Devices and methods of communication
By determining a set of QCL assumptions for a subset of resources, the terminal device optimizes AI/ML based beam prediction, addressing inefficiencies in QCL assumption consistency and reducing complexity.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- NEC CORP
- Filing Date
- 2025-01-26
- Publication Date
- 2026-07-30
AI Technical Summary
The implementation of quasi co-location (QCL) assumption consistency for AI/ML based beam prediction is unclear, particularly in determining QCL configurations for CSI-RS resources used in model inference and performance monitoring, leading to inefficiencies in beam measurement and increased complexity.
A terminal device determines a set of QCL assumptions for a first set of resources associated with performance monitoring, measures a subset of resources based on these assumptions, and evaluates the performance monitoring and model inference results, reducing the need to sweep all Rx beams and minimizing overhead.
This approach reduces complexity and overhead by allowing the terminal device to focus on a subset of resources, enhancing the efficiency of AI/ML based beam prediction.
Smart Images

Figure CN2025075309_30072026_PF_FP_ABST
Abstract
Description
DEVICES AND METHODS OF COMMUNICATIONTECHNICAL FIELD
[0001] Embodiments of the present disclosure generally relate to the field of telecommunication, and in particular, to methods, devices and computer storage media of communication for artificial intelligence (AI) / machine learning (ML) based beam prediction.BACKGROUND
[0002] Generally, quasi co-location (QCL) assumption is useful for a terminal device to determine large scale properties. TypeD QCL assumption is useful for a terminal device to determine a receiving (Rx) beam. For AI / ML based beam prediction, QCL assumption consistency needs to be maintained for a model training, a model inference and a performance monitoring. However, implementations of the QCL assumption consistency are still unclear.SUMMARY
[0003] In general, embodiments of the present disclosure provide methods, devices and computer storage media of communication for AI / ML based beam prediction.
[0004] In a first aspect, there is provided a terminal device. The terminal device comprises a processor configured to cause the terminal device to: receive, from a network device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; determine a set of QCL assumptions for a first set of resources, the first set of resources being associated with the performance monitoring; measure, based on the set of QCL assumptions, a subset of resources in the first set of resources; and determine a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.
[0005] In a second aspect, there is provided a network device. The network device comprises a processor configured to cause the network device to: transmit, to a terminal device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; and receive, from the terminal device, a report for a result of the performance monitoring, the result being based on a measurement for a subset of resources in a first set of resources associated with the performance monitoring, a result of the model inference, and a set of QCL assumptions for the first set of resources.
[0006] In a third aspect, there is provided a method of communication at a terminal device. The method comprises: receiving, from a network device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; determining a set of QCL assumptions for a first set of resources, the first set of resources being associated with the performance monitoring; measuring, based on the set of QCL assumptions, a subset of resources in the first set of resources; and determining a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.
[0007] In a fourth aspect, there is provided a method of communication at a network device. The method comprises: transmitting, to a terminal device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; and receiving, from the terminal device, a report for a result of the performance monitoring, the result being based on a measurement for a subset of resources in a first set of resources associated with the performance monitoring, a result of the model inference, and a set of QCL assumptions for the first set of resources.
[0008] In a fifth aspect, there is provided a computer readable medium having instructions stored thereon. The instructions, when executed on at least one processor, cause the at least one processor to perform the method according to the third or fourth aspect of the present disclosure.
[0009] Other features of the present disclosure will become easily comprehensible through the following description.BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Through the more detailed description of some 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:
[0011] FIG. 1A illustrates an example communication network in which some embodiments of the present disclosure can be implemented;
[0012] FIG. 1B illustrates a schematic diagram illustrating an example model inference for AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented;
[0013] FIG. 2A illustrates a schematic diagram illustrating an example scenario of AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented;
[0014] FIG. 2B illustrates a schematic diagram illustrating another example scenario of AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented;
[0015] FIG. 3 illustrates a signaling chart illustrating an example process of communication according to some embodiments of the present disclosure;
[0016] FIG. 4A illustrates a schematic diagram illustrating an example QCL assumption determination according to some embodiments of the present disclosure;
[0017] FIG. 4B illustrates a schematic diagram illustrating an example QCL assumption determination for a time-domain prediction according to some embodiments of the present disclosure;
[0018] FIG. 4C illustrates a schematic diagram illustrating an example determination of a subset of resources according to some embodiments of the present disclosure;
[0019] FIG. 4D illustrates a schematic diagram illustrating an example determination of a subset of resources for a time-domain prediction according to some embodiments of the present disclosure;
[0020] FIG. 4E illustrates a schematic diagram illustrating an example resource configuration according to some embodiments of the present disclosure;
[0021] FIG. 5 illustrates a flowchart illustrating an example method of communication implemented at a terminal device in accordance with some embodiments of the present disclosure;
[0022] FIG. 6 illustrates a flowchart illustrating an example method of communication implemented at a network device in accordance with some embodiments of the present disclosure; and
[0023] FIG. 7 is a simplified block diagram of a device that is suitable for implementing embodiments of the present disclosure.
[0024] Throughout the drawings, the same or similar reference numerals represent the same or similar element.DETAILED DESCRIPTION
[0025] Principle of the present disclosure will now be described with reference to some 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 limitations as to the scope of the disclosure. The disclosure described herein can be implemented in various manners other than the ones described below.
[0026] 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.
[0027] 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, device on vehicle for V2X communication where X means pedestrian, vehicle, or infrastructure / network, devices for integrated access and backhaul (IAB) , small data transmission (SDT) , mobility, multicast and broadcast services (MBS) , positioning, dynamic / flexible duplex in commercial networks, reduced capability (RedCap) , 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 has ‘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.
[0028] 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) , network-controlled repeaters, and the like.
[0029] The terminal device or the network device may have AI / ML 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.
[0030] The terminal or the network device may work on several frequency ranges, e.g. FR1 (410 MHz to 7125 MHz) , FR2 (24.25GHz to 71GHz) , FR3 (7125 MHz to 24.25GHz) , frequency band larger than 100GHz as well as Tera Hertz (THz) . It can further work on licensed / unlicensed / shared spectrum. The terminal device may have more than one connections with the network devices under MR-DC application scenario. The terminal device or the network device can work on full duplex, flexible duplex and cross division duplex modes.
[0031] The network device may have the function of network energy saving, self-organizing networks (SON) / minimization of drive tests (MDT) . The terminal device may have the function of power saving.
[0032] 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.
[0033] In one embodiment, 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 one embodiment, the first network device may be a first RAT device and the second network device may be a second RAT device. In one embodiment, 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 one embodiment, information A may be transmitted to the terminal device from the first network device and information B may be transmitted to the terminal device from the second network device directly or via the first network device. In one embodiment, 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.
[0034] 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. The term ‘and / or’ indicates that there may be three relationships. For example, A and / or B may indicate cases includes ‘only A’ , ‘both A and B’ , and ‘only B’ . The term ‘at least one of the following items’ or a similar expression thereof refers to any combination of these items, including any combination of a single item or a plurality of items. For example, the term ‘at least one of A, B, or C’ may represent A, B, C, ‘A and B’ , ‘A and C’ , ‘B and C’ , or ‘A, B and C’ . The term ‘a set of’ may be interchangeably used with ‘one or more’ . Other definitions, explicit and implicit, may be included below.
[0035] 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.
[0036] As known, a transmission configuration indictor (TCI) state is used for indicating a QCL assumption for a certain channel or reference signal (RS) . Especially for an aperiodic (AP) channel status information reference signal (CSI-RS) , a higher layer parameter ‘qcl-info’ is configured to indicate a QCL assumption. It is necessary because a terminal device has only one chance to measure an AP CSI-RS. However, for AI / ML based beam prediction, a QCL configuration for CSI-RS resources used for a model inference, a performance monitoring and a model training is not defined.
[0037] In view of this, embodiments of the present disclosure provide solutions of communication for AI / ML based beam prediction. In one aspect, a terminal device may receive, from a network device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring. The terminal device may determine a set of QCL assumptions for a first set of resources associated with the performance monitoring, and measure a subset of resources in the first set of resources based on the set of QCL assumptions. Then the terminal device may determine a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions. As such, by determining a set of QCL assumptions, which is usually a subset of all Rx beams of a terminal device, the terminal device may not need to sweep all its Rx beams to measure performance monitoring resources, which saves complexity of the terminal device. By determining a subset of resources, overhead of RS transmission and report may be reduced.
[0038] For convenience, definitions of some terms in the present disclosure may be listed as below. · AI / ML Model: 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: a generic term referring to delivery of an AI / ML model from one entity to another entity in any manner. An entity may mean a network node / function (e.g., gNB, location management function (LMF) , etc. ) , UE, proprietary server, etc. · AI / ML model inference: 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: a subprocess of training, to evaluate 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: a process to train an AI / ML model (e.g., by learning an input / output relationship) in a data driven manner and obtain the trained AI / ML model for inference. · AI / ML model transfer: a delivery of an AI / ML model over an air interface in a manner that is not transparent to the third generation partnership project (3GPP) signalling, either parameters of a model structure known at a receiving end or a new model with parameters. The delivery may contain a full model or a partial model. · AI / ML model validation: a subprocess of training, to evaluate quality of an AI / ML model using a dataset different from one used for model training, that helps selecting model parameters that generalize beyond a dataset used for model training. · data collection: a process of collecting data by a network node, management entity, or UE for the purpose of AI / ML model training, data analytics and inference. · federated learning / federated training: 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: the term ‘functionality’ may refer to UE-capability information / parameters e.g., AI / ML-specific feature groups (FGs) . This interpretation may be suitable when ‘supported functionalities’ is used. The term ‘functionality’ may refer to an information element (IE) ‘CSI-ReportConfig’ for inference configuration or a set of inference related parameters or information / parameters indicated by UE. This interpretation may be suitable when ‘applicable functionalities’ is used. The term ‘functionality’ may refer to configurations based on CSI framework. This interpretation may be suitable when ‘activated functionalities’ is used. Therefore, meaning and granularity of the term ‘functionality’ for applicable functionalities, activated functionalities and supported functionalities may or may not be the same. · functionality identification: a process / method of identifying an AI / ML functionality for the common understanding between the NW 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: enable an AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model activation is equivalent to activation of configurations for model inference. · model deactivation: disable an AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model deactivation is equivalent to deactivation of configurations for model inference. · model download: Model transfer from the network to UE. · model identification: a process / method of identifying an AI / ML model for the common understanding between the NW and the UE. 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: a procedure that monitors inference performance of an AI / ML model. In some embodiments, model monitoring is equivalent to performance monitoring, or actions according to configurations for performance monitoring. · model parameter update: a process of updating model parameters of a model. · model selection: a process of selecting an AI / ML model for activation among multiple models for the same AI / ML enabled feature. Model selection may or may not be carried out simultaneously with model activation. In some embodiments, model selection is equivalent to selection of configurations for model inference. · model switching: a process of deactivating a currently active AI / ML model and activating a different AI / ML model for a specific AI / ML-enabled feature. In some embodiments, model switching is equivalent to deactivating a currently active configuration for AI / ML model and activating a different configuration for AI / ML model for a specific AI / ML-enabled feature. · model update: a process of updating model parameters and / or model structure of a model. · model upload: model transfer from UE to the network. · NW-side (AI / ML) model: an AI / ML Model whose inference is performed entirely at the network. · offline field data: data collected from field and used for offline training of an AI / ML model. · offline training: 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. In some embodiments, model training is equivalent to actions according to configurations for model training. · online field data: data collected from field and used for online training of the AI / ML model. · online training: 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. The notion of (near) real-time vs. non real-time is context-dependent and is relative to the inference time-scale. 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. Fine-tuning / re-training may be done via online or offline training. · reinforcement learning (RL) : a process of training an AI / ML model from an input (also known as a state) and a feedback signal (also known as a reward) resulting from the model’s output (also known as an 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 unlabelled data. · supervised learning: a process of training a model from input and its corresponding labels. · two-sided (AI / ML) model: 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: an AI / ML Model whose inference is performed entirely at the UE. · unsupervised learning: 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: 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.
[0039] In the context of the present disclosure, the terms ‘model’ , ‘functionality’ , ‘model / functionality’ , ‘inference configuration’ , ‘a set of inference related parameters’ and ’A I / ML related information / parameters indicated by UE’ may be used interchangeably. The terms ‘feature’ and ‘feature group’ may be used interchangeably. The terms ‘model’ and ‘model group’ may be used interchangeably. The terms ‘functionality’ , ‘functionality group’ , ‘functionality set’ may be used interchangeably. The terms ‘ID’ , ‘index’ , ‘indicator’ and ‘identifier’ may be used interchangeably.
