Communication method, terminal device, network device, and program

Abstract

The embodiments of the present disclosure relate to a method, an apparatus, and a computer-readable medium for communication. According to the embodiments of the present disclosure, a terminal device receives a measurement configuration or a transmission configuration from a network device. The terminal device performs a measurement based on the measurement configuration, and reports a result of the measurement to the network device. Alternatively, if the terminal device performs a transmission based on the transmission configuration, the network device performs a measurement based on the transmission. The network device trains an artificial intelligence (AI) or machine learning (ML) based on the result of the measurement. Thus, the actual measurement result is used to build a dataset suitable for the AI ​​or ML model in the network device.

Description

[Technical field]

[0001] TECHNICAL FIELD Embodiments of the present disclosure relate generally to the field of telecommunications, and more particularly to communication methods, apparatus, and computer-readable media. [Background technology]

[0002] Several techniques have been proposed to improve communication performance. For example, a communication device may adopt an artificial intelligent / machine learning (AI / ML) model to improve communication quality. The AI / ML model can be applied to various scenarios to achieve better performance. Therefore, it is worth studying how to properly train the AI / ML model to ensure satisfactory communication performance. Summary of the Invention [Problem to be solved by the invention]

[0003] Overall, the exemplary embodiments of the present disclosure provide a solution for communications. [Means for solving the problem]

[0004] In a first aspect, a communication method is provided, comprising: receiving, at a terminal device, a first configuration from a network device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model at the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; and transmitting information to the network device regarding a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0005] In a second aspect, a communication method is provided, comprising: sending, in a network device, a first configuration to a terminal device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model in the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; and receiving information from the terminal device regarding a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0006] In a third aspect, a terminal device is provided, the terminal device comprising: a processing unit and a memory coupled to the processing unit storing instructions that, when executed by the processing unit, cause the terminal device to perform operations including receiving from a network device a first configuration indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model in the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference, and transmitting information to the network device regarding a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0007] In a fourth aspect, a network device is provided, the network device comprising: a processing unit and a memory coupled to the processing unit storing instructions, which when executed by the processing unit, cause the network device to perform operations including: sending a first configuration to a terminal device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model in the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; and receiving information from the terminal device regarding a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0008] In a fifth aspect, there is provided a computer readable medium storing instructions which, when executed on at least one processor, cause the at least one processor to perform a method according to the first or second aspect.

[0009] Other features of the present disclosure will become readily apparent from the following description. [Brief description of the drawings]

[0010] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description of several exemplary embodiments of the present disclosure in the accompanying drawings. [Figure 1] FIG. 1 is a schematic diagram of a communication environment in which embodiments of the present disclosure can be implemented. [Figure 2A] FIG. 1 is a schematic diagram of an AI / ML model that can be implemented in a network device. [Figure 2B] A schematic diagram of the dataset constructed for the AI / ML model. [Diagram 3] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 4A] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 4B] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 4C] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 5A] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 5B] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 5C] FIG. 2 illustrates a signaling flow for communication in accordance with some embodiments of the present disclosure. [Figure 6A]FIG. 2 is a schematic diagram of inputs and outputs of an AI / ML model, in accordance with some embodiments of the present disclosure. [Figure 6B] FIG. 2 is a schematic diagram of inputs and outputs of an AI / ML model, in accordance with some embodiments of the present disclosure. [Figure 6C] FIG. 2 is a schematic diagram of inputs and outputs of an AI / ML model, in accordance with some embodiments of the present disclosure. [Figure 6D] FIG. 2 is a schematic diagram of inputs and outputs of an AI / ML model, in accordance with some embodiments of the present disclosure. [Figure 7A] FIG. 1 is a schematic diagram of a data set in accordance with some embodiments of the present disclosure. [Figure 7B] FIG. 1 is a schematic diagram of a data set in accordance with some embodiments of the present disclosure. [Figure 8] 1 is a flowchart of an exemplary method according to an embodiment of the present disclosure. [Figure 9] 1 is a flowchart of an exemplary method according to an embodiment of the present disclosure. [Figure 10] FIG. 1 is a schematic block diagram of an apparatus suitable for implementing embodiments of the present disclosure.

[0011] In the drawings, the same or similar reference numbers represent the same or similar elements. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The principles of the present disclosure will now be described with reference to some exemplary embodiments. It should be understood that these embodiments are provided for illustrative purposes only, to assist those skilled in the art in understanding and implementing the present disclosure, and do not imply any limitations on the scope of the present disclosure. The disclosure described herein can be implemented in various ways different from those described below.

[0013] In the following description and claims, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0014] As used herein, the term "terminal device" refers to any device having wireless or wired communication capabilities. Examples of terminal devices include User Equipment (UE), personal computers, desktops, mobile phones, cellular phones, smartphones, personal digital assistants (PDAs), portable computers, tablets, wearable devices, Internet of Things (IoT) devices, Ultra-Reliable Low Latency Communication (URLLC) devices, Internet of Everything (IoE) devices, machine type communication (MTC) devices, vehicle-mounted devices for V2X communications where X stands for pedestrian, vehicle, or infrastructure / network, devices for integrated access and integrated access and backhaul (IAB), satellite-based or airborne vehicles in a non-terrestrial network (NTN) including High Altitude Platforms (HAPs) including satellites and Unmanned Aircraft Systems (UASs), extended reality (XR) devices including different types of reality such as augmented reality (AR), mixed reality (MR), virtual reality (VR), unmanned aerial vehicles (UAVs), which are aircraft without a human pilot and commonly referred to as drones. "Terminal Device" includes, but is not limited to, a vehicle (V2X), a device on a high speed train (HST), or an image capture device such as a digital camera, a sensor game device, a music storage and playback device, or an Internet appliance that allows wireless or wired Internet access and browsing, etc. The "Terminal Device" may further have "multicast / broadcast" capabilities to support V2X applications, transparent IPv4 / IPv6 multicast distribution, IPTV, smart TV, wireless services, over-the-air software distribution, group communication, and IoT applications that are public safety and mission critical. It may also incorporate one or more Subscriber Identity Modules (SIMs), known as multi-SIMs.The term "terminal" may be used interchangeably with UE, mobile station, subscriber station, mobile terminal, user terminal, or wireless device. In the following description, the terms "terminal", "communications device", "terminal", "user equipment" and "UE" may be used interchangeably.

[0015] The terminal device or network device may have the capability of artificial intelligence (AI) or machine learning. Generally, it includes a model trained from a large amount of data collected for a specific function and can be used to predict some information.

[0016] The terminal device or network device may operate on several frequency ranges, such as FR1 (410 MHz-7125 MHz), FR2 (24.25 GHz-71 GHz), frequency bands greater than 100 GHz, and Tera Hertz (THz). It can also operate on licensed / unlicensed / shared spectrum. The terminal device may have two or more connections with the network device under Multi-Radio Dual Connectivity (MR-DC) application scenarios. The terminal device or network device can operate in full duplex, flexible duplex, and cross division duplex modes.

[0017] The term "network device" refers to a device capable of providing or hosting a cell or coverage in which a terminal device can communicate. Examples of network devices include, but are not limited to, a Node B (NodeB or NB), an evolved Node B (eNodeB or eNB), a next generation Node B (gNB), a transmit / receive point (TRP), a remote radio unit (RRU), a radio head (RH), a remote radio head (RRH), an IAB node, a femto node, a pico node, a low power node such as a reconfigurable intelligent surface (RIS), and a network management entity such as an Operations, Administration and Maintenance (OAM) entity.

[0018] In one embodiment, the terminal device can connect to 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 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 an eNB and the second RAT device is a gNB. Information about the different RATs may be transmitted from at least one of the first network device and the second network device to the terminal device. In one embodiment, the first information may be transmitted from the first network device to the terminal device, and the second information may be transmitted from the second network device to the terminal device directly or via the first network device. In one embodiment, information about the configuration of the terminal device configured by the second network device may be transmitted from the second network device via the first network device. Information regarding the reconfiguration of the terminal device configured by the second network device may be transmitted from the second network device directly to the terminal device or via the first network device.

[0019] The communications described herein may conform to any suitable standard, including, but not limited to, New Radio Access (NR), Long Term Evolution (LTE), LTE-Evolution, LTE-Advanced (LTE-A), Wideband Code Division Multiple Access (WCDMA), Code Division Multiple Access (CDMA), cdma2000, Global System for Mobile Communications (GSM), and the like. Furthermore, the communications may be performed according to any generation of communication protocols now known or developed in the future. Examples of communication protocols include, but are not limited to, first generation (1G), second generation (2G), 2.5G, 2.85G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), and sixth generation (6G) communication protocols. The techniques described herein may be used for the wireless networks and radio technologies mentioned above, as well as other wireless networks and radio technologies. The embodiments of the present disclosure may be performed according to any generation of communication protocols now known or developed in the future. Examples of communication protocols include, but are not limited to, first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G) communication protocols, 5.5G, 5G-Advanced networks, or sixth generation (6G) networks.

[0020] The term "circuitry" as used herein may refer to a hardware circuit and / or a combination of a hardware circuit and software. For example, a circuit may be a combination of analog and / or digital hardware circuitry and software / firmware. As yet another example, a circuit may be any portion of a hardware processor having software, including a digital signal processor, software, and one or more memories, that cooperate to cause a device, such as a terminal device or a network device, to perform various functions. In yet another example, a circuit may be a hardware circuit and / or a processor, such as a microprocessor or a portion thereof, that requires software / firmware for operation, but the software may not be present if not required for operation. As used herein, the term "circuitry" also includes an implementation of only a hardware circuit or one or more processors, or a portion of a hardware circuit or one or more processors and its (or their) associated software and / or firmware.

[0021] As used herein, the singular forms "a," "an," and "said" include the plural forms unless the context clearly indicates otherwise. The term "comprises" and variations thereof should be understood as open-ended terms meaning "including, but not limited to." The term "based on" should be understood as "based at least in part on." The terms "one embodiment" and "embodiment" should be understood as "at least one embodiment." The term "another embodiment" should be understood as "at least one other embodiment." Terms such as "first," "second," and the like may refer to different or the same object. The following may include other explicit and implicit definitions.

[0022] In some instances, values, procedures, or devices are referred to as "best," "lowest," "highest," "minimum," "maximum," etc. It should be understood that such descriptions are intended to illustrate that selections can be made from among many functional alternatives used, and that such selections are not necessarily better, smaller, higher, or otherwise more preferred than other selections.

[0023] As mentioned above, the AI / ML model can be applied to various scenarios to achieve better performance. In some embodiments, the AI / ML model may be implemented on the network device side. Alternatively, the AI / ML model may be implemented on the terminal device side. In other embodiments, the AI / ML model may be implemented on both the network device side and the terminal device side.

[0024] For example, the terminal device can perform beam management on an AI / ML model. In this case, the terminal device can measure a portion of candidate beam pairs and estimate the quality for all candidate beam pairs using AI or ML. Massive MIMO (mMIMO) and beamforming are widely used in the communications industry. The terms "beamforming" and "mMIMO" may be used interchangeably. Generally speaking, beamforming uses multiple antennas to control the direction of a wavefront by appropriately weighting the amplitude and phase of individual antenna signals in an array of multiple antennas. The most commonly seen definition is that mMIMO is a system in which the number of antennas exceeds the number of users. Coverage is beam-based, not cell-based, in 5G. There is no cell-level reference channel on which the coverage of a cell can be measured. Instead, each cell has one or more Synchronization Signal Block (SSB) beams. SSB beams are static or semi-static and always point in the same direction. They form a grid of beams that cover the entire cell area. A user equipment (UE) searches for and measures beams and maintains a set of candidate beams, which may include beams from multiple cells. With 5G millimeter wave (mmWave) enabling directional communication with more antenna elements and providing additional beamforming gain, efficient beam management becomes important where the UE and gNB periodically identify the best beam to work with at any given time.

