Beam management method and system based on digital twinning

Abstract

The application provides a beam management method based on digital twinning, which comprises the following steps: in a QoS reporting period, reporting the measurement quality of all beams and QoS requirements to a base station as reporting data at a user end; feeding back the reporting data to a digital twinning network (DTN) by the base station; inputting the reporting data into a beam measurement sampling rate prediction model for inference by the DTN, determining the beam measurement sampling rate and the corresponding beam prediction model, and delivering them to the base station, which then delivers them to the user end; performing quality measurement on the corresponding beams based on the beam measurement sampling rate at the user end, and performing beam prediction based on the beam measurement quality by using the beam prediction model at the user end or the base station to obtain a plurality of optimal measurement quality beams; integrating all optimal measurement quality beams reported by the user end served by all base stations by the base station, selecting an optimal beam by using a current resource scheduling algorithm, and indicating the optimal beam to the user end. The application adopting the above scheme can flexibly adapt the beam measurement mode according to customer requirements.

Description

Technical Field

[0001] This application relates to the field of wireless network technology, and in particular to a beam management method and system based on digital twins. Background Technology

[0002] To ensure coverage performance of wireless networks in the millimeter-wave band, base stations and users interact through narrow-angle shaped beams. Beam management measures beam pairs in different directions and selects the optimal beam pair to guarantee the quality of interaction between the base station and the user. 5G NR significantly improves wireless network coverage performance in the millimeter-wave band through beam management technology. To further reduce terminal overhead while ensuring beam management performance, the beam management mechanism has become an important research topic.

[0003] To better standardize 5G NR beam management technology, 3GPP initiated a research project on beam management. One basic component of beam management was standardized, specifically including the following aspects:

[0004] Beam scanning: Beams from different directions are time-division multiplexed to cover a specific area. Each beam carries signals such as CSI-RS and SSB. Through beam scanning, users can obtain the signals carried by beams from different directions.

[0005] Beam measurement: The user measures the reference signal carried by the received beam and obtains the beam quality in that direction by calculating the signal quality of the reference signal.

[0006] Beam Reporting: Users report measurement information of the reference signal carried by the beam. The measurement information should include at least the reference signal ID and the corresponding measurement quality. Measurement quality includes L1-RSRP or L1-SINR.

[0007] Beam determination: The base station and the user select the transmit / receive beams. In connected mode, the base station should determine the transmit beam based on the measurement information fed back by the user and indicate the beam to the user.

[0008] With the continuous development and evolution of wireless networks, the introduction of digital twin technology into the communications industry to support the management and optimization of the entire network lifecycle has become a global industry consensus.

[0009] In existing beam management schemes, the model is trained on the user or terminal side by inputting a uniform number of beam pairs. Because the model is fixed, existing schemes cannot adapt to users flexibly adjusting beam measurement methods, requiring users or base stations to retrain the model, resulting in significant training overhead and failing to meet diverse service needs. Furthermore, the fixed model cannot handle diverse channel conditions. Since beam pair measurement quality may have different data characteristics under different channel conditions, a large number of beam pairs may be required to ensure the performance of model derivation when the user's channel conditions are poor. Summary of the Invention

[0010] This application aims to at least partially address one of the technical problems in the related art.

[0011] Therefore, the first objective of this application is to propose a beam management method based on digital twins, which solves the technical problem that existing methods cannot meet user needs. It can allocate models with different measurement overheads to users under different channel conditions in the communication network for beam prediction, effectively balancing the accuracy of beam prediction and beam measurement overhead under different channel conditions, and flexibly adapting the beam measurement method according to user needs.

[0012] The second objective of this application is to propose a beam management system based on digital twins.

[0013] To achieve the above objectives, a first aspect of this application proposes a beam management method based on digital twins, comprising:

[0014] Step S1: During the QoS reporting period, generate reporting data at the user end and report the reporting data to the base station. The reporting data includes the measurement quality and QoS requirements of all beams.

[0015] Step S2: The reported data is fed back to the digital twin network (DTN) via the base station;

[0016] Step S3: Input the reported data into the beam measurement sampling rate prediction model through DTN for inference, determine the beam measurement sampling rate, determine the beam prediction model corresponding to the beam measurement sampling rate, and send the beam measurement sampling rate and the corresponding beam prediction model to the base station;

[0017] Step S4: Send the beam measurement sampling rate to the user terminal via the base station, or send the beam measurement sampling rate and the corresponding beam prediction model to the user terminal via the base station;

[0018] Step S5: At the user end, perform quality measurement on the corresponding beam based on the beam measurement sampling rate, and at the user end or base station, use the beam prediction model to perform beam prediction based on the beam measurement quality to obtain at least one beam with optimal measurement quality.