[0040] In the context of the present disclosure, the term ‘NW’ herein may refer to ‘operations, administration and maintenance (OAM) ’ , ‘server’ , or ‘advanced mobile location (AML) / LMF’ . The term ‘occasion’ herein may be interchangeably used with ‘report occasion’ .
[0041] In the context of the present disclosure, the terms ‘beam’ , ‘precoder’ , ‘precoding’ , ‘precoding matrix’ , ‘spatial relation information’ , ‘spatial relation info’ , ‘precoding information’ , ‘precoding information and number of layers’ , ‘precoding matrix indicator (PMI) ’ , ‘precoding matrix indicator’ , ‘transmission precoding matrix indication’ , ‘precoding matrix indication’ , ‘TCI state’ , ‘uplink (UL) TCI state’ , ‘joint TCI state’ , ‘transmission configuration indicator’ , ‘QCL’ , ‘quasi-co-location’ , ‘QCL parameter’ , ‘QCL assumption’ , ‘QCL relationship’ and ‘spatial relation’ can be used interchangeably. A beam may refer to a downlink (DL) beam, UL beam, transmitting (Tx) beam, Rx beam, beam pair, RS resource, RS resource set, antenna port, antenna port group, antenna element (s) , antenna array (s) , or beam group. The RS resource may refer to a DL RS resource, UL RS resource, CSI-RS resource, synchronization signal (SS) / physical broadcast channel (PBCH) block, or SRS resource. The RS resource set may refer to a DL RS resource set, UL RS resource set, CSI-RS resource set, SS / PBCH block set, or SRS resource set.
[0042] In the context of the present disclosure, QCL types corresponding to each DL RS may be given by a higher layer parameter ‘qcl-Type’ in QCL-Info and may take one of the following values: - 'typeA' : {Doppler shift, Doppler spread, average delay, delay spread} - 'typeB' : {Doppler shift, Doppler spread} - 'typeC' : {Doppler shift, average delay} - 'typeD' : {Spatial Rx parameter} .
[0043] In the context of the present disclosure, TypeD may correspond to Rx beam herein.
[0044] The terms ‘reference signal received power (RSRP) ’ , ‘layer 1 (L1) -RSRP’ , ‘layer 3 (L3) -RSRP’ , ‘filtered RSRP’ herein may be used interchangeably. If ‘RSRP’ is used as a beam quality metric, methods are readily extended to other metrics like ‘signal to interference and noise ratio (SINR) ’ , ‘reference signal received quality (RSRQ) ’ , ‘received signal strength indicator (RSSI) ’ , etc. The term ‘CSI-RS for beam management’ and ‘CSI-RS configured in a resource set with higher layer parameter ‘repetition” may be used interchangeably. The term ‘CSI-RS for channel acquisition’ and ‘CSI-RS configured in a resource set without higher layer parameter ‘repetition’ and without higher layer parameter ‘trs-info’ may be used interchangeably.
[0045] In the context of the present disclosure, the term ‘for model inference’ herein may refer to ‘for the report configuration for model inference’ or ‘for the report configuration with report quantities configured with predicted results such as predicted CSI-RS resource indicator (CRI) / SS / PBCH block resource indicator (SSBRI) , predicted RSRP, predicted CSI, etc. ’ The term ‘for performance monitoring’ herein may refer to ‘for the report configuration for performance’ or ‘for the report configuration with report quantities configured with performance monitoring metric such as prediction accuracy, RSRP difference between prediction and measurement, difference between prediction and ground truth, etc. ’ The term ‘for model training herein may refer to ‘for the report configuration for model training’ or ‘for the report configuration with report quantities configured with both model inputs related results and model output related results such as RSRP measurement results of both Set B and Set A, RSRP measurement results of Set B, and beam information of Set A, etc. ’
[0046] In the context of the present disclosure, the term ‘historical measurement result’ may refer to one of the following options: · Option 1: based on number of measurements (denoted as Pt) , number of RSs (denoted as Mt) and prediction time (denoted as T2) . T2 means a time duration for prediction. Mt means the number of time instances for measurement as AI / ML inputs with a periodicity of Tper. Pt means the number of time instance (s) for prediction with a periodicity of Tper in T2. · Option 2: based on a periodicity (denoted as T) of required reference signals for measurements. For every T=Y ms, reference signals for measurements are needed. · Option 3: based on times (denoted as Z) of a given minimal periodicity Tper of reference signals for measurements. UE may measure reference signals for model inputs every Z times of Tper. · Option 4: based on an observation window (e.g., number / distance) , e.g., 5 / 5ms (e.g., 5 times observations and 5ms between two observations, or 5 times observations in 5ms) , or 10 / 5ms (e.g., 10 times observations and 5ms between two observations, or 10 times observations in 5ms) .
[0047] It is to be noted that the term ‘periodicity of Tper’ , ‘T’ , or ‘distance’ mentioned above may be called as an interval or time interval between two historical measurements, or between two measurements for historical results. The time duration related to Mt×Tper, Y ×Tper, or ‘distance’ , or ‘observation window’ mentioned above may be called as a measurement window for historical measurement results. The measurement window for historical measurement results may be configured by a network device, or reported by a terminal device, or pre-defined or parameters of the AI / ML model.
[0048] In the context of the present disclosure, the term ‘future time instance’ may refer to one of the following options: · Option 1: N future time instance (s) that is based on an output of AI / ML model inference. The future time instance may include information about a timestamp, or an interval between two joint time instances. · Option 2: based on number of measurements (denoted as Pt) , number of RSs (denoted as Mt) and prediction time (denoted as T2) . T2 means a time duration for beam prediction. Mt means the number of time instances for measurement as AI / ML inputs with a periodicity of Tper. Pt means the number of time instance (s) for prediction with a periodicity of Tper in T2. · Option 3: based on a periodicity (denoted as T) of required reference signals for measurements. For every T=Y ms, reference signals for measurements are needed. · Option 4: based on times (denoted as Z) of a given minimal periodicity Tper of reference signals for measurements. UE may measure reference signals for model inputs every Z times of Tper. Prediction time is defined as the time from each measurement instance to the latest prediction instance before the next measurement instance. · Option 5: based on a prediction window (e.g., number / distance between prediction instances / distance from the last observation instance to the first or starting prediction instance) , e.g., 1 / 5ms / 5ms.
[0049] It is to be noted that the term ‘periodicity of Tper’ or ‘distance’ mentioned above may be called as an interval or time interval between two future time instances. The time duration related to Pt×Tper, ‘Z’ , ‘distance’ , or number / distance may be called as prediction window for future time instances. The prediction window for future time instances may be configured by a network device, or reported by a terminal device, or pre-defined or parameters of the AI / ML model.
[0050] 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, a dataset ID, a set of AI / ML related parameters, or an ID of a set of AI / ML related parameters. As a result, the above terms may be used interchangeably.
[0051] In some embodiments, the model may be represented by or associated with a channel, a resource, a resource set, a RS resource, a RS resource set, a RS port, a set of RS ports, a RS port ID, or a set of RS port IDs. In some embodiments, the model may comprise a set of weight values that may be learned during training, for example for a specific architecture or configuration, where a set of weight 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, or L1-SINR) of a set of beams of a set of candidate cells.
[0052] In some embodiments, an input of the model (i.e., AI / ML 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 the model (i.e., AI / ML output) may refers to the output of a model and indicate result (s) outputted by the model, which is equivalent to label / data.
[0053] In some embodiments, a ground truth label of data (or ground-truth label) for monitoring or training the model (i.e., AI / ML output) may refers to the authoritative, accepted data, or true answer or outcome for the model. In some embodiments, ‘ground truth’ , ‘ground truth label’ , ‘ground truth label of data’ , ‘input label’ , ‘input data’ and ‘data’ can be used interchangeably. In some embodiments, the ground truth can be interpreted as actual / factual (i.e. actual / factual measured) data / values / results / collections / parameters, which can be used as reference, compared to prediction or inference.
[0054] Principles and implementations of the present disclosure will be described in detail below with reference to the figures.EXAMPLE OF COMMUNICATION NETWORK
[0055] FIG. 1A illustrates an example communication network 100A in which embodiments of the present disclosure can be implemented. As shown in FIG. 1A, the communication network 100A may comprise a terminal device 110 and a network device 120 served by the terminal device 110.
[0056] As shown in FIG. 1A, the terminal device 110 may have a plurality of beams, and the network device 120 may have a plurality of beams. A channel (or called as a sub-channel) may be formed between one of the beams of the terminal device 110 and one of the beams of the network device 120. The terminal device 110 may transmit information to the network device 120 or receive information from the network device 120 via one or more of the sub-channels.
[0057] It is to be understood that the number of devices and beams in FIG. 1A is given for the purpose of illustration without suggesting any limitations to the present disclosure. The communication network 100A may include any suitable number of network devices and / or terminal devices and / or beams adapted for implementing implementations of the present disclosure.
[0058] The communications in the communication network 100A 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.
[0059] Communication in a direction from the terminal device 110 towards the network device 120 is referred to as UL communication, while communication in a reverse direction from the network device 120 towards the terminal device 110 is referred to as DL communication. A 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) .
[0060] FIG. 1B illustrates a schematic diagram 100B illustrating an example model inference for AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented. As shown in FIG. 1B, in an inference procedure for beam management (BM) , measurements based on a set of beams (also referred to as Set B herein) are used as a model input of an AI / ML model 130 to predict information of another set of beams (also referred to as Set A herein) . In addition, beam ID information may be also provided as an input to an AI / ML model. Based on model output (e.g., probability of each beam in Set A to be the Top 1 beam, predicted L1-RSRP) of the AI / ML model 130, Top 1 / Top K beam (s) among Set A of beams can be predicted and / or potentially with predicted L1-RSRPs (depending on a labeling) . For BM-Case 1, the measurements of Set B are used as a model input to predict Top 1 / Top K beams from Set A. For BM-Case 2, the measurements of Set B at historic time instance (s) are used as a model input for temporal DL beam prediction of beams from Set A. The case that Set A and Set B are different (Set B is not a subset of Set A) , and Set B is a subset of Set A for both BM-Case 1 and BM-Case 2, and the case that Set A and Set B are the same for BM-Case 2 are considered. Performance of DL Tx beam prediction and DL Tx-Rx beam pair prediction may be evaluated.
[0061] For both BM-Case 1 and BM-Case 2, a terminal device may report a prediction result to NW based on an output of a UE-sided model, or NW may predict the Top 1 / Top K beam (s) based on the reported measurements of Set B for a NW-sided model. BM-Case 1 may also be called as a spatial-domain prediction, and BM-Case 2 may also be called as a time-domain prediction herein.
[0062] As mentioned above, for AI / ML based beam prediction, a QCL configuration for CSI-RS resources used for a model inference, a performance monitoring and a model training has not been defined. It is still unclear how to determine whether a prediction is accurate or not upon consideration of a QCL assumption of a resource.
[0063] It is assumed that a QCL assumption corresponding to a predicted beam is the same as a QCL assumption corresponding to a measured beam. FIG. 2A illustrates a schematic diagram 200A illustrating an example scenario of AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented.
[0064] In some scenarios, as shown by a reference sign 210 in FIG. 2A, Top 1 predicted beam 211 in Set A matches a Rx beam 212 which may be not used during a model inference. Top 1 measured beam 213 matches a Rx beam 214 which is aligned with the Rx beam 212. The Top 1 measured beam 213 corresponds to the Top 1 predicted beam 211. Thus, the prediction is an accurate prediction if the performance matric is based on the accuracy which is defined as whether Top 1 predicted beam is the Top 1 measured beam.
[0065] In some scenarios, as shown by a reference sign 220 in FIG. 2A, Top 1 predicted beam 221 in Set A matches a Rx beam 222 which may be not used during a model inference. Top 1 measured beam 223 matches a Rx beam 224 which is aligned with the Rx beam 222. The Top 1 measured beam 223 does not correspond to the Top 1 predicted beam 221. Thus, the prediction is an inaccurate prediction if the performance matric is based on the accuracy which is defined as whether Top 1 predicted beam is the Top 1 measured beam.
[0066] That is, if a QCL assumption corresponding to a predicted beam is the same as a QCL assumption corresponding to a measured beam, it is clear to determine whether a prediction is accurate or not.
[0067] It is assumed that a QCL assumption corresponding to a predicted beam is different from a QCL assumption corresponding to a measured beam. FIG. 2B illustrates a schematic diagram 200B illustrating another example scenario of AI / ML based beam prediction in which some embodiments of the present disclosure can be implemented.
[0068] In some scenarios, as shown by a reference sign 230 in FIG. 2B, Top 1 predicted beam 231 in Set A matches a Rx beam 232 which is not known at NW. Top 1 measured beam 233 matches a Rx beam 234 which is not aligned with the Rx beam 232. The Top 1 measured beam 233 does not correspond to the Top 1 predicted beam 231. In this case, it is unclear whether the prediction is an inaccurate prediction.