[0025] Additionally, the terminal device may perform CSI feedback based on the AI / ML model. In this case, the original CSI information may be compressed by an AI encoder located in the terminal device and restored by an AI decoder located in the network device. The AI / ML model may also be used for reference signal (RS) overhead reduction. For example, the terminal device may use a new RS pattern, e.g., less dense DMRS, less CSI-RS ports.

[0026] In many studies, 3GPP channel models (statistical models) are used to generate all samples. The samples in the training / validation / test datasets are naturally independent and identically distributed. The AI / ML models may be trained and deployed by the network alone, without real measurements. However, in the field, the actual channel characteristics and real realizations are different from the statistical models. The statistical models may not be suitable for the real environment. Therefore, it is worth proposing new solutions on constructing a suitable dataset for the AI / ML models.

[0027] To solve at least some of the above problems and other potential problems, a solution is proposed for improving the AI / ML model. According to an embodiment of the present disclosure, a terminal device receives a measurement configuration or a transmission configuration from a network device. The terminal device performs a measurement based on the measurement configuration and reports a result of the measurement to the network device. Alternatively, if the terminal device performs a transmission based on the transmission configuration, the network device performs a measurement based on the transmission. The network device trains an AI / ML model based on the result of the measurement. Thus, the actual measurement result is used to build a dataset suitable for the AI / ML model in the network device.

[0028] 1 is a schematic diagram of a communication system in which embodiments of the present disclosure may be implemented. The communication system 100, which is part of a communication network, comprises terminal devices 110-1, 110-2, ..., 110-N, which may be collectively referred to as "terminal devices 110." The number N may be any suitable integer. The terminal devices 110 may communicate with each other.

[0029] The communication system 100 further includes a network device. In the communication system 100, the network device 120 and the terminal device 110 can communicate data and control information with each other. The number of terminal devices shown in FIG. 1 is shown for illustrative purposes only and does not imply any limitation.

[0030] Communications in the communications system 100 may be achieved according to any suitable communications protocol, including, but not limited to, cellular communications protocols, such as first generation (1G), second generation (2G), third generation (3G), fourth generation (4G), and fifth generation (5G), wireless local area network communications protocols, such as Institute for Electrical and Electronics Engineers (IEEE) 802.11, and / or any other protocols now known or developed in the future. Furthermore, communications may utilize any suitable wireless communications technology, including, but not limited to, code division multiple access (CDMA), frequency division multiple access (FDMA), time division multiple access (TDMA), frequency division duplexer (FDD), time division duplexer (TDD), multiple input multiple output (MIMO), orthogonal frequency division multiple access (OFDMA), and / or any other technology now known or developed in the future.

[0031] The embodiments of the present disclosure may be applied to any suitable scenario. For example, the embodiments of the present disclosure may be implemented on a reduced capability NR device. Alternatively, the embodiments of the present disclosure may be implemented within one of NR multiple input multiple output (MIMO), NR sidelink enhancements, NR systems at frequencies higher than 52.6 GHz, enhanced NR operation up to 71 GHz, Narrowband Internet of Things (NB-IOT) / enhanced machine type communications (eMTC) over non-terrestrial based networks (NTN), NTN, UE power saving enhancements, NR coverage enhancements, NB-IOT and LTE-MTC, integrated access and backhaul (IAB), NR multicast and broadcast services, or multi-radio dual connectivity enhancements.

[0032] The term "slot" as used herein means a dynamic scheduling unit. A slot includes a predetermined number of symbols. The term "downlink (DL) subslot" may refer to a virtual subslot built on an uplink (UL) subslot. A DL subslot may include fewer symbols than one DL slot. A slot as used herein may refer to a normal slot including a predetermined number of symbols and a subslot including fewer symbols than the predetermined number of symbols. The terms "beam", "beam pair", and "beam pair link" as used herein refer to a communication link and may be represented by "resource", "resource set", or "resource setting". The term "beam quality" as used herein refers to "RSRP", "SINR", "RSSI", or "RSRQ" measured on a corresponding resource or measured via a corresponding beam. The terms "precoder", "precoding", "precoding matrix", "beam", "spatial relationship information", "spatial domain transmit filter", "spatial domain filter", "spatial parameters", "spatial relationship information", "spatial relationship info", "TPMI", "precoding information", "precoding information and number of layers", "precoding matrix indicator (PMI)", "precoding matrix indicator", "transmit precoding matrix indication", "precoding matrix indication", "TCI state", "transmit configuration indicator", "quasi co-location (QCL)", "quasi-co-location", "QCL parameters" and "spatial relationship" may be used interchangeably. The terms "SRI", "SRS resource set index", "UL TCI", "UL spatial domain filter", "UL beam", "combined TCI" may be used interchangeably. The terms "candidate", "predicted", "target" and "potential" may be used interchangeably.

[0033] FIG. 2A shows a schematic diagram of an AI / ML model. The AI / ML model 200 may be implemented on the network device 120 side. As used herein, an "AI / ML model" may refer to a program or algorithm that utilizes a set of data that allows it to recognize certain patterns, thereby allowing it to reach a conclusion or make a prediction when provided with sufficient information. Generally speaking, an AI / ML model may be a mathematical algorithm that is "trained" with data and the input of a human expert to reproduce the decision that the expert would make when provided with the same information.

[0034] According to some embodiments of the present disclosure, the AI / ML model 200 may be used to learn how to generate an output based on an input for beam management. For example, in this case, the input of the AI / ML model may include at least a subset K1 of beams and possibly their RSRPs, and the output may include at least a restored set K of beams (K>K1) and their RSRPs, or a best beam and its RSRP. In some embodiments, time domain measurements are considered. In this case, the input of the AI / ML model may be at time nT, and the output of the AI / ML model may be at time n. Alternatively, according to some embodiments of the present disclosure, the AI / ML model 200 may learn how to find a narrow beam based on wide beam information. In this case, the input of the AI / ML model may include at least a set K′ of wide beams and possibly their RSRPs, and the output of the AI / ML model may include at least a restored set K of narrow beams and their RSRPs, or an Alt2 best beam in K and possibly its RSRP. The wide beam may be the beam used in initial access, and the narrow beam may be the beam used for UE-specific data transmission. In some embodiments, different AI / ML models may be used to select different subsets K1.

[0035] As shown in FIG. 2A, the AI / ML model 200 may include data collection 210, model training 220, model inference 230, and actors 240. Data collection 210 is a function that provides input data to model training 220 and model inference 230. AI / ML algorithm-specific data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) is not performed in data collection 210. Examples of input data may include measurements from UEs or different network entities, feedback from actors 240, and output from the AI / ML model. As used herein, the term "training data" may refer to data required as input for model training 220. As used herein, the term "inference data" may refer to data required as input for model inference 230.

[0036] Model training 220 is a function that performs ML model training, validation, and testing, which may generate model performance metrics as part of the model testing procedure. Model training 220 is also responsible for data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) based on the training data delivered by data collection 210, if necessary. As shown in FIG. 2B, there may be three datasets for model training 220: a training dataset, a validation dataset, and a test dataset. Model training 220 may use model deployment / update to initially deploy the trained, validated, and tested AI / ML model to model inference 230 or to deliver an updated model to model inference 230. The classification of different datasets follows a common practice in machine learning. The training dataset may be used to directly improve the parameters of the model. The validation dataset may be used to evaluate the performance of the model while optimizing the hyperparameters of the model. The test dataset may be used to evaluate the model after the hyperparameter optimization is completed.

[0037] Model inference 230 is a function that provides the AI / ML model inference output (e.g., prediction or decision). It may provide model performance feedback to model training 220. Model inference 230 is also responsible for data preparation (e.g., data pre-processing and cleaning, formatting and transformation) based on inference data delivered by data collection 210, as needed.

[0038] An actor 210 is a function that receives output from model reasoning 230 and triggers or executes a corresponding action. An actor 210 may trigger actions on other entities or on itself.

[0039] First, reference is made to FIG. 3. FIG. 3 illustrates a signaling diagram illustrating a process 300 between a terminal device and a network device according to some exemplary embodiments of the present disclosure. For illustrative purposes only, the process 300 is described with reference to FIG. 1. The process 300 may involve the terminal device 110-1 and the network device 120 of FIG. 1. The process 300 may be used to collect field data for AI / ML model training. Collecting field data for AI / ML model training may be aperiodic or semi-persistent configuration / measurement / reporting and may have lower priority than data traffic transmission and other regular configuration / measurement / reporting. Collecting field data for model training may be done within a time window / interval or duty cycle that the terminal device proposes or requests based on its capabilities.

[0040] The network device 120 transmits (3010) a first configuration to the terminal device 110-1. The first configuration indicates at least one resource subset from a first set of resources. In some embodiments, the first set of resources may include resources for beam measurement and reporting. The at least one resource subset is for constructing a first dataset for training the AI / ML model 200 in the network device 120. The first dataset includes one or more of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference.

[0041] In some embodiments, the first configuration may be a first measurement configuration. In this case, the terminal device 110-1 may perform measurements on the at least one resource subset based on the first measurement configuration. For example, the first configuration may indicate a first candidate data sample for constructing the first data set. The first candidate data sample may include a resource subset as an input of an AI / ML model and a target resource as an output of the AI / ML model. Note that the input and / or output of the AI / ML model may also include other parameters not limited to the resource subset and the target resource. Alternatively, the first configuration may indicate the terminal device 110-1 to perform measurements on a first set of resources. In some embodiments, the terminal device 110-1 may determine a quality based on the result of the measurement. For example, the terminal device 110-1 may determine a reference signal received power (RSRP) based on the measurement. Alternatively, the terminal device 110-1 may determine a signal-to-interference-and-noise ratio (SINR) based on the measurement.

[0042] In another embodiment, the first configuration may be a first transmission configuration. In this case, the terminal device may perform transmission based on the first transmission configuration. For example, the first configuration may indicate a set of reference signal resources for a sounding reference signal (SRS). Alternatively or additionally, the first configuration may indicate a reference signal resource subset from the set of reference signal resources. In another embodiment, the first configuration may indicate a target reference signal resource.

[0043] The terminal device 110-1 transmits (3020) the information to the network device 120. In some embodiments, if the first configuration is the first measurement configuration, the terminal device 110-1 may transmit a result of the measurement to the network device 120. Alternatively, the first configuration is the first transmission configuration, and the terminal device 110-1 may transmit the sounding reference signal to the network 120. In this case, the network device 120 may measure the sounding reference signal.

[0044] The network device 120 trains (3030) an AI / ML model. The network device 120 may construct a first data set based on the results of the measurements. The AI / ML model may be trained based on the first data set. In this manner, the AI / ML model may be optimized based on the actual measurement results.

[0045] The network device 120 may transmit an indication of the updated resource subset to the terminal device 110-1 (3040). As shown in FIG. 6D, the updated resource subset may be an output from the AI / ML model. In some embodiments, the updated resource subset may be an input for AI / ML model training. Alternatively or additionally, the updated resource subset may be used to collect data for AI / ML model training. In other embodiments, the updated resource subset may be used to collect data for AI / ML model inference. In some other embodiments, the updated resource subset may be used for regular beam measurement and reporting. In this way, randomly selected measurement subsets may be avoided, which may provide better prediction performance.

