the consistency of Set B beams and Set A beams includes at least one or more of the following:

● Set size consistency for Set B, Set A: consistency in number of beams and/or associated resources for Set B and Set A, across training and inference;

● Order/indexing consistency: consistency in ordering of resources (e.g., resource index consistency) for Set B beams and Set A beams, across training and inference;

● QCL consistency: consistency in QCL relationship of Set A beams with respect to Set B beams, across training and inference;

● Beam shape consistency: relative pointing direction and beamwidth difference between physical beams with respect to Set A and Set B resources across training and inference should be under predefined tolerances.

Merely for a better understanding, reference is now made to FIG. 2B and FIG. 2C, which illustrates example sets of beams 200B and 200C across different ML stages.

In the example of FIG. 2B, two stages for Set A and Set B beams based on different possible gNB codebooks (Set B is not a subset of Set A) are illustrated, where Set A and Set B beams during different stages are different.

In the example of FIG. 2C, two cases for Set A and Set B beams based on different possible gNB codebooks (Set B is a subset of Set A) are illustrated, where Set A and Set B beams during different stages are different.

In addition, the consistency of set A and set B beams is difficult to achieve because beamforming algorithms may be dramatically different across different network device/UE vendors. Even for a network device/UE vendors, it is up to network device /UE implementation on whether to use the same beamforming weights all the time. Further, the mapping between “physical” beams and “logical” resources are up to configuration/implementation too.

Merely for a better understanding, reference is now made to FIG. 2D, which illustrates example mappings 200D of sets of beams, where resource IDs 1 to 64 are corresponding to the Set A. In FIG. 2D, implementation (a) shows that resource IDs {18, 20, 22, 24, 42, 44, 46, 48} are corresponding to the Set B. By contrast, in the implementation (b) , resource IDs {18, 20, 22, 24, 42, 44, 48} are associated with other beams, which causes inconsistency. Actually, resource IDs {11, 27, 43, 59, 14, 30, 46, 62} in the implementation (b) is consistent with the resource IDs {18, 20, 22, 24, 42, 44, 46, 48} in the implementation (a) .

According to the example of FIG. 2D, it may be seen that even for same beam shape, the relationship still may be different if only depends on the resource IDs.

Due to the above reasons, beam IDs/resource IDs for Set B beams and Set A beams cannot guarantee the required consistency for the relationship between Set B beams and Set A beams.

By far, no detailed solution which may ensure consistency of Set B beams and Set A beams across training and inference has been proposed.

According to the example processed discussed in the following, at least the above issue may be avoided.

According to some embodiments, a solution for enabling consistency of model-related or functionality-related information is proposed. In particular, a reference (beam) pattern is introduced to enable the consistency for beams across different devices and across different ML stages (i.e., AI/ML lifecycle management (LCM) stages) .

Reference is made to FIG. 3, which illustrates a signaling flow 300 for communicating information about the number of predicted beams in accordance with some embodiments of the present disclosure. For the purposes of discussion, the signaling flow 300 will be discussed with reference to FIG. 1, for example, by using the device 110 and the further device 120.

It is to be understood that the operations at the device 110 and the further device 120 should be coordinated. In other words, the further device 120 and the device 110 should have common understanding about configurations, parameters and so on. Such common understanding may be implemented by any suitable interactions between the further device 120 and the device 110 or both the further device 120 and the device 110 applying the same rule/policy. In the following, although some operations are described from a perspective of the device 110, it is to be understood that the corresponding operations should be performed by the further device 120. Similarly, although some operations are described from a perspective of the further device 120, it is to be understood that the corresponding operations should be performed by the device 110. Merely for brevity, some of the same or similar contents are omitted here.

In the following embodiments, the device 110 may be described as a terminal device or a network device (including an access network device or a core network device) , and the further device 120 may be described as a terminal device or a network device (including an access network device or a core network device) .

The terms “consistency” herein may refer to the consistency between conditions/additional conditions/implementation/algorithms applied during model training and conditions/additional conditions/implementation/algorithms applied during model inference, for example, the consistency between beams applied during model training and beams applied during model inference. In addition, it may be extended to other LCM stages.

In operation, the device 110 determines 310 information used for enabling the consistency of model-related or functionality-related information between a first ML stage of a model or a functionality and a second ML stage of the model or the functionality.

In the following, details about the information used for enabling the consistency will be discussed.

In some embodiments, the information used for enabling the consistency may indicate at least one of the following:

● first information indicating at least one used or supported reference pattern,

● second information indicating at least one used or supported reference value,

● third information indicating at least one used or supported assumption,

● fourth information indicating whether the consistency is needed,

● fifth information indicating a tolerance range associated with the information.

In some embodiments, the information used for enabling the consistency (also referred to as calibration information or consistency information) may the following:

● reference pattern information (i.e., the first information) ; Further, for AI/ML model developed for beam management, the reference pattern is a reference beam pattern;

● other calibration information, or consistency information include the reference values of and/or whether the consistency assumptions hold on parameters like transmit power, receiver type, speed, rotation, and so on, i.e., the second and third information.

In some embodiment, a reference pattern (first information) for an AI/ML model for beam management may be represented by the following: antenna modelling information, beamforming information, the relationship among beamforming information and resource IDs for Set B beams and corresponding Set A beams.

In some embodiment, one model may correspond to one piece of antenna modelling information/beamforming information, or may correspond to one relationship between the beamforming information and at least one resource. In other words, the mapping between the model and the antenna modelling information/beamforming information/relationship between the beamforming information and at least one resource is a one-to-one mapping. It is clarified that although the one-to-one mapping is discussed, in some cases, the mapping may be changed, such as, the mapping may be a multi-to-one mapping, a one-to-multi mapping or a multi to multi mapping.

In summary, one or more reference pattern (first information) may be specific to one or more models or functionalities.

In some embodiments, the first information may comprise at least one of the following:

● antenna modelling information associated with at least one of the device 110 or the further device 120,

● beamforming information associated with at least one of the device 110 or the further device 120,

● orientation information between the device 110 and the further device 120, or

● a relationship between the beamforming information and at least one resource, each of the at least one resource being associated with a beam.

In some embodiments, the first information may comprise NW-UE orientation information, for example, the NW panel and UE panel are parallel and pointing towards broad-side to each other.

In some embodiments, the first information may comprise the relationship or transformation between local and global coordinate systems, which may be used for defining the radiation pattern, pointing direction, azimuth/elevation informationNW-UE orientation information, and so on.

In some embodiments, the antenna modelling information may comprise at least one of the following:

● the number of antenna elements in a first spatial dimension, represented as M;

● the number of antenna elements in a second spatial dimension, represented as N;

● the number of panels in a first spatial dimension, represented as M_g,

● the number of panels in a second spatial dimension, represented as N_g,

● an antenna element spacing in a first spatial dimension, represented as d_H,

● an antenna element spacing in a second spatial dimension, represented as d_v,

● a panel spacing in a first spatial dimension, represented as d_g, V,

● a panel spacing in a second spatial dimension, represented as d_g, H,

● the number of polarizations, represented as P,

● an antenna array type,

● an antenna numbering rule,

● a polarization slant angle,

● a polarization type,

● an antenna element radiation pattern in a first spatial dimension,

● an antenna element radiation pattern in a second spatial dimension,

● a combining method for a multi-dimension antenna element pattern,

● the maximum directional gain of an antenna element,

● at least one complex weight for an antenna element in an elevation, or,

● antenna height.

Merely for a better understanding, some example embodiments about the antenna modelling information will be discussed as below.

In some embodiments, the antenna modelling information may comprise atenna modelling information for NW, for example, including (M, N, P, M_g, N_g) and (d_V, d_H) , in addition (d_g, V, d_g, H) , where N is the number of columns (or number of horizontal antenna elements) , M is the number of antenna elements with the same polarization in each column, P is the number of polarizations, M_g, N_g is the number of vertical panels and horizontal panels, respectively. Antenna elements are uniformly spaced in the horizontal direction with a spacing of d_H and in the vertical direction with a spacing of d_V, in addition, Antenna panels are uniformly spaced in the horizontal direction and the vertical direction with (d_g, V, d_g, H) .

Alternatively, in in addition, in some embodiments, the antenna modelling information may comprise one or many of the following information/parameters:

cross-polarized array, uniform linear array, and so on,

antenna numbering, e.g., assumes observation of the antenna array from the front (with x-axis pointing towards broad-side and increasing y-coordinate for increasing column number) ,

polarization slant angle,

polarization type: Linear (cross, vertical, horizontal) , Circular, Elliptical, and so on,

antenna element vertical radiation pattern (dB) ,

antenna element horizontal radiation pattern (dB) ,

combining method for 3D antenna element pattern (dB) ,

maximum directional gain of an antenna element,

complex weight for antenna element in elevation, or

antenna height, and so on.