[0069] For example, some example predictions and measurements are shown in Table 1 below. Table 1
[0070] It can be seen that, if measured Top 1 beam obtained with first Rx beam is beam 1, the prediction may be treated as accurate. If measured Top 1 beam obtained with second Rx beam is beam 2, the prediction may be treated as inaccurate. Thus, it is unclear whether the prediction is an inaccurate prediction.
[0071] In some scenarios, as shown by a reference sign 240 in FIG. 2B, Top 1 predicted beam 241 in Set A matches a Rx beam 242 which is not known at NW. Top 1 measured beam 243 matches a Rx beam 244 which is not aligned with the Rx beam 242. The Top 1 measured beam 243 corresponds to the Top 1 predicted beam 241. In this case, it is unclear whether the prediction is an accurate prediction.
[0072] For example, some example predictions and measurements are shown in Table 2 below. Table 2
[0073] It can be seen that, if a performance metric is a RSRP difference, |RSRP_p -RSRP_1| < |RSRP_p -RSRP_2|, then with first Rx beam, the prediction may be treated as accurate, and with second Rx beam, the prediction may be treated as inaccurate. Thus, it is unclear whether the prediction is an accurate prediction.
[0074] That is, if a QCL assumption corresponding to a predicted beam is different from a QCL assumption corresponding to a measured beam, it is unclear how to determine whether a prediction is accurate or not.
[0075] Embodiments of the present disclosure provide solutions of a performance monitoring for AI / ML based beam prediction so as to overcome the above and other potential issues. Detailed description will be given with reference to FIGs. 3 to 4E below.EXAMPLE IMPLEMENTATION OF PERFORMANCE MONITORING
[0076] FIG. 3 illustrates a signaling chart illustrating an example process 300 of communication according to some embodiments of the present disclosure. For the purpose of discussion, the process 300 will be described with reference to FIG. 1A. The process 300 may involve the terminal device 110 and the network device 120 as illustrated in FIG. 1A. It is to be understood that the steps and the order of the steps in FIG. 3 are merely for illustration, and not for limitation. For example, the order of the steps may be changed. Some of the steps may be omitted or any other suitable additional steps may be added.
[0077] As shown in step 310 of FIG. 3, the terminal device 110 may receive, from the network device 120, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality. The configuration of the model inference is associated with the configuration of the performance monitoring.
[0078] In some embodiments, the configuration of the performance monitoring may indicate a set of resources (for convenience, also referred to as a first set of resources or a monitoring resource set herein) associated with the performance monitoring (i.e., for measurement) .
[0079] In some embodiments, the configuration of the model inference may indicate a set of resources (for convenience, also referred to as a second set of resources herein) associated with a prediction for the model inference. It is to be understood that the second set of resources corresponds to Set A described above. In some embodiments, the configuration of the model inference may indicate a set of resources (for convenience, also referred to as a third set of resources herein) associated with a measurement for the model inference. It is to be understood that the third set of resources corresponds to Set B described above. In some embodiments, the terminal device 110 may be provided with the configuration of model training which may indicate two set of resources associated with a prediction for the model inference. The configuration of model training is associated with the configuration of the model inference and / or the configuration of the performance monitoring. It is to be understood that the two set of resources correspond to Set B and Set A used for model training.
[0080] It is to be noted that the configuration of the performance monitoring and the configuration of the model inference may comprise any suitable information, and may be associated with each other in any suitable ways. The present disclosure does not limit these aspects.
[0081] As shown in step 320 of FIG. 3, the terminal device 110 may determine a set of QCL assumptions for the first set of resources. In other words, the terminal device 110 may determine a limited set of Rx beams to measure one or more resources for the performance monitoring. The other QCL assumptions (e.g., one or more Rx beams not in the limited set of Rx beams of the terminal device 110) may not be used to measure the one or more resources for the performance monitoring. There may be several methods to determine the set of QCL assumptions.
[0082] In some embodiments, the configuration of the performance monitoring may comprise a configuration for the set of QCL assumptions. That is, the set of QCL assumptions may be explicitly configured for the performance monitoring. In this case, the terminal device 110 may determine the set of QCL assumptions based on a configuration of a QCL assumption for a resource in the first set of resources.
[0083] It is to be noted that the configuration for the set of QCL assumptions may be dependent on NW implementation. In some embodiments, the configuration for the set of QCL assumptions may provide QCL information per resource or per resource set. In some embodiments, if the QCL information is configured per resource, the set of QCL assumptions may include those configured for the second set of resources (e.g., predicted Top K beams based on the model inference) .
[0084] In some embodiments, the terminal device 110 may determine the set of QCL assumptions based on one or more associated configurations.
[0085] For example, the terminal device 110 may determine the set of QCL assumptions based on an associated model inference configuration, for which an ID of an inference report configuration is configured in the configuration for the performance monitoring. For example, the terminal device 110 may determine the set of QCL assumptions based on a QCL assumption for a resource in the third set of resources. In other words, the set of QCL assumptions may be determined during the model inference, e.g., based on the measurement of resources for Set B.
[0086] In another example, the terminal device 110 may determine the set of QCL assumptions based on the associated model inference configuration first, and then based on the associated model training configuration. For example, the terminal device 110 may determine the second set of resources (e.g., predicted Top K beams) based on model inference, then determine the set of QCL assumptions based on a QCL assumption for a resource in the second set of resources used in the model training. In some embodiments, the model inference configuration may be linked to the model training via an associated ID. In some embodiments, the set of QCL assumptions may be associated with those configured or determined during the model training, e.g., measurement of resources in Set B / Set A. In some embodiments, the terminal device 110 may need to provide information on QCL assumptions for model training during data collection. In some embodiments, the terminal device 110 may need to provide information on QCL assumptions (e.g., best Rx beam) of resources in Set B / Set A after model training. In some embodiments, the terminal device 110 may provide information on QCL assumptions of resources (e.g., best Rx beam) in Set B / Set A after model training.
[0087] In some embodiments, resources in the first set of resources and resources for a model training for the functionality may be configured with same QCL information. In other words, NW may configure the same ‘qcl-info’ for resources in the monitoring resource set and resources for the model training. For example, the set of QCL assumptions may be determined during the model training, e.g., the measurement of resources for Set B as model inputs.
[0088] In some embodiments, the resources in the first set of resources and resources in a second set of resources associated with a prediction for the model inference may be configured with same QCL information. In other words, NW may configure the same ‘qcl-info’ for resources in the monitoring resource set and resources in Set A for the model inference. For example, the set of QCL assumptions may be configured for Set A for the model inference, and the configuration may be based on QCL assumptions determined during the model training.
[0089] In some embodiments, the terminal device 110 may align a QCL assumption in the set of QCL assumptions to a QCL assumption or a potential QCL assumption of a predicted beam during the model inference. That is, a QCL assumption of a measurement resource during performance monitoring needs to be aligned to the (potential) QCL assumption of the predicted beam during model inference. For example, the QCL assumption (e.g., Rx beam of the terminal device 110) of the predicted beam may not be applied during the model inference, or the predicted beam may be within the set of measured beams during the model inference (i.e., in Set B) . If the predicted beam is within the set of measured beams during the model inference (i.e., in Set B) , the terminal device 110 may align a QCL assumption in the set of QCL assumptions to a QCL assumption of the predicted beam during the model inference. If the QCL assumption of the predicted beam is not be applied during the model inference (i.e., in Set A but not in Set B) , the terminal device 110 may align a QCL assumption in the set of QCL assumptions to a QCL assumption configured for the predicted beam (e.g., potential QCL assumption of a resource in Set A) during the model inference.
[0090] If the predicted beam is not transmitted during the model inference or the QCL assumption of the predicted beam is not applied during the model inference, the terminal device 110 may align the QCL assumption in the set of QCL assumptions to a QCL assumption of a resource corresponding to the predicted beam during the model training. That is, if the predicted beam is not transmitted during the model inference or the QCL assumption of the predicted beam is not applied during the model inference, the QCL assumption of the measurement resource needs to be aligned to the QCL assumption of the corresponding resources determined during the model training.
[0091] FIG. 4A illustrates a schematic diagram 400A illustrating an example QCL assumption determination according to some embodiments of the present disclosure. As shown in FIG. 4A, Top 1 predicted beam 411 in Set A matches a Rx beam 412 which may be not used during the model inference. Top 1 measured beam 413 matches a Rx beam 414. If the Rx beam 414 is not used during the model inference, the Rx beam 414 may need to be linked to model training. The matched Rx beam 412 of the Top 1 predicted beam 411 in Set A has been used during the model training, and thus the Rx beam 414 of the Top 1 measured beam 413 is linked to the Rx beam 412 which has been used during the model training. As such, a QCL assumption for a Top 1 measured beam may be determined correctly.
[0092] In some embodiments, the set of QCL assumptions may be associated with a set of future time instances. In some embodiments, for time domain prediction with predicted results for more than one future time instance, for a same resource, QCL assumptions may be different for different future time instances. In some embodiments, same QCL assumption may be needed for different future time instances. In some embodiments, the terminal device 110 may align a QCL assumption of a future time instance during the performance monitoring to a QCL assumption of the future time instance during the model inference or model training. For example, the QCL assumption corresponding to f-th further instance during the performance monitoring needs to be aligned to the QCL assumption of f-th further instance during the model inference or model training.
[0093] FIG. 4B illustrates a schematic diagram 400B illustrating an example QCL assumption determination for a time-domain prediction according to some embodiments of the present disclosure. As shown in FIG. 4B, during the model inference, Top 1 predicted beam 421 for a first future time instance matches a Rx beam 422, and Top 1 predicted beam 423 for a second future time instance matches a Rx beam 424. During the performance monitoring, Top 1 measured beam 425 for the first future time instance matches a Rx beam 426, and Top 1 measured beam 427 for the second future time instance matches a Rx beam 428. In this case, the terminal device 110 may use the Rx beam 422 for the first future time instance during the model inference to check whether the Top 1 measured beam 425 matches the Top 1 predicted beam 421, or check whether a L1-RSRP difference between the Top 1 predicted beam 421 and the Top 1 measured beam 425 is lower than or equal to a threshold. And the terminal device 110 may use the Rx beam 424 for the second future time instance during the model inference to check whether the Top 1 measured beam 427 matches the Top 1 predicted beam 423, or check whether a L1-RSRP difference between the Top 1 predicted beam 423 and the Top 1 measured beam 427 is lower than or equal to a threshold.
[0094] In some embodiments, prediction results may be Top K beams, for each of the Top K beams, the QCL assumptions may be different. In some embodiments, same QCL assumption may be needed for different future time instances. In some embodiments, the terminal device 110 may align a QCL assumption corresponding to a first measured beam in a first number of best measured beams (e.g., Top K measured beams) during the performance monitoring to a QCL assumption of a first predicted beam in the first number of best predicted beams (e.g., Top K predicted beams) during the model inference.
[0095] In some embodiments, if a ranking order of Top K beams is also a part of prediction, a performance monitoring metric needs to consider accuracy of the ranking order predicted and the QCL assumptions of the k-th measured beam.
[0096] In some embodiments, the predicted ranking order may be based on predicted RSRP. For example, the predicted ranking order may be based on predicted RSRP with the same QCL assumptions. In another example, the predicted ranking order may be based on predicted RSRP with the same or different QCL assumptions.
[0097] In some embodiments, the predicted ranking order may be based on predicted beams with probability to be Top 1 or Top K. For example, the predicted ranking order may be based on predicted beams with probability to be Top 1 or Top K with the same QCL assumptions. In another example, the predicted ranking order may be based on predicted beams with probability to be Top 1 or Top K with the same or different QCL assumptions.
[0098] In some embodiments, the measured ranking order may be based on measured RSRP. For example, the measured ranking order may be based on measured RSRP with the same QCL assumptions. In another example, the measured ranking order may be based on measured RSRP with the same or different QCL assumptions.
[0099] In some embodiments, the terminal device 110 may align an order of QCL assumptions in the first number of best measured beams (e.g., Top K measured beams) during the performance monitoring to an order of QCL assumptions in the first number of best predicted beams (e.g., Top K predicted beams) during the model inference. If the first predicted beam is not transmitted during the model inference or the QCL assumption of the first predicted beam is not applied during the model inference, the terminal device 110 may align the QCL assumption corresponding to the first measured beam to a QCL assumption of a resource corresponding to the first predicted beam during the model training.
[0100] In other words, the QCL assumption corresponding to the k-th top measured beam during the performance monitoring needs to be aligned to the QCL assumption of the k-th top predicted beam during the model inference. If the k-th top predicted beam is not transmitted during the model inference, or the QCL assumption of the k-th top predicted beam is not applied during the model inference, the QCL assumption corresponding to the k-th top measured beam needs to be linked to the QCL assumption of the corresponding resources determined during the model training.