[0046] In some embodiments, explicit signaling may be required to inform the terminal device 110-1 about the updated measurement subset K1′. For example, the input of the AI / ML model training may be updated based on the updated resource subset. In other embodiments, the updated resource subset may be transmitted within a first measurement configuration, which means that the configuration for collecting data for the AI / ML model training may be updated. In this case, the RRC IEs “RS-subset-to-measure-for-training” and “RS-to-compare-for-training” may be used to inform the updated resource subset. Alternatively, the updated resource subset may be transmitted within a second measurement configuration, which means that the configuration for collecting data for the AI / ML model inference may be updated. In this case, a new signaling “RS-subset-measure-for-inference” may be used to inform the updated resource subset. As mentioned above, different AI / ML models may be used to select different subsets K1. In other words, multiple AI / ML models (model i) can learn how to generate K from different versions of K1_i. In this case, instead of using K1' as output, the AI / ML model ID i may be used as output, where one version of the K1 choices corresponds to one AI / ML model.

[0047] In some embodiments, inputs from the terminal device 110-1 for the first AI / ML mode may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs from the network device 120 for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic condition, transmission scheme, wide beam used in initial access. Inputs from other terminal devices for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs from other network devices 120 for the AI / ML model may include one or more of: number of Tx / Rx beams / panels, beamforming gain, beam width, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic conditions, transmission scheme, wide beam used in initial access. Inputs from other AI / ML models for the AI / ML model may include one or more of: UE location prediction, UE trajectory prediction, handover prediction, initial access prediction, channel state information (CSI) prediction.

[0048] In some embodiments, the output from the AI / ML model for the terminal device 110-1 may include one or more of transmit power, beam switching decisions, active subset of Tx / Rx beams, transmission scheme. The output from the AI / ML model for the network device 120 may include one or more of transmit power, handover, active subset of Tx / Rx beams, scheduling, beam switching decisions, multi-user scheduling decisions, traffic conditions, transmission scheme. The output from the AI / ML model may be used as input to other AI / ML models for predicting UE position / trajectory, predicting handover / load balancing / energy saving decisions, predicting CSI.

[0049] In some embodiments, when AI / ML model inference is applied, an additional parameter regarding the number of reduced UE Rx beams may be used to determine the number of configured RS resources in the RS resource set, the measurement period requirement, and the measurement accuracy requirement. In some embodiments, the parameter may be indicated by the network device 120. Alternatively, the parameter may be reported by the terminal device 110-1. Alternatively, the parameter may be reported by the terminal device 110-1 in a capability report. In other embodiments, the parameter may also be an output from the AI / ML model. In this way, it is possible to guide the configuration and set performance requirements when AI / ML model inference is applied. Tables 1-3 show the updated descriptions based on the reduced number of receive beams. [Table 1] [Table 2] [Table 3]

[0050] In some other embodiments, the number of occupied CSI processing units (CPUs) and the CSI computation time may depend on the number of resources in the first set of resources that are actually measured for beam quality. For example, the number of occupied CPUs may be NCPU (i.e., the UE capability on the maximum supported number of simultaneous CSI computations) or K1+1 (i.e., the number of actually measured resources in the resource set including the target resource). Alternatively or additionally, for the time span from the CSI reference to the last symbol carrying an RS in K1, the number of occupied CPUs may be NCPU or K1 (i.e., the number of actually measured resources in the resource set). The number of occupied CPUs may be:

number

[0051] An exemplary embodiment of the dataset construction of the present disclosure will be described with reference to FIGS. 4A-5C.

[0052] FIG. 4A is a signaling chart illustrating a process between terminal device 110-1 and network device 120 according to some example embodiments of the present disclosure, where the first configuration is a measurement configuration.

[0053] The network device 120 may transmit a first measurement configuration to the terminal device 110-1 (4010). The first measurement configuration may indicate resources for measurement. In some embodiments, the first measurement configuration may indicate an index of the resource for measurement. In this case, for example, the first measurement configuration may include a synchronization signal / physical broadcast channel block index. Alternatively, the first measurement configuration may include a channel state information (CSI) reference signal (RS) resource index. The first measurement configuration may indicate the terminal device 110-1 to perform measurements on all resources. In other words, the network device 120 may obtain information about a complete set of beam pairs and beam quality.

[0054] In some embodiments, the first measurement configuration may include a report configuration for model training (ReportConfigForModelTraining). Additionally, the first measurement configuration may include a resource configuration for model training (ResourceConfigForModelTraining).

[0055] The terminal device 110-1 may perform a first measurement based on the first measurement configuration (4020). The terminal device 110-1 may perform a first measurement on the resource. The terminal device 110-1 may determine a quality on the resource based on a result of the first measurement. For example, the terminal device 110-1 may determine an RSRP on the resource. Alternatively, the terminal device 110-1 may determine an SINR on the resource.

[0056] The terminal device 110-1 may transmit a first measurement report to the network device 120 (4030). The first measurement report may include indexes of resources and their qualities. For example, the first measurement report may include K beam pairs and their qualities, such as K SSBRI / CRI+K L1-RSRP. The number of resources in the first measurement report may be based on the number of beams in the network device 120 and the number of beams in the terminal device 110-1. FIG. 6A shows one data sample reported by the terminal device 110-1 based on the first measurement result. As shown in FIG. 6A, the first measurement report includes a complete set of beam pairs and their RSRPs. The number of rows in FIG. 6A corresponds to the number of UE receiving beams. The number of columns in FIG. 6A corresponds to the number of network transmitting beams. Each element in FIG. 6A corresponds to one resource. As shown in FIG. 7A, the measurement results in the first measurement report (i.e., the reported qualities of the resources) may be part of the samples in the test data set. Similarly, the measurements may also be made part of the samples in the training or validation data sets.

[0057] In some embodiments, the terminal device 110-1 may not report the index of the resource. In this case, in the first measurement report, the RSRP of the resource may be sorted in ascending / descending order of the index of the resource. Alternatively or additionally, the terminal device 110-1 may not report resources having poor quality. For example, if one resource has an RSRP lower than a threshold RSRP, the terminal device 110-1 may not report the index of such resource in the first measurement report. In other embodiments, the terminal device 110-1 may not report the receive beam information. For example, the terminal device 110-1 may report only information about the transmit beam, assuming that each transmit beam is measured through its optimal UE receive beam. In some embodiments, the AI / ML model is limited per wide beam pair identified in the initial access phase. In some other embodiments, the network device 120 may transmit a reference quality of the resource to the terminal device 110-1. In this case, the terminal device 110-1 may transmit a first measurement report including a differential quality of the first set of resources relative to the reference quality. For example, the network device 120 may transmit a map of (RS ID+RSRP) to the terminal device 110-1. In this case, the terminal device 110-1 may feed back a difference map for the version of the NW. More specifically, the network device 120 may transmit a map of (RS ID+RSRP) to the terminal device 110-1. K ), and the terminal device 110-1 measures RS and calculates the difference for IDs 1 to K, thereby

number

[0058] The network device 120 may train an AI / ML model based on the results of the measurements (4040). In some embodiments, the output of the AI / ML model may recover all resources and all RSRPs. Alternatively, the output of the AI / ML model may be the best resource / beam and the RSRP of the best resource. With reference to FIG. 6B, the input of the AI / ML model may be the resource subset K1 and the corresponding RSRPs, and the output of the AI / ML model may recover all resources / beams and their RSRPs. With reference to FIG. 6C, the input of the AI / ML model may be the resource subset K1 and the corresponding RSRPs, and the output of the AI / ML model may be the best beam in set K that maximizes the RSRP.

number

[0059] In some embodiments, the network device 120 may transmit (4050) a second measurement configuration to the terminal device 110-1. The second measurement configuration may indicate a resource for the second measurement. In some embodiments, the second measurement configuration may indicate an index of the resource for the second measurement. In this case, for example, the second measurement configuration may include an SSB index. Alternatively, the second measurement configuration may include a CSI-RS resource index.

[0060] The terminal device 110-1 may perform a second measurement based on the second measurement configuration (4060). The terminal device 110-1 may perform a second measurement on the resource. The terminal device 110-1 may determine a quality on the resource based on a result of the second measurement. For example, the terminal device 110-1 may determine an RSRP on the resource. Alternatively, the terminal device 110-1 may determine an SINR on the resource.

[0061] The terminal device 110-1 may transmit a second measurement report to the network device 120 (4070). The second measurement report may include indexes of the resources and their qualities. The second measurement report may be used as inference data for real-time UE beam management.

[0062] The network device 120 may infer using the AI / ML model based on the second measurement report (4080). In other words, the network device 120 may perform model inference according to the AI / ML model. For example, the AI / ML model may output a prediction including, for example, a complete set of beam pairs and beam qualities, a best beam for subsequent transmission, and an updated subset for the second measurement.

[0063] FIG. 4B is a signaling chart illustrating a process between terminal device 110-1 and network device 120 in accordance with some example embodiments of the present disclosure in which the first configuration is a transmit configuration.

[0064] The network device 120 may transmit a first transmission configuration to the terminal device 110-1 (4011). The first transmission configuration may indicate resources for transmitting the sounding reference signal. In some embodiments, the first transmission configuration may indicate an index of a resource for transmitting the sounding reference signal. The first transmission configuration may indicate to the terminal device 110-1 to perform transmission on all resources. In some embodiments, the first transmission configuration may include an information element "SRS resource set usage" indicating that the resources are used to train an AI / ML model.

[0065] The terminal device 110-1 may perform a first measurement based on the first transmission configuration (4021). The terminal device 110-1 may transmit a sounding reference signal on the resource based on the first transmission configuration, thus not placing an extra burden on the terminal device 110-1.

[0066] The network device 120 may perform a first measurement based on the sounding reference signal (4031). The network device 120 may determine a quality on the received sounding reference signal. For example, the network device 120 may determine an RSRP of the received sounding reference signal. Alternatively, the network device 120 may determine a SINR of the received sounding reference signal. As shown in FIG. 7A, the measurement results (i.e., the measured quality of the resource) in the first measurement report may be made part of the samples in the test data set. Similarly, the measurement results may also be made part of the samples in the training data set or the validation data set.

[0067] The network device 120 may train an AI / ML model based on the results of the first measurement (4041). In some embodiments, the output of the AI / ML model may recover all resources and all RSRPs. Alternatively, the output of the AI / ML model may be the best resource / beam and the RSRP of the best resource. Thus, the AI / ML model may be optimized based on the actual measurement results.

[0068] In some embodiments, the network device 120 may transmit a second measurement configuration to the terminal device 110-1 (4051). The second measurement configuration may indicate a resource for the second measurement. In some embodiments, the second measurement configuration may indicate an index of the resource for the second measurement. In this case, for example, the second measurement configuration may include an SSB index. Alternatively, the second measurement configuration may include a CSI-RS resource index.

[0069] The terminal device 110-1 may perform a second measurement based on the second measurement configuration (4061). The terminal device 110-1 may perform a second measurement on the resource. The terminal device 110-1 may determine a quality on the resource based on a result of the second measurement. For example, the terminal device 110-1 may determine an RSRP on the resource. Alternatively, the terminal device 110-1 may determine an SINR on the resource.

[0070] The terminal device 110-1 may transmit a second measurement report to the network device 120 (4071). The second measurement report may include indexes of resources and their qualities. The second measurement report may be used as inference data for real-time UE beam management.

[0071] The network device 120 may infer using the AI / ML model based on the second measurement report (4081). In other words, the network device 120 may perform model inference according to the AI / ML model. For example, the AI / ML model may output a prediction including, for example, a complete set of beam pairs and beam qualities, a best beam for subsequent transmission, and a subset for the second measurement.