Reference is now made to FIG. 4, which illustrates a 2-D planar antenna structure 400 where each column is a cross-polarized array. It should be noted that although discussions are made by assuming 2-dimension (2D) planar antenna array structure, i.e., antenna elements are placed in the vertical and horizontal direction, it may also be 1-dimension (1D) linear array or other antenna structure. The present disclosure is not limited in this regard.

Alternatively, or in addition, in some embodiments, the antenna modelling information also may comprise antenna modelling information for a UE, which is similar with the antenna modelling information for an NW. For brevity, the same or similar contents are omitted.

In some embodiments, the antenna modelling information may also be different for Set B beams and corresponding Set A beams.

In some embodiments, multiple antenna model parameters sets may be associated with one AI/ML model/functionality.

As discussed above, the mapping the model and the antenna modelling information also may be a multi-to-one mapping, a one-to-multi mapping or a multi to multi mapping. In view of this, in some embodiments, one or multiple antenna modelling parameters sets may be associated with one AI/ML model/functionality. Alternatibely, in some embodiments, one antenna modelling parameters sets may be associated with one or multiple AI/ML model (s) /functionality (ies) .

In some embodiments, the beamforming information comprises at least one of the following:

● a beamforming type,

● the number of transceiver units (TXRUs)

● the number of beams, or

● at least one beamforming weights.

In some embodiments, the beamforming type is one of the following: an analog beamforming, a digital beamforming, a hybrid beamforming.

In some embodiments, the number of TXRUs may be a total number of TXRUs, or may be associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle.

In some embodiments, the number of beams may be a total number of beams, or may be associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle.

In some embodiments the at least one beamforming weights comprises: at least one discrete fourier transform (DFT) weights, a set of weights associated with a specific bean set, a correspondence between beams and a weights matrix.

Merely for a better understanding, some example embodiments about the beamforming information will be discussed as below.

In some embodiments, the beamforming information for NW is discussed as below.

In some embodiments, the beamforming information may comprise beamforming type: such as, analog, digital, hybrid, and so on.

In some embodiments, the beamforming information may comprise the number of TXRUs, in addition, may include the number of antenna elements per TXRU, such as,

● the total number of TXRU, a number of TXRU for beams in set A, a number of TXRU for beams in Set B, a number of TXRU for first type of beams (e.g., wide beams or SSB beams) , a number of TXRU for second type of beams (e.g., narrow beams or CSI-RS beams) ,

● the number of TXRU for horizontal/vertical beams respectively, for azimuth/elevation beams respectively,

● the number of TXRU in a first/second dimension (e.g., horizontal/vertical) for a beam, respectively,

● In some embodiments, the beamforming information may comprise the number of beams, in addition, may be one or many of the following:

● the total number of NW Tx beams, number of beams in set A, number of beams in Set B, number of first type of beams (e.g., wide beams or SSB beams) , number of second type of beams (e.g., narrow beams or CSI-RS beams) ,

● the number of horizontal beams, number of vertical beams, or

● the number of azimuth beams (N_az) , number of elevation beams (N_el) .

In some embodiments, the beamforming information may comprise beamforming weights (or TXRU weights mapping, codebook, and son on) of each beam, for example, the 2D DFT weights (weights matrix W with N_az azimuth beams and N_el elevation beams) , may be shown in the formulas:

whereindenotes Kronecker product, W is weights matrix, w_az and w_el are beamforming vectors, N_az is the number of azimuth beams, N_el is the number of elevation beams, w_N (θ) is a base vector, is a vector w_N (θ) applied with N_az and θ_i andis a vector w_N (θ) applied with N_el andθ_i is an azimuth value andis an elevation value.

It should be understood that except for the above DFT method, weights matrix W also may be defined based on non-DFT based methods.

In some embodiments, the beamforming weights also may comprise weights for beams in Set B and Set A, where Set A may be a subset of all NW beams. Further, the relationship between Set B and Set A may be determined if the weights of beams in Set B and Set A are known.

In some embodiments, the beamforming weights also may comprise order/indexing for beam in Set B and Set A, where each beam may be referred by the following:

● which weight vectors for generating the beam,

● information on positions in beamforming weight matrix W, e.g., i-th column and j-th row, or in beamforming vectors w_az, w_el respectively, e.g., i-th element in w_az and j-th element in w_el respectively,

●information, e.g., azimuth and elevation angles of the beam, or the position/order of thein the set. For example, it may also be azimuth of departure angles (AOD, maybe after scaling) , zenith of departure angles (ZOD, maybe after scaling) , for Rx Beam at UE side, may use azimuth of arrival angles (AOA) , zenith of arrival angles (ZOA) .

In some embodiments, the beamforming weights also may comprise further optimization of weights, for example, orthogonality of beams, oversampling DFT, antenna gain pattern shaping, side lobe/back lobe control weights, gating control weights, which may also include beam shape information: pointing direction, beam width, and so on.

Alternatively, or in addition, in some embodiments, the beamforming information may comprise the beamforming information for the UE, which is similar with the beamforming information for the NW. For brevity, the same or similar contents are omitted.

In some embodiments, the beamforming information may also be different for Set B beams and corresponding Set A beams.

In some embodiments, multiple Beamforming information parameters sets may be associated with one AI/ML model/functionality.

As discussed above, the mapping the model and the antenna modelling information also may be a multi-to-one mapping, a one-to-multi mapping or a multi to multi mapping. In view of this, in some embodiments, one or multiple beamforming information parameters sets may be associated with one AI/ML model/functionality. Alternatively, in some embodiments, one beamforming information parameters sets may be associated with one or multiple AI/ML model (s) /functionality (ies) .

In some embodiments, the beamforming information may be a set of weights, and the relationship between the beamforming information and the at least one resource may be a mapping between the set of weights and the at least one resource.

In some embodiments, each element in the set of weights may correspond to: a first value in the first spatial dimension and a second value in the second spatial dimension.

Alternatively, in some embodiments, each element in the set of weights may correspond to a specific azimuth value and a specific elevation.

In some embodiments, the relationship between the beamforming information and the at least one resource may be represented by an order used for mapping the at least one resource to a subset in the set of weights, the order comprising an ascending or descending order of at least one of the following:

● first a first spatial dimension then a second spatial dimension,

● first the second spatial dimension then the first spatial dimension,

● first an azimuth dimension then an elevation dimension,

● first the elevation dimension then the azimuth dimension, or

● an order of a plurality of beam groups, the at least one beam groups being divided into the plurality of beam groups.

Alternatively, or in addition, in some embodiments, the relationship between the beamforming information and the at least one resource may be represented by information indicating the subset in the set of weights, the information comprising at least one of the following:

● an indication indicating a start position of the subset,

● the number of rows in the subset,

● the number of columns in the subset,

● a first step size in the rows if elements in the subset are not adjacent weights, or

● a second step size in the columns if the elements in the subset are not adjacent weights.

In summary, the reference pattern may be represented by a relationship between beamforming weights and resource IDs.

Reference is now made to FIG. 5, which illustrates example mappings 500 among the resources and the weights. Implementations (1) and (2) are resource IDs, while implementations (a) , (b) and (c) are weights.

In some embodiments, the relationship between the beamforming information and the at least one resource may be represented by resource ID numbering, e.g., first horizontal/azimuth then vertical/elevation, or first vertical/elevation then horizontal/azimuth, as shown in FIG. 5.

In some embodiments, each resource ID may correspond to an element in W (as implementations (a) , (b) in FIG. 5) , or a (as implementations (c) in FIG. 5) .

Alternatively, or in addition, grouping of beams may be applied first and resource ID numbering as above is performed within a group, as implementations (2) in FIG. 5. The groups may be ordered in first horizontal/azimuth, then vertical/elevation, or vice versa.

In addition, the numbering may be across groups: e.g., first a same first relative index within each group across groups and then a same second relative index within each group across groups.

In some embodiments, the relationship between the beamforming information and the at least one resource may be represented by an indication of this relationship, which may indicate W_i, _j orfor each resource ID or for Set A/B, for example, the start position (row/column, horizontal/vertical) and the number of elements per (row/column, horizontal/vertical) , in addition, step size information if the beamforming weights are not adjacent weights (as the first step size and the second step size in FIG. 5) .

In the following, how to determine/exchange the information used for enabling the consistency will be discussed.

In the present disclosure, any of the device 110 and the further device 120 may determine the information used for enabling the consistency and transmit the information used for enabling the consistency to the other device. Thus, in some embodiments, the device 110 may receive 310-1 a first message comprising the information used for enabling consistency, and then may determine the information used for enabling consistency based on the first message.

Alternatibely, in some embodiments, the device 110 may determine the information used for enabling the consistency by itself, and then transmits 310-2 a second message comprising the information used for enabling the consistency to the further device 120.

