Positioning method and apparatus, device, program product, and storage medium

Abstract

The application discloses a positioning method and device, equipment, program product and storage medium. The method comprises the following steps: receiving first information sent by a service consumer, wherein the first information is used for requesting to subscribe to a positioning analysis service; performing positioning analysis based on a first network model according to the first information, obtaining a positioning result, and sending the positioning result to the service consumer.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a positioning method, apparatus, device, program product, and storage medium. Background Technology

[0002] In recent years, with the rapid development of the low-altitude industry, the demand for high-precision positioning in connected terminal devices such as intelligent driving vehicles and drones has become increasingly urgent. However, along with this growing demand, the challenges are also becoming more prominent. For example, how can we maximize the positioning accuracy of connected terminals in complex environments? How can we effectively assess the reliability of positioning results and ensure the stability of positioning capabilities in various scenarios? Therefore, achieving terminal positioning remains a problem that urgently needs improvement. Summary of the Invention

[0003] To address the aforementioned technical problems, this application provides a positioning method, apparatus, device, program product, and storage medium.

[0004] In a first aspect, the positioning method provided in this application is applied to a first network function, the method comprising:

[0005] Receive the first message sent by the service consumer, which is used to request subscription to the location analysis service;

[0006] Based on the first information and the first network model, a location analysis is performed to obtain the location result, which is then sent to the service consumer.

[0007] Secondly, the positioning method provided in this application is applied to a second network function, and the method includes:

[0008] Receive the second information sent by the first network function, the second information being used to request subscription to the training of the first network model;

[0009] The first network model is trained using the dataset, and the trained first network model is sent to the first network function.

[0010] Thirdly, the positioning method provided in this application is applied to a third network function, and the method includes:

[0011] Receive the third information sent by the second network function. The third information is used to request subscription to the dataset. The dataset is used to train the first network model. The first network model is used for localization analysis.

[0012] Determine whether the dataset needs to be collected based on the third piece of information, and obtain the determination result.

[0013] Fourthly, the positioning device provided in this application is applied to a first network function, and the device includes:

[0014] The first communication unit is used to receive first information sent by the service consumer, the first information being used to request subscription to the location analysis service;

[0015] The first processing unit is used to perform localization analysis based on the first information and the first network model to obtain the localization result.

[0016] The first communication unit is used to send the positioning result to the service consumer.

[0017] Fifthly, the positioning device provided in this application is applied to a second network function, and the device includes:

[0018] The second communication unit is used to receive the second information sent by the first network function. The second information is used to request subscription to the training of the first network model.

[0019] The second processing unit is used to train the first network model using the dataset;

[0020] The second communication unit is used to send the trained first network model to the first network function.

[0021] Sixthly, the positioning device provided in this application is applied to a third network function, and the device includes:

[0022] The third communication unit is used to receive third information sent by the second network function. The third information is used to request subscription to the dataset. The dataset is used to train the first network model. The first network model is used for localization analysis.

[0023] The third processing unit is used to determine whether the dataset needs to be collected based on the third information and to obtain the determination result.

[0024] Seventhly, the positioning device provided in this application includes: a processor and a memory, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to execute any of the above-described positioning methods.

[0025] Eighthly, this application provides a computer program product comprising: a computer program that, when executed by a processor, implements any of the above-described positioning methods.

[0026] Ninthly, the computer-readable storage medium provided in this application is used to store a computer program that causes a computer to perform any of the above-described positioning methods.

[0027] In the technical solution of this application, a first network function receives first information sent by a service consumer, the first information being used to request subscription to a location analysis service; based on the first information and a first network model, it performs location analysis to obtain a location result, and then sends the location result back to the service consumer. Thus, the first network function performs location analysis through the first network model, obtains the location result and the reliable range of the location result, thereby improving the accuracy of the location. Attached Figure Description

[0028] The accompanying drawings, which are provided to further illustrate this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application.

[0029] Figure 1 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 1 ;

[0030] Figure 2 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 2 ;

[0031] Figure 3 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 3 ;

[0032] Figure 4 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 4 ;

[0033] Figure 5 This is a schematic diagram of the BNN model architecture provided in the embodiments of this application;

[0034] Figure 6 This is a schematic diagram of the structural composition of the positioning device provided in the embodiments of this application. Figure 1 ;

[0035] Figure 7 This is a schematic diagram of the structural composition of the positioning device provided in the embodiments of this application. Figure 2 ;

[0036] Figure 8 This is a schematic diagram of the structural composition of the positioning device provided in the embodiments of this application. Figure 3 ;

[0037] Figure 9 This is a schematic structural diagram of a positioning device provided in an embodiment of this application;

[0038] Figure 10 This is a schematic structural diagram of the chip according to an embodiment of this application. Detailed Implementation

[0039] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0040] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0041] It should also be noted that the terms "first," "second," and "third" used in the embodiments of this application are only used to distinguish similar objects and do not represent a specific order of objects. It is understood that "first," "second," and "third" can be interchanged in a specific order or sequence where permissible, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. It should also be understood that the "instruction" mentioned in the embodiments of this application can be a direct instruction, an indirect instruction, or an indication of an association relationship. For example, A instructing B can mean that A directly instructs B, for example, B can be obtained through A; it can also mean that A indirectly instructs B, for example, A instructs C, and B can be obtained through C; or it can mean that there is an association relationship between A and B. It should also be understood that the term "correspondence" mentioned in the embodiments of this application may indicate a direct or indirect correspondence between the two, or an association between the two, or a relationship of instruction and being instructed, configuration and being configured, etc.