[0019] Step S6: The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

[0020] Optionally, in one embodiment of this application, step S1 specifically includes:

[0021] The base station sends QoS reporting and beam measurement reporting period start information to the user terminal.

[0022] Determine the scanning and measurement sequence of the beam pair at the user end;

[0023] The user receives the period start information sent by the base station and measures the reference signals carried in all beams sent by the base station according to the scanning and measurement sequence of the beam pairs to obtain the measurement results of all beams.

[0024] On the user's end, based on the currently executed business, the user's QoS requirements for the current period are generated;

[0025] The user terminal reports all beam measurement results and the user's QoS requirements for the current period to the base station.

[0026] Optionally, in one embodiment of this application, before step S2, the following steps are further included:

[0027] The physical network of base stations and user terminals is simulated using DTN.

[0028] Optionally, in one embodiment of this application, before step S3, the following steps are further included:

[0029] The measurement quality and QoS requirements of all beams received within a preset time period are processed by the DTN to form the first training dataset;

[0030] The input portion of the first training dataset is sampled according to a specified sampling rate using DTN, and the first training dataset is divided into groups with different numbers of inputs according to the sampling rate, forming training datasets with different sampling rates.

[0031] The structure and parameters of the beam prediction model are determined by DTN, and the beam prediction model at different sampling rates is trained using training datasets grouped by different sampling rates.

[0032] The prediction accuracy of the beam prediction model within a preset time period is statistically analyzed using DTN to determine the optimal beam prediction model for each sampling rate under the corresponding QoS requirements, and a second training dataset is generated using the optimal beam prediction model.

[0033] The structure and parameters of the beam measurement sampling rate prediction model are determined by DTN, and a beam measurement sampling rate prediction model based on QoS requirements and channel conditions is trained based on the second training dataset.

[0034] Optionally, in one embodiment of this application, step S3 specifically includes:

[0035] By using DTN, the reported data in the current period is used as input data to the trained beam measurement sampling rate prediction model for inference, to predict the beam measurement sampling rate, and to determine the beam prediction model under that sampling rate.

[0036] The beam measurement sampling rate and the beam prediction model at that sampling rate are sent to the base station via the DTN.

[0037] Optionally, in one embodiment of this application, step S5 specifically includes:

[0038] At the user end, the quality of the beam pair is measured according to the beam measurement sampling rate specified by the sampling rate to obtain the measurement quality of the beam pair;

[0039] At the user end or base station, a beam prediction model is used to predict the minimum transmit beam ID with the best measurement quality based on the measurement quality of the beam pair.

[0040] Beam prediction at the user end includes:

[0041] The measurement quality of the beam pair is input into the beam prediction model for inference, predicting at least one transmit beam ID with optimal measurement quality, and reporting the predicted at least one transmit beam ID with optimal measurement quality to the base station; beam prediction is performed at the base station, including:

[0042] During the QoS reporting period, the user end reports the measurement quality of the beam pair to the base station. The base station then inputs the measurement quality of the beam pair into the beam prediction model for inference, predicting the minimum transmit beam ID with the best measurement quality.

[0043] To achieve the above objectives, a second aspect of the present invention proposes a beam management system based on digital twins, including a user terminal, a base station, and a DTN, wherein,

[0044] During the QoS reporting period, the user terminal generates reporting data and reports the data to the base station. The reporting data includes the measurement quality and QoS requirements of all beams.

[0045] The base station will report the data back to the digital twin network (DTN);

[0046] The DTN inputs the reported data into the beam measurement sampling rate prediction model for inference, determines the beam measurement sampling rate, determines the beam prediction model corresponding to the beam measurement sampling rate, and sends the beam measurement sampling rate and the corresponding beam prediction model to the base station.

[0047] The base station sends the beam measurement sampling rate to the user terminal, or the base station sends the beam measurement sampling rate and the corresponding beam prediction model to the user terminal;

[0048] The user terminal performs quality measurement on the corresponding beam based on the beam measurement sampling rate. The user terminal or base station uses the beam prediction model to perform beam prediction based on the beam measurement quality to obtain at least one beam with optimal measurement quality.

[0049] The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

[0050] Optionally, in one embodiment of this application, the DTN simulates the physical entity network of the base station and the user terminal.