In some embodiments the first message may be received or the second message may be transmitted during at least one of the following:

● a dataset delivery procedure for the model or the functionality,

● a model delivery procedure for the model or the functionality,

● a data collection procedure for the model or the functionality,

● a model inference procedure for the model or the functionality,

● a model monitoring procedure for the model or the functionality,

● a model training procedure for the model or the functionality,

● a registration or identification procedure for the model or the functionality, or

● the calibration procedure.

In some embodiments, the information may be comprised in at least one of the following:

● a model description,

● a feature or feature group,

● capability-related information,

● supported condition information associated with the model or the functionality, or

● additional condition information associated with the model or the functionality.

In some embodiments, the information used for enabling consistency of model-related or functionality-related information, especially the reference pattern information (i.e., the first information) , may be indicated during at least one of the following: dataset transfer/delivery, data collection for model training, model inference, model monitoring, model transfer/delivery, model/functionality registration/identification, UE capability reporting or feature/feature group reporting. Alternatively, or in addition, the information used for enabling consistency of model-related or functionality-related information may be part of condition/additional condition associated with an AI/ML model/functionality. In this way, calibration information may be exchanged among different devices.

Beacuse delivering the information representing the reference pattern may require a large data size, and thus it may be more suitable using RRC message, MAC CE and other suitable signaling.

In some embodiments, if more than one reference patterns (also may be other calibration information) is available/can be used, reference pattern ID/index may be provided for each configured/reported/activated reference pattern.

In some embodiments, for a respective AI/ML model/functionality, the reference patterns (also may be other calibration information) may be indicated in at least one of the following:

● dataset delivery/exchange procedure: reference pattern information is indicated together with the dataset, or the mapping relationship between reference pattern ID/index and dataset ID/index is predefined, NW configured or UE reported; Alternatively, the reference pattern may be implied by the dataset, i.e., indicated implicitly. In some embodiments, the reference pattern information may be associated with the dataset, the reference pattern information may be part of the dataset or the reference pattern information may be reflected by the dataset information,

● data collection procedure: reference pattern information is indicated together with the trigger information of data collection or configuration of data collection; Alternatively, the reference pattern may be implied by the trigger information of data collection or configuration of data collection, i.e., indicated implicitly. In some embodiments, the reference pattern information may be associated with the trigger information of data collection or configuration of data collection, the reference pattern information may be part of the trigger information of data collection or configuration of data collection or the reference pattern information may be reflected by the trigger information of data collection or configuration of data collection,

● model transfer/delivery procedure: reference pattern information is indicated together with the model, for example, AI/ML Model is transferred from NW to UE, or 3^rd party to UE.Alternatively, the reference pattern may be implied by the model information, i.e., indicated implicitly. In some embodiments, the reference pattern information may be associated with the model information, the reference pattern information may be part of the model information or the reference pattern information may be reflected by the model information,

● model/functionality registration/identification procedure; For model-ID based AI/ML lifecycle management, reference pattern information can be indicated in model ID or model description for each model; For functionality based AI/ML lifecycle management, reference pattern information can be indicated for each AI/ML feature/feature group. Alternatively, the reference pattern may be implied by the model ID or the model description, i.e., indicated implicitly. In some embodiments, the reference pattern information may be associated with the model ID or the model description, the reference pattern information may be part of the model ID or the model description or the reference pattern information may be reflected by the model ID or the model description,

● alternatively, or in addition, the calibration information especially the reference pattern information can be part of condition/additional condition associated with an AI/ML model/functionality, for example, as one of assistance information. In addition, if more than one parameter set of conditions/additional conditions is available/can be used, condition/additional condition ID/index can be provided for each parameters set of condition/additional condition. Alternatively, the reference pattern may be implied by the condition/additional condition, i.e., indicated implicitly. In some embodiments, the reference pattern information may be associated with the condition/additional condition, the reference pattern information may be part of the condition/additional condition or the reference pattern information may be reflected by the condition/additional condition.

In some embodiments, for UE side model, the UE could provide the reference patterns (also may be other calibration information) via UE capability reporting, or via feature/feature group. Further, the reference patterns (also may be other calibration information) may be used for online model training, for model inference, for model monitoring.

In some embodiments, the UE capability reporting or feature/feature group reporting may provide: whether the UE needs the consistency between model training and model inference, for activating/applying an AI/ML model or functionality.

In some embodiments, the UE capability reporting or feature/feature group reporting may provide: whether the UE needs the consistency between offline (or non-real-time) model training and online (or real-time) model training (and/or fine-tuning) , and/or the consistency between model inference and model monitoring.

In some embodiments, the UE capability reporting or feature/feature group reporting may provide the supported reference patterns, such as,

● the patterns may be those used in model training or offline model training, or

● the number of reference patterns may be same or different for different AI/ML model/functionality.

As one example, for a supported AI/ML model/functionality spatial domain beam prediction with Set A has 64 beams and Set B has 1/8 of the size of Set A, the supported reference patterns may be: 1) Set A with 8 *8 beams based on 8*8 2D DFT weights and Set B with 8 beams selected uniformly as in the implementation (a) or (b) in FIOG. 2D; 2) Set A with 16*4 beams based on 16*4 2D DFT weights and Set B, and so on.

In some embodiments, the UE capability reporting or feature/feature group reporting may provide the supported tolerance range/threshold for the difference from the reference pattern, where within the tolerance range, the model inference can still work. For example, for the Set A and Set B in a reference pattern, the tolerated beam in Set B may satisfy ||w_{beam in SetB}-w_{tolerated beam}||²≤ε, where ε is the tolerance range/threshold, w_{beam in SetB} are weights of beams in Set B, w_{tolerated beam} are weights of the reference beams.

For NW side model, usually, NW may not need to provide reference patterns, at least for DL Tx beam prediction. However, to control UE side beamforming used for measurement and report, NW may also provide reference patterns , which can be done via QCL/Rx beam indication. In some embodiments, the NW may request the UE to apply a specific QCL assumption/Rx beamforming weights for measurement and report or simply ask UE to apply the same QCL assumption/Rx beam.

In the following, how to utilize the information used for enabling the consistency will be discussed.

In some embodiments, the device 110 may perform 320 the first ML stage with the further device 120 based on the information used for enabling consistency. Alternatibely, or in addition, in some embodiments, the device 110 may perform 330 the second ML stage with the further device 120 based on the information used for enabling consistency.

In some embodiments, in case that the device 110 is a network device 110 and the further device 120 is a terminal device 110, the device 110 may transmit, to the further device 120 and based on the information used for enabling the consistency, at least one of the following: a measurement configuration to be used by the further device 120 for data collection, or information about updated beamforming information.

In some embodiments, in case that the device 110 is a terminal device 110 and the further device 120 is a network device 110, the device 110 may perform, based on the information used for enabling consistency, measurements on at least one beam, and may transmit measurement results of at least part of the at least one beam to the further device 120.

For better understanding, reference is now made to FIG. 6A and 6B, which illustrate signaling flows 600A and 600B of communication in accordance with some embodiments of the present disclosure.

In the examples of FIG. 6A and 6B, the calibration information especially the reference pattern may be used for data collection for model training, model inference, model monitoring and other ML stages, such that the calibration at the UE/NW side for inputs collected from different NW devices is achieved.

In some embodiments, for UE side model, based on the reference pattern, NW may decide/adjust its resource configuration and the transmit beamforming weight for each resource for model training, inference, monitoring at UE side respectively.

As illustrated in FIG. 6A, in some embodiments, the NW may decide/adjust its resource configuration (i.e., Set B) or the transmit beamforming weight for each resource for model inference (i.e., for UE to predict the best beam in Set A) .

For model training, model monitoring or other LCM, NW may decide/adjust its resource configuration and/or the transmit beamforming weight for each resource for both set B and set A, as well as the relationship between Set B and Set A.

It is noted that when different UEs has different reference patterns, it is difficult for NW device to meet all UE requirements, which means it would be almost impossible to adjust beamforming weights, but it is still possible to adjust resource configuration.

For NW side model, NW may provide reference pattern information to UE to collect data from UE used for model training, model inference, model monitoring at NW side.

As illustrated in FIG. 6B, in some embodiments, the NW may ask UE to consider reference Rx beam to measure Set B beams and to provide beam report for NW side model inference (i.e., for NW to predict the best beam in Set A based on UE report Set B measurement results) .

In some embodiments, for model training, model monitoring or other LCM, NW may ask UE to adjust its Rx beamforming weight for each resource for both set B and set A too.

In some embodiments, for NW side model or third part model, the reference pattern can be used with exchanging data among different NW/UE devices, e.g., different gNBs, cells, and so on.

As discussed above, the calibration information, especially reference pattern information, may be used to keep consistency, which may be provided via UE capability or feature/feature group reporting and may be used for data collection for model training, model inference and model monitoring, such that the consistency during different stages of AI/ML model LCM (such as, data collection from different devices) is achieved.

In some embodiments, a calibration procedure may be used to check whether consistency holds, to calibrate if consistency does not hold, based on the calibration information especially reference pattern.