[0042] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and they all fall within the protection scope of the embodiments of this application.

[0043] In recent years, with the rapid development of the low-altitude industry, the demand for high-precision positioning in connected terminal devices such as intelligent driving vehicles and drones has become increasingly urgent. However, along with this growing demand, the challenges are also becoming more prominent: how to improve the positioning accuracy of connected terminals in complex environments as much as possible? How to effectively assess the reliability of positioning results and ensure the stability of positioning capabilities in various scenarios? Currently, there are already a variety of connected terminal positioning technologies available.

[0044] One category is traditional positioning methods, such as Real-Time Kinematic (RTK) platforms, sensing base stations, LiDAR, and millimeter-wave radar. These methods directly acquire data and perform positioning through sensors, providing high accuracy and real-time performance. However, the accuracy of these technologies may decrease when faced with signal interference, complex terrain, or environmental obstructions.

[0045] Another type is the artificial intelligence (AI)-based positioning method. This type of method typically relies on target detection technology and uses image data or radar data collected by connected terminals for positioning. AI technology analyzes sensor data through deep learning algorithms, which can extract complex features from multi-dimensional information, thereby improving the accuracy and adaptability of positioning, especially performing well in non-line-of-sight or high-interference environments.

[0046] However, current methods for locating connected terminals have the following main problems:

[0047] (1) Most AI-based network terminal positioning technologies are deployed locally on the terminal, which is limited by the scarcity of data resources and the computing power of the equipment. Terminal devices face significant computing pressure when processing large amounts of data, which not only affects the training efficiency of the model, but also hinders the continuous optimization and deployment of the model;

[0048] (2) Existing network-connected terminal positioning technologies typically only provide fixed positioning results and cannot effectively assess the reliability of these results, lacking a systematic analysis of errors and uncertainties. This limitation of lacking reliability assessment may lead to significant positioning deviations in complex or dynamic environments.

[0049] Therefore, improving the accuracy of positioning becomes a problem that needs to be considered. To this end, the following technical solutions based on embodiments of this application are proposed.

[0050] To facilitate understanding of the technical solutions of the embodiments of this application, the technical solutions of this application are described in detail below through specific embodiments. The above-mentioned related technologies are optional solutions and can be arbitrarily combined with the technical solutions of the embodiments of this application, all of which fall within the protection scope of the embodiments of this application. The embodiments of this application include at least some of the following contents.

[0051] Figure 1 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 1 Applied to the first network function, such as Figure 1 As shown, the positioning method includes the following steps:

[0052] Step 101: Receive the first message sent by the service consumer. The first message is used to request subscription to the location analysis service.

[0053] Step 102: Based on the first information and the first network model, perform location analysis to obtain the location result, and send the location result to the service consumer.

[0054] In some implementations, the first network function is the Network Data Analytics Function (NWDAF) responsible for the Analytics Logical Function (AnLF).

[0055] Specifically, the NWDAF responsible for AnLF receives the first information sent by the service consumer. The first information is used to request subscription to the location analysis service. Based on the first information and various detection data, the NWDAF performs location analysis based on the first model to obtain the location result and sends the location result to the service consumer.

[0056] In some implementations, the first information includes service identification information and location result format information.

[0057] In some implementations, the location result format information is used to indicate the format of the location result, which includes the location data and the confidence interval of the location data.

[0058] In some implementations, the first information is Nnwdaf_AnalyticsSubscription_Subscribe.

[0059] Specifically, service consumers subscribe to location analytics services from NWDAF by invoking the Nnwdaf_AnalyticsSubscription_Subscribe service operation. The reliability of location data is improved by including a specific service identifier in the request message Nnwdaf_AnalyticsSubscription_Subscribe to indicate a specific service and by supporting a result delivery method that provides location data plus a trusted interval.

[0060] In some implementations, the NWDAF responsible for AnLF sends the location results to the service consumer via Nnwdaf_AnalyticsSubscription_Notify.

[0061] In some implementations, the first network model is a Bayesian Neural Network (BNN) model.

[0062] In some implementations, the location analysis service is an uncertain location analysis service, and correspondingly, the first network model is also an uncertain location model.

[0063] In some implementations, when a first network model exists within the first network function, the first network model is directly used to perform location analysis services and generate analysis results. These analysis results are the location results, including location data and the confidence interval of the location data.

[0064] In some implementations, when the first network model is not present in the first network function, the method further includes: sending second information to the second network function, the second information being used to subscribe to the training of the first network model; and receiving the trained first network model sent by the second network function.

[0065] In some implementations, the second network function is the Network Data Analytics Function (NWDAF) responsible for the Model Training Logical Function (MTLF).

[0066] In some implementations, the second information is Nnwdaf_BNNTraining_Subscribe.

[0067] Specifically, the NWDAF responsible for AnLF subscribes to the training of the uncertainty localization model corresponding to the uncertainty localization analysis service from the NWDAF responsible for MTLF by calling the Nnwdaf_BNNTraining_Subscribe service. The uncertainty localization model is the first network model, also known as the BNN model. The analysis results obtained from the uncertainty localization analysis service include the localization data and the confidence interval of the localization data; that is, uncertainty is the confidence interval of the localization data.

[0068] In some implementations, once the second network function has received the trained first network model, it sends the trained first network model to the first network function.

[0069] In some implementations, the NWDAF responsible for MTLF calls the Nnwdaf_BNNTraining_Notify service operation to provide the trained first network model to the NWDAF responsible for AnLF for localization analysis services.