[0072] FIG. 4C is a signaling chart illustrating a process between terminal device 110-1 and network device 120 in accordance with some example embodiments of the present disclosure, where the first configuration is a transmit configuration.

[0073] The network device 120 may transmit a first transmission configuration to the terminal device 110-1 (4012). The first transmission configuration may indicate resources for transmitting the sounding reference signal. In some embodiments, the first transmission configuration may indicate an index of a resource for transmitting the sounding reference signal. The first transmission configuration may indicate to the terminal device 110-1 to perform transmission on all resources. In some embodiments, the first transmission configuration may include an information element "SRS resource set usage" indicating that the resources are used to train an AI / ML model.

[0074] The terminal device 110-1 may perform a first measurement based on the first transmission configuration (4022). The terminal device 110-1 may transmit a sounding reference signal on the resource based on the first transmission configuration, thus not placing an extra burden on the terminal device 110-1.

[0075] The network device 120 may perform a first measurement based on the sounding reference signal (4032). The network device 120 may determine a quality on the received sounding reference signal. For example, the network device 120 may determine an RSRP of the received sounding reference signal. Alternatively, the network device 120 may determine a SINR of the received sounding reference signal. As shown in FIG. 7A, the measurement results (i.e., the measured quality of the resource) in the first measurement report may be made part of the samples in the test data set. Similarly, the measurement results may also be made part of the samples in the training data set or the validation data set.

[0076] The network device 120 may train an AI / ML model based on the results of the first measurement (4042). In some embodiments, the output of the AI / ML model may recover all resources and all RSRPs. Alternatively, the output of the AI / ML model may be the best resource / beam and the RSRP of the best resource. Thus, the AI / ML model may be optimized based on the actual measurement results.

[0077] The network device 120 may transmit the second transmission configuration to the terminal device 110-1 (4052). The second transmission configuration may indicate resources for transmitting the sounding reference signal. In some embodiments, the second transmission configuration may indicate an index of a resource for transmitting the sounding reference signal. In some embodiments, the second transmission configuration may include an information element "SRS resource set usage" indicating that the resource is to be used to infer an AI / ML model.

[0078] The terminal device 110-1 may perform a second measurement based on the second transmission configuration 4062. The terminal device 110-1 may transmit a sounding reference signal on the resource based on the second transmission configuration.

[0079] The network device 120 may perform a second measurement based on the sounding reference signal (4072). The network device 120 may determine a quality on the received sounding reference signal. For example, the network device 120 may determine an RSRP of the received sounding reference signal. Alternatively, the network device 120 may determine a SINR of the received sounding reference signal.

[0080] The network device 120 may infer using the AI / ML model based on the results of the second measurement (4082). In other words, the network device 120 may perform model inference according to the AI / ML model. For example, the AI / ML model may output a prediction including, for example, a complete set of beam pairs and beam qualities, a best beam for subsequent transmission, and a subset for the second measurement.

[0081] According to some embodiments, the terminal device 110-1 may be used as a classifier to classify training data samples generated by the network device 120. The terminal device 110-1 may transmit the classification result to the network device 120. The embodiments will be described with reference to Figures 5A to 5C in which the terminal device 110-1 functions as a classifier.

[0082] FIG. 5A is a signaling chart illustrating a process between terminal device 110-1 and network device 120 according to some example embodiments of the present disclosure, where the first configuration is a measurement configuration.

[0083] The network device 120 may transmit a first measurement configuration to the terminal device 110-1 (5010). The first measurement configuration may indicate a resource subset for measurements.

[0084] The network device 120 may notify the terminal device 110-1 of a first candidate data sample for constructing a first data set (5020). In some embodiments, the first candidate data sample may include a resource subset as an input of the AI / ML model and a target resource as an output of the AI / ML model. As shown in FIG. 7B, the potential data sample for training the AI / ML model may be transmitted to the terminal device 110-1 and measured by the terminal device 110-1.

[0085] The first candidate data sample may be indicated to the terminal device 110-1 via appropriate signaling. For example, the first candidate data sample may be indicated via radio resource control (RRC) signaling. Alternatively, the first candidate data sample may be indicated within a medium access control (MAC) control element (CE). In some embodiments, the first candidate data sample may be indicated within downlink control information (DCI). In some embodiments, the first measurement configuration may include the first candidate data sample for constructing the first data set.

[0086] In some embodiments, the RRC IEs "RS-subset-to-measure-for-training" and "RS-to-compare-for-training" may be used to indicate the first candidate data sample. The size of the RS-subset-to-measure-for-training may depend on the number of beams at the NW side and the number of beams at the UE side. Additionally, a time offset may be added between the last symbol carrying the RS in the subset K1 of resources / beams and the target resource / beam.

number

number

[0087] Additionally, the first measurement configuration may include a reporting configuration for the terminal device 110-1. The terminal device 110-1 may determine which reporting amount to report to the network device 120 based on the reporting configuration. In some embodiments, the first reporting amount may indicate whether the first candidate data sample is suitable for constructing a first data set for the AI / ML model. Alternatively, the first reporting amount may indicate a result of a comparison of a quality of the resource subset and a quality of the target resource.

[0088] The terminal device 110-1 may perform a first measurement based on the first measurement configuration (5030). For example, the terminal device 110-1 may perform a first measurement on a resource subset. The terminal device 110-1 may determine a quality of the resource subset based on the first measurement on the resource subset. For example, the terminal device 110-1 may determine an RSRP on the resource subset. Alternatively, the terminal device 110-1 may determine an SINR on one resource subset.

[0089] The terminal device 110-1 may also perform a first measurement on the target resource. The terminal device 110-1 may determine the quality of the target resource based on the first measurement on the target resource. For example, the terminal device 110-1 may determine the RSRP on the target resource. Alternatively, the terminal device 110-1 may determine the SINR on the target resource.

[0090] The terminal device 110-1 may transmit a first measurement report to the network device 120 (5040). The first measurement report may include a first reporting amount. In some embodiments, the first reporting amount may indicate an identification result indicating whether the first candidate data sample is suitable for constructing a first data set for an AI / ML model. In some embodiments, the first reporting amount may indicate whether the first candidate data sample is suitable for constructing a first data set for an AI / ML model. Alternatively, the first reporting amount may indicate a result of a comparison between a quality of the resource subset and a quality of a target resource. For example, the terminal device 110-1 may compare a quality of the resource subset with a quality of the target resource. In this case, if the quality of at least one of the resource subsets is better than the quality of the target resource, the first reporting amount may indicate that the first candidate data sample is not suitable for constructing a first data set for an AI / ML model. For example, if the RSRP of at least one of the K1 RSs is

number

number

number

number

[0091] In some embodiments, the first measurement report may include 1 bit for one candidate data sample, or N bits for N candidate data samples. Alternatively, the first measurement report may include 1 bit for N candidate data samples. The bit width for the resource index may be ceil(log2(K1+1)) or ceil(log2(K1)) or 0, where K1 represents the number of resources in the resource subset.

[0092] In some embodiments, the first reporting amount may include an index of the resource and a quality of the resource. For example, the first reporting amount may include a first index of the target resource and a quality of the target resource. Alternatively or additionally, the first reporting amount may include an index of the resource subset and a quality of the resource subset. In other embodiments, the first reporting amount may include a first index of the target resource having a quality of the target resource and an index of the resource subset having a quality of the resource subset. By way of example only, the first reporting amount may include the SSBRI / CRI of the target resource, the RSRP / SINR of the target resource. Alternatively or additionally, the first reporting amount may also include the SSBRI / CRI of the resource subset and the RSRP / SINR of the resource subset. In this way, overhead can be reduced.

[0093] Alternatively, the first reporting amount may include a second index of the best resource determined by the terminal device 110-1. For example, the first reporting amount may include the SSBRI / CRI of the best resource. In this case, the terminal device 110-1 may recommend the best resource or the best beam. In some embodiments, the first reporting amount may include an index of the measurement resource subset determined by the terminal device 110-1. For example, the first reporting amount may include the SSBRI / CRI of the subset of the measurement resources. In this case, the terminal device 110-1 may recommend the measured resource or beam. In other embodiments, the first reporting amount may include a second index of the best resource and an index of the measurement resource subset determined by the terminal device 110-1.

[0094] In some embodiments, the first reporting quantity may be the quality of the resource. For example, the first reporting quantity may include the quality of the best resource / beam. By way of example only, the first reporting quantity may include the RSRP or SINR of the best resource / beam. Alternatively, the first reporting quantity may be the quality of a subset of the measurement resources / beams. In this case, the quality of the subset of the measurement resources / beams may be sorted in ascending / descending order of the resource index. In this way, it is possible to reduce the overhead of the first report.

[0095] The network device 120 may train the AI / ML model based on the first measurement report (5050). In some embodiments, if the first measurement report indicates that the first candidate data sample is not suitable for constructing a dataset, the network device 120 may not use the first candidate data sample to train the AI / ML model. Alternatively, if the first measurement report indicates that the first candidate data sample is suitable for constructing a dataset, the network device 120 may use the first candidate data sample to train the AI / ML model. As shown in FIG. 7B, the network device 120 may construct a test dataset based on the identification result in the first measurement report.

[0096] FIG. 5B is a signaling chart illustrating a process between terminal device 110-1 and network device 120 in accordance with some example embodiments of the present disclosure in which the first configuration is a transmit configuration.

[0097] The network device 120 may transmit a first measurement configuration to the terminal device 110-1 (5011). The first measurement configuration may indicate a resource subset for measurements. The first measurement configuration may include two associated reporting configurations. For example, the first measurement configuration may include a first reporting configuration for resources in K1 and

number

[0098] The network device 120 may inform the terminal device 110-1 of a first candidate data sample for constructing a first data set (5021). In some embodiments, the first candidate data sample may include a resource subset as an input for the AI / ML model. As shown in FIG. 7B, potential data samples for training the AI / ML model may be transmitted to and measured by the terminal device 110-1. The network device 120 may transmit a reference signal on the resource subset.

[0099] The first candidate data sample may be indicated to the terminal device 110-1 via appropriate signaling. For example, the first candidate data sample may be indicated via radio resource control (RRC) signaling. Alternatively, the first candidate data sample may be indicated within a medium access control (MAC) control element (CE). In some embodiments, the first candidate data sample may be indicated within downlink control information (DCI). In some embodiments, the first measurement configuration may include the first candidate data sample for constructing the first data set.

[0100] In some embodiments, the RRC IE "RS-subset-to-measure-for-training" may be used to indicate the first candidate data sample. The size of the RS-subset-to-measure-for-training may depend on the number of beams on the NW side and the number of beams on the UE side. The added IE to indicate the resource subset K1 to be measured may be a bit string (01..01), where 1 means to measure and 0 means not to measure, and the length of the bit string is equal to the number of resources in the resource set. Additionally, a time offset is added between the last symbol carrying the RS in the subset K1 of resources / beams and the target resource / beam.

number

[0101] The terminal device 110-1 may perform a first measurement based on the first measurement configuration (5031). For example, the terminal device 110-1 may perform a first measurement on a resource subset. The terminal device 110-1 may determine a quality of the resource subset based on the first measurement on the resource subset. For example, the terminal device 110-1 may determine an RSRP on the resource subset. Alternatively, the terminal device 110-1 may determine an SINR on the resource subset.

[0102] The terminal device 110-1 may transmit a first measurement report to the network device 120 (5041). In some embodiments, the first report amount may be the quality of a resource. For example, the first report amount may be the quality of a subset of measurement resources / beams. In this case, the quality of the subset of measurement resources / beams may be sorted in ascending / descending order of resource index. In this way, it is possible to reduce the overhead of the first report.