In some embodiments, the calibration procedure may be used to during model inference phase (also may be other ML stage, such as, before data collection for model training, before or after model update ) , or during model monitoring phase, for example, as a solution to handle model failure or performance degradation.

In some embodiments, the calibration procedure may be used together with other procedures too, for example, before/after cell selection, handover, beam training, and so on. Further, the calibration procedure may be performed periodically or may be triggered by signaling or based on pre-defined event and the calibration procedure may be initiated by UE or NW.

Further, it is clarified that the calibration procedure may be used together with or may be part of other ML procedures or in other LCM stages, for example, the calibration procedure may be used together with or may be part of dataset transfer/delivery, data collection for model training, model inference, and model monitoring, model transfer/delivery, model/functionality registration/identification, UE capability reporting or feature/feature group reporting. More details are discussed as below.

In some embodiments, the device 110 may perform a calibration procedure with the further device 120 based on the information used for enabling consistency.

In the present disclosure, any of the device 110 and the further device 120 may initiate the calibration procedure. Thus, in some embodiments, the device 110 may transmit 340-2 a request for initiating the calibration procedure to the further device 120.

Alternatibely, in some embodiments, the device 110 may receive 340-1 the request for initiating the calibration procedure from the further device 120.

In the present disclosure, the transmission of the request may be performed based on trigger event (s) or initiating the calibration procedure, as discussed below.

In some embodiments, the device 110 may transmit the request in response to detecting at least one trigger event for initiating the calibration procedure, wherein the at least one trigger event may comprise at least one of the following:

● a cell handover,

● a network device 110 switch,

● a transmit receive point (TRP) switch,

● a performance deterioration of the model or the functionality, or

● a performance deterioration of channel quality between the device 110 and the further device 120.

In the following, details about the request will be discussed.

In some embodiments, the request may comprise at least one of the following:

● a first indication for triggering the calibration procedure,

● a second indication used for requesting for a resource allocation or requesting for an update of configured resources,

● an identity of the model or the functionality,

● an identity of a functionality associated with the model or the functionality, or

● the information used for enabling the consistency.

In some embodiments, the request may be based on the information used for enabling the consistency (i.e., calibration information) , especially based on the reference pattern information. In some embodiments, the request may comprise the information used for enabling the consistency, especially comprise the reference pattern information.

More specifically, the request may comprise the calibration information (including the reference pattern information) used in the first stage and/or in the second stage, such as, the calibration information (including the reference pattern information) used in model training and/or in model inference/monitoring.

In some embodiments, the request may be comprised in any suitable message, including but not limited to, the following: a dedicated signaling, a handover request, a handover command, a beam switch request, or a beam switch command.

In some embodiments, after receiving the request, the device 110 may transmit a response of the request to the further device 120, where the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device 110, an indication indicating whether the calibration information (including the reference pattern information) used in the first stage and/or in the second stage by the device 110 is the same as the reference calibration information which has been exchanged with the further device 120, a difference between the calibration information (including the reference pattern information) used by the device 110 and the reference calibration information which has been exchanged with the further device 120.

In some embodiments, the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device 110, an indication indicating whether the calibration information (including the reference pattern information) used in the first stage and/or in the second stage by the device 110 is the same as the reference calibration information, a difference between the calibration information (including the reference pattern information) used by the device 110 and the reference calibration information which has been exchanged with the further device 120.

In some embodiments, the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device 110, an indication indicating whether the calibration information (including the reference pattern information) used in model training and/or in model inference/monitoring by the device 110 is the same as the reference calibration information, a difference between the calibration information (including the reference pattern information) used by the device 110 in model training and/or in model inference/monitoring and the reference calibration information which has been exchanged with the further device 120.

In some embodiments, the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device 110, an indication indicating whether the pattern used by the device 110 is the same as the reference pattern, a difference between the pattern used by the device 110 and the reference pattern, or adjust beamforming information used for communicating with the further device 120.

Additionally, in some embodiments, in case that the device 110 is a network device 110, after receiving the request, transmit a response of the request to the further device 120, the response indicating an updated resource configuration.

Additionally, in some embodiments, in case that the device 110 is a terminal device 110, after receiving the request, transmit a response of the request to the further device 120, the response comprising a second indication used for requesting for a resource allocation or requesting for an update of configured resources.

Some example embodiments are discussed with reference to FIG. 7A, which illustrates a signaling flow 700A of communication in accordance with some embodiments of the present disclosure. In the example ofFIG. 7A, the UE initiates the calibration procedure to request the changes needed to keep consistency.

In some embodiments, an event-based (or, condition-based) method may be used to enable UE initiated calibration procedure, for example, the event or the condition may be defined based on that:

● UE switches to a new TRP/gNB/cell; In this scenario, the new NW device may or may not be able to follow the reference pattern to do beamforming and to configure Set B and/or Set A;

● For UE side model monitoring, the monitoring result shows that the AI/ML model performance is worse than a threshold, which implies that NW configuration for model inference (e.g., Set B configuration) may not be aligned with the trained model.

In some embodiments, the UE may initiate a calibration procedure by sending a calibration request.

In some embodiments, the calibration request may be one simple field to trigger calibration for the currently applied AI/ML model/functionality. Alternatively, or in addition, the calibration request may include the model/functionality information.

In some embodiments, the calibration request may be based on the reference pattern, for example, the calibration request may also include reference pattern information.

In some embodiments, the calibration request may include information to request resources/configurations, or to request some update on the current configurations, for example, add/remove/update resources in Set B for measurement, via resource IDs, or via weights/angles as described in the reference pattern, update on the number of predicted beams to report, update on the number of historical measurements for future beam prediction.

In some embodiments, the calibration request may be signaled via a dedicated signaling, or may be carried in other signaling like handover request/command, beam switch request/command, and so on.

In some embodiments, the NW may send a response to the request by providing confirmation information, which can be a simple message to indicate whether consistency holds.

In some embodiments, the confirmation information may also be based on the reference pattern associated with the currently applied AI/ML model/functionality, e.g., which reference pattern is assumed at NW, if multiple patterns are associated, whether the applied pattern is the same as the reference pattern, or within the tolerance range compared to the reference pattern, or the difference from the reference pattern.

In some embodiments, the NW may adjust the configuration, based on UE request and/or the reference pattern, e.g., add/remove/update resources configurations for measurements, for model inference, monitoring, ground truth reporting, and/or model training and so on, add/remove/update report configurations, for model inference, monitoring, ground truth reporting, and/or model training and so on.

In some embodiments, the adjusting of Tx beamforming weights may be performed based on UE request and/or the reference pattern, e.g., adjusting the relationship between Set B beams and Set A beams, for model inference, monitoring, ground truth reporting, and/or model training and so on, or adjusting the Tx beamforming weights.

In some embodiments, UE capability reporting on whether UE can support to initiate the calibration may be performed.

In some embodiments, configuration information about the pre-defined or NW configured events or conditions that triggers calibration or pre-defined or NW configured periodic calibration may be provided to the UE.

In some embodiments, configuration information about the resources (such as, dedicated UL resources like dedicated PUSCH, dedicated PUCCH resource, and so on) used for sending the calibration request may be provided to the UE.

In the example of FIG. 7B, the NW initiates the calibration procedure to keep consistency.

In FIG. 7B, the NW may initiate a calibration procedure by sending a calibration request.

In some embodiments, the calibration request may be one simple field to trigger calibration for the currently applied AI/ML model/functionality, alternatively, it can include the model/functionality information, in addition, it may be based on the reference pattern, for example, it may also include reference pattern information.

In some embodiments, the UE may send a response to the request by providing confirmation information, which can be a simple message to indicate whether consistency holds.

In some embodiments, the confirmation information may also be based on the reference pattern associated with the currently applied AI/ML model/functionality, e.g.,

● which reference pattern is assumed at UE, if multiple patterns are associated;

● whether the applied pattern is the same as the reference pattern, or within the tolerance range compared to the reference pattern;

● the difference from the reference pattern.

In some embodiments, the UE may adjust Rx beamforming weights, based on NW request and/or the reference pattern, e.g., adjusting the Rx beamforming weights for measurement and report for Set B beams and Set A beams, for model inference, monitoring, ground truth reporting, and/or model training, and so on.

In some embodiments, the UE may send a response to further request resources/configurations, or to request some update on the current configurations, based on NW request and/or the reference pattern, e.g.,

● request to add/remove/update resources and report configurations for measurements, for model inference, monitoring, ground truth reporting, and/or model training, and so on,

● request to update on the number of predicted beams to report,

● request to update on the number of historical measurements for future beam prediction.

In some embodiments, UE capability reporting on whether UE can support NW initiated calibration may be performed.

In some embodiments, configuration information about the resources (such as, dedicated UL resources like dedicated PUSCH, dedicated PUCCH resource and so on) used for sending the confirmation or the further request may be provided to the UE.