[0070] In some implementations, localization analysis is performed based on a first network model, including: performing localization analysis based on a trained first network model.

[0071] In some implementations, when the first network function receives the trained first network model, it performs localization analysis based on the trained first network model to obtain the localization analysis result, i.e., the localization result.

[0072] In some implementations, the location analysis results include, but are not limited to, information such as altitude, longitude, latitude, altitude uncertainty, longitude uncertainty, and latitude uncertainty. Here, uncertainty refers to the confidence interval for a particular indicator generated by the Bayesian neural network.

[0073] In some implementations, location services can be provided for connected terminals. For example, the location results can be applied to scenarios such as drone flight path planning / autonomous vehicle road planning and obstacle avoidance.

[0074] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers; the localization analysis based on the first network model includes: extracting multiple features from data of multiple dimensions through multiple feature extraction layers; fusing multiple features through the feature fusion layer to obtain fused features; and analyzing the fused features through the logistic regression layer to obtain localization data and the confidence interval of the localization data.

[0075] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and Bayesian logistic regression layers.

[0076] Specifically, the feature extraction layer utilizes various feature extraction networks, including but not limited to VGG16, Unet, and ResNet50, to extract features from data across different dimensions. This data includes, but is not limited to, image data and radar data. By selecting appropriate network models for different types of data, important information within the data can be efficiently captured to support subsequent deep learning processing. The feature fusion layer effectively fuses data features from different dimensions. By integrating features from multiple data types (such as image features and radar features), the feature fusion layer can comprehensively utilize various types of information, enhancing the model's expressive power and decision accuracy. This process can be accomplished through various methods, including but not limited to weighted averaging, concatenation, and attention mechanisms, thereby achieving a global understanding of complex scenarios and improving the system's performance in multimodal data. The Bayesian logistic regression layer, based on the fused features, judges and analyzes the positioning results and uncertainties of the connected terminal. Through Bayesian logistic regression, not only can the predicted value of the positioning result be output, but the uncertainty of the result can also be quantified. The structure of the Bayesian logistic regression layer differs from traditional neural networks, where the weights of each network node are not fixed values ​​but random variables following a Gaussian distribution. There are various ways to construct a Bayesian logistic regression layer, including but not limited to techniques such as Reparameterization and Flipout.

[0077] The technical solution of this application embodiment receives first information sent by a service consumer, the first information being used to request subscription to a location analysis service; based on the first information and a first network model, location analysis is performed to obtain a location result, which is then sent to the service consumer. In this way, the first network function performs location analysis through the first network model to obtain the location result and the reliable range of the location result, thereby improving the accuracy of the location.

[0078] Figure 2 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 2 It is used in the second network function, such as Figure 2 As shown, the positioning method includes the following steps:

[0079] Step 201: Receive the second information sent by the first network function. The second information is used to request subscription to the training of the first network model.

[0080] Step 202: Train the first network model using the dataset, and send the trained first network model to the first network function.

[0081] In some implementations, the first network function is responsible for the NWDAF of AnLF, and the second network function is responsible for the NWDAF of MTLF.

[0082] In some implementations, the second information is Nnwdaf_BNNTraining_Subscribe.

[0083] In some implementations, the first network model is a BNN model.

[0084] In some implementations, the location analysis service is an uncertain location analysis service, and correspondingly, the first network model is also an uncertain location model.

[0085] In some implementations, the method further includes: sending third information to a third network function, the third information being used to request subscription to a dataset; and receiving the dataset sent by the third network function. Here, the third network function may store the dataset.

[0086] In some implementations, the third network function is the Data Collection Coordination Function (DCCF), and the third information is Ndccf_DataManagement_Subscribe.

[0087] In some implementations, third information is sent to a third network function, wherein the third information is used to acquire a dataset, receive the dataset sent by the third network function, train a first network model using the dataset, and send the trained first network model to the first network function.

[0088] In some implementations, if the dataset is stored in DCCF, MTLF subscribes to the dataset from DCCF, and DCCF sends the dataset to MTLF; if the dataset does not exist in DCCF, it needs to be obtained from the data source.

[0089] Specifically, the NWDAF responsible for AnLF subscribes to the training of the uncertainty localization model corresponding to the uncertainty localization analysis service from the NWDAF responsible for MTLF by calling the Nnwdaf_BNNTraining_Subscribe service operation. The NWDAF responsible for MTLF subscribes to the data required by the uncertainty localization model from the DCCF by calling the Ndccf_DataManagement_Subscribe service operation. The DCCF transmits the data to the model training network function MTLF, trains the first network model using the dataset, and sends the trained first network model to the NWDAF responsible for AnLF. Here, the uncertainty localization model is the first network model, also known as the BNN model. The analysis results obtained by the uncertainty localization analysis service include the localization data and the confidence interval of the localization data, that is, the uncertainty is the confidence interval of the localization data.

[0090] In some implementations, the NWDAF responsible for MTLF calls the Nnwdaf_BNNTraining_Notify service operation to provide the trained first network model to the NWDAF responsible for AnLF for localization analysis services.

[0091] In some implementations, the DCCF sends data to the NWDAF responsible for the MTLF via the Ndccf_DataManagement_Fetch response.

[0092] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

[0093] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and Bayesian logistic regression layers.