[0103] In some embodiments, the first reporting amount may include an index of a resource and a quality of the resource. For example, the first reporting amount may include an index of a resource subset and a quality of the resource subset. For example, the first reporting amount may also include an SSBRI / CRI of the resource subset and an RSRP / SINR of the resource subset. Thus, the network device determines whether the target resource has the best quality.

[0104] The network device 120 may notify the terminal device 110-1 of the target resources (5051). The network device 120 may transmit a reference signal on the target resources. The target resources may be indicated to the terminal device 110-1 via appropriate signaling. For example, the target resources may be indicated via radio resource control (RRC) signaling. Alternatively, the target resources may be indicated within a medium access control (MAC) control element (CE). In some embodiments, the target resources may be indicated within downlink control information (DCI).

[0105] In some embodiments, the RRC IE "RS-to-compare-for-training" may be used to indicate the first candidate data sample. Alternatively or additionally, the predicted best beam / resource.

number

[0106] In some embodiments, the first measurement report may include 1 bit for one candidate data sample, or N bits for N candidate data samples. Alternatively, the first measurement report may include 1 bit for N candidate data samples. The bit width for the resource index may be ceil(log2(K1+1)) or ceil(log2(K1)) or 0, where K1 represents the number of resources in the resource subset.

[0107] The terminal device 110-1 may perform a second measurement on the target resource (5061). For example, the terminal device 110-1 may determine an RSRP on the target resource. Alternatively, the terminal device 110-1 may determine an SINR on the target resource.

[0108] The terminal device 110-1 may transmit a second measurement report to the network device 120 (5071). The second measurement report may include the quality of the target resource. For example, the second measurement report may include the RSRP of the target resource.

[0109] Alternatively or additionally, the second measurement report may include a first reporting quantity. In some embodiments, the first reporting quantity may indicate whether the first candidate data sample is suitable for constructing a first data set for an AI / ML model. For example, the terminal device 110-1 may compare the quality of the resource subsets with the quality of the target resource. In this case, if the quality of at least one of the resource subsets is better than the quality of the target resource, the first reporting quantity may indicate that the first candidate data sample is not suitable for constructing a first data set for an AI / ML model. Alternatively, if the quality of the target resource is better than all of the qualities of the resource subsets, the first reporting quantity may indicate that the first candidate data sample is suitable for constructing a first data set for an AI / ML model. In some embodiments, a comparison offset may be applied. For example, an RSRP offset may be added to the left or right side of the inequality: RSRPs of all K1 RSs. <RSRP

number

number

[0110] The network device 120 may train the AI / ML model based on the second measurement report (5081). In some embodiments, if the second measurement report indicates that the first candidate data sample is not suitable for constructing a dataset, the network device 120 may not use the first candidate data sample to train the AI / ML model. Alternatively, if the second measurement report indicates that the first candidate data sample is suitable for constructing a dataset, the network device 120 may use the first candidate data sample to train the AI / ML model. As shown in FIG. 7B, the network device 120 may construct a test dataset based on the identification result in the first measurement report.

[0111] FIG. 5C is a signaling chart illustrating a process between terminal device 110-1 and network device 120 in accordance with some example embodiments of the present disclosure in which the first configuration is a transmit configuration.

[0112] The network device 120 may transmit the first transmission configuration to the terminal device 110-1 (5012). The first transmission configuration may indicate a resource subset for transmitting the sounding reference signal. In some embodiments, the first transmission configuration may indicate an index of the resource subset for transmitting the sounding reference signal. Alternatively, the first transmission configuration may indicate an index of the target resource.

[0113] The resource subset may be indicated to the terminal device 110-1 via appropriate signaling. For example, the resource subset may be indicated via radio resource control (RRC) signaling. For example, the RRC IEs "RS-subset-to-transmit-for-training" and "RS-to-compare-for-training" may be added to the RRC signaling. Alternatively, the resource subset may be indicated in a medium access control (MAC) control element (CE). In some embodiments, the resource subset may be indicated in downlink control information (DCI). Additionally, a time offset may be provided between the last symbol carrying an RS in the subset K1 of resources / beams and the target resource / beam.

number

[0114] The terminal device 110-1 may perform a first measurement based on the first transmission configuration (5022). The terminal device 110-1 may transmit a sounding reference signal on the resource subset based on the first transmission configuration. In some embodiments, the terminal device 110-1 may also transmit a sounding reference signal on the target resource, thus not placing an additional burden on the terminal device 110-1.

[0115] In some other embodiments, the network device 120 may send the second transmission configuration to the terminal device 110-1. The second transmission configuration may indicate an index of the target resource.

[0116] The network device 120 may perform a first measurement based on the sounding reference signal (5032). The network device 120 may determine a quality of the sounding reference signal received on the resource subset. For example, the network device 120 may determine an RSRP of the received sounding reference signal. Alternatively, the network device 120 may determine a SINR of the received sounding reference signal. In some embodiments, the network device 120 may determine a quality on the sounding reference signal received on the target resource.

[0117] The network device 120 may train the AI / ML model based on the measurement results (5042). In some embodiments, if the measurement results indicate that the quality on the target resource is better than the quality on the resource subset, the network device 120 may use the first candidate data sample including the resource subset as an input to the AI / ML model and the target resource as an input to the AI / ML model to train the AI / ML model. Alternatively, if the measurement results indicate that at least one resource in the resource subset has better quality than the target resource, the network device 120 may not use the first candidate data sample to train the AI / ML model.

[0118] It should be noted that the above-described embodiments may be implemented separately. Alternatively, the above-described embodiments may be realized in any suitable combination.

[0119] 8 is a flow chart of an example method 700 according to an embodiment of the present disclosure. The method 700 may be implemented in any suitable device. For illustrative purposes only, the method 700 may be implemented in a terminal device 110-1 as shown in FIG.

[0120] At block 710, the terminal device 110-1 receives a first configuration from the network device 120. The first configuration indicates at least one resource subset from a first set of resources. In some embodiments, the first set of resources may include resources for beam measurement and reporting. The at least one resource subset is for constructing a first dataset for training the AI / ML model 200 in the network device 120. The first dataset includes one or more of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference.

[0121] In some embodiments, the first configuration may be a first measurement configuration. In this case, the terminal device 110-1 may perform measurements on the at least one resource subset based on the first measurement configuration. For example, the first configuration may indicate a first candidate data sample for constructing the first data set. The first candidate data sample may include a resource subset as an input of an AI / ML model and a target resource as an output of the AI / ML model. Alternatively, the first configuration may indicate the terminal device 110-1 to perform measurements on a first set of resources. In some embodiments, the terminal device 110-1 may determine a quality based on a result of the measurement. For example, the terminal device 110-1 may determine a reference signal received power (RSRP) based on the measurement. Alternatively, the terminal device 110-1 may determine a signal-to-interference-and-noise ratio (SINR) based on the measurement.

[0122] In other embodiments, the first configuration may be a first transmission configuration. In this case, the terminal device may perform transmission based on the first transmission configuration. For example, the first configuration may indicate a set of reference signal resources for a sounding reference signal (SRS). Alternatively or additionally, the first configuration may indicate a reference signal resource subset from the set of reference signal resources. In some embodiments, the first configuration may indicate a target reference signal resource.

[0123] At block 720, terminal device 110-1 transmits information to network device 120. In some embodiments, if the first configuration is the first measurement configuration, terminal device 110-1 may transmit a result of the measurement to network device 120. Alternatively, if the first configuration is the first transmission configuration, terminal device 110-1 may transmit a sounding reference signal to network device 120. In this case, network device 120 may measure the sounding reference signal.

[0124] In some embodiments, the terminal device 110-1 may receive an indication of the updated resource subset from the network device 120. In some embodiments, the updated resource subset may be an input for AI / ML model training. Alternatively or additionally, the updated resource subset may be used to collect data for AI / ML model training. In other embodiments, the updated resource subset may be used to collect data for AI / ML model inference. In some other embodiments, the updated resource subset may be used for regular beam measurements and reporting. In this way, randomly selected measurement subsets may be avoided.

[0125] In some embodiments, explicit signaling may be required to inform the terminal device 110-1 about the updated measurement subset K1. For example, the input of the AI / ML model training may be updated based on the updated resource subset. In other embodiments, the updated resource subset may be transmitted within a first measurement configuration, which means that the configuration for collecting data for AI / ML model training may be updated. In this case, the RRC IEs "RS-subset-to-measure-for-training" and "RS-to-compare-for-training" may be used to inform the updated resource subset. Alternatively, the updated resource subset may be transmitted within a second measurement configuration, which means that the configuration for collecting data for AI / ML model inference may be updated. In this case, a new signaling "RS-subset-measure-for-inference" may be used to inform the updated resource subset. As mentioned above, different AI / ML models may be used to select different subsets K1. In other words, multiple AI / ML models (model i) can learn how to generate K from different versions of K1_i. In this case, instead of using K1' as output, the AI / ML model ID i may be used as output, where one version of the K1 choices corresponds to one AI / ML model.

[0126] In some embodiments, inputs from the terminal device 110-1 for the first AI / ML mode may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs from the network device 120 for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic condition, transmission scheme, wide beam used in initial access. Inputs from other terminal devices for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs for the AI / ML model from other network devices may include one or more of: number of Tx / Rx beams / panels, beamforming gain, beam width, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic condition, transmission scheme, wide beam used in initial access. Inputs for the AI / ML model from other AI / ML models may include one or more of: UE location prediction, UE trajectory prediction, handover prediction, initial access prediction, channel state information (CSI) prediction.

[0127] In some embodiments, the output from the AI / ML model for the terminal device 110-1 may include one or more of transmit power, beam switching decisions, active subset of Tx / Rx beams, transmission scheme. The output from the AI / ML model for the network device 120 may include one or more of transmit power, handover, active subset of Tx / Rx beams, scheduling, beam switching decisions, multi-user scheduling decisions, traffic conditions, transmission scheme. The output from the AI / ML model may be used as input to other AI / ML models for predicting UE position / trajectory, predicting handover / load balancing / energy saving decisions, predicting CSI.

[0128] In some embodiments, when AI / ML model inference is applied, an additional parameter regarding the number of reduced UE Rx beams may be used to determine the number of configured RS resources in the RS resource set, the measurement period requirement, and the measurement accuracy requirement. In some embodiments, the parameter may be indicated by the network device 120. Alternatively, the parameter may be reported by the terminal device 110-1. Alternatively, the parameter may be reported by the terminal device 110-1 in a capability report. In other embodiments, the parameter may also be an output of the AI / ML model. In this way, it is possible to guide the configuration and set performance requirements when AI / ML model inference is applied.

[0129] In some embodiments, the number of occupied CSI processing units (CPUs) and the CSI computation time may depend on the number of resources in the first set of resources that are actually measured for beam quality. For example, the number of occupied CPUs may be NCPU (i.e., the UE capability on the maximum supported number of simultaneous CSI computations) or K1+1 (i.e., the number of actually measured resources in the resource set including the target resource). Alternatively or additionally, for a time span from the CSI reference to the last symbol carrying an RS in K1, the number of occupied CPUs may be NCPU or K1 (i.e., the number of actually measured resources in the resource set). The number of occupied CPUs may be:

number

[0130] 9 is a flow chart of an example method 800 according to an embodiment of the present disclosure. The method 800 may be implemented in any suitable device. For illustrative purposes only, the method 800 may be implemented in a network device 120 as shown in FIG.

[0131] In block 810, the network device 120 transmits a first configuration to the terminal device 110-1. The first configuration indicates at least one resource subset from a first set of resources. In some embodiments, the first set of resources may include resources for beam measurement and reporting. The at least one resource subset is for constructing a first dataset for training the AI / ML model 200 in the network device 120. The first dataset includes one or more of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference.