[0101] For illustration, an example is shown in Table 3 below. Table 3
[0102] As shown in Case 1 of Table 3, measured Top K during performance monitoring may have a ranking order different from predicted Top K, but stick to the same order to apply the QCL assumption. In this case, the performance monitoring results suggest that the prediction may not be accurate.
[0103] As shown in Case 2 of Table 3, measured Top K during performance monitoring may have a ranking order same as predicted Top K, but the terminal device 110 may apply different QCL assumptions (or different order of applying QCL assumptions) . In this case, the performance monitoring results still suggest that the prediction may not be accurate.
[0104] Continuing to refer to FIG. 3, as shown in step 330, the terminal device 110 may measure, based on the set of QCL assumptions, a subset of resources in the first set of resources. In some embodiments, upon determination of the set of QCL assumptions in the step 320, the terminal device 110 may only apply the set of QCL assumptions to measure the subset of resources during the performance monitoring. This means that the terminal device 110 may only measure a subset of resources configured in the monitoring resource set.
[0105] In some embodiments, the subset of resources may comprise a resource configured with a same QCL assumption as a QCL assumption of a predicted beam in the first number of best predicted beams (e.g., Top K predicted beams) . In other words, the terminal device 110 may only measure the subset of resources with the same QCL assumptions as the Top K predicted beams. The terminal device 110 may not measure one or more monitoring resources with one or more QCL assumptions other than the set of QCL assumptions.
[0106] In some embodiments, the subset of resources may comprise a resource configured with a same QCL assumption as a QCL assumption of a beam in Set B, e.g., when a predicted beam in the Top K predicted beams belongs to Set B.
[0107] In some embodiments, the network device 120 may only transmit one or more RSs via the subset of resources correspondingly. The network device 120 may be aware of the configured QCL information per each resource in the whole monitoring resource set. In other words, the network device 120 may not transmit those resources not in the subset, or those resources configured with a different QCL assumption other than the Top K predicted beams. For example, the network device 120 may transmit other physical channels / signals in those resource blocks (RBs) / resource element (REs) configured for monitoring resources.
[0108] FIG. 4C illustrates a schematic diagram 400C illustrating an example determination of a subset of resources according to some embodiments of the present disclosure. As shown in FIG. 4C, during the model inference, Top 1 predicted beam 431 matches a Rx beam 432. During the performance monitoring, the terminal device 110 may only measure a subset of resources 433 with the same QCL assumption as the Top 1 predicted beam 431. For example, Top 1 measured beam 434 may match a Rx beam 435 which may need to be linked to model training.
[0109] In some embodiments where the set of QCL assumptions is associated with a set of future time instances, the terminal device 110 may measure a resource set associated with a future time instance based on a QCL assumption associated with the future time instance. That is, for time-domain prediction with predicted results for more than one future time instances, for different further time instances, the subset selection may be different. The terminal device 110 may need to measure different subsets of resources in the monitoring resource set based on different QCL assumption for different future time instances. In some embodiments, same QCL assumption may be needed for different future time instances.
[0110] FIG. 4D illustrates a schematic diagram 400D illustrating an example determination of a subset of resources for a time-domain prediction according to some embodiments of the present disclosure. As shown in FIG. 4D, during the model inference, Top 1 predicted beam 441 for a first future time instance matches a Rx beam 442, and Top 1 predicted beam 443 for a second future time instance matches a Rx beam 444. Resources 445 in the monitoring resource set match the Rx beam 442, and resources 446 in the monitoring resource set match the Rx beam 444. During the performance monitoring, the terminal device 110 may only measure the resources 445 for the first future time instance, and only measure the resources 446 for the second future time instance.
[0111] In some embodiments, the prediction results may be Top K beams. For each of the Top K beams, the set of QCL assumptions may be different. In some embodiments, same QCL assumption may be needed for all the Top K beams. In some embodiments, the terminal device 110 may need to measure more than one subsets of resources in the monitoring resource set. In total, the terminal device 110 may need to measure the union of each subset corresponding to the set of QCL assumptions.
[0112] In some embodiments, Top K predicted beams may be grouped by QCL assumptions. For example, the Top K predicted beams may be grouped into Top K1 predicted beams with QCL 1 and Top K2 predicted beams with QCL 2.
[0113] In some embodiments, measurement resources in the monitoring resource set may be grouped by QCL assumptions of the predicted beams. For example, resources with QCL 1 belongs to subset 1, and the terminal device 110 may need to determine Top K1 measured beams in subset 1. Resources with QCL 2 belongs to subset 2, and the terminal device 110 may need to determine Top K2 measured beams in subset 2. Resources with QCL assumptions different from QCL 1 and QCL 2 may not be measured for performance monitoring.
[0114] As such, by determining a subset of resources, overhead of RS resources for performance monitoring may be saved.
[0115] As shown in step 340 of FIG. 3, the terminal device 110 may determine a result of the performance monitoring (i.e., the performance monitoring metric) based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.
[0116] In some embodiments, the terminal device 110 may determine the result of the performance monitoring by: determining a second number of best measured beams (e.g., Top K-i measured beams) associated with a QCL assumption (e.g., i-th QCL assumption) ; determining the second number of best predicted beams (e.g., Top K-i predicted beams) associated with the QCL assumption; and determining a first part of the result associated with the QCL assumption based on the second number of best measured beams and best predicted beams. In other words, the performance monitoring metric may be calculated per QCL assumptions and subsets accordingly.
[0117] In some embodiments, the terminal device 110 may determine the first part of the result based on whether the second number of best predicted beams matches the second number of best measured beams. For example, for i-th QCL assumption, the performance monitoring metric may be based on whether the Top K-i predicted beams match the Top K-i measured beams.
[0118] In some embodiments, the terminal device 110 may determine the first part of the result based on whether a measurement result difference between the second number of best predicted beams and the second number of best measured beams is lower than or equal to a difference threshold. For example, for i-th QCL assumption, the performance monitoring metric may be based on a RSRP difference between predicted Top K-i beams and the measured Top K-i beams, or any other suitable ways. It is to be noted that the terminal device 110 may calculate and report the performance monitoring metric with all the different QCL assumptions.
[0119] For illustration, an example is described in connection with Tables 4 to 7 below. Table 4 shows an example prediction result (Top 4 predicted beam) and its QCL information. Table 4
[0120] Based on different QCL assumptions, Tables 5 and 6 may be derived from Table 4 and shown as below. Table 5 shows an example prediction result with QCL 1, and Table 6 shows an example prediction result with QCL 2. Table 5 Table 6
[0121] Table 7 shows an example measurement result (Top 4 measured beam) . In Table 7, a measurement result with QCL 1 and a measurement result with QCL 2 are also separately considered. Table 7
[0122] Based on a comparison between the prediction result for QCL 1 in Table 5 and the measurement result for QCL 1 in Table 7, the terminal device 110 may determine a monitoring result for QCL 1 indicating that the prediction is correct. Based on a comparison between the prediction result for QCL 2 in Table 6 and the measurement result for QCL 2 in Table 7, the terminal device 110 may determine a monitoring result for QCL 2 indicating the prediction is correct.
[0123] For illustration, some example embodiments for prediction accuracy will be described as below.
[0124] For BM-Case1 and BM-Case2 with a UE-sided AI / ML model, for UE-assisted performance monitoring, at least one of the following alternatives for monitoring metric may be used, including: - Alt 1: Top 1 or Top K beam prediction accuracy (with or without margin) by comparing the prediction results and the Top 1 or Top K beam based on the measurements from a resource set / resources for monitoring; - Alt 2: The L1-RSRP difference information based on actual measurement of the L1-RSRP of one or more of Top K predicted beam, and L1-RSRP measurements from a resource set / resources for monitoring; - Alt 3: The RSRP difference information between the predicted RSRP and measured L1- RSRP of corresponding beam (s) of a resource set / resources for monitoring; - Alt 4: The probability information of the predicted beam (s) to be the Top 1 or Top K beam.
[0125] In some embodiments, to calculate the accuracy (i.e., whether correct or incorrect) , the following options may be used to count whether a model inference (inference report / model inference instance) or a performance monitoring instance is accurate or not: - Option 1 (Top-1 / 1) : the Top-1 beam with largest measured value of the resource set (s) for monitoring is Top-1 predicted beam; - Option 2 (1 / Top-K) : the Top-1 beam with largest measured value of L1-RSRP of the resource set (s) for monitoring is one of the Top-K predicted beams, where K >1, or K is equal to the number of reported predicted beam configured by inference report configuration; - Option 3a (Top-K / M) : the Top-K predicted beams are among Top M beam (s) with largest M measured value (s) of L1-RSRP (s) of the resource set (s) for monitoring; - Option 3b (Top-K / M) : at least one of the Top-K predicted beams is among Top M beam (s) with largest M measured value (s) of L1-RSRP (s) of the resource set (s) for monitoring; - Option 4 (best of Top-K / 1 with margin) : The beam with largest measured value of L1- RSRP of Top-K predicted beams is within a margin X dB of largest measured value of L1-RSRP of the resource set (s) for monitoring. In above options, K may be >1, or K is equal to the number of reported predicted beam configured by inference report configuration; K, M, X are also configurable or may be reported by UE.
[0126] In some embodiments, the prediction accuracy may be the number of correct (or incorrect) monitoring instances (or model inference instances) divided the number of total monitoring instances (or model inference instances) in a monitoring window. In some examples, the prediction accuracy may also be the ratio of the number of correct (or incorrect) monitoring instances (or model inference instances) over the number of the total monitoring instances (or model inference instances) . In some examples, the prediction accuracy may also be the ratio of the number of correct monitoring instances (or model inference instances) over the number of the incorrect monitoring instances (or model inference instances) .
[0127] In some embodiments, one correct (or incorrect) monitoring instance may be defined as the number (or the ratio) of correct (or incorrect) beams in Top K predicted beams is higher (or lower) than a threshold, for the monitoring instance. A correct (or incorrect) predicted beam may be a beam in (or not in) the Top K measured beams, or the RSRP difference between predicted RSRP and measurement RSRP is lower (or higher) than a threshold.
[0128] In some embodiments, one correct (or incorrect) monitoring instance may be defined as the order of Top K predicted beams based on predicted RSRP (or probability) is the same (or different) as the order or Top K measured beams based on measured RSRP, or the correlation factor between the order of Top K predicted beams based on predicted RSRP (or probability) and the order or Top K measured beams based on measured RSRP is higher (or lower) than a threshold.
[0129] In some embodiments, the QCL assumption of each one of Top K predicted beams and the QCL assumption of each one of Top K measured beams may need to be aligned in a one-to-one manner.
[0130] In some embodiments, one correct (or incorrect) monitoring instance may be defined as the number (or the ratio) of correct (or incorrect) time instances in F predicted time instances is higher (or lower) than a threshold, for the monitoring instance. A correct (or incorrect) predicted time instance may be the predicted beam is (or is not) the Top 1 measured beams at the time instance, or the RSRP difference between predicted RSRP and measurement RSRP is lower (or higher) than a threshold.
[0131] In some embodiments, the QCL assumption of each one of F predicted time instances in model inference and the QCL assumption of each one of F predicted time instances in performance monitoring may need to be aligned in a one-to-one manner.
[0132] In some examples, a correct (or incorrect) predicted time instance may be defined as the number (or the ratio) of correct (or incorrect) beams in Top K predicted beams in the predicted time instance is higher (or lower) than a threshold.
[0133] In some examples, a correct (or incorrect) predicted time instance may be defined as the order of Top K predicted beams based on predicted RSRP (or probability) is the same (or different) as the order or Top K measured beams based on measured RSRP, or the correlation factor between the order of Top K predicted beams based on predicted RSRP (or probability) and the order or Top K measured beams based on measured RSRP is higher (or lower) than a threshold.
[0134] In some examples, the prediction accuracy may be the average of the prediction accuracy per monitoring instance (or model inference instances) in a monitoring window. If one monitoring instance (or model inference instances) is with no performance monitoring results (or model inference results) , the monitoring instance (or model inference instances) may be dropped in the average operation, or be counted as 0 or 1.
[0135] In some embodiments, the prediction accuracy per monitoring instance may be defined as the ratio of correct (or incorrect) beams in Top K predicted beams for a monitoring instance. A correct (or incorrect) predicted beam may be a beam in (or not in) the Top K measured beams, or the RSRP difference between predicted RSRP and measurement RSRP is lower (or higher) than a threshold. In some examples, a weighted ratio may be used that the Top i-th predicted beam is with a higher weight than that of Top j-th predicted beam if i < j.