[0094] Specifically, the feature extraction layer utilizes various feature extraction networks, including but not limited to VGG16, Unet, and ResNet50, to extract features from data across different dimensions. This data includes, but is not limited to, image data and radar data. By selecting appropriate network models for different types of data, important information within the data can be efficiently captured to support subsequent deep learning processing. The feature fusion layer effectively fuses data features from different dimensions. By integrating features from multiple data types (such as image features and radar features), the feature fusion layer can comprehensively utilize various types of information, enhancing the model's expressive power and decision accuracy. This process can be accomplished through various methods, including but not limited to weighted averaging, concatenation, and attention mechanisms, thereby achieving a global understanding of complex scenarios and improving the system's performance in multimodal data. The Bayesian logistic regression layer, based on the fused features, judges and analyzes the positioning results and uncertainties of the connected terminal. Through Bayesian logistic regression, not only can the predicted value of the positioning result be output, but the uncertainty of the result can also be quantified. The structure of the Bayesian logistic regression layer differs from traditional neural networks, where the weights of each network node are not fixed values ​​but random variables following a Gaussian distribution. There are various ways to construct a Bayesian logistic regression layer, including but not limited to techniques such as Reparameterization and Flipout.

[0095] In some implementations, the dataset includes radar data, sensing base station data, and image data, as well as location information of the terminals corresponding to the radar data, sensing base station data, and image data, respectively; wherein the location information includes longitude information, latitude information, and altitude information.

[0096] In some implementations, the dataset includes data and corresponding labels; training a first network model using the dataset includes: training the first network model based on a first loss function, the data, and the labels; wherein the first loss function is determined based on first information, the first information is sent by a service consumer to a first network function, and the first information is used to request subscription to a location analysis service.

[0097] In some implementations, the NWDAF responsible for MTLF trains a BNN model based on the loss function required by the task and the data and labels obtained from the DCCF when performing the corresponding task. Therefore, when the service consumer sends the first information to the first network function, the current task can be determined based on the first information. Since the first information is used to request subscription to the location analysis service, the current task is the location analysis service, and the first loss function is the model training loss function corresponding to the location analysis service.

[0098] Specifically, the NWDAF, responsible for MTLF, uses a loss function designed according to task requirements and data and labels obtained from DCCF to train the BNN model when performing uncertain localization tasks. During training, the model undergoes periodic performance evaluations, and training terminates once the evaluation results meet the expected standards.

[0099] In some implementations, due to the unique structure of BNN networks, the output of the Bayesian logistic regression layer may exhibit maxima or minima, leading to gradient explosion or vanishing gradients. Therefore, to stabilize the training process and ensure model convergence, measures are often needed to limit the range of gradients. These measures include, but are not limited to, gradient clipping and weight normalization.

[0100] In some implementations, the first information is Nnwdaf_AnalyticsSubscription_Subscribe.

[0101] In some implementations, the first information includes service identification information and location result format information.

[0102] In some implementations, the location result format information is used to indicate the format of the location result, which includes the location data and the confidence interval of the location data.

[0103] Specifically, service consumers subscribe to location analytics services from NWDAF by invoking the Nnwdaf_AnalyticsSubscription_Subscribe service operation. The reliability of location data is improved by including a specific service identifier in the request message Nnwdaf_AnalyticsSubscription_Subscribe to indicate a specific service and by supporting a result delivery method that provides location data plus a trusted interval.

[0104] In some implementations, before receiving the dataset sent by the third network function, the method further includes: receiving fourth information sent by the third network function, the fourth information being used to inform the second network function to obtain the dataset; and sending fifth information to the third network function based on the fourth information, the fifth information being used to obtain the dataset from the third network function.

[0105] In some implementations, the fourth message is Ndccf_DataManagement_Notify, and the fifth message is Ndccf_DataManagement_Fetch request.

[0106] Specifically, DCCF uses `Ndccf_DataManagement_Notify` to inform the NWDAF responsible for MTLF that data is ready, triggering a fetch command. Data sent to the NWDAF responsible for MTLF may be processed and formatted by DCCF. If `Ndccf_DataManagement_Notify` contains a fetch command, the NWDAF responsible for MTLF will send an `Ndccf_DataManagement_Fetch` request to fetch data from DCCF.

[0107] Understandably, DCCF processes the data obtained from the data source into a data format suitable for training.

[0108] In the technical solution of this application embodiment, the second network function receives second information sent by the first network function. The second information is used to request subscription to the training of the first network model. The second network function trains the first network model using a dataset and sends the trained first network model back to the first network function. In this way, the second network function sends the trained first network model back to the first network function, enabling the first network function to perform localization analysis using the first network model, obtain localization results and the confidence interval of the localization results, and improve the accuracy of localization.

[0109] Figure 3 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 3 Applications to third-party network functions, such as Figure 3 As shown, the positioning method includes the following steps:

[0110] Step 301: Receive the third information sent by the second network function. The third information is used to request subscription to the dataset. The dataset is used to train the first network model. The first network model is used for localization analysis.

[0111] Step 302: Determine whether the dataset needs to be collected based on the third information, and obtain the determination result.

[0112] In some implementations, the second network function is the NWDAF responsible for MTLF. The third network function is the DCCF. The third information is Ndccf_DataManagement_Subscribe.

[0113] Specifically, the NWDAF responsible for MTLF subscribes to the data required by the DCCF for the first network model by calling the Ndccf_DataManagement_Subscribe service operation. The first network model is used for localization analysis. The DCCF will determine whether a new data collection operation needs to be triggered for the current model; that is, the DCCF will determine whether the data requested for a new trigger has already been collected.

[0114] In some implementations, the first network model is a BNN model.

[0115] In some implementations, the location analysis service is an uncertain location analysis service, and correspondingly, the first network model is also an uncertain location model.