[0132] In some embodiments, the first configuration may be a first measurement configuration. In this case, the terminal device 110-1 may perform measurements on the at least one resource subset based on the first measurement configuration. For example, the first configuration may indicate a first candidate data sample for constructing the first data set. The first candidate data sample may include a resource subset as an input of an AI / ML model and a target resource as an output of the AI / ML model. Alternatively, the first configuration may indicate the terminal device 110-1 to perform measurements on a first set of resources. In some embodiments, the terminal device 110-1 may determine a quality based on a result of the measurement. For example, the terminal device 110-1 may determine a reference signal received power (RSRP) based on the measurement. Alternatively, the terminal device 110-1 may determine a signal-to-interference-and-noise ratio (SINR) based on the measurement.

[0133] In other embodiments, the first configuration may be a first transmission configuration. In this case, the terminal device may perform transmission based on the first transmission configuration. For example, the first configuration may indicate a set of reference signal resources for a sounding reference signal (SRS). Alternatively or additionally, the first configuration may indicate a reference signal resource subset from the set of reference signal resources. In some embodiments, the first configuration may indicate a target reference signal resource.

[0134] In block 830, the network device 120 receives information from the terminal device 110-1. In some embodiments, if the first configuration is the first measurement configuration, the terminal device 110-1 may transmit a result of the measurement to the network device 120. Alternatively, if the first configuration is the first transmission configuration, the terminal device 110-1 may transmit a sounding reference signal to the network device 120. In this case, the network device 120 may measure the sounding reference signal.

[0135] In some embodiments, the network device 120 may train an AI / ML model. The network device 120 may construct a first data set based on the results of the measurements. The AI / ML model may be trained based on the first data set. Thus, the AI / ML model may be optimized based on the actual measurement results.

[0136] The network device 120 may send an indication of the updated resource subset to the terminal device 110-1. In some embodiments, the updated resource subset may be an input for AI / ML model training. Alternatively or additionally, the updated resource subset may be used to collect data for AI / ML model training. In other embodiments, the updated resource subset may be used to collect data for AI / ML model inference. In some other embodiments, the updated resource subset may be used for regular beam measurements and reporting. In this way, randomly selected measurement subsets may be avoided.

[0137] In some embodiments, explicit signaling may be required to inform the terminal device 110-1 about the updated measurement subset K1. For example, the input of the AI / ML model training may be updated based on the updated resource subset. In other embodiments, the updated resource subset may be transmitted within a first measurement configuration, which means that the configuration for collecting data for AI / ML model training may be updated. In this case, the RRC IEs "RS-subset-to-measure-for-training" and "RS-to-compare-for-training" may be used to inform the updated resource subset. Alternatively, the updated resource subset may be transmitted within a second measurement configuration, which means that the configuration for collecting data for AI / ML model inference may be updated. In this case, a new signaling "RS-subset-measure-for-inference" may be used to inform the updated resource subset. As mentioned above, different AI / ML models may be used to select different subsets K1. In other words, multiple AI / ML models (model i) can learn how to generate K from different versions of K1_i. In this case, instead of using K1' as output, the AI / ML model ID i may be used as output, where one version of the K1 choices corresponds to one AI / ML model.

[0138] In some embodiments, inputs from the terminal device 110-1 for the first AI / ML mode may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs from the network device 120 for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic condition, transmission scheme, wide beam used in initial access. Inputs from other terminal devices for the AI / ML model may include one or more of the following: number of Tx / Rx beams / panels, beamforming gain, beamwidth, location information, trajectory, speed, orientation, direction of movement, transmit power, measurement subset of Tx / Rx beams, transmission scheme, wide beam used in initial access. Inputs from other network devices 120 for the AI / ML model may include one or more of: number of Tx / Rx beams / panels, beamforming gain, beam width, location information, transmit power, handover, measurement subset of Tx / Rx beams, scheduling information, beam switching decision, multi-user scheduling decision, traffic conditions, transmission scheme, wide beam used in initial access. Inputs from other AI / ML models for the AI / ML model may include one or more of: UE location prediction, UE trajectory prediction, handover prediction, initial access prediction, channel state information (CSI) prediction.

[0139] In some embodiments, the output from the AI / ML model for the terminal device 110-1 may include one or more of transmit power, beam switching decisions, active subset of Tx / Rx beams, transmission scheme. The output from the AI / ML model for the network device 120 may include one or more of transmit power, handover, active subset of Tx / Rx beams, scheduling, beam switching decisions, multi-user scheduling decisions, traffic conditions, transmission scheme. The output from the AI / ML model may be used as input to other AI / ML models for predicting UE position / trajectory, predicting handover / load balancing / energy saving decisions, predicting CSI.

[0140] In some embodiments, when AI / ML model inference is applied, an additional parameter regarding the number of reduced UE Rx beams may be used to determine the number of configured RS resources in the RS resource set, the measurement period requirement, and the measurement accuracy requirement. In some embodiments, the parameter may be indicated by the network device 120. Alternatively, the parameter may be reported by the terminal device 110-1. Alternatively, the parameter may be reported by the terminal device 110-1 in a capability report. In other embodiments, the parameter may also be an output of the AI / ML model. In this way, it is possible to guide the configuration and set performance requirements when AI / ML model inference is applied.

[0141] In some other embodiments, the number of occupied CSI processing units (CPUs) and the CSI computation time may depend on the number of resources in the first set of resources that are actually measured for beam quality. For example, the number of occupied CPUs may be NCPU (i.e., the UE capability on the maximum supported number of simultaneous CSI computations) or K1+1 (i.e., the number of actually measured resources in the resource set including the target resource). Alternatively or additionally, for the time span from the CSI reference to the last symbol carrying an RS in K1, the number of occupied CPUs may be NCPU or K1 (i.e., the number of actually measured resources in the resource set). The number of occupied CPUs may be:

number

[0142] In some embodiments, a terminal device comprises circuitry configured to: receive, at the terminal device, a first configuration from a network device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model at the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; and transmit information to the network device regarding a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0143] In some embodiments, the first set of resources includes at least resources for beam measurement and reporting. In some embodiments, the terminal device comprises circuitry configured to perform reference signal (RS) quality measurements on at least one resource subset. In some embodiments, the terminal device comprises circuitry configured to transmit information by determining RS quality on the at least one resource subset based on the RS quality measurements and transmitting information regarding the RS quality on the at least one resource subset to a network device.

[0144] In some embodiments, the first configuration further includes a first candidate data sample for constructing a first dataset and a first reporting configuration. In some embodiments, the terminal device comprises a circuit configured to transmit information by transmitting a first report to the network device, the first report including a first reporting quantity indicating whether the first candidate data sample is suitable for constructing a first dataset for the first AI / ML model.

[0145] In some embodiments, the first candidate data sample includes a resource subset as an input of the first AI / ML model and a target resource as an output of the first AI / ML model.

[0146] In some embodiments, the terminal device comprises circuitry configured to: determine a quality of the resource subset based on the measurements on the resource subset; determine a quality of the target resource based on the measurements on the target resource; and compare the quality of the resource subset to a quality of the target resource. In some embodiments, the terminal device comprises circuitry configured to transmit information by transmitting a first report to a network device, the first report quantity being indicative of a result of the comparison of the quality of the resource subset to a quality of the target resource.

[0147] In some embodiments, the first reporting amount further includes at least one of a first index of a target resource having a quality of the target resource and an index of a resource subset having a quality of the resource subset, a second index of a best resource determined by the terminal device, an index of a measurement resource subset determined by the terminal device, a measured quality of the best resource, or a measured quality of the measurement resource subset.

[0148] In some embodiments, the first configuration further includes a first candidate data sample for constructing a first data set, the first candidate data sample including a resource subset as an input of the first AI / ML model. In some embodiments, the terminal device comprises a circuit configured to: perform measurements on the resource subset; report to the network device a quality of the resource subset based on the measurements on the resource subset; receive from the network device a second configuration indicating a target resource as an output of the first AI / ML model; perform measurements on the target resource; and determine a quality of the target resource based on the measurements on the target resource.

[0149] In some embodiments, the terminal device comprises circuitry configured to transmit information by transmitting a first report to the network device including a quality of the target resource, or transmitting a first report to the network device including a first report quantity indicating a comparison result between the quality of the target resource and a quality of the resource subset.

[0150] In some embodiments, if a quality of at least one of the resource subsets is better than a quality of the target resource, the first reporting quantity indicates that the first candidate data sample is not suitable for constructing a first dataset for the first AI / ML model, and if the quality of the target resource is better than all of the qualities of the resource subsets, the first reporting quantity indicates that the first candidate data sample is suitable for constructing a first dataset for the first AI / ML model.

[0151] In some embodiments, the first set of resources includes a set of reference signal resources, the at least one resource subset includes a reference signal resource subset from the set of reference signal resources, and the first configuration further indicates a target reference signal resource. In some embodiments, the terminal device comprises circuitry configured to transmit information by transmitting the set of sounding reference signals to the network device on the set of reference signal resources and transmitting the target sounding reference signal to the network device on the target reference signal resources.

[0152] In some embodiments, the terminal device is configured to have a time offset between the last symbol carrying a set of sounding reference signals and the first symbol carrying a target sounding reference signal.

[0153] In some embodiments, the first set of resources includes a set of reference signal resources and the at least one resource subset includes a reference signal resource subset from the set of reference signal resources. In some embodiments, the terminal device comprises a circuit configured to transmit information by transmitting the set of sounding reference signals to a network device on the set of reference signal resources. In some embodiments, the terminal device comprises a circuit configured to receive from the network device a second configuration indicating target reference signal resources and transmit the target sounding reference signal to the network device on the target reference signal resources.

[0154] In some embodiments, the first configuration indicates to the terminal device to perform measurements on the first set of resources. In some embodiments, the terminal device comprises circuitry configured to perform measurements on the first set of resources and determine a quality of the first set of resources based on the measurements on the first set of resources. In some embodiments, the terminal device comprises circuitry configured to transmit information by transmitting a second report to the network device indicative of the quality of the first set of resources.

[0155] In some embodiments, the second report includes at least one of an index of the first set of resources and a quality of the first set of resources, an index of a resource subset from the first set of resources and a quality of the resource subset that exceeds a predetermined threshold, or a difference in quality of the first set of resources relative to a reference quality of the first set of resources.

[0156] In some embodiments, the first set of resources includes a set of reference signal resources, and the first configuration indicates to the terminal device to transmit a sounding reference signal on the set of reference signal resources. In some embodiments, the terminal device comprises circuitry configured to transmit information to the network device by transmitting the sounding reference signal on the set of reference signal resources.

[0157] In some embodiments, the terminal device comprises a circuit configured to perform: receiving an indication of an updated resource subset from the first set of resources from the network device, the updated resource subset being output from the first AI / ML model. In some embodiments, the terminal device comprises a circuit configured to receive the indication by receiving a first configuration from the network device including the indication of the updated resource subset or receiving a second configuration from the network device for inference of the data processing model including the indication of the updated resource subset.

[0158] In some embodiments, at least one of the number of reference signal resources set in the reference signal resource set, the measurement period requirement, or the measurement accuracy requirement is determined based on the number of receive beams applied during inference of the AI / ML model, and the number of receive beams applied during inference of the first AI / ML model is indicated by a network device or determined by a terminal device, or the number of receive beams applied during inference of the first AI / ML model is output by the AI / ML model.

[0159] In some embodiments, the number of channel state information (CSI) processing units and the CSI calculation time depend on the number of resources measured for beam quality in the first set of resources.