In the present disclosure, the NW may initiate early calibration to prepare consistency for candidate cells, such that the calibration at UE side for measurement and report to support mobility is achieved. Details are discussed as below.

In some embodiments, further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in a candidate cell also may be exchanged or utilized. Such processes will be discussed.

In some embodiments, the calibration procedure for the candidate cell may be indicated by either the source cell or the candidate cell.

In some embodiments, the device 110 is a network device 110 providing a source cell, and the device 110 may determine further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in a candidate cell, and then transmit, to the further device 120, the request for initiating the calibration procedure for the candidate cell.

After receiving the request for initiating the calibration procedure for the candidate cell, the further device 120 may responds a response of the request for the candidate cell to the device 110. Then, the device 110 may transmit the response to the candidate cell.

In some embodiments, the device 110 is a terminal device 110, and the device 110 may receive, from a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell. Then, the device 110 may transmit, based on further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in the candidate cell and to the source cell or the candidate cell, a response of the request for the candidate cell.

Reference is now made to FIG. 8, which illustrates a signaling flow 800 of communication in accordance with some embodiments of the present disclosure.

In the example of FIG. 8, the NW may initiate a calibration procedure for other cells by sending a calibration request, where the other cells may refer to neighbor cells, cells with different cell ID, different PCI, and so on.

Further, the other cell may be configured as candidate cells, or may be the target cell for handover/switch.

In some embodiments, the source/serving cell and candidate/target cell may exchange the calibration information including the reference pattern.

As the further information is introduced, it needs to further distinguish the information for the source cell and the further information for the candidate cell.

In some embodiments, the request for initiating the calibration procedure for the candidate cell and the response of the request for the candidate cell indicate an identity associated with the further information. In some embodiment, the identity associated with the further information may be an index of the further information. Alternatively, the identity associated with the further information may be implied by any suitable identity which has correspondence with the further information. Such as, the information corresponds to the source cell, the further information corresponds to the candidate cell. In this event, the identity of the candidate cell may imply/indicate the further information for the candidate cell.

In addition, the mapping between Cell ID and reference pattern ID can be established and can be notified to UE, and candidate cell configuration can be used to provide reference pattern assumed at other cells.

In some embodiments, the reference pattern ID can be implicitly signaled via the signaling of cell ID too.

In some embodiments, the UE may send a response to the request by providing confirmation information to serving cell/other cells, which can be a simple message to indicate whether consistency holds.

In some embodiments, the confirmation information may also be based on the reference pattern of multiple cells associated with the currently applied AI/ML model/functionality, e.g.,

● which cell is assumed, e.g., cell ID,

● which reference pattern is assumed at UE, if multiple patterns are associated,

● whether the applied pattern is the same as the reference pattern, or within the tolerance range compared to the reference pattern, or

● the difference from the reference pattern.

In some embodiments, the source cell and candidate/target cell may exchange the UE conformation information.

In some embodiments, the UE may adjust Rx beamforming weights, based on NW request and/or the reference pattern, e.g.,

● adjusting the Rx beamforming weights for measurement and report for Set B beams and Set A beams for source cell/serving cell, for model inference, monitoring, ground truth reporting, and/or model training, and so on; or

● adjusting the Rx beamforming weights for measurement and report for Set B beams and Set A beams for other cells, for model inference, monitoring, ground truth reporting, and/or model training, and so on. The model may be used to predict beams for other cells.

In some embodiments, the UE may send a response for either serving cell or other cells to: request resources/configurations, or to request some update on the current configurations, based on NW request and/or the reference pattern, e.g.,

● request to add/remove/update resources and report configurations for measurements, for model inference, monitoring, ground truth reporting, and/or model training, and so on;

● request to update on the number of predicted beams to report; or

In some embodiments, UE capability reporting on whether UE may NW initiated calibration for cells different from the serving cell may be performed.

In the present disclosure, the UE may initiate early calibration to prepare consistency for candidate cells, such that the calibration at UE side for measurement and report to support mobility is achieved.

In some embodiments, the calibration procedure for the candidate cell may be indicated by the terminal device, as discussed below.

In some embodiments, the device 110 is a terminal device 110, and the device 110 may transmit, to a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell, and then the device 110 may receive, from the source cell or the candidate cell, a response of the request for the candidate cell.

In some embodiments, the device 110 is a network device 110 providing a source cell and the further device 120 is a terminal device 110, and t the device 110 may receive, from the further device 120, a request for initiating the calibration procedure for a candidate cell, and then may transmit, to the candidate cell, the request for initiating the calibration procedure for the candidate cell.

As the candidate cell, the candidate cell may respond the request for the candidate cell with a response of the request for the candidate cell. As a result, after receiving the response of the request for the candidate cell, the device 110 may transmit the response to the further device 120.

Reference is now made to FIG. 9, which illustrates a signaling flow 900 of communication in accordance with some embodiments of the present disclosure.

In the example of FIG. 9, the UE may initiate a calibration procedure for other cells by sending a calibration request, where the other cells may refer to neighbor cells, cells with different cell ID, different PCI, and so on.

In some embodiments, the mapping between Cell ID and reference pattern ID can be established, and candidate cell configuration can be used to provide reference pattern assumed at other cells.

In some embodiments, the calibration request may include information to request resources/configurations, or to request some update on the current configurations, for serving cell and/or for other cells, e.g., cell ID may be included in the request.

In some embodiments, the NW may send a response to the request by providing confirmation information for serving cell/other cells, which can be a simple message to indicate whether consistency holds.

In some embodiments, the confirmation information may also be based on the reference pattern of multiple cells associated with the currently applied AI/ML model/functionality, e.g., which cell is assumed, e.g., cell ID is included in the response/confirmation information.

In some embodiments, the source cell and candidate/target cell may adjust the configuration for serving cell/other cells, based on UE request and/or the reference pattern.

In some embodiments, the source cell and candidate/target cell may adjust the Tx beamforming weights for serving cell/other cells, based on UE request and/or the reference pattern.

In some embodiments, the source cell and candidate/target cell may exchange the information on adjusted weights or adjusted configuration.

In some embodiments, UE capability reporting on whether UE can support to initiate the calibration for other cells may be provided to the NW.

As discussed above, in the example of FIG. 8 and FIG, 9, capability-related information and configuration information may be exchanged among the device 110 and the further device 120. For a better understanding, example capability-related information and configuration information are summarized as below.

Additionally, in some embodiments, capability-related information may be exchanged among the device 110 and the further device 120. In this way, the corresponding device may well understand the capability-related information of the other device.

In some embodiments, in case that the device 110 is a terminal device 110 and the further device 120 is a network device 110 providing a source cell, the device 110 may transmit to the further device 120 capability-related information indicating at least one of the following:

● whether the device 110 supports to initiate the calibration procedure for a source cell or a candidate cell, or

● whether the device 110 supports the further device 120 initiates the calibration procedure.

In some embodiments, the device 110 may receive related information which may be used during the calibration procedure. For example, the device 110 may receive, from the further device 120, configuration information indicating at least one of the following:

● at least one trigger event for initiating the calibration procedure,

● a periodicity used for initiating the calibration procedure,

● resources used by the device 110 for transmitting the request, or

● resources used by the device 110 for transmitting a response of the request.

According to the above processes, the consistency of model-related or functionality-related information among different ML stages is ensured.

Example methods

FIG. 9 illustrates a flowchart of a communication method 900 implemented at a device in accordance with some embodiments of the present disclosure. For the purpose of discussion, the method 900 will be described from the perspective of the device 110 in FIG. 1.

At block 910, the device determines information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality.

At block 920, the device performs, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage.

In some example embodiments, the processor is further configured to cause the device to perform one of the following: determining the information used for enabling the consistency based on a first message comprising the information, the first message being transmitted by the further device; or after determining the information used for enabling the consistency, transmitting a second message comprising the information to the further device, wherein the information indicates at least one of the following: first information indicating at least one used or supported reference pattern, second information indicating at least one used or supported reference value, third information indicating at least one used or supported assumption, fourth information indicating whether the consistency is needed, fifth information indicating a tolerance range associated with the information.

In some example embodiments, the first information comprises at least one of the following: antenna modelling information associated with at least one of the device or the further device, beamforming information associated with at least one of the device or the further device, orientation information between the device and the further device, or a relationship between the beamforming information and at least one resource, each of the at least one resource being associated with a beam.

In some example embodiments, the antenna modelling information comprises at least one of the following: the number of antenna elements in a first spatial dimension, the number of antenna elements in a second spatial dimension, the number of panels in a first spatial dimension, the number of panels in a second spatial dimension, an antenna element spacing in a first spatial dimension, an antenna element spacing in a second spatial dimension, a panel spacing in a first spatial dimension, a panel spacing in a second spatial dimension, the number of polarizations, an antenna array type, an antenna numbering rule, a polarization slant angle a polarization type, an antenna element radiation pattern in a first spatial dimension, an antenna element radiation pattern in a second spatial dimension, a combining method for a multi-dimension antenna element pattern, the maximum directional gain of an antenna element, at least one complex weight for an antenna element in an elevation or, antenna height.