[0116] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

[0117] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and Bayesian logistic regression layers.

[0118] Specifically, the feature extraction layer utilizes various feature extraction networks, including but not limited to VGG16, Unet, and ResNet50, to extract features from data across different dimensions. This data includes, but is not limited to, image data and radar data. By selecting appropriate network models for different types of data, important information within the data can be efficiently captured to support subsequent deep learning processing. The feature fusion layer effectively fuses data features from different dimensions. By integrating features from multiple data types (such as image features and radar features), the feature fusion layer can comprehensively utilize various types of information, enhancing the model's expressive power and decision accuracy. This process can be accomplished through various methods, including but not limited to weighted averaging, concatenation, and attention mechanisms, thereby achieving a global understanding of complex scenarios and improving the system's performance in multimodal data. The Bayesian logistic regression layer, based on the fused features, judges and analyzes the positioning results and uncertainties of the connected terminal. Through Bayesian logistic regression, not only can the predicted value of the positioning result be output, but the uncertainty of the result can also be quantified. The structure of the Bayesian logistic regression layer differs from traditional neural networks, where the weights of each network node are not fixed values ​​but random variables following a Gaussian distribution. There are various ways to construct a Bayesian logistic regression layer, including but not limited to techniques such as Reparameterization and Flipout.

[0119] In some implementations, when the determination result is that collection is not required, the method further includes: sending a fourth message to a second network function, the fourth message being used to inform the second network function to obtain the dataset; receiving a fifth message sent by the second network function, the fifth message being used by the second network function to obtain the dataset from the third network function; and sending the dataset to the second network function.

[0120] In some implementations, the fourth message is Ndccf_DataManagement_Notify, and the fifth message is Ndccf_DataManagement_Fetch request.

[0121] Specifically, when the determination result indicates that data collection is not required, DCCF uses Ndccf_DataManagement_Notify to inform the NWDAF responsible for MTLF that the data is ready, triggering a fetch instruction. The data sent to the NWDAF responsible for MTLF may be processed and formatted by DCCF. If Ndccf_DataManagement_Notify contains a fetch instruction, the NWDAF responsible for MTLF will send an Ndccf_DataManagement_Fetch request to fetch data from DCCF.

[0122] In some implementations, the DCCF sends data to the NWDAF responsible for the MTLF via the Ndccf_DataManagement_Fetch response.

[0123] In some implementations, when the determination result indicates that data collection is required, the method further includes: sending a sixth message to the data source, the sixth message being used to request subscription to the dataset; and receiving a seventh message sent by the data source, the seventh message including the dataset.

[0124] In some implementations, the dataset includes radar data, sensing base station data, and image data, as well as terminal location information corresponding to the radar data, sensing base station data, and image data, respectively; wherein the location information includes longitude information, latitude information, and altitude information.

[0125] In some implementations, the sixth piece of information is Nnf_EventExposure_Subscribe. The seventh piece of information is Nnf_EventExposure_Notify.

[0126] Specifically, DCCF uses the Nnf_EventExposure_Subscribe service operation to subscribe to data from different DataSources. The subscribed data includes, but is not limited to, radar data, sensing base station data, image data, and the location information of the corresponding connected terminals. Location information includes, but is not limited to, longitude, latitude, and altitude. When new output data becomes available, the data source uses Nnf_EventExposure_Notify to send the data to DCCF.

[0127] It is understandable that when DCCF uses Nnf_EventExposure_Subscribe to subscribe to data from the Data Source, the Data Source will then use Nnf_EventExposure_Notify to send the data to DCCF.

[0128] In some implementations, the third network function subscribes to data from the data source and then updates the first network model based on the subscribed data.

[0129] In some implementations, when the determination result indicates that data collection is required, the method further includes: sending a fourth message to a second network function, the fourth message being used to inform the second network function to acquire the dataset; receiving a fifth message sent by the second network function, the fifth message being used by the second network function to acquire the dataset from the third network function; and sending the dataset to the second network function.

[0130] In some implementations, after the DCCF obtains the dataset from the data source, it needs to send the dataset to the second network function.

[0131] Specifically, DCCF uses `Ndccf_DataManagement_Notify` to inform the NWDAF responsible for MTLF that data is ready, triggering a fetch command. Data sent to the NWDAF responsible for MTLF may be processed and formatted by DCCF. If `Ndccf_DataManagement_Notify` contains a fetch command, the NWDAF responsible for MTLF will send an `Ndccf_DataManagement_Fetch` request to fetch data from DCCF.

[0132] In some implementations, the DCCF sends data to the NWDAF responsible for the MTLF via the Ndccf_DataManagement_Fetch response.

[0133] In some implementations, after receiving the third information sent by the second network function, the method further includes: obtaining user authorization information of the dataset from user data management.

[0134] In some implementations, the NWDAF responsible for MTLF subscribes to the data required by the DCCF for the uncertain location model by calling the Ndccf_DataManagement_Subscribe service operation. If the data collection involves users, the DCCF must detect user subscriptions and obtain user authorization information in advance from the User Data Management (UDM).

[0135] In the technical solution of this application embodiment, the third network function receives third information sent by the second network function. This third information is used to request subscription to a dataset, which is used to train a first network model. The first network model is used for localization analysis. Based on the third information, the third function determines whether the dataset needs to be collected and obtains a judgment result. Thus, based on the judgment result, it is determined whether the dataset needs to be collected, and the dataset is sent to the second network function. The second network function uses the dataset to train the first network model and sends the trained first network model back to the first network function. The first network function performs localization analysis using the first network model to obtain localization results and the confidence interval of the localization results, thereby improving the accuracy of localization.