[0160] In some embodiments, the network device comprises circuitry configured to: send a first configuration to a terminal device at the network device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) / machine learning (ML) model at the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; and receive information from the terminal device relating to a quality decision based on the at least one resource subset, the information for constructing the first dataset.

[0161] In some embodiments, the first set of resources includes at least resources for beam measurement and reporting. In some embodiments, the network device comprises a circuit configured to receive information by receiving information from a terminal device regarding reference signal (RS) quality on at least one resource subset. In some embodiments, the network device comprises a circuit configured to train an AI / ML model based on RS quality of a set of beam pairs.

[0162] In some embodiments, the first configuration further includes a first candidate data sample for constructing the first dataset. In some embodiments, the network device comprises a circuit configured to receive information by receiving a first report from the terminal device, the first report including a first report quantity indicating whether the first candidate data sample is suitable for constructing the first dataset for the first AI / ML model.

[0163] In some embodiments, the first candidate data sample includes a resource subset as an input of the first AI / ML model and a target resource as an output of the first AI / ML model.

[0164] In some embodiments, the network device comprises circuitry configured to receive the information by receiving a first report from the terminal device, the first report including a first report quantity indicative of a result of a comparison between a quality of the resource subset and a quality of the target resource.

[0165] In some embodiments, the first reporting quantity further includes at least one of a first index of a target resource having a quality of the target resource and an index of a resource subset having a quality of the resource subset, a second index of a best resource determined by the terminal device, an index of a measurement resource subset determined by the terminal device, a measured quality of the best resource, or a measured quality of the measurement resource subset.

[0166] In some embodiments, the first configuration further includes a first candidate data sample for constructing a first data set, the first candidate data sample including a resource subset as an input of the first AI / ML model. In some embodiments, the network device comprises a circuit, the circuit configured to perform: sending a second configuration to the terminal device, the second configuration indicating the target resource as an output of the first AI / ML model.

[0167] In some embodiments, the network device comprises circuitry configured to receive information by receiving a first report from a terminal device including a quality of the target resource, or receiving a first report from the terminal device including a first report quantity indicating a comparison result between the quality of the target resource and the quality of the resource subset.

[0168] In some embodiments, if a quality of at least one of the resource subsets is better than a quality of the target resource, the first reporting quantity indicates that the first candidate data sample is not suitable for constructing a first dataset for the first AI / ML model, and if the quality of the target resource is better than all of the qualities of the resource subsets, the first reporting quantity indicates that the first candidate data sample is suitable for constructing a first dataset for the first AI / ML model.

[0169] In some embodiments, the first set of resources includes a set of reference signal resources, the at least one resource subset includes a reference signal resource subset from the set of reference signal resources, and the first configuration further indicates a target reference signal resource. In some embodiments, the network device comprises circuitry configured to receive information by receiving a set of sounding reference signals from a terminal device on the set of reference signal resources and receiving a target sounding reference signal from the terminal device on the target reference signal resources.

[0170] In some embodiments, the first set of resources includes a set of reference signal resources and the at least one resource subset includes a reference signal resource subset from the set of reference signal resources. In some embodiments, the network device comprises circuitry configured to receive information by receiving a set of sounding reference signals from a terminal device on the set of reference signal resources. In some embodiments, the network device comprises circuitry configured to: transmit a second configuration to the terminal device indicating the target reference signal resources; and receive the target sounding reference signal from the terminal device on the target reference signal resources.

[0171] In some embodiments, the first configuration indicates to the terminal device to perform measurements on the first set of resources, hi some embodiments, the network device comprises circuitry configured to receive information by receiving a second report from the terminal device indicative of a quality of the first set of resources.

[0172] In some embodiments, the second report includes at least one of an index of the first set of resources and a quality of the first set of resources, an index of a resource subset from the first set of resources and a quality of the resource subset that exceeds a predetermined threshold, or a difference in quality of the first set of resources relative to a reference quality of the first set of resources.

[0173] In some embodiments, the first set of resources includes a set of reference signal resources, and the first configuration indicates to the terminal device to transmit a sounding reference signal on the set of reference signal resources. In some embodiments, the network device comprises circuitry configured to receive information by receiving a sounding reference signal from the terminal device on the set of reference signal resources.

[0174] In some embodiments, the network device comprises a circuit configured to obtain an updated resource subset from an output of the first AI / ML model and to send an indication of the updated resource subset to the terminal device. In some embodiments, the network device comprises a circuit configured to send the indication by sending a first configuration to the terminal device including the indication of the updated resource subset or sending a second configuration for inference of the data processing model including the indication of the updated resource subset to the terminal device.

[0175] In some embodiments, at least one of the number of reference signal resources set in the reference signal resource set, the measurement period requirement, or the measurement accuracy requirement is determined based on the number of receive beams applied during inference of the AI / ML model, and the number of receive beams applied during inference of the first AI / ML model is indicated by a network device or determined by a terminal device, or the number of receive beams applied during inference of the first AI / ML model is output by the AI / ML model.

[0176] 10 is a schematic block diagram of an apparatus 900 suitable for implementing embodiments of the present disclosure. The apparatus 900 may be considered as another exemplary implementation of the terminal device 110 as shown in FIG. 1. Thus, the apparatus 900 may be implemented in the terminal device 110 or as at least a part thereof. Alternatively, the apparatus 900 may be considered as another exemplary implementation of the network device 120 as shown in FIG. 1. Thus, the apparatus 900 may be implemented in the network device 120 or as at least a part thereof.

[0177] As shown, the apparatus 900 comprises a processor 910, a memory 920 coupled to the processor 910, a suitable transmitter (TX) and receiver (RX) 940 coupled to the processor 910, and a communication interface coupled to the TX / RX 940. The memory 920 stores at least a portion of a program 930. The TX / RX 940 is used for bidirectional communication. The TX / RX 940 has at least one antenna to facilitate communication, although the access nodes referred to herein may in practice have multiple antennas. The communication interface may represent any interface required for communication with other network elements, such as an X2 interface for bidirectional communication between eNBs, an S1 interface for communication between a mobility management entity (MME) / serving gateway (S-GW) and an eNB, an Un interface for communication between an eNB and a relay node (RN), or an Uu interface for communication between an eNB and a terminal device.

[0178] The program 930 is assumed to include program instructions that, when executed by the associated processor 910, enable the device 900 to operate according to embodiments of the present disclosure, as described herein with reference to Figures 3-8. The embodiments of the present disclosure may be realized by computer software executable by the processor 910 of the device 900, or by hardware, or by a combination of software and hardware. The processor 910 may be configured to implement various embodiments of the present disclosure. Furthermore, the combination of the processor 910 and the memory 920 may form a processing means 950 suitable for implementing various embodiments of the present disclosure.

[0179] The memory 920 may be of any type suitable for a local technology network and may be implemented using any suitable data storage technology, such as, by way of non-limiting example, non-transitory computer-readable storage media, semiconductor-based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory. Although only one memory 920 is shown in the device 900, there may be several physically different memory modules in the device 900. The processor 910 may be of any type suitable for a local technology network and may include, by way of non-limiting example, one or more of a general purpose computer, a special purpose computer, a microprocessor, a digital signal processor (DSP) and a processor based on a multi-core processor architecture. The device 900 may have multiple processors, for example application specific integrated circuit chips time-slaved to a clock that synchronizes the main processor.

[0180] Overall, 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 that may be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of the present disclosure have been illustrated and described using block diagrams, flow charts, or some other pictorial representations, it should be understood that the blocks, devices, systems, techniques, or methods described herein may be implemented, by way of non-limiting examples, in hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing device, or any combination thereof.

[0181] 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 instructions included in a program module, that execute in a device on a target real or virtual processor to perform a process or method as described above with reference to Figures 2-9. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform particular tasks or implement particular abstract data types. In various embodiments, the functionality of the program modules may be combined or split between program modules as appropriate. The machine-executable instructions of the program modules may be executed in local or distributed devices. In a distributed device, the program modules may be located in both local and remote storage media.

[0182] Program codes for carrying out the 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, a special purpose computer, or other programmable data processing device, and when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may run entirely on the machine, partially on the machine, as a separate software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0183] The above program code may be implemented on a machine-readable medium, which may be any tangible medium that can contain or store a program used by or associated with an instruction execution system, device, or apparatus. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the aforementioned media. More specific examples of the machine-readable storage medium may include an electrical connection having one or more wires, a portable computer disk, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the aforementioned.

[0184] It should be noted that although operations have been described in a particular order, it should not be understood that such operations are required to be performed in the particular order shown, or in any sequential order, or to perform all of the operations described, in order to achieve desired results. In some cases, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the disclosure, but rather as descriptions of features that may be specific to certain embodiments. Some features described in the context of individual embodiments may be realized in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may be implemented in multiple embodiments separately or in any suitable subcombination.

[0185] Although the present disclosure has been described in language specific to structural features and / or methodological acts, it should be understood that the present disclosure, as 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.

[0186] As used herein, the term "terminal device" refers to any device having wireless or wired communication capabilities. Examples of terminal devices include user equipment (UE), personal computers, desktops, mobile phones, cellular phones, smartphones, personal digital assistants (PDAs), portable computers, tablets, wearable devices, Internet of Things (Iota) devices, Ultra-Reliable Low Latency Communications (URLLC) devices, Internet of Anything (Iowa) devices, machine type communication (MTC) devices, vehicle-mounted devices for V2X communications where X stands for pedestrian, vehicle, or infrastructure / network, devices for integrated access and integrated access and backhaul (IAB), satellite-based or airborne vehicles in non-terrestrial networks (NTN) including High Altitude Platforms (HAPs) including satellites and Unmanned Aircraft Systems (UASs), extended reality (XR) devices including different types of reality such as augmented reality (AR), mixed reality (MR), virtual reality (VR), unmanned aerial vehicles (UAVs), which are aircraft without a human pilot and commonly referred to as drones, high speed trains (HSTs), and other types of real-world applications. These include, but are not limited to, devices on a train, image capture devices such as digital cameras, sensor game devices, music storage and playback devices, or Internet appliances that allow wireless or wired Internet access and browsing. The "Terminal Device" may further have "multicast / broadcast" capabilities to support public safety and mission-critical V2X applications, transparent IPv4 / IPv6 multicast distribution, IPTV, smart TV, wireless services, over-the-air software distribution, group communication, and Iota applications. It may also incorporate one or more Subscriber Identity Modules (SIMs), known as multi-SIMs.The term "terminal equipment" may be used interchangeably with UE, mobile station, subscriber station, mobile terminal, user terminal, or wireless equipment.

[0187] The term "network device" refers to a device capable of providing or hosting a cell or coverage in which a terminal device can communicate. Examples of network devices include, but are not limited to, a Node B (node ​​or NB), an evolved node (node ​​or eNB), a next generation node (gNB), a transmit / receive point (TRP), a remote radio unit (RRU), a radio head (RH), a remote radio head (RRH), an IAB node, a femto node, a pico node, a low power node such as a reconfigurable intelligent surface (RIS), etc.

[0188] The terminal device or network device may have the capability of artificial intelligence (AI) or machine learning. Generally, it includes a model trained from a large amount of data collected for a specific function and can be used to predict some information.

[0189] The terminal device or network device may operate on several frequency ranges, such as FR1 (410 MHz-7125 MHz), FR2 (24.25 GHz-71 GHz), frequency bands greater than 100 GHz, and Tera Hertz (THz). It can also operate on licensed / unlicensed / shared spectrum. The terminal device may have two or more connections with the network device under Multi-Radio Dual Connectivity (MR-DC) application scenarios. The terminal device or network device can operate in full duplex, flexible duplex, and cross division duplex modes.