[0136] In some embodiments, the prediction accuracy per monitoring instance may be related to a correlation factor (e.g., a normalized factor between 0 and 1) between the order of Top K predicted beams based on predicted RSRP (or probability) and the order or Top K measured beams based on measured RSRP.
[0137] In some embodiments, the QCL assumption of each one of Top K predicted beams and the QCL assumption of each one of Top K measured beams may need to be aligned in a one-to-one manner.
[0138] In some embodiments, the prediction accuracy per monitoring instance may be defined as the ratio of correct (or incorrect) time instances in F predicted time instances for the monitoring instance. A correct (or incorrect) predicted time instance may be the predicted beam is (or is not) the Top 1 measured beams at the time instance, or the RSRP difference between predicted RSRP and measurement RSRP is lower (or higher) than a threshold. In some examples, a weighted ratio may be used that the i-th time instance is with a higher weight than that of j-th time instance if i < j.
[0139] In some examples, a correct (or incorrect) predicted time instance may be defined as the number (or the ratio) of correct (or incorrect) beams in Top K predicted beams in the predicted time instance is higher (or lower) than a threshold.
[0140] In some examples, a correct (or incorrect) predicted time instance may be defined as the order of Top K predicted beams based on predicted RSRP (or probability) is the same (or different) as the order or Top K measured beams based on measured RSRP, or the correlation factor between the order of Top K predicted beams based on predicted RSRP (or probability) and the order or Top K measured beams based on measured RSRP is higher (or lower) than a threshold.
[0141] In some embodiments, for each of F predicted time instances, the prediction accuracy per each predicted instance is defined as the ratio of correct (or incorrect) beams in Top K predicted beams in the predicted time instance. And the prediction accuracy per monitoring instance is the average of prediction accuracy per each predicted instance for all or the F’ out F predicted instances. In other words, the prediction accuracy per monitoring instance is the number of correct beams divided by the number of future time instances (F) times the number of predicted beams (K) .
[0142] In some embodiments, the QCL assumption of each of F predicted time instances in model inference and the QCL assumption of each of F predicted time instances in performance monitoring may need to be aligned in a one-to-one manner.
[0143] As shown in step 350 of FIG. 3, the terminal device 110 may report the result of the performance monitoring to the network device 120. In some embodiments, the terminal device 110 may determine a size of an uplink signaling for the report of the result based on number of QCL assumptions in the set of QCL assumptions, and transmit the result of the performance monitoring based on the size of the uplink signaling.
[0144] In other words, the size of uplink signaling may be varied from the number of QCL assumptions applied. For example, the size of uplink signaling may be an uplink control information (UCI) payload size. It is to be noted that any other suitable sizes may also be feasible.
[0145] In some embodiments, the terminal device 110 may indicate the number of QCL assumptions used for performance monitoring metric calculation. Alternatively, whether to indicate the number of QCL assumptions may be based on NW configuration. For example, the terminal device 110 may receive, from the network device 120, a configuration indicating whether the terminal device 110 reports the number of QCL assumptions. If the configuration indicates that the terminal device 110 reports the number of QCL assumptions, the terminal device 110 may indicate the number of QCL assumptions to the network device 120.
[0146] In some embodiments, default rules of QCL assumption determination may not be applied to some CSI-RS resources, at least for CSI-RS resources for model inference and / or performance monitoring.
[0147] For example, as to ‘qcl-info’ configuration for model inference, two CSI resource configuration IDs are configured for Set A and Set B separately. In some examples, only one CSI resource configuration ID is configured for Set B, and Set A is not configured.
[0148] In some embodiments, the terminal device 110 may need to measure resources in Set B. In some embodiments, at least for an AP CSI-RS, ‘qcl-info’ configuration may be needed. In some embodiments, only Set B is included in a trigger state configuration. In some embodiments, the configuration of the performance monitoring may comprise QCL assumption information associated with a resource setting configured for Set B. For example, a higher layer parameter ‘resourceSet’ points to the resource setting configured for Set B, e.g., with a CSI resource configuration ID for Set B. In some embodiments, several ‘qcl-info’ configuration alternatives may be considered: ‘qcl-info’ configuration is present; ‘qcl-info’ configuration is absent; or ‘qcl-info’ configuration is ignored.
[0149] In some embodiments, the trigger state configuration may follow a QCL assumption determined for Set B during model training. In some examples, the terminal device 110 may maintain QCL assumption for resources in Set B after model training. In some example, the terminal device 110 may report QCL assumption for resources in Set B after model training. In some examples, the terminal device 110 may obtain the QCL assumption by the associated ID and the corresponding model training. In some examples, QCL TypeC may only refer to a synchronization signal and physical broadcast channel block (SSB) .
[0150] For illustration, if only Set B is configured in the AP trigger state, an example procedure may be described as below. - Each trigger state in CSI-AperiodicTriggerStateList contains a list of associated CSI-ReportConfigs indicating the Resource Set IDs for channel measurement. - For each aperiodic CSI-RS resource in a CSI-RS resource set of the first resource setting associated with each CSI triggering state, the UE is indicated the quasi co-location configuration of quasi co-location RS source (s) and quasi co-location type (s) , through higher layer signaling of qcl-info which contains a list of references to TCI-State's for the aperiodic CSI-RS resources associated with the CSI triggering state. If a State referred to in the list is configured with a reference to an RS configured with qcl-Type set to 'typeD' , that RS may be an SS / PBCH block located in the same or different CC / DL BWP or a CSI-RS resource configured as periodic or semi-persistent located in the same or different CC / DL BWP.
[0151] For illustration, an example trigger state configuration may be described as below. CSI-AssociatedReportConfigInfo : : = SEQUENCE { reportConfigId CSI-ReportConfigId, resourcesForChannel CHOICE { nzp-CSI-RS SEQUENCE { resourceSet INTEGER (1.. maxNrofNZP-CSI-RS-ResourceSetsPerConfig) , qcl-info SEQUENCE (SIZE (1.. maxNrofAP-CSI-RS-ResourcesPerSet) ) OF TCI-StateId OPTIONAL --Cond Aperiodic }, csi-SSB-ResourceSet INTEGER (1.. maxNrofCSI-SSB- ResourceSetsPerConfig) }
[0152] For illustration, an example configuration ‘resourceSet’ may be described as below. resourceSet NZP-CSI-RS-ResourceSet for Set B channel measurements. Entry number in nzp-CSI-RS-ResourceSetList in the CSI-ResourceConfig (if two CSI-ResourceConfigIds, it means the CSI-ResourceConfigId for Set B) indicated by resourcesForChannelMeasurement in the CSI-ReportConfig indicated by reportConfigId above (value 1 corresponds to the first entry, value 2 to the second entry, and so on) .
[0153] In some embodiments, the terminal device 110 may not measure resources in Set A if configured. In some embodiments, ‘qcl-info’ configuration may not be necessary. For an AP CSI report, both Set B and Set A may be included in a trigger state configuration. In some embodiments, the configuration of the performance monitoring may comprise first QCL assumption information associated with a resource setting (also referred to as a first resource setting herein) configured for Set B and second QCL assumption information associated with a resource setting (also referred to as a second resource setting herein) configured for Set A. For example, the higher layer parameter ‘resourceSet’ points to the resource setting configured for Set B, e.g., with a CSI resource configuration ID for Set B. The higher layer parameter ‘resourceSet2’ points to the resource setting configured for Set A, e.g., with a CSI resource configuration ID for Set A.
[0154] In some embodiments, several ‘qcl-info’ configuration for Set A alternatives may be considered: ‘qcl-info’ configuration is present; ‘qcl-info’ configuration is absent; or ‘qcl-info’ configuration may be ignored.
[0155] In some embodiments, the trigger state configuration may follow the QCL assumption determined for Set A during model training. In some examples, the terminal device 110 may need to provide QCL assumption for resources in Set A after model training. In some examples, QCL TypeC can only refer to SSB. In some examples, the higher layer parameter ‘resourceSet’ may point to the resource setting configured for Set B, e.g., with a CSI resource configuration ID for Set B.
[0156] For illustration, if both Set B and Set A are configured in the AP trigger state, an example procedure may be described as below. - Each trigger state in CSI-AperiodicTriggerStateList contains a list of associated CSI-ReportConfigs indicating the Resource Set IDs for channel measurement and optionally the Resource Set IDs for prediction. - For each aperiodic CSI-RS resource in a CSI-RS resource set of the first resource setting associated with each CSI triggering state, the UE is indicated the quasi co-location configuration of quasi co-location RS source (s) and quasi co-location type (s) , through higher layer signaling of qcl-info which contains a list of references to TCI-State's for the aperiodic CSI-RS resources associated with the CSI triggering state. If a State referred to in the list is configured with a reference to an RS configured with qcl-Type set to 'typeD' , that RS may be an SS / PBCH block located in the same or different CC / DL BWP or a CSI-RS resource configured as periodic or semi-persistent located in the same or different CC / DL BWP. - For each aperiodic CSI-RS resource in a CSI-RS resource set of the second resource setting associated with each CSI triggering state, the UE is not indicated the quasi co-location configuration of quasi co-location RS source (s) and quasi co-location type (s) .
[0157] For illustration, an example trigger state configuration may be described as below. CSI-AssociatedReportConfigInfo : : = SEQUENCE { reportConfigId CSI-ReportConfigId, resourcesForChannel CHOICE { nzp-CSI-RS SEQUENCE { resourceSet INTEGER (1.. maxNrofNZP-CSI-RS-ResourceSetsPerConfig) , resourceSet2 INTEGER (1.. maxNrofNZP-CSI-RS-ResourceSetsPerConfig) , qcl-info SEQUENCE (SIZE (1.. maxNrofAP-CSI-RS-ResourcesPerSet) ) OF TCI-StateId OPTIONAL --Cond Aperiodic qcl-info2 SEQUENCE (SIZE (1.. maxNrofAP-CSI-RS-ResourcesPerSet) ) OF TCI-StateId OPTIONAL --Cond Aperiodic }, csi-SSB-ResourceSet INTEGER (1.. maxNrofCSI-SSB- ResourceSetsPerConfig) }
[0158] For illustration, an example configuration ‘resourceSet’ may be described as below. resourceSet NZP-CSI-RS-ResourceSet for Set B channel measurements. Entry number in nzp-CSI-RS-ResourceSetList in the CSI-ResourceConfig for Set B indicated by resourcesForChannelMeasurement in the CSI-ReportConfig indicated by reportConfigId above (value 1 corresponds to the first entry, value 2 to the second entry, and so on) . resourceSet2 NZP-CSI-RS-ResourceSet for Set A. Entry number in nzp-CSI-RS- ResourceSetList in the CSI-ResourceConfig for Set A indicated by resourcesForChannelMeasurement in the CSI-ReportConfig indicated by reportConfigId above (value 1 corresponds to the first entry, value 2 to the second entry, and so on) .
[0159] For illustration, an example configuration ‘qcl-info2’ may be described as below. qcl-info2 When this field is absent for aperiodic CSI RS, and applyIndicatedTCI-State or applyIndicatedTCI-State2 is not configured, the UE shall use QCL information included in the "indicated" DL only / Joint TCI state, (except CSI-RS for model inference, CSI-RS for performance monitoring) .
[0160] Conventionally, if a unified TCI state is indicated, the unified TCI state provides a reference signal for QCL for a demodulation reference signal (DM-RS) of PDSCH and DM-RS of PDCCH in a bandwidth part (BWP) / component carrier (CC) , for CSI-RS and provides a RS with qcl-Type set to 'typeD' , if applicable, for determining UL TX spatial filter for dynamic-grant and configured-grant based PUSCH and PUCCH resource in a BWP / CC, and a sounding reference signal (SRS) . However, if a terminal device applies the unified TCI state to the CSI-RS for model inference (or performance monitoring, model training) , since the TCI state is changed very fast, the terminal device may use different QCL assumptions (including Rx beams) for each model inference. This may cause inconsistency issue among each model inference, model training, and / or performance monitoring.
[0161] In view of this, in some embodiments, if a unified TCI state is indicated to be applied to a RS, the terminal device 110 may skip the applying of the unified TCI state to the RS. For example, the unified TCI state cannot apply to the CSI-RS for model inference, for CSI-RS for performance monitoring. In some examples, the unified TCI state also cannot apply to the CSI-RS for model training.
[0162] in some embodiments, if the unified TCI state is indicated to be applied to the RS, the terminal device 110 may drop a report for the result of the performance monitoring.
[0163] in some embodiments, if the unified TCI state is indicated to be applied to the RS, the terminal device 110 may skip a measurement of the RS.
[0164] in some embodiments, if the unified TCI state is indicated to be applied to the RS, the terminal device 110 may skip an update of the report. That is, the terminal device 110 may not be required to update the report if the unified TCI state is indicated to be applied to the RS.