[0051] The beam management method and system based on digital twins in this application, while ensuring beam management performance, processes the channel state and QoS reported by users through a digital twin network (DTN). Based on the different user states and requirements, it determines the beam prediction model to be used by the user, effectively reducing beam management overhead and latency while comprehensively considering the user's channel state and communication requirements. In this embodiment, based on the existing basic beam measurement process, the physical network consisting of the user and the base station is constructed as a DTN model. In each QoS requirement reporting cycle, the beam pair measurement quality and QoS requirement reporting information generated in the physical network are input into the DTN, and processed by the constructed DTN to determine the beam prediction model that the user should use under the given channel state and communication quality requirements. During this cycle, the user performs beam scanning and measurement according to the sampling rate corresponding to the model, and selects the optimal beam pair and its corresponding reference signal ID as the downlink transmission beam indication to the user based on the model. This embodiment uses a model that assigns different measurement overheads to users under different channel conditions in a communication network for beam prediction. It effectively balances the accuracy of beam prediction and beam measurement overhead under different channel conditions, and flexibly adapts the beam measurement method according to user needs.

[0052] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0053] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0054] Figure 1 This is a flowchart illustrating a beam management method based on digital twins provided in Embodiment 1 of this application.

[0055] Figure 2 This is a schematic diagram of the beam pair ID table of the beam management method based on digital twins according to an embodiment of this application;

[0056] Figure 3 This is a schematic diagram illustrating the training process of the beam prediction model in the DTN functional model of the beam management method based on digital twins in this application embodiment;

[0057] Figure 4 This is a schematic diagram illustrating the training process of the beam measurement sampling rate generation model based on QoS requirements and channel status in the DTN functional model of the beam management method based on digital twins in this application embodiment.

[0058] Figure 5 This is a schematic diagram illustrating the reasoning process of the beam measurement sampling rate generation model based on QoS requirements and channel status in the DTN functional model of the beam management method based on digital twin in this application embodiment.

[0059] Figure 6 This is a schematic diagram illustrating the inference process of the beam prediction model in the DTN functional model of the beam management method based on digital twins in this application embodiment;

[0060] Figure 7 This is a schematic diagram of a beam management system based on digital twin, provided as an embodiment of this application. Detailed Implementation

[0061] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0062] The beam management method and system based on digital twins according to embodiments of this application are described below with reference to the accompanying drawings.

[0063] Figure 1This is a flowchart illustrating a beam management method based on digital twins provided in Embodiment 1 of this application. This embodiment only considers the situation within a single user QoS requirement reporting cycle. The length of the QoS requirement reporting cycle may be related to the user's mobility, the user's location, etc. This embodiment does not discuss issues related to QoS requirement reporting cycle management.

[0064] like Figure 1 As shown, the beam management method based on digital twins includes the following steps:

[0065] Step S1: During the QoS reporting period, generate reporting data at the user end and report the reporting data to the base station. The reporting data includes the measurement quality and QoS requirements of all beams.

[0066] In this embodiment, step S1 includes:

[0067] Step S101: The base station sends QoS reporting and beam measurement reporting cycle start information to the user. After synchronization, the QoS reporting cycle begins, and the user is ready to start beam scanning.

[0068] Step S102: The user determines the scanning and measurement sequence of the beam pair;

[0069] In one embodiment, the user's measurement process for beam pairs involves traversing all received beams, and under each transmitted beam (i.e., the reference signal), traversing all received beams, and arranging all beam pairs in sequence to form a specific beam pair ID table.

[0070] In one embodiment, reference Figure 2 This is a schematic diagram of a beam pair ID table for a beam management method based on digital twins provided in an embodiment of this application. In this table, the beams are grouped according to the transmitted beams. Under each transmitted beam, all received beams are traversed in sequence to form beam pair IDs. Each beam pair ID corresponds to a specific transmitted beam and a specific received beam. At the same time, each beam pair ID corresponds to a beam pair measurement quality.

[0071] In step S103, the user receives the period start information sent by the base station, measures the reference signals carried in all beams sent by the base station, and obtains the measurement results of all beams.

[0072] In one embodiment, beam quality measurement is based on CSI-RS / SSB reference signals. During the base station's transmit beam scanning process, the base station sends CSI-RS / SSB reference signals to the user. The user measures the reference signals and obtains the beam quality of the transmit beam direction by calculating the signal quality of the reference signals.

[0073] In one embodiment, the user selects L1-RSRP or L1-SINR as the criterion for evaluating the quality of the reference signal.

[0074] Step S104: The user generates the user's QoS requirements for this cycle based on the currently executed service.

[0075] In one embodiment, the user selects either L1-RSRP or L1-SINR of the beam as an indicator of QoS.

[0076] In one embodiment, the user selects their own throughput as a QoS metric.

[0077] In step S105, the user summarizes the measurement information of the beam pairs and reports the measurement information and the user's QoS requirement information to the base station based on the QoS requirements generated by the service.