[0190] Embodiments of the present disclosure may be implemented in test equipment, such as, for example, a signal generator, a signal analyzer, a spectrum analyzer, a network analyzer, a test terminal equipment, a test network equipment, a channel emulator, and the like.

[0191] Embodiments of the present disclosure may be performed in accordance with any generation of communication protocols now known or developed in the future, including, but not limited to, first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G) communication protocols, 5.5G, 5G-Advanced networks, or sixth generation (6G) networks.

Claims

1. 1. A method of communication comprising: receiving, at a terminal device, a first configuration from a network device indicating at least one resource subset from a first set of resources, the at least one resource subset for constructing a first dataset for training a first artificial intelligence (AI) or machine learning (ML) model at the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; transmitting information to the network device regarding a quality decision based on the at least one resource subset, the information being for constructing the first data set; Including, method.

2. the first set of resources includes at least a resource for beam measurement and reporting; The method comprises: performing a reference signal (RS) quality measurement on the at least one resource subset; Transmitting the information includes: determining an RS quality on the at least one resource subset based on the RS quality measurements; transmitting the information regarding the RS quality on the at least one resource subset to the network device; Including, The method of claim 1.

3. The first configuration further includes a first candidate data sample for constructing the first data set and a first reporting configuration; Transmitting the information includes: sending a first report to the network device, the first report including a first report quantity indicating whether the first candidate data sample is suitable for constructing the first data set for the first AI or ML model; Including, The method of claim 1.

4. the first candidate data sample includes a resource subset as an input of the first AI or ML model and a target resource as an output of the first AI or ML model; The method according to claim 3.

5. determining a quality of the resource subset based on measurements on the resource subset; determining a quality of the target resource based on measurements on the target resource; comparing the quality of the resource subset with the quality of the target resource; Further comprising: Transmitting the information includes: transmitting a first report to the network device, the first report including a first report quantity indicative of a comparison between the quality of the resource subset and the quality of the target resource; Including, The method according to claim 4.

6. The first report amount is: a first index of the target resource having the quality of the target resource and an index of the resource subset having the quality of the resource subset; A second index of the best resource determined by the terminal device; an index of the measurement resource subset determined by the terminal device; the measured quality of said best resource; or a measured quality of said measurement resource subset; Further comprising at least one of: The method according to claim 5.

7. The first configuration further includes a first candidate data sample for constructing the first data set, the first candidate data sample including a resource subset as an input for the first AI or ML model; The method comprises: performing measurements on said resource subset; reporting to the network device a quality of the resource subset based on the measurements on the resource subset; receiving a second configuration from the network device indicating a target resource as an output of the first AI or ML model; performing measurements on the target resource; determining a quality of the target resource based on the measurements on the target resource; Further comprising: The method of claim 1.

8. Transmitting the information includes: sending a first report to the network device comprising a quality of the target resource; or sending a first report to the network device, the first report including a report quantity indicative of a comparison between the quality of the target resource and the quality of the resource subset; Including, The method of claim 7.

9. If a quality of at least one of the resource subsets is better than the quality of the target resource, the first reporting amount indicates that the first candidate data sample is not suitable for constructing the first data set for the first AI or ML model; If the quality of the target resource is better than all of the qualities of the resource subset, the first reporting amount indicates that the first candidate data sample is suitable for constructing the first data set for the first AI or ML model.

9. The method according to claim 6 or 8.

10. the first set of resources comprises a set of reference signal resources, the at least one resource subset comprises a reference signal resource subset from the set of reference signal resources, and the first configuration further indicates a target reference signal resource; Transmitting the information includes: transmitting a set of sounding reference signals to the network device on the set of reference signal resources; transmitting a target sounding reference signal to the network device on the target reference signal resource; Including, The method of claim 1.

11. the terminal device is configured to have a time offset between a last symbol carrying the set of sounding reference signals and a first symbol carrying the target sounding reference signal. The method of claim 10.

12. the first set of resources comprises a set of reference signal resources, and the at least one resource subset comprises a reference signal resource subset from the set of reference signal resources; Transmitting the information includes: transmitting a set of sounding reference signals to the network device on the set of reference signal resources; Including, The method comprises: receiving a second configuration from the network device indicating a target reference signal resource; transmitting a target sounding reference signal to the network device on the target reference signal resource; Further comprising: The method of claim 1.

13. The first configuration indicates to the terminal device to perform measurements on the first set of resources; The method comprises: performing measurements on the first set of resources; determining a quality of the first set of resources based on measurements on the first set of resources; Further comprising: Transmitting the information includes: sending a second report to the network device indicating the quality of the first set of resources; Including, The method of claim 1.

14. The second report states: an index of the first set of resources and the quality of the first set of resources; an index of a resource subset from the first set of resources and a quality of the resource subset that exceeds a predefined threshold; or a difference in quality of the first set of resources relative to a reference quality of the first set of resources; At least one of: The method of claim 13.

15. The first set of resources includes a set of reference signal resources, and the first configuration indicates to the terminal device to transmit a sounding reference signal on the set of reference signal resources; Transmitting the information includes: transmitting the sounding reference signal to the network device on the set of reference signal resources; Including, The method of claim 1.

16. receiving an indication of an updated subset of resources from the first set of resources from the network device; the updated resource subset is output from the first AI or ML model; Receiving the instruction includes: receiving the first configuration from the network device, the first configuration including the indication of the updated resource subset; or receiving from the network device a second configuration for inference of the AI ​​or ML model including the indication of the updated resource subset; Including, The method of claim 1.

17. Based on the number of receive beams applied during inference of the first AI or ML model, The number of reference signal resources configured in the reference signal resource set; Measurement period requirements, or a measurement accuracy requirement is determined; The number of receive beams applied during the inference of the first AI or ML model is indicated by the network device or determined by the terminal device, or the number of receive beams applied during the inference of the first AI or ML model is output by the first AI or ML model.

17. The method according to any one of claims 1 to 16.

18. The number of channel state information (CSI) processing units and the CSI calculation time depend on the number of resources whose beam quality is actually measured in the first set of resources. The method of claim 1.

19. 1. A method of communication comprising: Sending a first configuration to a terminal device in a network device, the first configuration indicating at least one resource subset from a first set of resources, the at least one resource subset being for constructing a first dataset for training a first artificial intelligence (AI) or machine learning (ML) model in the network device, the first dataset including at least one of a training dataset for model training, a validation dataset for model training, a test dataset for model training, or a dataset for model inference; receiving information from the terminal device regarding a quality decision based on the at least one resource subset, the information being for constructing the first data set; Including, method.

20. the first set of resources includes at least a resource for beam measurement and reporting; Receiving the information includes: receiving from the terminal device the information regarding a reference signal (RS) quality on the at least one resource subset; The method comprises: training the first AI or ML model based on the RS quality of a set of beam pairs; Further comprising:

20. The method of claim 19.

21. The first set further includes a first candidate data sample for constructing the first data set; Receiving the information includes: receiving a first report from the terminal device, the first report including a first report quantity indicating whether the first candidate data sample is suitable for constructing the first data set for the first AI or ML model; Including, 20. The method of claim 19.

22. the first candidate data sample includes a resource subset as an input of the first AI or ML model and a target resource as an output of the first AI or ML model; 22. The method of claim 21.

23. Receiving the information includes: receiving a first report from the terminal device, the first report including a first report quantity indicative of a result of a comparison between a quality of the resource subset and a quality of the target resource; Including, 23. The method of claim 22.

24. The first report amount is: a first index of the target resource having the quality of the target resource and an index of the resource subset having the quality of the resource subset; A second index of the best resource determined by the terminal device; an index of the measurement resource subset determined by the terminal device; the measured quality of said best resource; or a measured quality of said measurement resource subset; Further comprising at least one of:

24. The method of claim 23.

25. The first configuration further includes a first candidate data sample for constructing the first data set, the first candidate data sample including a resource subset as an input for the first AI or ML model; The method comprises: sending a second configuration to the terminal device indicating target resources as an output of the first AI or ML model; Further comprising:

20. The method of claim 19.

26. Receiving the information includes: receiving a first report from the terminal device comprising a quality of the target resource; or receiving from the terminal device a first report including a first report quantity indicative of a comparison result between the quality of the target resource and the quality of the resource subset; Including, 26. The method of claim 25.

27. If a quality of at least one of the resource subsets is better than the quality of the target resource, the first reporting amount indicates that the first candidate data sample is not suitable for constructing the first data set for the first AI or ML model; If the quality of the target resource is better than all of the qualities of the resource subset, the first reporting amount indicates that the first candidate data sample is suitable for constructing the first data set for the first AI or ML model.

27. The method of claim 24 or 26.

28. the first set of resources comprises a set of reference signal resources, the at least one resource subset comprises a reference signal resource subset from the set of reference signal resources, and the first configuration further indicates a target reference signal resource; Receiving the information includes: receiving a set of sounding reference signals from the terminal device on the set of reference signal resources; receiving a target sounding reference signal from the terminal device on the target reference signal resource; Including, 20. The method of claim 19.

29. the first set of resources comprises a set of reference signal resources, and the at least one resource subset comprises a reference signal resource subset from the set of reference signal resources; Receiving the information includes: receiving a set of sounding reference signals from the terminal device on the set of reference signal resources; The method comprises: transmitting a second configuration to the terminal device indicating a target reference signal resource; receiving a target sounding reference signal from the terminal device on the target reference signal resource; Further comprising:

20. The method of claim 19.

30. The first configuration indicates to the terminal device to perform measurements on the first set of resources; Receiving the information includes: receiving a second report from the terminal device indicating a quality of the first set of resources; Including, 20. The method of claim 19.

31. The second report states: an index of the first set of resources and the quality of the first set of resources; an index of a resource subset from the first set of resources and a quality of the resource subset that exceeds a predefined threshold; or a difference in quality of the first set of resources relative to a reference quality of the first set of resources; At least one of: The method of claim 30.

32. The first set of resources includes a set of reference signal resources, and the first configuration indicates to the terminal device to transmit a sounding reference signal on the set of reference signal resources; Receiving the information includes: receiving the sounding reference signal from the terminal device on the set of reference signal resources; Including, 20. The method of claim 19.

33. Obtaining an updated resource subset from an output of the first AI or ML model; transmitting an indication of the updated resource subset to the terminal device; Further comprising: Transmitting the instruction includes: transmitting the first configuration to the terminal device, the first configuration including the indication of the updated resource subset; or transmitting to the terminal device a second configuration for inference of the first AI or ML model, the second configuration including the indication of the updated resource subset; Including, 20. The method of claim 19.

34. Based on the number of receive beams applied during inference of the first AI or ML model, The number of reference signal resources configured in the reference signal resource set; Measurement period requirements, or a measurement accuracy requirement is determined; The number of receive beams applied during the inference of the first AI or ML model is indicated by the network device or determined by the terminal device, or the number of receive beams applied during the inference of the first AI or ML model is output by the first AI or ML model.

34. A method according to any one of claims 1 to 33.

35. A processor; a memory coupled to the processor and storing instructions; A terminal device comprising: The instructions, when executed by the processor, perform operations including the method of any one of claims 1 to 18. Terminal device.

36. A processor; a memory coupled to the processor and storing instructions; A network device comprising: The instructions, when executed by the processor, perform operations including the method of any one of claims 19 to 34. Network device.

37. storing instructions which, when executed on at least one processor, cause said at least one processor to perform a method according to any one of claims 1 to 18 or any one of claims 19 to 34; Computer-readable medium.

Images