In some example embodiments, the beamforming information comprises at least one of the following: a beamforming type, the number of transceiver units (TXRUs) the number of beams, or at least one beamforming weights.

In some example embodiments, the beamforming type is one of the following: an analog beamforming, a digital beamforming, a hybrid beamforming, the number of TXRUs is a total number of TXRUs or is associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle, the number of beams is a total number of beams or is associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle, the at least one beamforming weights comprises: at least one discrete fourier transform (DFT) weights, a set of weights associated with a specific bean set, a correspondence between beams and a weights matrix.

In some example embodiments, the beamforming information is a set of weights, and the relationship between the beamforming information and the at least one resource is a mapping between the set of weights and the at least one resource, and wherein the relationship between the beamforming information and the at least one resource is represented by at least one of the following: an order used for mapping the at least one resource to a subset in the set of weights, the order comprising an ascending or descending order of at least one of the following: first a first spatial dimension then a second spatial dimension, first the second spatial dimension then the first spatial dimension, first an azimuth dimension then an elevation dimension, first the elevation dimension then the azimuth dimension, or an order of a plurality of beam groups, the at least one beam groups being divided into the plurality of beam groups; or information indicating the subset in the set of weights, the information comprising at least one of the following: an indication indicating a start position of the subset, the number of rows in the subset, the number of columns in the subset, a first step size in the rows if elements in the subset are not adjacent weights, or a second step size in the columns if the elements in the subset are not adjacent weights.

In some example embodiments, each element in the set of weights corresponds to: a first value in the first spatial dimension and a second value in the second spatial dimension, or a specific azimuth value and a specific elevation.

In some example embodiments, the first message is received or the second message is transmitted during at least one of the following: a dataset delivery procedure for the model or the functionality, a model delivery procedure for the model or the functionality, a data collection procedure for the model or the functionality, a model inference procedure for the model or the functionality, a model monitoring procedure for the model or the functionality, a model training procedure for the model or the functionality, a registration or identification procedure for the model or the functionality, or the calibration procedure.

In some example embodiments, the information used for enabling the consistency is comprised in at least one of the following: a model description, a feature or feature group, capability-related information, supported condition information associated with the model or the functionality, or additional condition information associated with the model or the functionality.

In some example embodiments, the device is a network device and the further device is a terminal device, the device may transmit, to the further device and based on the information used for enabling the consistency, at least one of the following: a measurement configuration to be used by the further device for data collection, or information about updated beamforming information; and wherein the device is a terminal device and the further device is a network device, the processor is further configured to cause the device to: perform, based on the information used for enabling consistency, measurements on at least one beam; and transmit measurement results of at least part of the at least one beam to the further device.

In some example embodiments, the device may perform one of the following: transmit, to the further device, a request for initiating the calibration procedure, or receive, from the further device, the request for initiating the calibration procedure.

In some example embodiments, the device may transmit the request in response to detecting at least one trigger event for initiating the calibration procedure, the at least one trigger event comprising at least one of the following: a cell handover, a network device switch, a transmit receive point (TRP) switch, a performance deterioration of the model or the functionality, or a performance deterioration of channel quality between the device and the further device.

In some example embodiments, the request comprises at least one of the following: a first indication for triggering the calibration procedure, a second indication used for requesting for a resource allocation or requesting for an update of configured resources, an identity of the model or the functionality, an identity of a functionality associated with the model or the functionality, or the information used for enabling the consistency.

In some example embodiments, the request is one of the following: a dedicated signaling, a handover request, a handover command, a beam switch request, or a beam switch command.

In some example embodiments, after receiving the request, the device may transmit a response of the request to the further device, the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device, an indication indicating whether the pattern used by the device is the same as the reference pattern, a difference between the pattern used by the device and the reference pattern, or adjust beamforming information used for communicating with the further device.

In some example embodiments, in case that the device is a network device, after receiving the request, the device may transmit a response of the request to the further device, the response indicating an updated resource configuration, or in case that the device is a terminal device, after receiving the request, transmit a response of the request to the further device, the response comprising a second indication used for requesting for a resource allocation or requesting for an update of configured resources.

In some example embodiments, the device is a network device providing a source cell, the device may determine further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in a candidate cell; transmit, to the further device, the request for initiating the calibration procedure for the candidate cell; receive, from the further device, a response of the request for the candidate cell; and transmit the response to the candidate cell.

In some example embodiments, the device is a terminal device, and the processor is further configured to cause the device to: receive, from a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell; transmit, based on further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in the candidate cell and to the source cell or the candidate cell, a response of the request for the candidate cell.

In some example embodiments, the device is a terminal device, and the processor is further configured to cause the device to: transmit, to a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell; and receive, from the source cell or the candidate cell, a response of the request for the candidate cell.

In some example embodiments, the device is a network device providing a source cell and the further device is a terminal device, the device may receive, from the further device, a request for initiating the calibration procedure for a candidate cell; transmit, to the candidate cell, the request for initiating the calibration procedure for the candidate cell; receive, from the candidate cell, a response of the request for the candidate cell; and transmit the response to the further device.

In some example embodiments, the request for initiating the calibration procedure for the candidate cell and the response of the request for the candidate cell indicate an identity associated with the further information.

In some example embodiments, the device is a terminal device and the further device is a network device providing a source cell, and the device may transmit to the further device capability-related information indicating at least one of the following: whether the device supports to initiate the calibration procedure for a source cell or a candidate cell, or whether the device supports the further device initiates the calibration procedure.

In some example embodiments, the device may receive, from the further device, configuration information indicating at least one of the following: at least one trigger event for initiating the calibration procedure, a periodicity used for initiating the calibration procedure, resources used by the device for transmitting the request, or resources used by the device for transmitting a response of the request.

In some example embodiments, the device is a terminal device or a network device, and the further device is a terminal device or a network device.

FIG. 10 is a simplified block diagram of a device 1000 that is suitable for implementing embodiments of the present disclosure. The device 1000 can be considered as a further example implementation of any of the devices as shown in FIG. 1. Accordingly, the device 1000 can be implemented at or as at least a part of the device 110.

As shown, the device 1000 includes a processor 1010, a memory 1020 coupled to the processor 1010, a suitable transceiver 1040 coupled to the processor 1010, and a communication interface coupled to the transceiver 1040. The memory 1020 stores at least a part of a program 1030. The transceiver 1040 may be for bidirectional communications or a unidirectional communication based on requirements. The transceiver 1040 may include at least one of a transmitter 1042 and a receiver 1044. The transmitter 1042 and the receiver 1044 may be functional modules or physical entities. The transceiver 1040 has at least one antenna to facilitate communication, though in practice an Access Node mentioned in this application may have several ones. The communication interface may represent any interface that is necessary for communication with other network elements, such as X2/Xn interface for bidirectional communications between eNBs/gNBs, S1/NG interface for communication between a Mobility Management Entity (MME) /Access and Mobility Management Function (AMF) /SGW/UPF and the eNB/gNB, Un interface for communication between the eNB/gNB and a relay node (RN) , or Uu interface for communication between the eNB/gNB and a terminal device.

The program 1030 is assumed to include program instructions that, when executed by the associated processor 1010, enable the device 1000 to operate in accordance with the embodiments of the present disclosure, as discussed herein with reference to FIGS. 1 to 9. The embodiments herein may be implemented by computer software executable by the processor 1010 of the device 1000, or by hardware, or by a combination of software and hardware. The processor 1010 may be configured to implement various embodiments of the present disclosure. Furthermore, a combination of the processor 1010 and memory 1020 may form processing means 1050 adapted to implement various embodiments of the present disclosure.

The memory 1020 may be of any type suitable to the local technical network and may be implemented using any suitable data storage technology, such as a non-transitory computer readable storage medium, semiconductor based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory, as non-limiting examples. While only one memory 1020 is shown in the device 1000, there may be several physically distinct memory modules in the device 1000. The processor 1010 may be of any type suitable to the local technical network, and may include one or more of general purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) and processors based on multicore processor architecture, as non-limiting examples. The device 1000 may have multiple processors, such as an application specific integrated circuit chip that is slaved in time to a clock which synchronizes the main processor.

According to embodiments of the present disclosure, a device comprising a circuitry is provided. The circuitry is configured to: determine information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and perform, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage. According to embodiments of the present disclosure, the circuitry may be configured to perform any method implemented by the device as discussed above.