[0165] For illustration, an example procedure may be described as below. - The UE can be configured with a list of up to 128 TCI-State configurations, within the higher layer parameter dl-OrJointTCI-StateList in PDSCH-Config for providing a reference signal for the quasi co-location for DM-RS of PDSCH and DM-RS of PDCCH in a BWP / CC, for CSI-RS (except CSI-RS for model inference, CSI-RS for performance monitoring) , and to provide a reference signal with qcl-Type set to 'typeD' , if applicable, for determining UL TX spatial filter for dynamic-grant and configured-grant based PUSCH and PUCCH resource in a BWP / CC, and SRS.
[0166] For illustration, an example configuration ‘qcl-info’ or ‘qcl-info2’ may be described as below. qcl-info, qcl-info2 When this field is absent for aperiodic CSI RS, and applyIndicatedTCI-State or applyIndicatedTCI-State2 is not configured, the UE shall use QCL information included in the "indicated" DL only / Joint TCI state, (except CSI-RS for model inference, CSI-RS for performance monitoring) .
[0167] For illustration, an example configuration ‘applyIndicatedTCI-State’ or ‘applyIndicatedTCI-State2’ may be described as below. applyIndicatedTCI-State, applyIndicatedTCI-State2 This field indicates, for an aperiodic CSI-RS resource set (perSet) or for CSI- RS resource (perResource) , if UE applies the first or the second "indicated" DL only TCI or joint TCI as specified in TS 38.214
[0019] , clause 5.2.1.5.1. The applyIndicatedTCI-State is for ResourcesForChannel, and applyIndicatedTCI-State2 is for ResourcesForChannels2. When applyIndicatedTCI-State and applyIndicatedTCI-State2 are absent, the UE shall use qcl-info for ResourcesForChannel and use qcl-info2 for ResourcesForChannel2. (except CSI-RS for model inference, CSI-RS for performance monitoring) .
[0168] Conventionally, a default QCL assumption of a CSI-RS may be based on a default control resource set (CORESET) , or a QCL of other DL signal if any on the same OFDM symbol. For example, for an aperiodic CSI-RS scheduled with an offset less than a UE reported threshold ‘beamSwitchTiming’ . But for a CSI-RS for model inference (or performance monitoring, or model training) , the default Rx beam may not be the matched Rx beam, the measurement may cause low performance or even wrong model inference.
[0169] In view of this, in some embodiments, if a QCL assumption of a DL signal or a QCL assumption of a CORESET with a lowest CORESET ID in a latest slot within an active BWP of a cell is to be applied to a RS, the terminal device 110 may skip the applying of the QCL assumption of the DL signal to the RS. For example, QCL assumption of the other DL signal on the same symbol cannot apply to the CSI-RS.
[0170] The other DL signal on the same symbol may refer to a PDSCH scheduled with an offset larger than or equal to the threshold ‘timeDurationForQCL’ ; a periodic CSI-RS, semi-persistent CSI-RS, aperiodic CSI-RS in a non-zero power (NZP) CSI-RS resource set scheduled with offset larger than or equal to a UE reported threshold ‘beamSwitchTiming’ when the reported value is one of the values {14, 28, 48} and when IE ‘enableBeamSwitchTiming’ is not provided or the NZP CSI-RS resource set is configured with a higher layer parameter ‘trs-Info’ ; an aperiodic CSI-RS in a NZP CSI-RS resource set configured with a higher layer parameter ‘repetition’ set to ‘off’ or configured without the higher layer parameters ‘repetition’ and ‘trs-Info’ scheduled with offset larger than or equal to 48 when the UE provides beamSwitchTiming-r16 and enableBeamSwitchTiming is provided; an aperiodic CSI-RS in a NZP CSI-RS resource set configured with the higher layer parameter ‘repetition’ set to ‘on’ scheduled with offset larger than or equal to the UE reported threshold beamSwitchTiming-r16 and enableBeamSwitchTiming is provided.
[0171] In some embodiments, if a QCL assumption of a DL signal or a QCL assumption of a CORESET with a lowest CORESET ID in a latest slot within an active BWP of a cell is to be applied to a RS, the terminal device 110 may skip the applying of the QCL assumption the CORESET to the RS. That is, QCL assumption of the CORESET with the lowest CORESET ID in the latest slot within the active BWP of the cell cannot apply to the CSI-RS.
[0172] In some embodiments, if a QCL assumption of a DL signal or a QCL assumption of a CORESET with a lowest CORESET ID in a latest slot within an active BWP of a cell is to be applied to a RS, the terminal device 110 may drop a report for the result of the performance monitoring.
[0173] In some embodiments, if a QCL assumption of a DL signal or a QCL assumption of a CORESET with a lowest CORESET ID in a latest slot within an active BWP of a cell is to be applied to a RS, the terminal device 110 may skip a measurement of the RS.
[0174] In some embodiments, if a QCL assumption of a DL signal or a QCL assumption of a CORESET with a lowest CORESET ID in a latest slot within an active BWP of a cell is to be applied to a RS, the terminal device 110 may skip an update of the report.
[0175] In some embodiments, the terminal device 110 does not expect that a scheduling offset between a last symbol of a physical downlink control channel (PDCCH) carrying downlink control information (DCI) triggering a report of the result and a first symbol of one or more aperiodic RSs is smaller than a beam switch timing.
[0176] For illustration, an example procedure may be described as below. - If there is any other DL signal with an indicated TCI state in the same symbols as the CSI-RS, UE applies the QCL assumption of the other DL signal also when receiving the aperiodic CSI-RS. (except CSI-RS for model inference, CSI-RS for performance monitoring) .
[0177] For illustration, an example procedure may be described as below. - UE applies the first one of TCI states indicated for the CORESET with the lowest CORESET ID in the latest slot within the active BWP of the cell in which the CSI-RS is to be received when receiving the aperiodic CSI-RS (except a CSI-RS for model inference, a CSI-RS for performance monitoring) , if two TCI states are activated for the CORESET. Otherwise, the UE applies the single activated TCI state of the CORESET with the lowest CORESET ID in the latest slot within the active BWP of the cell in which the CSI-RS is to be received, when receiving the aperiodic CSI-RS (except a CSI-RS for model inference, a CSI-RS for performance monitoring) .
[0178] For illustration, an example procedure may be described as below. - For CSI-RS for model inference, CSI-RS for performance monitoring, the UE does not expect that the scheduling offset between the last symbol of the PDCCH carrying the triggering DCI and the first symbol of the aperiodic CSI-RS resources is smaller than beamSwitchTiming + in CSI-RS symbols, where beamSwitchTiming is UE reported value, the reported value is one of the values of and the beam switching timing delay d is defined if μPDCCH < μCSIRS, else d is zero.
[0179] Conventionally, if a CSI-RS is on the same OFDM symbol with other signals (PDCCH DMRS, or SS / PBCH, etc. ) , UE assumes they are quasi co-located with ‘typeD’ , if ‘typeD’ is applicable. But for CSI-RSs for model inference (or performance monitoring, or model training) , the CSI-RSs are usually considered as different from the beams used to transmit PDCCH DMRS or SSB, it is then difficult to guarantee that they are quasi co-located with 'typeD’ .
[0180] One solution is to not define the same QCL (e.g., quasi co-located with 'typeD’ ) requirement. In some embodiments, if a RS for a measurement and a first DL signal with different QCL assumptions are located on one or more overlapping symbols, the terminal device 110 may skip an applying of QCL assumptions to the first DL signal and the reference signal.
[0181] In some embodiments, if the RS for a measurement and the first DL signal with different QCL assumptions are located on one or more overlapping symbols, the terminal device 110 may drop a report for the result of the performance monitoring.
[0182] In some embodiments, if the RS for the measurement and the first DL signal with different QCL assumptions are located on one or more overlapping symbols, the terminal device 110 may skip the measurement of the RS.
[0183] In some embodiments, if the RS for the measurement and the first DL signal with different QCL assumptions are located on one or more overlapping symbols, the terminal device 110 may skip an update of the report.
[0184] In some embodiments, the first DL signal may be a PDCCH transmission (e.g., DM-RS) or SSB. In some embodiments, QCL assumption of the CORESET cannot apply to the CSI-RS for model inference. In some embodiments, QCL assumption of the SSB cannot apply to the CSI-RS for model inference. In some embodiments, the terminal device 110 may drop the report, or skip the measurement, or be not required to update the report, unless the terminal device 110 indicates the support of simultaneous reception of CSI-RS for layer 1 measurement and PDSCH with different QCL TypeD on overlapping OFDM symbols and simultaneous layer 1 measurement of CSI-RS overlapping with another CSI-RS with different QCL TypeD on overlapping OFDM symbol (s) .
[0185] The other solution is to not allow the same symbol configuration. In some embodiments, resources of the RS for the functionality and the first DL signal are configured to be not located on a same symbol. In some embodiments, the terminal device 110 may transmit, to the network device 120, an indication indicating that the terminal device 110 supports simultaneous reception of the RS and the first DL signal with different QCL assumptions on one or more overlapping symbols.
[0186] For example, PDCCH DMRS and CSI-RS resources for AI / ML cannot be located on the same OFDM symbols. For example, SSB and CSI-RS resources for AI / ML cannot be located on the same OFDM symbols. For example, the terminal device 110 may drop the report, or skip the measurement, or be not required to update the report unless the terminal device 110 indicates the support of simultaneous reception of CSI-RS for layer 1 measurement and PDSCH with different QCL TypeD on overlapping OFDM symbols and simultaneous layer 1 measurement of CSI-RS overlapping with another CSI-RS with different QCL TypeD on overlapping OFDM symbol (s) .
[0187] In some embodiments, the first DL signal is a further RS for the functionality. For example, CSI-RS resources for AI / ML cannot be located on the same OFDM symbols. FIG. 4E illustrates a schematic diagram 400E illustrating an example resource configuration according to some embodiments of the present disclosure. As shown in FIG. 4E, if two CSI-RS resources are with different QCL TypeD, it is not possible for a terminal device to measure L1-RSRP for both of them accurately. In some embodiments, CSI-RS resources may be with the same number of port, same density, e.g., with 1-port and density 3, and same bandwidth.
[0188] In some embodiments, CSI-RS resources for AI / ML cannot be located on the same OFDM symbols, unless the terminal device 110 indicates the support of simultaneous layer 1 measurement of CSI-RS overlapping with another CSI-RS with different QCL Type-D on overlapping OFDM symbol (s) . In some embodiments, the terminal device 110 may also report the maximum different QCL Type-D assumptions for CSI-RS resources on the overlapping OFDM symbols.
[0189] In some embodiments, the terminal device 110 may indicate the CSI-RS resources can be located on overlapping OFDM symbol (s) with same QCL assumptions. In some embodiments, the terminal device 110 may provide such information after model training, to let NW be aware of the which resources can be on the same symbols.
[0190] In some scenarios, if a lot of resources cannot share the same QCL, it may cause that a resource set spans across more than one slot. In this case, a configuration of periodicity and / or offset, or a trigger offset may need to be clear on which slot the configuration is indicated (usually it is the first slot) .
[0191] That is, if CSI-RS resources for AI / ML are located on the same OFDM symbol but with different QCL assumptions and the terminal device 110 cannot support simultaneous reception of different Rx beams, some priority rule of defining which QCL assumptions to be used first may be provided.
[0192] In some embodiments, the terminal device 110 may receive, from the network device 120, a configuration indicating a time offset for a first slot in a set of slots associated with an aperiodic RS resource set. If the aperiodic reference signal resource set is triggered, the terminal device 110 may determine the set of slots based on the time offset and a timing of the triggering of the aperiodic RS resource set.
[0193] For example, for an aperiodic CSI-RS resource set if triggered, and if the associated periodic CSI-RS resource set is configured with more than one slots, a higher layer parameter ‘aperiodicTriggeringOffset’ indicates the triggering offset for the first slot for the first resource in the set.
[0194] In another example, for a periodic or semi-persistent (P / SP) CSI-RS resource, periodicityAndOffset defines the CSI-RS periodicity and slot offset for periodic / semi-persistent CSI-RS. All the CSI-RS resources within one set are configured with the same periodicity, while the slot offset can be same or different for different CSI-RS resources.
[0195] In some embodiments, a time duration between a PDCCH trigger and a corresponding report for AI / ML model inference, or performance monitoring is large. In order to not prevent other scheduling and reporting, the terminal device 110 may be scheduled with another UL transmission or CSI report and transmit in between the PDCCH trigger and the corresponding report. It can be treated as ‘out-of-order’ since the another UL transmission or CSI report is scheduled later but is transmitted earlier, compared to the PDCCH trigger and a corresponding report for AI / ML model inference, or performance monitoring. In some embodiments, the terminal device 110 may provide capability information to indicate that the terminal device 110 can support the above mentioned ‘out-of-order’ scheduling and report during AI / ML model inference, and / or performance monitoring.