[0136] The technical solutions of the embodiments of this application are illustrated below with specific application implementation examples.

[0137] Based on the foregoing embodiments, the first network function is NWDAF responsible for AnLF, the second network function is NWDAF responsible for MTLF, the third network function is DCCF, and the first network model is a BNN model, i.e., an uncertain localization model. Based on this, the localization method provided by the embodiments of this application will be further described.

[0138] To address the aforementioned issues, this application proposes a BNN-based terminal uncertainty localization method in mobile communication networks. It novelly defines an uncertainty localization capability for connected terminals within the NWDAF framework. Based on consumer analysis needs, the NWDAF retrieves suitable data from the data source via DCCF to train and test the model, and then sends the final analysis results to the consumer. This capability is implemented based on the BNN algorithm, comprehensively analyzing detection data from multiple sensors to infer the location result of the connected terminal and its corresponding uncertainty. This output is applicable to scenarios such as UAV flight path / autonomous vehicle road planning and obstacle avoidance.

[0139] Figure 4 This is a flowchart illustrating the positioning method provided in the embodiments of this application. Figure 4 ,like Figure 4 As shown, this includes the service consumer, NWDAF containing AnLF (i.e., the NWDAF responsible for AnLF), NWDAF containing MTLF (i.e., the NWDAF responsible for MTLF), DCCF, UDM, and data source. For ease of description, NWDAF containing AnLF will be referred to as AnLF, and NWDAF containing MTLF will be referred to as MTLF. The specific uncertainty localization capability in NWDAF signaling flow, model training, and uncertainty localization acquisition includes the following steps:

[0140] Step 401: The service consumer sends Nnwdaf_AnalyticsSubscription_Subscribe to AnLF.

[0141] Service consumers subscribe to the location analytics service from NWDAF by invoking the Nnwdaf_AnalyticsSubscription_Subscribe service operation. The reliability of the location data is improved by introducing a specific service identifier in the request message Nnwdaf_AnalyticsSubscription_Subscribe to indicate a specific service and supporting the provision of location data plus a trusted interval as the result. If a relevant inference model exists in AnLF, step 413 is executed. If no relevant inference model exists in AnLF, step 402 is executed.

[0142] Step 402: AnLF sends Nnwdaf_BNNTraining_Subscribe to MTLF.

[0143] If the AnLF inference module does not have a relevant inference model, the NWDAF responsible for AnLF subscribes to the training of the uncertainty localization model corresponding to the uncertainty localization analysis service from the NWDAF responsible for MTLF by calling the Nnwdaf_BNNTraining_Subscribe service operation. The uncertainty localization model is a BNN model.

[0144] Step 403: MTLF sends Ndccf_DataManagement_Subscribe to DCCF.

[0145] The NWDAF responsible for MTLF subscribes to the data required by the DCCF for the uncertain location model by calling the Ndccf_DataManagement_Subscribe service operation. If user authorization is required, step 404 is executed; otherwise, step 405 is executed.

[0146] Step 404: Check the user consent information in the UDM. Subscribe to changes to the user consent information.

[0147] If the data collection involves users, it is necessary to detect user contracts and obtain user authorization information from the UDM in advance.

[0148] Step 405: Determine whether the requested data can be obtained in DCCF.

[0149] The DCCF will determine whether a new data collection operation needs to be triggered for the current model; that is, the DCCF will determine whether the data requested for the new trigger has already been collected. If it cannot be obtained directly in the DCCF, then proceed to step 406; otherwise, proceed to step 408.

[0150] Step 406: DCCF sends Nnf_EventExposure_Subscribe to Data Source.

[0151] If the data requested in step 403 is not yet available or has not been collected, DCCF will use the Nnf_EventExposure_Subscribe service operation to subscribe to data from different Data Sources. The subscribed data includes, but is not limited to, radar data, sensing base station data, image data, and the location information of the corresponding connected terminals. Location information includes, but is not limited to, longitude, latitude, and altitude.

[0152] Step 407: Data Source sends Nnf_EventExposure_Notify to DCCF.

[0153] When new output data becomes available, the data source sends the data to the DCCF using Nnf_EventExposure_Notify. It's understood that the Data Source will only send data to the DCCF after the DCCF sends Nnf_EventExposure_Subscribe to the Data Source. This data can be used to train the uncertain localization model and also to update the uncertain localization model.

[0154] Step 408: DCCF sends Ndccf_DataManagement_Notify to MTLF.

[0155] DCCF uses Ndccf_DataManagement_Notify to inform all notification endpoints (i.e., MTLFs) specified in step 403 that data is ready, triggering a fetch instruction. The data sent to the notification endpoints may be processed and formatted by DCCF. Understandably, DCCF processes the data into a format suitable for training the uncertain localization model.

[0156] Step 409: MTLF sends an Ndccf_DataManagement_Fetch Request to DCCF.

[0157] If Ndccf_DataManagement_Notify contains a fetch instruction, the notification endpoint will send an Ndccf_DataManagement_Fetch request to fetch data from DCCF.

[0158] Step 410: DCCF sends an Ndccf_DataManagement_Fetch response to MTLF.

[0159] DCCF transmits data to the MTLF function, which is used to train the network for the notification model.

[0160] Step 411: Train the BNN model.