[0078] In one embodiment, the user also reports the locally accessible beam prediction model index, so that the DTN can subsequently decide whether to issue a new model to the user.

[0079] Step S2: The reported data is fed back to the digital twin network (DTN) via the base station;

[0080] Step S3: Input the reported data into the beam measurement sampling rate prediction model through DTN for inference, determine the beam measurement sampling rate, determine the beam prediction model corresponding to the beam measurement sampling rate, and send the beam measurement sampling rate and the corresponding beam prediction model to the base station;

[0081] In this embodiment, step S3 includes:

[0082] In step S301, the base station feeds back the user-reported data to the DTN. The DTN constructs a beam measurement quality dataset and groups it according to the sampling rate. Each group trains a model for beam prediction.

[0083] Furthermore, step S301 includes the following steps, see reference. Figure 3 This is a schematic diagram illustrating the training process of the beam prediction model in the DTN functional model of a beam management method based on digital twins, provided in an embodiment of this application. The process includes the following steps:

[0084] Step S3011: The DTN processes the user beam measurement quality information and corresponding QoS requirement information collected over a period of time to form a model training dataset.

[0085] In one embodiment, the DTN network comprises physical sectors served by multiple base stations, and the DTN's database contains data collected by all base stations in the network. The DTN processes the reported beam measurement quality information and QoS requirement information collected by the base stations to form a dataset for model training and derivation. The dataset includes at least one of the following: user ID, measurement timestamp, beam pair ID table, measurement quality corresponding to the beam pair ID, and QoS requirement information. No distinction is made between base stations for user information collected from different base stations.

[0086] In one embodiment, the user ID represents the user performing beam measurement, the measurement timestamp represents the time when the user performs beam measurement, the beam pair ID table represents the relative position of each beam pair among all the user's beam pairs, the measurement quality corresponding to the beam pair ID represents the reference signal quality measured by the user on each beam pair, and the QoS requirement represents the QoS requirement generated by the user in the current period based on the current service.

[0087] In step S3012, DTN samples the input portion of the dataset according to the specified sampling rate, and divides the dataset into groups with different numbers of inputs according to the sampling rate, forming multiple training datasets.

[0088] In one embodiment, the DTN, based on the beampair ID table, identifies measurement beampairs belonging to the same user. Assuming there are a total of n measurement beampairs for a user, the DTN divides the beam measurement quality dataset into m groups according to different desired sampling rates (assuming there are m sampling rates), with each group having a sampling rate of k. i (0 <k i <1) Each group contains nk i Each beam measurement information group's beam measurement quality data is sampled at equal intervals from all beam pair measurement quality data.

[0089] In step S3013, DTN determines the model structure and model parameters, and trains beam prediction models at different sampling rates using training datasets grouped by different sampling rates.

[0090] In one embodiment, for each group, the DTN trains a beam prediction neural network model to recover the measurement quality of all transmitted beams based on the measurement quality of a subset of beam pairs. The DTN can determine the model structure of the training model based on the sampling rate of beam measurements (i.e., the number of beam pair measurement quality values ​​input to the model) and the specific training task characteristics and requirements of each group.

[0091] In one embodiment, the model structure of the training model can be determined as follows.

[0092] For determining the number of layers and nodes in the model, you can refer to setting the number of input layer nodes as A, representing the quantity of measurement beam pairs in the input model. This value is related to the sampling rate k of each group model. i The number of nodes in the input layer is related to the total number of beam pairs N. The higher the sampling rate, the more beam pairs the user measures, and the larger the number of nodes in the input layer should be. The number of nodes in the output layer is set to B, which depends on the total number of transmitted beams B. The larger the total number of transmitted beams, the larger the number of nodes in the output layer should be. The number of nodes in each hidden layer and the number of hidden layers need to take into account factors such as model size and model generalization ability.

[0093] For determining the inter-layer connection method, you can refer to the following: if the hidden layer and the input layer are fully connected, the activation function can be the ReLU function; if the hidden layers are fully connected, the activation function can be the ReLU function; if the hidden layer and the output layer are fully connected, the activation function can be the ReLU function.

[0094] The process of determining the loss function can include using the mean squared error (MSE) loss function, the mean absolute error (MAE) loss function, the Huber loss function, etc.

[0095] For determining the hyperparameters of the network model, you can refer to setting the number of learning epochs to T. The setting of the number of learning epochs needs to consider the impact on model training speed, training cost, and model training accuracy; the learning rate is set to l; and random weight initialization is selected as the weight initialization method.

[0096] In one embodiment, the base station calculates the training loss value based on the output of each group model and the label value information, for example, using the mean squared error loss function to calculate the training loss value:

[0097]

[0098] Where I represents the amount of training data for the model, and y i This represents the output result of the model after data i is processed. This represents the label value of data i.