[0196] In some embodiments, if a first resource for the functionality with a first QCL assumption and a second resource for the functionality with a second QCL assumption are located on a same symbol and the terminal device 110 does not support simultaneous reception on receiving beams, the terminal device 110 may determine whether the first resource corresponds to a best predicted beam. If the first resource corresponds to a best predicted beam, the terminal device 110 may apply the first QCL assumption firstly for the first set of resources.
[0197] For example, the terminal device 110 may apply QCL of Top 1 / Top K predicted beam first, or, the terminal device 110 may measure a resource corresponding to the Top 1 / Top K predicted beam first, e.g., at least for CSI-RS resources for performance monitoring.
[0198] For example, on one OFDM symbol, a first CSI-RS resource is within the Top K predicted beam, and a second CSI-RS resource is not within the Top K predicted beam. Then the terminal device 110 may apply the QCL assumptions of the first CSI-RS resource.
[0199] In some embodiments, if a first resource for the functionality with a first QCL assumption and a second resource for the functionality with a second QCL assumption are located on a same symbol and the terminal device 110 does not support simultaneous reception on receiving beams, the terminal device 110 may determine whether the first resource and the second resource correspond to a first predicted beam and a second predicted beam in the first number of best predicted beams and the first predicted beam is ranked before the second predicted beam. If the first resource and the second resource correspond to the first predicted beam and the second predicted beam, and the first predicted beam is ranked before the second predicted beam, the terminal device 110 may apply the first QCL assumption firstly for the first set of resources. In other words, an order of applied QCL assumptions is based on a ranking of predicted beams. For example, on one OFDM symbol, a first CSI-RS resource is the predicted Top 1 beam, a second CSI-RS resource is the predicted Top 2 beam, then the terminal device 110 may apply a QCL assumption of the first CSI-RS resource.
[0200] So far, a solution of performance monitoring for AI / ML based beam prediction is described in connection with the process 300. It is to be noted that operations or steps described in the process 300 may be carried out separately or in any suitable combinations.
[0201] For example, the determination of the set of QCL assumptions as described in the step 320 and the determination or measurement of the subset of monitoring resources as described in the step 330 may be implemented separately or jointly.
[0202] For example, the determination of the set of QCL assumptions as described in the step 320 and the determination or measurement of the subset of monitoring resources as described in the step 330 may be implemented in an opposite order. That is, the terminal device 110 may determine a subset of resources in the first set for resources (a subset of resource for performance monitoring, e.g., predicted Top K beams) , and then based on the subset of resources, the terminal device 110 may determine a set of QCL assumptions.
[0203] For example, the determination of the set of QCL assumptions as described in the step 320 and the determination or measurement of the subset of monitoring resources as described in the step 330 may be enabled by NW configuration, reported by UE capability reporting, or UE assistant information (UAI) , pre-defined, or be part of the AI / ML model / functionality.
[0204] For example, the determination of the set of QCL assumptions as described in the step 320 and the determination or measurement of the subset of monitoring resources as described in the step 330 may be valid during a monitoring window.EXAMPLE IMPLEMENTATION OF METHODS
[0205] Corresponding to the above process, embodiments of the present disclosure provide methods of communication implemented at a terminal device and a network device. These methods will be described below with reference to FIGs. 5 and 6.
[0206] FIG. 5 illustrates a flowchart illustrating an example method 500 of communication implemented at a terminal device in accordance with some embodiments of the present disclosure. For example, the method 500 may be performed at the terminal device 110 as shown in FIG. 1A. For the purpose of discussion, in the following, the method 500 will be described with reference to FIG. 1A. It is to be understood that the method 500 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0207] At block 510, the terminal device 110 may receive, from the network device 120, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring.
[0208] At block 520, the terminal device 110 may determine a set of QCL assumptions for a first set of resources. The first set of resources is associated with the performance monitoring.
[0209] In some embodiments, the terminal device 110 may determine the set of QCL assumptions by: determining the set of QCL assumptions based on a configuration of a QCL assumption for a resource in the first set of resources; or determining the set of QCL assumptions based on a QCL assumption for a resource in a second set of resources, the second set of resources being associated with a prediction for the model inference.
[0210] In some embodiments, the terminal device 110 may determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by: aligning a QCL assumption in the set of QCL assumptions to a QCL assumption of a predicted beam during the model inference; or in accordance with a determination that the predicted beam is not transmitted during the model inference or the QCL assumption of the predicted beam is not applied during the model inference, aligning the QCL assumption in the set of QCL assumptions to a QCL assumption of a resource corresponding to the predicted beam during a model training for the functionality.
[0211] In some embodiments, the set of QCL assumptions is associated with a set of future time instances. The terminal device 110 may determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by: aligning a QCL assumption of a future time instance during the performance monitoring to a QCL assumption of the future time instance during the model inference or a model training for the functionality.
[0212] In some embodiments, the terminal device 110 may determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by: at least one of aligning a QCL assumption corresponding to a first measured beam in a first number of best measured beams during the performance monitoring to a QCL assumption of a first predicted beam in the first number of best predicted beams during the model inference, or aligning an order of QCL assumptions in the first number of best measured beams during the performance monitoring to an order of QCL assumptions in the first number of best predicted beams during the model inference; or in accordance with a determination that the first predicted beam is not transmitted during the model inference or the QCL assumption of the first predicted beam is not applied during the model inference, aligning the QCL assumption corresponding to the first measured beam to a QCL assumption of a resource corresponding to the first predicted beam during a model training for the functionality.
[0213] In some embodiments, resources in the first set of resources and resources for a model training for the functionality are configured with same QCL information. In some embodiments, the resources in the first set of resources and resources in a second set of resources associated with a prediction for the model inference are configured with same QCL information.
[0214] In some embodiments, the subset of resources comprises a resource configured with a same QCL assumption as a QCL assumption of a predicted beam in a first number of best predicted beams.
[0215] At block 530, the terminal device 110 may measure, based on the set of QCL assumptions, a subset of resources in the first set of resources.
[0216] In some embodiments, the set of QCL assumptions is associated with a set of future time instances, and the subset of resources comprises one or more resource set associated with the set of future time instances. The terminal device 110 may measure the subset of resources by: measuring a resource set associated with a future time instance based on a QCL assumption associated with the future time instance.
[0217] At block 540, the terminal device 110 may determine a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.
[0218] In some embodiments, the terminal device 110 may determine the result of the performance monitoring by: determining a second number of best measured beams associated with a QCL assumption; determining the second number of best predicted beams associated with the QCL assumption; and determining a first part of the result associated with the QCL assumption based on the second number of best measured beams and best predicted beams.
[0219] In some embodiments, the terminal device 110 may determine the first part of the result by: determining the first part of the result based on whether the second number of best predicted beams matches the second number of best measured beams; or determining the first part of the result based on whether a measurement result difference between the second number of best predicted beams and the second number of best measured beams is lower than or equal to a difference threshold.
[0220] In some embodiments, the terminal device 110 may report, to the network device 120, the result of the performance monitoring by: determining a size of an uplink signaling for the report of the result based on number of QCL assumptions in the set of QCL assumptions; and transmitting the result of the performance monitoring based on the size of the uplink signaling.
[0221] In some embodiments, the configuration of the performance monitoring comprises QCL assumption information associated with a resource setting configured for a third set of resources, the third set of resources being associated with a measurement for the model inference. In some embodiments, the configuration for the performance monitoring comprises first QCL assumption information associated with a first resource setting configured for the third set of resources and second QCL assumption information associated with a second resource setting configured for a second set of resources, the second set of resources being associated with a prediction for the model inference.
[0222] In some embodiments, in accordance with a determination that a unified TCI state is indicated to be applied to a reference signal, the terminal device 110 may perform an operation comprising at least one of the following: skipping the applying of the unified TCI state to the reference signal; dropping a report for the result of the performance monitoring; skipping a measurement of the reference signal; or skipping an update of the report.
[0223] In some embodiments, in accordance with a determination that a QCL assumption of a downlink signal or a QCL assumption of a CORESET with a lowest CORESET identity in a latest slot within an active BWP of a cell is to be applied to a reference signal, the terminal device 110 may perform an operation comprising at least one of the following: skipping the applying of the QCL assumption of the downlink signal or the QCL assumption of the CORESET to the reference signal; dropping a report for the result of the performance monitoring; skipping a measurement of the reference signal; or skipping an update of the report.
[0224] In some embodiments, the terminal device 110 does not expect that a scheduling offset between a last symbol of a PDCCH carrying DCI triggering a report of the result and a first symbol of one or more aperiodic reference signals is smaller than a beam switch timing.
[0225] In some embodiments, in accordance with a determination that a reference signal for a measurement and a first downlink signal with different QCL assumptions are located on one or more overlapping symbols, the terminal device 110 may perform an operation comprising at least one of the following: skipping an applying of QCL assumptions to the first downlink signal and the reference signal; dropping a report for the result of the performance monitoring; skipping a measurement of the reference signal; or skipping an update of the report.
[0226] In some embodiments, resources of a reference signal for the functionality and a first downlink signal are configured to be not located on a same symbol. In some embodiments, the terminal device 110 may transmit, to the network device 120, an indication indicating that the terminal device supports simultaneous reception of the reference signal and the first downlink signal with different QCL assumptions on one or more overlapping symbols.
[0227] In some embodiments, the first downlink signal is a PDCCH transmission or SSB. In some embodiments, the first downlink signal is a further reference signal for the functionality.
[0228] In some embodiments, the terminal device 110 may receive, from the network device 120, a configuration indicating a time offset for a first slot in a set of slots associated with an aperiodic reference signal resource set. In some embodiments, in accordance with a determination that the aperiodic reference signal resource set is triggered, the terminal device 110 may determine the set of slots based on the time offset and a timing of the triggering of the aperiodic reference signal resource set.
[0229] In some embodiments, in accordance with a determination that a first resource for the functionality with a first QCL assumption and a second resource for the functionality with a second QCL assumption are located on a same symbol and the terminal device does not support simultaneous reception on receiving beams, the terminal device 110 may perform an operation comprising at least one of the following: in accordance with a determination that the first resource corresponds to a best predicted beam, applying the first QCL assumption firstly for the first set of resources; or in accordance with a determination that the first resource and the second resource correspond to a first predicted beam and a second predicted beam in a first number of best predicted beams, and the first predicted beam is ranked before the second predicted beam, applying the first QCL assumption firstly for the first set of resources.
[0230] With the method 500, by determining a set of QCL assumptions, which is usually a subset of all Rx beams of a terminal device, a terminal device may not need to sweep all its Rx beams to measure performance monitoring resources, which saves complexity of the terminal device. By determining a subset of resources, overhead of RS transmission and report may be reduced.
[0231] FIG. 6 illustrates a flowchart illustrating an example method 600 of communication implemented at a network device in accordance with some embodiments of the present disclosure. For example, the method 600 may be performed at the network device 120 as shown in FIG. 1A. For the purpose of discussion, in the following, the method 600 will be described with reference to FIG. 1A. It is to be understood that the method 600 may include additional blocks not shown and / or may omit some blocks as shown, and the scope of the present disclosure is not limited in this regard.
[0232] At block 610, the network device 120 may transmit, to the terminal device 110, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring.
[0233] At block 620, the network device 120 may receive, from the terminal device 110, a report for a result of the performance monitoring. The result is based on a measurement for a subset of resources in a first set of resources associated with the performance monitoring, a result of the model inference, and a set of QCL assumptions for the first set of resources.
[0234] In some embodiments, the set of QCL assumptions is associated with a set of future time instances. In some embodiments, the result of the performance monitoring comprises parts of the result associated with QCL assumptions.
[0235] In some embodiments, resources in the first set of resources and resources for a model training for the functionality are configured with same QCL information. In some embodiments, the resources in the first set of resources and resources in a second set of resources associated with a prediction for the model inference are configured with same QCL information.
[0236] In some embodiments, the network device 120 is further caused to at least one of the following: skip transmitting one or more reference signals on one or more resources in the first set of resources other than the subset of resources; or transmit, to the terminal device 110, a configuration indicating a time offset for a first slot in a set of slots associated with an aperiodic reference signal resource set.
[0237] In some embodiments, the network device 120 may receive the report by: determining a size of an uplink signaling for the report based on number of QCL assumptions in the set of QCL assumptions; and receiving the report based on the size of the uplink signaling.
[0238] In some embodiments, the configuration of the performance monitoring comprises a configuration of the set of QCL assumptions. In some embodiments, the configuration of the performance monitoring comprises QCL assumption information associated with a resource setting configured for a third set of resources, the third set of resources being associated with a measurement for the model inference. In some embodiments, the configuration of the performance monitoring comprises first QCL assumption information associated with a first resource setting configured for the third set of resources and second QCL assumption information associated with a second resource setting configured for a second set of resources, the second set of resources being associated with a prediction for the model inference.