[0161] When performing uncertain localization tasks, the NWDAF, responsible for MTLF, uses a loss function designed according to task requirements and data and labels obtained from DCCF to train the BNN model (Uncertainty Localization Model). During training, the model undergoes periodic performance evaluation, and training terminates once the evaluation results meet the expected standards. It's important to note that due to the unique nature of the network structure, the output of the Bayesian logistic regression layer may exhibit maxima or minima, leading to gradient explosion or vanishing gradients. Therefore, to stabilize the training process and ensure model convergence, measures are often taken to limit the range of gradients. These measures include, but are not limited to, gradient clipping and weight normalization.

[0162] Step 412: MTLF sends Nnwdaf_BNNTraining_Notify to AnLF.

[0163] The NWDAF responsible for MTLF calls the Nnwdaf_BNNTraining_Notify service to provide the trained uncertainty localization model to the NWDAF responsible for AnLF for analysis.

[0164] Step 413: AnLF generates analysis results.

[0165] The NWDAF team responsible for AnLF uses the trained model to perform analysis services and generate analysis results. These results include, but are not limited to, information such as the altitude, longitude, latitude, altitude uncertainty, longitude uncertainty, and latitude uncertainty of the connected terminal. It is understood that uncertainty refers to the confidence interval for a certain indicator generated by the Bayesian neural network.

[0166] Step 414: AnLF sends Nnwdaf_AnalyticsSubscription_Notify to the service consumer.

[0167] The NWDAF responsible for AnLF sends the analysis results to service consumers via Nnwdaf_AnalyticsSubscription_Notify.

[0168] Figure 5 This is a schematic diagram of the BNN model architecture provided in an embodiment of this application. Figure 5 As shown, data 1, data 2, and data 3 are input into feature extraction layers 1, 2, and 3, respectively, where data 1, 2, and 3 are data of different dimensions. The output of the feature extraction layers is used as the input to the feature fusion layer, and the output of the feature fusion layer is used as the input to the Bayesian logistic regression layer, finally yielding the localization result and its uncertainty.

[0169] Specifically, the feature extraction layer utilizes various feature extraction networks, including but not limited to VGG16, Unet, and ResNet50, to extract features from data across different dimensions. This data includes, but is not limited to, image data and radar data. By selecting appropriate network models for different types of data, important information within the data can be efficiently captured to support subsequent deep learning processing. The feature fusion layer effectively fuses features from data across different dimensions. By integrating features from multiple data types (such as image features and radar features), the feature fusion layer can comprehensively utilize various types of information, enhancing the model's expressive power and decision-making accuracy. This process can be accomplished through various methods, including but not limited to weighted averaging, concatenation, and attention mechanisms, thereby achieving a global understanding of complex scenarios and improving the system's performance in multimodal data. The Bayesian logistic regression layer, based on the fused features, judges and analyzes the positioning results and uncertainties of the connected terminal. Through Bayesian logistic regression, not only can the predicted value of the positioning result be output, but the uncertainty of the result can also be quantified. The structure of the Bayesian logistic regression layer differs from traditional neural networks, where the weights of each network node are not fixed values ​​but random variables following a Gaussian distribution. There are various ways to construct a Bayesian logistic regression layer, including but not limited to techniques such as Reparameterization and Flipout.

[0170] The technical solution of this application embodiment includes a core network function that supports providing uncertain positioning services. This is achieved by introducing specific service identifiers to indicate specific services and supporting a result provision method that combines positioning data with a trusted interval, thereby improving the reliability of the positioning data. The model training network function should support accepting training requests for uncertain positioning models and triggering the acquisition of external data sources. This data includes, but is not limited to, radar data, sensing base station data, image data, and the location information of the network-connected terminals corresponding to these data. Location information includes, but is not limited to, longitude, latitude, and altitude. The machine learning model proposed in this application embodiment extracts different features from multi-dimensional data through a feature extraction network and effectively integrates these features using various fusion strategies, including but not limited to weighted averaging, feature concatenation, and attention mechanisms, maximizing the advantages of each dimension of data and improving the overall performance of the model. The machine learning model proposed in this application embodiment analyzes the fused features through a Bayesian logistic regression layer, which not only provides accurate positioning results for network-connected terminals but also generates an uncertainty assessment of the positioning results through the probabilistic prediction capabilities of the BNN, quantifying the uncertainty of the positioning results. The NWDAF inference network function should support subscription to uncertain positioning capabilities and provide inference results to relevant service consumers in the form of a subscription service.

[0171] Therefore, a terminal uncertainty localization method based on BNN is proposed, which can achieve accurate terminal localization by fusing data features from different dimensions and provide uncertainty assessment of the localization results. By implementing the aforementioned method in the NWDAF of the 5G core network (5G Core, 5GC), it is possible not only to effectively utilize the resource advantages of 5GC to train a more accurate and robust model, but also to improve computational efficiency and reduce energy consumption by leveraging centralized management and resource optimization at the network level.

[0172] Figure 6 This is a schematic diagram of the structural composition of the positioning device 600 provided in the embodiments of this application. Figure 1 Applied to the first network function, such as Figure 6 As shown, the positioning device includes:

[0173] The first communication unit 601 is used to receive first information sent by a service consumer, the first information being used to request subscription to the location analysis service;

[0174] The first processing unit 602 is used to perform localization analysis based on the first information and the first network model to obtain the localization result.

[0175] The first communication unit 601 is used to send the positioning result to the service consumer.

[0176] In some implementations, the first communication unit 601 is used to send second information to the second network function, the second information being used to subscribe to the training of the first network model; and to receive the trained first network model sent by the second network function.