[0099] In one embodiment, the base station updates the parameters of each beamp-pair grouping model based on the training loss value, the model update method, and the selected hyperparameters, such as stochastic gradient descent (SGD) or Adam, and updates the parameters of the specific model layer, for example, using the SGD algorithm to update the beamp-pair grouping model parameters.

[0100]

[0101] in, This represents the beam prediction model parameters to be updated in round t. This represents the beam pair grouping model parameters after the t-th round update. This represents the gradient of the training loss value calculated in round t. Let t represent the learning rate in round t.

[0102] Step S3014: Based on the prediction accuracy during model training, DTN generates a dataset that can correctly predict beams for each sampling rate under the corresponding QoS requirements.

[0103] In one embodiment, the DTN trains m beam prediction models at different sampling rates. During training, for a sample, if the model at that sampling rate can successfully predict the correct optimal beam, and the measurement results of the optimal beam meet the user's QoS requirements, then the model corresponding to that sampling rate is considered to be included in the dataset label of the correctly predicted beam model. A new sample is then generated, with the user's QoS requirements as input data, and the measurement results of the top L optimal beams obtained from beam scanning. The output label of the sample is a 0 / 1 vector of size m, indicating whether the beam prediction model at different sampling rates correctly predicts and meets the user's QoS requirements.

[0104] In step S3015, the DTN determines whether to distribute the model and where to distribute it based on the request situation in the physical network.

[0105] In one embodiment, the DTN saves the model trained in step 204 above, and predicts the sampling rate model that the reporting user needs to use based on the sampling rate prediction model described later. If the user does not have a beam prediction model corresponding to the sampling rate, or has released the model, the beam prediction model corresponding to the sampling rate is sent to the user, and the sampling rate used by the user in the beam scanning and measurement process is updated.

[0106] In step S302, the DTN performs statistical analysis on the prediction accuracy of the beam prediction model, generates a sampling rate prediction model dataset, and trains the sampling rate prediction model.

[0107] Furthermore, step S302 includes the following steps, see reference. Figure 4 This is a schematic diagram illustrating the training process of a beam measurement sampling rate generation model based on QoS requirements and channel state in a beam management method based on digital twins provided in this application embodiment. The process includes the following steps:

[0108] Step S3021: The DTN performs statistical analysis on the prediction accuracy of the beam prediction model over a period of time, and generates a dataset that can correctly predict the beam under the corresponding QoS requirements for each sampling rate.

[0109] In one embodiment, step S3021 is the same as step S3014 described in step S301 above.

[0110] In step S3022, the DTN determines the model structure and model parameters, and trains a beam measurement sampling rate generation model based on QoS requirements and channel conditions using the dataset generated by the beam prediction model.

[0111] In one embodiment, the model structure and parameters of the training model can be determined as follows.

[0112] For determining the number of layers and nodes in the model, you can refer to setting the number of nodes in the input layer to L+1, representing the measurement quality of the optimal L measurement beam pairs in the input model and the QoS requirements reported by one user; the number of nodes in the output layer is set to m, which depends on the total number m of beam prediction models with different sampling rates. The larger the number of distinctions between sampling rates, the larger the number of nodes in the output layer should be. The number of nodes in each hidden layer and the number of hidden layers need to take into account factors such as model size and model generalization ability.

[0113] The process of determining the inter-layer connection method can be based on the following: the hidden layer and the input layer are fully connected, and the activation function can be the ReLU function; the hidden layers are fully connected, and the activation function can be the ReLU function; the hidden layer and the output layer are fully connected, and the activation function can be the Sigmoid function. This will output the probability that the model can successfully predict the correct beam and meet the user's QoS requirements for each sampling rate.

[0114] The process of determining the loss function can employ methods such as the binary cross-entropy loss function.

[0115] For determining the hyperparameters of the network model, you can refer to setting the number of learning epochs to T. The setting of the number of learning epochs needs to consider the impact on model training speed, training cost, and model training accuracy; the learning rate is set to l; and random weight initialization is selected as the weight initialization method.

[0116] In one embodiment, the base station calculates the training loss value based on the output of each group model and the label value information, for example, using the binary cross-entropy loss function to calculate the training loss value:

[0117]

[0118] Where I represents the amount of training data for the model, y represents the true label of data i, which takes the value of 0 or 1, and p represents the probability value of data i predicted by the model.

[0119] In one embodiment, the base station updates the parameters of each beamp-pair grouping model based on the training loss value, the model update method, and the selected hyperparameters, such as stochastic gradient descent (SGD) or Adam, and updates the parameters of the specific model layer, for example, using the SGD algorithm to update the beamp-pair grouping model parameters.