[0239] In some embodiments, resources of a reference signal for the functionality and a first downlink signal are configured to be not located on a same symbol. In some embodiments, the network device 120 may receive, from the terminal device 110, an indication indicating that the terminal device supports simultaneous reception of the reference signal and the first downlink signal with different QCL assumptions on one or more overlapping symbols.
[0240] In some embodiments, the first downlink signal is a PDCCH transmission or SSB. In some embodiments, the first downlink signal is a further reference signal for the functionality.
[0241] With the method 600, overhead of RS transmission and report may be reduced.
[0242] It is to be understood that operations of the methods 500 and 600 correspond to that described in connection with FIGs. 1A to 4E, and other details are omitted here for conciseness.EXAMPLE IMPLEMENTATION OF DEVICES
[0243] FIG. 7 is a simplified block diagram of a device 700 that is suitable for implementing embodiments of the present disclosure. The device 700 can be considered as a further example implementation of the terminal device 110 or the network device 120 as shown in FIG. 1A. Accordingly, the device 700 can be implemented at or as at least a part of the terminal device 110 or the network device 120.
[0244] As shown, the device 700 includes a processor 710, a memory 720 coupled to the processor 710, a suitable transceiver 740 coupled to the processor 710, and a communication interface coupled to the transceiver 740. The memory 710 stores at least a part of a program 730. The transceiver 740 may be for bidirectional communications or a unidirectional communication based on requirements. The transceiver 740 may include at least one of a transmitter 742 or a receiver 744. The transmitter 742 and the receiver 744 may be functional modules or physical entities. The transceiver 740 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.
[0245] The program 730 is assumed to include program instructions that, when executed by the associated processor 710, enable the device 700 to operate in accordance with the embodiments of the present disclosure, as discussed herein with reference to FIGs. 1A to 6. The embodiments herein may be implemented by computer software executable by the processor 710 of the device 700, or by hardware, or by a combination of software and hardware. The processor 710 may be configured to implement various embodiments of the present disclosure. Furthermore, a combination of the processor 710 and memory 720 may form processing means 750 adapted to implement various embodiments of the present disclosure.
[0246] The memory 720 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 720 is shown in the device 700, there may be several physically distinct memory modules in the device 700. The processor 710 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 700 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.
[0247] In some embodiments, a device comprises a circuitry configured to perform the method 500 or 600. 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.
[0248] 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.
[0249] 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. 1A to 6. 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.
[0250] 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.
[0251] 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.
[0252] 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.
[0253] 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 terminal device, comprising:a processor configured to cause the terminal device to:receive, from a network device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring;determine a set of quasi co-location (QCL) assumptions for a first set of resources, the first set of resources being associated with the performance monitoring;measure, based on the set of QCL assumptions, a subset of resources in the first set of resources; anddetermine a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.The terminal device of claim 1, wherein the terminal device is caused to determine the set of QCL assumptions by:determining the set of QCL assumptions based on a configuration of a QCL assumption for a resource in the first set of resources; ordetermining the set of QCL assumptions based on a QCL assumption for a resource in a second set of resources, the second set of resources being associated with a prediction for the model inference.The terminal device of claim 2, wherein the terminal device is caused to determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by:aligning a QCL assumption in the set of QCL assumptions to a QCL assumption of a predicted beam during the model inference; orin accordance with a determination that the predicted beam is not transmitted during the model inference or the QCL assumption of the predicted beam is not applied during the model inference, aligning the QCL assumption in the set of QCL assumptions to a QCL assumption of a resource corresponding to the predicted beam during a model training for the functionality.The terminal device of claim 2, wherein the set of QCL assumptions is associated with a set of future time instances, and wherein the terminal device is caused to determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by:aligning a QCL assumption of a future time instance during the performance monitoring to a QCL assumption of the future time instance during the model inference or a model training for the functionality.The terminal device of claim 2, wherein the terminal device is caused to determine the set of QCL assumptions based on the QCL assumption for the resource in the second set of resources by:at least one of aligning a QCL assumption corresponding to a first measured beam in a first number of best measured beams during the performance monitoring to a QCL assumption of a first predicted beam in the first number of best predicted beams during the model inference, or aligning an order of QCL assumptions in the first number of best measured beams during the performance monitoring to an order of QCL assumptions in the first number of best predicted beams during the model inference; orin accordance with a determination that the first predicted beam is not transmitted during the model inference or the QCL assumption of the first predicted beam is not applied during the model inference, aligning the QCL assumption corresponding to the first measured beam to a QCL assumption of a resource corresponding to the first predicted beam during a model training for the functionality.The terminal device of claim 1, wherein resources in the first set of resources and resources for a model training for the functionality are configured with same QCL information; orwherein the resources in the first set of resources and resources in a second set of resources associated with a prediction for the model inference are configured with same QCL information.The terminal device of claim 1, wherein the subset of resources comprises a resource configured with a same QCL assumption as a QCL assumption of a predicted beam in a first number of best predicted beams.The terminal device of claim 1, wherein the set of QCL assumptions is associated with a set of future time instances, and the subset of resources comprises one or more resource set associated with the set of future time instances, and wherein the terminal device is caused to measure the subset of resources by:measuring a resource set associated with a future time instance based on a QCL assumption associated with the future time instance.The terminal device of claim 1, wherein the terminal device is caused to determine the result of the performance monitoring by:determining a second number of best measured beams associated with a QCL assumption;determining the second number of best predicted beams associated with the QCL assumption; anddetermining a first part of the result associated with the QCL assumption based on the second number of best measured beams and best predicted beams.The terminal device of claim 9, wherein the terminal device is caused to determine the first part of the result by:determining the first part of the result based on whether the second number of best predicted beams matches the second number of best measured beams; ordetermining the first part of the result based on whether a measurement result difference between the second number of best predicted beams and the second number of best measured beams is lower than or equal to a difference threshold.The terminal device of claim 1, wherein the terminal device is further caused to:report, to the network device, the result of the performance monitoring by:determining a size of an uplink signaling for the report of the result based on number of QCL assumptions in the set of QCL assumptions; andtransmitting the result of the performance monitoring based on the size of the uplink signaling.The terminal device of claim 1, wherein the configuration of the performance monitoring comprises QCL assumption information associated with a resource setting configured for a third set of resources, the third set of resources being associated with a measurement for the model inference; orwherein the configuration for the performance monitoring comprises first QCL assumption information associated with a first resource setting configured for the third set of resources and second QCL assumption information associated with a second resource setting configured for a second set of resources, the second set of resources being associated with a prediction for the model inference.The terminal device of claim 1, wherein the terminal device is further caused to:in accordance with a determination that a unified transmission configuration indicator (TCI) state is indicated to be applied to a reference signal, perform an operation comprising at least one of the following:skipping the applying of the unified TCI state to the reference signal;dropping a report for the result of the performance monitoring;skipping a measurement of the reference signal; orskipping an update of the report.The terminal device of claim 1, wherein the terminal device is further caused to:in accordance with a determination that a QCL assumption of a downlink signal or a QCL assumption of a control resource set (CORESET) with a lowest CORESET identity in a latest slot within an active bandwidth part (BWP) of a cell is to be applied to a reference signal, perform an operation comprising at least one of the following:skipping the applying of the QCL assumption of the downlink signal or the QCL assumption of the CORESET to the reference signal;dropping a report for the result of the performance monitoring;skipping a measurement of the reference signal; orskipping an update of the report.The terminal device of claim 1, wherein the terminal device does not expect that a scheduling offset between a last symbol of a physical downlink control channel (PDCCH) carrying downlink control information (DCI) triggering a report of the result and a first symbol of one or more aperiodic reference signals is smaller than a beam switch timing.The terminal device of claim 1, wherein the terminal device is further caused to:in accordance with a determination that a reference signal for a measurement and a first downlink signal with different QCL assumptions are located on one or more overlapping symbols, perform an operation comprising at least one of the following:skipping an applying of QCL assumptions to the first downlink signal and the reference signal;dropping a report for the result of the performance monitoring;skipping a measurement of the reference signal; orskipping an update of the report.The terminal device of claim 1, wherein resources of a reference signal for the functionality and a first downlink signal are configured to be not located on a same symbol; orwherein the terminal device is further caused to: transmit, to the network device, an indication indicating that the terminal device supports simultaneous reception of the reference signal and the first downlink signal with different QCL assumptions on one or more overlapping symbols.The terminal device of claim 16 or 17, wherein the first downlink signal is a physical downlink control channel (PDCCH) transmission or synchronization signal and physical broadcast channel block (SSB) ; orwherein the first downlink signal is a further reference signal for the functionality.The terminal device of claim 1, wherein the terminal device is further caused to:receive, from the network device, a configuration indicating a time offset for a first slot in a set of slots associated with an aperiodic reference signal resource set; andin accordance with a determination that the aperiodic reference signal resource set is triggered, determine the set of slots based on the time offset and a timing of the triggering of the aperiodic reference signal resource set.The terminal device of claim 1, wherein the terminal device is further caused to:in accordance with a determination that a first resource for the functionality with a first QCL assumption and a second resource for the functionality with a second QCL assumption are located on a same symbol and the terminal device does not support simultaneous reception on receiving beams, perform an operation comprising at least one of the following:in accordance with a determination that the first resource corresponds to a best predicted beam, applying the first QCL assumption firstly for the first set of resources; orin accordance with a determination that the first resource and the second resource correspond to a first predicted beam and a second predicted beam in a first number of best predicted beams, and the first predicted beam is ranked before the second predicted beam, applying the first QCL assumption firstly for the first set of resources.A network device, comprising:a processor configured to cause the network device to:transmit, to a terminal device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; andreceive, from the terminal device, a report for a result of the performance monitoring, the result being based on a measurement for a subset of resources in a first set of resources associated with the performance monitoring, a result of the model inference, and a set of quasi co-location (QCL) assumptions for the first set of resources.The network device of claim 21, wherein the set of QCL assumptions is associated with a set of future time instances, orwherein the result of the performance monitoring comprises parts of the result associated with QCL assumptions.The network device of claim 21, wherein resources in the first set of resources and resources for a model training for the functionality are configured with same QCL information; orwherein the resources in the first set of resources and resources in a second set of resources associated with a prediction for the model inference are configured with same QCL information.The network device of claim 21, wherein the network device is further caused to at least one of the following:skip transmitting one or more reference signals on one or more resources in the first set of resources other than the subset of resources; ortransmit, to the terminal device, a configuration indicating a time offset for a first slot in a set of slots associated with an aperiodic reference signal resource set.The network device of claim 21, wherein the network device is caused to receive the report by:determining a size of an uplink signaling for the report based on number of QCL assumptions in the set of QCL assumptions; andreceiving the report based on the size of the uplink signaling.The network device of claim 21, wherein the configuration of the performance monitoring comprises a configuration of the set of QCL assumptions; orwherein the configuration of the performance monitoring comprises QCL assumption information associated with a resource setting configured for a third set of resources, the third set of resources being associated with a measurement for the model inference; orwherein the configuration of the performance monitoring comprises first QCL assumption information associated with a first resource setting configured for the third set of resources and second QCL assumption information associated with a second resource setting configured for a second set of resources, the second set of resources being associated with a prediction for the model inference.The network device of claim 21, wherein resources of a reference signal for the functionality and a first downlink signal are configured to be not located on a same symbol; orwherein the network device is further caused to: receive, from the terminal device, an indication indicating that the terminal device supports simultaneous reception of the reference signal and the first downlink signal with different QCL assumptions on one or more overlapping symbols.The network device of claim 27, wherein the first downlink signal is a physical downlink control channel (PDCCH) transmission or synchronization signal and physical broadcast channel block (SSB) ; orwherein the first downlink signal is a further reference signal for the functionality.A method of communication at a terminal device, comprising:receiving, from a network device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring;determining a set of quasi co-location (QCL) assumptions for a first set of resources, the first set of resources being associated with the performance monitoring;measuring, based on the set of QCL assumptions, a subset of resources in the first set of resources; anddetermining a result of the performance monitoring based on a measurement for the subset of resources, a result of the model inference, and the set of QCL assumptions.A method of communication at a network device, comprising:transmitting, to a terminal device, a configuration of a performance monitoring for a functionality and a configuration of a model inference for the functionality associated with the configuration of the performance monitoring; andreceiving, from the terminal device, a report for a result of the performance monitoring, the result being based on a measurement for a subset of resources in a first set of resources associated with the performance monitoring, a result of the model inference, and a set of quasi co-location (QCL) assumptions for the first set of resources.