The term “circuitry” used herein may refer to hardware circuits and/or combinations of hardware circuits and software. For example, the circuitry may be a combination of analog and/or digital hardware circuits with software/firmware. As a further example, the circuitry may be any portions of hardware processors with software including digital signal processor (s) , software, and memory (ies) that work together to cause an apparatus, such as a terminal device or a network device, to perform various functions. In a still further example, the circuitry may be hardware circuits and or processors, such as a microprocessor or a portion of a microprocessor, that requires software/firmware for operation, but the software may not be present when it is not needed for operation. As used herein, the term circuitry also covers an implementation of merely a hardware circuit or processor (s) or a portion of a hardware circuit or processor (s) and its (or their) accompanying software and/or firmware.

According to embodiments of the present disclosure, an apparatus is provided. The apparatus comprises means for determining information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and means for performing, based on the information, at least one of the following: means for a calibration procedure with a further device, means for the first ML stage, or means for the second ML stage. In some embodiments, the first apparatus may comprise means for performing the respective operations of the method 900. In some example embodiments, the first apparatus may further comprise means for performing other operations in some example embodiments of the method 900. The means may be implemented in any suitable form. For example, the means may be implemented in a circuitry or software module.

In summary, embodiments of the present disclosure provide the following aspects.

In an aspect, it is proposed a device comprising: a processor configured to cause the device to: determine information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and perform, based on the information, at least one of the following: a calibration procedure with a further device, the first ML stage, or the second ML stage.

In some embodiments, the processor is further configured to cause the device to perform one of the following: determining the information used for enabling the consistency based on a first message comprising the information, the first message being transmitted by the further device; or after determining the information used for enabling the consistency, transmitting a second message comprising the information to the further device, wherein the information indicates at least one of the following: first information indicating at least one used or supported reference pattern, second information indicating at least one used or supported reference value, third information indicating at least one used or supported assumption, fourth information indicating whether the consistency is needed, fifth information indicating a tolerance range associated with the information.

In some embodiments, the first information comprises at least one of the following: antenna modelling information associated with at least one of the device or the further device, beamforming information associated with at least one of the device or the further device, orientation information between the device and the further device, or a relationship between the beamforming information and at least one resource, each of the at least one resource being associated with a beam.

In some embodiments, the antenna modelling information comprises at least one of the following: the number of antenna elements in a first spatial dimension, the number of antenna elements in a second spatial dimension, the number of panels in a first spatial dimension, the number of panels in a second spatial dimension, an antenna element spacing in a first spatial dimension, an antenna element spacing in a second spatial dimension, a panel spacing in a first spatial dimension, a panel spacing in a second spatial dimension, the number of polarizations, an antenna array type, an antenna numbering rule, a polarization slant angle a polarization type, an antenna element radiation pattern in a first spatial dimension, an antenna element radiation pattern in a second spatial dimension, a combining method for a multi-dimension antenna element pattern, the maximum directional gain of an antenna element, at least one complex weight for an antenna element in an elevation or, antenna height.

In some embodiments, the beamforming information comprises at least one of the following: a beamforming type, the number of transceiver units (TXRUs) the number of beams, or at least one beamforming weights.

In some embodiments, the beamforming type is one of the following: an analog beamforming, a digital beamforming, a hybrid beamforming, the number of TXRUs is a total number of TXRUs or is associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle, the number of beams is a total number of beams or is associated with at least one of the following: a beam set with a specific type, a specific beam type, a specific spatial dimension, or a specific spatial angle, the at least one beamforming weights comprises: at least one discrete fourier transform (DFT) weights, a set of weights associated with a specific bean set, a correspondence between beams and a weights matrix.

In some embodiments, the beamforming information is a set of weights, and the relationship between the beamforming information and the at least one resource is a mapping between the set of weights and the at least one resource, and wherein the relationship between the beamforming information and the at least one resource is represented by at least one of the following: an order used for mapping the at least one resource to a subset in the set of weights, the order comprising an ascending or descending order of at least one of the following: first a first spatial dimension then a second spatial dimension, first the second spatial dimension then the first spatial dimension, first an azimuth dimension then an elevation dimension, first the elevation dimension then the azimuth dimension, or an order of a plurality of beam groups, the at least one beam groups being divided into the plurality of beam groups; or information indicating the subset in the set of weights, the information comprising at least one of the following: an indication indicating a start position of the subset, the number of rows in the subset, the number of columns in the subset, a first step size in the rows if elements in the subset are not adjacent weights, or a second step size in the columns if the elements in the subset are not adjacent weights.

In some embodiments, each element in the set of weights corresponds to: a first value in the first spatial dimension and a second value in the second spatial dimension, or a specific azimuth value and a specific elevation.

In some embodiments, the first message is received or the second message is transmitted during at least one of the following: a dataset delivery procedure for the model or the functionality, a model delivery procedure for the model or the functionality, a data collection procedure for the model or the functionality, a model inference procedure for the model or the functionality, a model monitoring procedure for the model or the functionality, a model training procedure for the model or the functionality, or a registration or identification procedure for the model or the functionality, or the calibration procedure.

In some embodiments, the information used for enabling the consistency is comprised in at least one of the following: a model description, a feature or feature group, capability-related information, supported condition information associated with the model or the functionality, or additional condition information associated with the model or the functionality.

In some embodiments, the device is a network device and the further device is a terminal device, the processor is further configured to cause the device to: transmit, to the further device and based on the information used for enabling the consistency, at least one of the following: a measurement configuration to be used by the further device for data collection, or information about updated beamforming information; and wherein the device is a terminal device and the further device is a network device, the processor is further configured to cause the device to: perform, based on the information used for enabling consistency, measurements on at least one beam; and transmit measurement results of at least part of the at least one beam to the further device.

In some embodiments, the processor is further configured to cause the device to perform one of the following: transmit, to the further device, a request for initiating the calibration procedure, or receive, from the further device, the request for initiating the calibration procedure.

In some embodiments, the processor is further configured to cause the device to: transmit the request in response to detecting at least one trigger event for initiating the calibration procedure, the at least one trigger event comprising at least one of the following: a cell handover, a network device switch, a transmit receive point (TRP) switch, a performance deterioration of the model or the functionality, or a performance deterioration of channel quality between the device and the further device.

In some embodiments, the request comprises at least one of the following: a first indication for triggering the calibration procedure, a second indication used for requesting for a resource allocation or requesting for an update of configured resources, an identity of the model or the functionality, an identity of a functionality associated with the model or the functionality, or the information used for enabling the consistency.

In some embodiments, the request is one of the following: a dedicated signaling, a handover request, a handover command, a beam switch request, or a beam switch command.

In some embodiments, the processor is further configured to cause the device to: after receiving the request, transmit a response of the request to the further device, the response indicating confirmation information comprising at least one of the following: an indication indicating whether the consistency is hold, an identity of a reference pattern used by the device, an indication indicating whether the pattern used by the device is the same as the reference pattern, a difference between the pattern used by the device and the reference pattern, or adjust beamforming information used for communicating with the further device.

In some embodiments, the processor is further configured to cause the device to: in case that the device is a network device, after receiving the request, transmit a response of the request to the further device, the response indicating an updated resource configuration, or in case that the device is a terminal device, after receiving the request, transmit a response of the request to the further device, the response comprising a second indication used for requesting for a resource allocation or requesting for an update of configured resources.

In some embodiments, the device is a network device providing a source cell, and the processor is further configured to cause the device to: determine further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in a candidate cell; transmit, to the further device, the request for initiating the calibration procedure for the candidate cell; receive, from the further device, a response of the request for the candidate cell; and transmit the response to the candidate cell.

In some embodiments, the device is a terminal device, and the processor is further configured to cause the device to: receive, from a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell; transmit, based on further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in the candidate cell and to the source cell or the candidate cell, a response of the request for the candidate cell.

In some embodiments, the device is a terminal device, and the processor is further configured to cause the device to: transmit, to a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell; and receive, from the source cell or the candidate cell, a response of the request for the candidate cell.

In some embodiments, the device is a network device providing a source cell and the further device is a terminal device, and the processor is further configured to cause the device to: receive, from the further device, a request for initiating the calibration procedure for a candidate cell; transmit, to the candidate cell, the request for initiating the calibration procedure for the candidate cell; receive, from the candidate cell, a response of the request for the candidate cell; and transmit the response to the further device.

In some embodiments, the request for initiating the calibration procedure for the candidate cell and the response of the request for the candidate cell indicate an identity associated with the further information.

In some embodiments, the device is a terminal device and the further device is a network device providing a source cell, and wherein the processor is further configured to cause the device to: transmit to the further device capability-related information indicating at least one of the following: whether the device supports to initiate the calibration procedure for a source cell or a candidate cell, or whether the device supports the further device initiates the calibration procedure.

In some embodiments, the processor is further configured to cause the device to: receive, from the further device, configuration information indicating at least one of the following: at least one trigger event for initiating the calibration procedure, a periodicity used for initiating the calibration procedure, resources used by the device for transmitting the request, or resources used by the device for transmitting a response of the request.

In some embodiments, the device is a terminal device or a network device, and the further device is a terminal device or a network device.