[0120]

[0121] in, This represents the sampling rate prediction model parameters to be updated in round t. This represents the beam pair grouping model parameters after the t-th round update. This represents the gradient of the training loss value calculated in round t. Let t represent the learning rate in round t.

[0122] Step S3023: The DTN decides whether to distribute the model based on the request situation in the physical network.

[0123] In one embodiment, the DTN saves the model trained in step S3022 above and calls the model at the beginning of each QoS reporting period. Based on the beam measurement information and QoS requirements reported by the user, the DTN predicts the sampling rate model that the reporting user needs to use. If the user does not have a model corresponding to the sampling rate, or has released the model, the DTN sends the beam prediction model corresponding to the sampling rate to the user and updates the sampling rate used by the user in the beam scanning and measurement process.

[0124] Having completed the above steps, the data set specifications and generation process for the beam prediction model and sampling rate prediction model in this embodiment have been described. The training process for both models has also been explained. After the models are trained and deployed, the subsequent steps will describe the model inference process, i.e., its usage, in detail.

[0125] In this embodiment, step S3 further includes:

[0126] In step S303, the DTN uses the measurement quality and QoS requirements reported by the user to infer the beam scanning and measurement sampling rate for the current period and sends out the beam prediction model.

[0127] Furthermore, step S303 includes the following steps, see reference. Figure 5 This is a flowchart illustrating the reasoning process of the beam measurement sampling rate generation model based on QoS requirements and channel state in the DTN functional model of a beam management method based on digital twins provided in this application embodiment. The flowchart includes the following steps:

[0128] In step S3031, the user performs quality measurements on all beam pairs at the start of its QoS reporting period, generates QoS requirements based on the current service status, and reports both to the base station.

[0129] In step S3032, the base station transfers the reported data to the DTN. The DTN calls the sampling rate generation function model, uses the model to infer the sampling rate for the user based on the reported data, and instructs the inferred sampling rate to the base station.

[0130] In one embodiment, the DTN processes the data to be processed reported by the base station according to the method in step S3011 to generate the input data to be inferred for the sampling rate prediction model; the processed data is sent into the sampling rate prediction model for inference to generate the sampling rate that the user should use and the beam prediction model corresponding to the sampling rate; and the sampling rate obtained by inference is indicated to the base station.

[0131] In step S3033, the base station determines the sampling rate used by the user during beam scanning and measurement according to the DTN instruction, and instructs the user on the sampling rate.

[0132] In one embodiment, model deployment and inference are performed on the user side. The base station sends the beam measurement sampling rate used in that period, as indicated by the DTN, to the user. Furthermore, the base station checks the user-reported local beam prediction model index; if the user has the model locally, it does not send it. If the user does not have the model locally, the base station sends the model to the user.

[0133] In one embodiment, model deployment and inference are performed at the base station. The base station indicates the beam measurement sampling rate for that period to the user, without needing to distribute the model.

[0134] Step S4: Send the beam measurement sampling rate to the user terminal via the base station, or send the beam measurement sampling rate and the corresponding beam prediction model to the user terminal via the base station;

[0135] Step S5: At the user end, perform quality measurement on the corresponding beam based on the beam measurement sampling rate, and at the user end or base station, use the beam prediction model to perform beam prediction based on the beam measurement quality to obtain at least one beam with optimal measurement quality.

[0136] In this embodiment, step S5 includes the following steps, refer to Figure 6 This is a flowchart illustrating the beam prediction model inference process in the DTN functional model of a beam management method based on digital twins, provided by an embodiment of the present invention. The flowchart includes the following steps:

[0137] In step S501, the user performs quality measurements on a portion of the beam pairs specified by the determined sampling rate to obtain the measurement quality of the portion of the beam pairs.

[0138] In one embodiment, the user measures all transmitted beams (reference signals) but samples from the received beams. Measuring only certain received beams corresponding to all transmitted beams reduces measurement overhead and provides measurement quality for the corresponding beam pairs.

[0139] In one embodiment, the user measures a portion of the transmitted beam (reference signal) but all received beams. Measuring only the portion of the transmitted beams corresponding to all received beams reduces measurement overhead and provides measurement quality for the corresponding beam pairs.

[0140] In one embodiment, the user measures a portion of the transmitted beam (reference signal) and samples from the received beam. Measuring both the portion of the transmitted beam and the corresponding portion of the received beam reduces measurement overhead and yields the measurement quality of the corresponding beam pair.

[0141] Step S502: Based on the DTN model distribution, perform beam prediction on the user side or base station side using the trained model.

[0142] In one embodiment, the model deployment and inference are completed on the user side. After each beam measurement in the QoS reporting cycle, the user substitutes the beam measurement quality into the corresponding beam prediction model for inference, predicting at least one transmit beam ID (reference signal) with the best measurement quality.

[0143] In one embodiment, the model deployment and inference are completed at the base station. After each beam measurement in the QoS reporting cycle, the user reports the beam measurement quality to the base station. Upon receiving the beam measurement quality data, the base station substitutes it into the corresponding beam prediction model for inference, predicting at least one transmit beam ID (reference signal) with the best measurement quality for the user.

[0144] In step S503, the base station obtains at least one reference signal ID with the best user measurement quality.

[0145] In one embodiment, model deployment and inference are completed on the user side. The user then reports at least one transmit beam ID (reference signal) with the best predicted measurement quality to the base station. The base station records the predicted reference signal ID reported by the user.

[0146] In one embodiment, the model deployment and inference are completed at the base station side. Then, after the base station locally predicts at least one transmit beam ID that provides the best measurement quality for the user, it records at least one reference signal ID that provides the best measurement quality for that user.

[0147] Step S6: The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

[0148] In this embodiment, step S6 includes:

[0149] The base station selects beams based on the deployed resource scheduling algorithm to maximize system performance. It also indicates the optimal beam (reference signal) ID to each user and schedules users according to the algorithm in each time slot.

[0150] The beam management method based on digital twins in this application involves the DTN training a unified sampling rate prediction neural network model for users serving multiple physical base stations, based on user beam measurement quality data and QoS requirement data. This model predicts the beam scanning sampling rate that the user should use in a given period based on the QoS requirements reported by the user and the measured quality of the L optimal beams. This reduces the resource overhead of beam scanning while ensuring beam prediction accuracy. Simultaneously, a set of beam prediction neural network models divided by different sampling rates are trained to predict the ID of at least one transmit beam (reference signal) with the best measurement quality at the corresponding sampling rate, based on the measurement quality of some beam pairs. This avoids the large reference signal resource consumption and significant latency caused by users measuring all beam pairs, effectively reducing beam management overhead while maintaining performance. This embodiment can flexibly adapt to user measurement methods and adjust the model's input and output to meet diverse service requirements.

[0151] In this embodiment, the base station trains a neural network model based on AI technology. Users only need to measure the beam quality of a few beam pairs to use the neural network model to predict the beam quality of all beam pairs, thereby reducing the overhead and latency of beam management.

[0152] In this embodiment, the base station is based on digital twin technology, using AI models to manage AI models, considering the actual QoS requirements of users, and using the models to transform the requirements into beam scanning sampling rates and corresponding beam prediction models. This flexibly meets user needs and accuracy while reducing beam management overhead.

[0153] To implement the above embodiments, this application also proposes a beam management system based on digital twins.

[0154] Figure 7 This is a schematic diagram of a beam management system based on digital twin, provided as an embodiment of this application.

[0155] like Figure 7 As shown, this digital twin-based beam management system includes a user terminal, a base station, and a DTN, wherein...

[0156] During the QoS reporting period, the user terminal generates reporting data and reports the data to the base station. The reporting data includes the measurement quality and QoS requirements of all beams.

[0157] The base station will report the data back to the digital twin network (DTN);

[0158] The DTN inputs the reported data into the beam measurement sampling rate prediction model for inference, determines the beam measurement sampling rate, determines the beam prediction model corresponding to the beam measurement sampling rate, and sends the beam measurement sampling rate and the corresponding beam prediction model to the base station.

[0159] The base station sends the beam measurement sampling rate to the user terminal, or the base station sends the beam measurement sampling rate and the corresponding beam prediction model to the user terminal;

[0160] The user terminal performs quality measurement on the corresponding beam based on the beam measurement sampling rate. The user terminal or base station uses the beam prediction model to perform beam prediction based on the beam measurement quality to obtain at least one beam with optimal measurement quality.

[0161] The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

[0162] Furthermore, in this embodiment of the application, the DTN simulates the physical entity network of the base station and the user terminal.

[0163] It should be noted that the foregoing explanation of the embodiment of the beam management method based on digital twins also applies to the beam management system based on digital twins in this embodiment, and will not be repeated here.