In an aspect, a device comprises: at least one processor; and at least one memory coupled to the at least one processor and storing instructions thereon, the instructions, when executed by the at least one processor, causing the device to perform the method implemented by the device discussed above.

In an aspect, a computer readable medium having instructions stored thereon, the instructions, when executed on at least one processor, causing the at least one processor to perform the method implemented by the device discussed above.

In an aspect, a computer program comprising instructions, the instructions, when executed on at least one processor, causing the at least one processor to perform the method implemented by the device discussed above.

Generally, various embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, logic or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software which may be executed by a controller, microprocessor or other computing device. While various aspects of embodiments of the present disclosure are illustrated and described as block diagrams, flowcharts, or using some other pictorial representation, it will be appreciated that the blocks, apparatus, systems, techniques or methods described herein may be implemented in, as non-limiting examples, hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof.

The present disclosure also provides at least one computer program product tangibly stored on a non-transitory computer readable storage medium. The computer program product includes computer-executable instructions, such as those included in program modules, being executed in a device on a target real or virtual processor, to carry out the process or method as described above with reference to FIGS. 1 to 10. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, or the like that perform particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or split between program modules as desired in various embodiments. Machine-executable instructions for program modules may be executed within a local or distributed device. In a distributed device, program modules may be located in both local and remote storage media.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowcharts and/or block diagrams to be implemented. The program code may execute entirely on a machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

The above program code may be embodied on a machine readable medium, which may be any tangible medium that may contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine readable medium may be a machine readable signal medium or a machine readable storage medium. A machine readable medium may include but not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine readable storage medium would include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM) , a read-only memory (ROM) , an erasable programmable read-only memory (EPROM or Flash memory) , an optical fiber, a portable compact disc read-only memory (CD-ROM) , an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are contained in the above discussions, these should not be construed as limitations on the scope of the present disclosure, but rather as descriptions of features that may be specific to particular embodiments. Certain features that are described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment may also be implemented in multiple embodiments separately or in any suitable sub-combination.

Although the present disclosure has been described in language specific to structural features and/or methodological acts, it is to be understood that the present disclosure defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

A device comprising:

a processor configured to cause the device to:

determine information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and

perform, based on the information, at least one of the following:

a calibration procedure with a further device,

the first ML stage, or

the second ML stage.
The device of claim 1, wherein the processor is further configured to cause the device to perform one of the following:

determining the information used for enabling the consistency based on a first message comprising the information, the first message being transmitted by the further device; or

after determining the information used for enabling the consistency, transmitting a second message comprising the information to the further device,

wherein the information indicates at least one of the following:

first information indicating at least one used or supported reference pattern,

second information indicating at least one used or supported reference value,

third information indicating at least one used or supported assumption,

fourth information indicating whether the consistency is needed, or

fifth information indicating a tolerance range associated with the information.
The device of claim 2, wherein the first information comprises at least one of the following:

antenna modelling information associated with at least one of the device or the further device,

beamforming information associated with at least one of the device or the further device,

orientation information between the device and the further device, or

a relationship between the beamforming information and at least one resource, each of the at least one resource being associated with a beam.
The device of claim 3, wherein the beamforming information is a set of weights, and the relationship between the beamforming information and the at least one resource is a mapping between the set of weights and the at least one resource,

and wherein the relationship between the beamforming information and the at least one resource is represented by at least one of the following:

an order used for mapping the at least one resource to a subset in the set of weights, the order comprising an ascending or descending order of at least one of the following:

first a first spatial dimension then a second spatial dimension,

first the second spatial dimension then the first spatial dimension,

first an azimuth dimension then an elevation dimension,

first the elevation dimension then the azimuth dimension, or

an order of a plurality of beam groups, the at least one beam groups being divided into the plurality of beam groups; or

information indicating the subset in the set of weights, the information comprising at least one of the following:

an indication indicating a start position of the subset,

a number of rows in the subset,

a number of columns in the subset,

a first step size in the rows if elements in the subset are not adjacent weights, or

a second step size in the columns if the elements in the subset are not adjacent weights.
The device of claim 4, wherein each element in the set of weights corresponds to:

a first value in the first spatial dimension and a second value in the second spatial dimension, or

a specific azimuth value and a specific elevation.
The device of claim 2, wherein the first message is received or the second message is transmitted during at least one of the following:

a dataset delivery procedure for the model or the functionality,

a model delivery procedure for the model or the functionality,

a data collection procedure for the model or the functionality,

a model inference procedure for the model or the functionality,

a model monitoring procedure for the model or the functionality,

a model training procedure for the model or the functionality,

a registration or identification procedure for the model or the functionality, or

the calibration procedure.
The device of claim 2, wherein the information used for enabling the consistency is comprised in at least one of the following:

a model description,

a feature or feature group,

capability-related information,

supported condition information associated with the model or the functionality, or

additional condition information associated with the model or the functionality.
The device of claim 1, wherein the device is a network device and the further device is a terminal device, and the processor is further configured to cause the device to:

transmit, to the further device and based on the information used for enabling the consistency, at least one of the following:

a measurement configuration to be used by the further device for data collection, or

information about updated beamforming information;

and wherein the device is a terminal device and the further device is a network device, and the processor is further configured to cause the device to:

perform, based on the information used for enabling consistency, measurements on at least one beam; and

transmit measurement results of at least part of the at least one beam to the further device.
The device of claim 1, wherein the processor is further configured to cause the device to perform one of the following:

transmitting, to the further device, a request for initiating the calibration procedure, or

receiving, from the further device, the request for initiating the calibration procedure.
The device of claim 9, wherein the request comprises at least one of the following:

a first indication for triggering the calibration procedure,

a second indication used for requesting for a resource allocation or requesting for an update of configured resources,

an identity of the model or the functionality,

an identity of a functionality associated with the model or the functionality, or

the information used for enabling the consistency.
The device of claim 9, wherein the processor is further configured to cause the device to:

after receiving the request, transmit a response of the request to the further device, the response indicating confirmation information comprising at least one of the following:

an indication indicating whether the consistency is hold,

an identity of a reference pattern used by the device,

an indication indicating whether the pattern used by the device is the same as the reference pattern,

a difference between the pattern used by the device and the reference pattern, or adjust beamforming information used for communicating with the further device.
The device of claim 9 wherein the processor is further configured to cause the device to:

in case that the device is a network device, after receiving the request, transmit a response of the request to the further device, the response indicating an updated resource configuration, or

in case that the device is a terminal device, after receiving the request, transmit a response of the request to the further device, the response comprising a second indication used for requesting for a resource allocation or requesting for an update of configured resources.
The device of claim 9, wherein the device is a network device providing a source cell, and the processor is further configured to cause the device to:

determine further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in a candidate cell;

transmit, to the further device, the request for initiating the calibration procedure for the candidate cell;

receive, from the further device, a response of the request for the candidate cell; and

transmit the response to the candidate cell.
The device of claim 9, wherein the device is a terminal device, and the processor is further configured to cause the device to:

receive, from a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell;

transmit, based on further information used for enabling consistency of model-related or functionality-related information between different ML stages of a further model or a further functionality used in the candidate cell and to the source cell or the candidate cell, a response of the request for the candidate cell.
The device of claim 9, wherein the device is a terminal device, and the processor is further configured to cause the device to:

transmit, to a source cell or a candidate cell, the request for initiating the calibration procedure for the candidate cell; and

receive, from the source cell or the candidate cell, a response of the request for the candidate cell.
The device of claim 9, wherein the device is a network device providing a source cell and the further device is a terminal device, and the processor is further configured to cause the device to:

receive, from the further device, a request for initiating the calibration procedure for a candidate cell;

transmit, to the candidate cell, the request for initiating the calibration procedure for the candidate cell;

receive, from the candidate cell, a response of the request for the candidate cell; and

transmit the response to the further device.
The device of any of claims 13-16, wherein the request for initiating the calibration procedure for the candidate cell and the response of the request for the candidate cell indicate an identity associated with the further information.
The device of claim 9, wherein the device is a terminal device and the further device is a network device providing a source cell,

and wherein the processor is further configured to cause the device to:

transmit to the further device capability-related information indicating at least one of the following:

whether the device supports to initiate the calibration procedure for a source cell or a candidate cell, or

whether the device supports the further device initiates the calibration procedure.
The device of claim 9, wherein the processor is further configured to cause the device to:

receive, from the further device, configuration information indicating at least one of the following:

at least one trigger event for initiating the calibration procedure,

a periodicity used for initiating the calibration procedure,

resources used by the device for transmitting the request, or

resources used by the device for transmitting a response of the request.
A communication method implemented at a device, comprising:

determining information used for enabling consistency of model-related or functionality-related information between a first machine-learning (ML) stage of a model or a functionality and a second ML stage of the model or the functionality; and

performing, based on the information, at least one of the following:

a calibration procedure with a further device,

the first ML stage, or

the second ML stage.