[0164] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0165] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0166] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0167] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0168] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0169] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0170] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0171] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A beam management method based on digital twins, characterized in that, include: Step S1: During the QoS reporting period, generate reporting data at the user end and report the reporting data to the base station, wherein the reporting data includes the measurement quality and QoS requirements of all beams; Step S2: The reported data is fed back to the digital twin network (DTN) through the base station; Step S3: Input the reported data into the beam measurement sampling rate prediction model through DTN for inference, determine the beam measurement sampling rate, determine the beam prediction model corresponding to the beam measurement sampling rate, and send the beam measurement sampling rate and the corresponding beam prediction model to the base station; Step S4: The beam measurement sampling rate is sent to the user terminal via the base station, or the beam measurement sampling rate and the corresponding beam prediction model are sent to the user terminal via the base station; Step S5: At the user end, quality measurement is performed on the corresponding beam based on the beam measurement sampling rate, and at the user end or base station, beam prediction is performed based on the beam measurement quality using the beam prediction model to obtain at least one beam with optimal measurement quality. Step S6: The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

2. The method as described in claim 1, characterized in that, Step S1 specifically includes: The base station sends QoS reporting and beam measurement reporting period start information to the user terminal. Determine the scanning and measurement sequence of the beam pair at the user end; The user receives the period start information sent by the base station and measures the reference signals carried in all beams sent by the base station according to the scanning and measurement sequence of the beam pairs to obtain the measurement results of all beams. On the user's end, based on the currently executed business, the user's QoS requirements for the current period are generated; The user terminal reports all beam measurement results and the user's QoS requirements for the current period to the base station.

3. The method as described in claim 1, characterized in that, Before step S2, the following is also included: The physical network of base stations and user terminals is simulated using DTN.

4. The method as described in claim 1, characterized in that, Before step S3, the following is also included: The measurement quality and QoS requirements of all beams received within a preset time period are processed by the DTN to form the first training dataset; The input portion of the first training dataset is sampled at a specified sampling rate using DTN, and the first training dataset is divided into groups with different numbers of inputs according to the sampling rate, forming training datasets with different sampling rates. The structure and parameters of the beam prediction model are determined by DTN, and the beam prediction model at different sampling rates is trained using the training datasets grouped by the different sampling rates. The prediction accuracy of the beam prediction model within a preset time period is statistically analyzed using DTN to determine the optimal beam prediction model for each sampling rate under the corresponding QoS requirements, and a second training dataset is generated using the optimal beam prediction model. The structure and parameters of the beam measurement sampling rate prediction model are determined by DTN, and a beam measurement sampling rate prediction model based on QoS requirements and channel conditions is trained based on the second training dataset.

5. The method as described in claim 4, characterized in that, Step S3 specifically includes: By using DTN, the reported data in the current period is used as input data to the trained beam measurement sampling rate prediction model for inference, to predict the beam measurement sampling rate, and to determine the beam prediction model under that sampling rate. The beam measurement sampling rate and the beam prediction model at that sampling rate are sent to the base station via the DTN.

6. The method as described in claim 1, characterized in that, Step S5 specifically includes: At the user end, the quality of the beam pair is measured according to the beam measurement sampling rate specified by the sampling rate to obtain the measurement quality of the beam pair; At the user end or base station, the beam prediction model is used to perform beam prediction based on the measurement quality of the beam pair, and the minimum transmit beam ID with the best measurement quality is predicted. Beam prediction at the user end includes: The measurement quality of the beam pair is input into the beam prediction model for inference, predicting at least one transmit beam ID with optimal measurement quality, and reporting the predicted at least one transmit beam ID with optimal measurement quality to the base station; beam prediction is performed at the base station, including: During the QoS reporting period, the user end reports the measurement quality of the beam pair to the base station. The base station then inputs the measurement quality of the beam pair into the beam prediction model for inference, predicting the minimum transmit beam ID with the best measurement quality.

7. A beam management system based on digital twin, characterized in that, This includes the user terminal, base station, and DTN, among which, During the QoS reporting period, the user terminal generates reporting data and reports the reporting data to the base station. The reporting data includes the measurement quality and QoS requirements of all beams. The base station feeds back the reported data to the digital twin network (DTN); The DTN inputs the reported data into the beam measurement sampling rate prediction model for inference, determines the beam measurement sampling rate, determines the beam prediction model corresponding to the beam measurement sampling rate, and sends the beam measurement sampling rate and the corresponding beam prediction model to the base station. The base station sends the beam measurement sampling rate to the user terminal, or the base station sends the beam measurement sampling rate and the corresponding beam prediction model to the user terminal; The user terminal performs quality measurement on the corresponding beam based on the beam measurement sampling rate. The user terminal or base station uses the beam prediction model to perform beam prediction based on the beam measurement quality to obtain at least one beam with optimal measurement quality. The base station integrates the best measurement quality beams reported by all base station service users, selects the best beam using the current resource scheduling algorithm, and indicates the best beam to the user.

8. The system as described in claim 7, characterized in that, DTN simulates the physical network of base stations and user terminals.

Images