[0177] In some implementations, the first processing unit 602 is used to perform localization analysis based on the trained first network model.

[0178] In some implementations, the first information includes service identification information and location result format information.

[0179] In some implementations, the location result format information is used to indicate the format of the location result, which includes the location data and the confidence interval of the location data.

[0180] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers; the first processing unit 602 is used to extract multiple features from data of multiple dimensions through multiple feature extraction layers; fuse multiple features through the feature fusion layer to obtain fused features; and analyze the fused features through the logistic regression layer to obtain location data and the confidence interval of the location data.

[0181] Those skilled in the art should understand that Figure 6 The functions of each unit in the positioning device shown can be understood by referring to the relevant description of the aforementioned method. Figure 6 The functions of each unit in the positioning device shown can be implemented by a program running on a processor or by specific logic circuits.

[0182] Figure 7 This is a schematic diagram of the structural composition of the positioning device 700 provided in the embodiments of this application. Figure 2 It is used in the second network function, such as Figure 7 As shown, the positioning device includes:

[0183] The second communication unit 701 is used to receive second information sent by the first network function, the second information being used to request subscription to the training of the first network model;

[0184] The second processing unit 702 is used to train the first network model using the dataset;

[0185] The second communication unit 703 is used to send the trained first network model to the first network function.

[0186] In some embodiments, the second communication unit 703 is used to send third information to the third network function, the third information being used to request subscription to a dataset; and to receive the dataset sent by the third network function.

[0187] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

[0188] In some implementations, the dataset includes data and corresponding labels; the second processing unit 702 is used to train a first network model based on a first loss function, the data, and the labels; wherein the first loss function is determined based on first information, the first information is sent by the service consumer to the first network function, and the first information is used to request subscription to the location analysis service.

[0189] In some implementations, the first information includes service identification information and location result format information.

[0190] In some implementations, the location result format information is used to indicate the format of the location result, which includes the location data and the confidence interval of the location data.

[0191] In some implementations, before receiving the dataset sent by the third network function, the second communication unit 703 is configured to receive fourth information sent by the third network function, the fourth information being used to inform the second network function to obtain the dataset; and to send fifth information to the third network function based on the fourth information, the fifth information being used to obtain the dataset from the third network function.

[0192] In some implementations, the dataset includes radar data, sensing base station data, and image data, as well as location information of the terminals corresponding to the radar data, sensing base station data, and image data, respectively; wherein the location information includes longitude information, latitude information, and altitude information.

[0193] Those skilled in the art should understand that Figure 7 The functions of each unit in the positioning device shown can be understood by referring to the relevant description of the aforementioned method. Figure 7 The functions of each unit in the positioning device shown can be implemented by a program running on a processor or by specific logic circuits.

[0194] Figure 8 This is a schematic diagram of the structural composition of the positioning device 800 provided in the embodiments of this application. Figure 3 Applications to third-party network functions, such as Figure 8 As shown, the positioning device includes:

[0195] The third communication unit 801 is used to receive third information sent by the second network function. The third information is used to request subscription to the dataset. The dataset is used to train the first network model. The first network model is used for localization analysis.

[0196] The third processing unit 802 is used to determine whether the dataset needs to be collected based on the third information and to obtain the determination result.

[0197] In some implementations, when the determination result indicates that data collection is required, the third communication unit 801 is used to send sixth information to the data source, the sixth information being used to request subscription to the dataset; and to receive seventh information sent by the data source, the seventh information including the dataset.

[0198] In some embodiments, the third communication unit 801 is configured to send fourth information to the second network function, the fourth information being used to inform the second network function to acquire a dataset; receive fifth information sent by the second network function, the fifth information being used by the second network function to acquire a dataset from the third network function; and send the dataset to the second network function.

[0199] In some implementations, when the determination result is that collection is not required, the third communication unit 801 is used to send fourth information to the second network function, the fourth information being used to inform the second network function to obtain the dataset; receive fifth information sent by the second network function, the fifth information being used by the second network function to obtain the dataset from the third network function; and send the dataset to the second network function.

[0200] In some implementations, after receiving the third information sent by the second network function, the third processing unit 802 is used to obtain user authorization information of the dataset from the user data management.

[0201] In some implementations, the dataset includes radar data, sensing base station data, and image data, as well as location information of the terminals corresponding to the radar data, sensing base station data, and image data, respectively; wherein the location information includes longitude information, latitude information, and altitude information.

[0202] In some implementations, the first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

[0203] Those skilled in the art should understand that Figure 8 The functions of each unit in the positioning device shown can be understood by referring to the relevant description of the aforementioned method. Figure 8 The functions of each unit in the positioning device shown can be implemented by a program running on a processor or by specific logic circuits.

[0204] Figure 9 This is a schematic structural diagram of a positioning device 900 provided in an embodiment of this application. The positioning device may have a first network function, a second network function, or a third network function. Figure 9 The positioning device 900 shown includes a processor 910, which can call and run computer programs from memory to implement the methods in the embodiments of this application.

[0205] Optionally, such as Figure 9 As shown, the positioning device 900 may further include a memory 920. The processor 910 can retrieve and run computer programs from the memory 920 to implement the methods described in this embodiment.

[0206] The memory 920 can be a separate device independent of the processor 910, or it can be integrated into the processor 910.

[0207] Optionally, such as Figure 9 As shown, the positioning device 900 may also include a transceiver 930, which the processor 910 can control to communicate with other devices. Specifically, it can send information or data to other devices or receive information or data sent by other devices.

[0208] The transceiver 930 may include a transmitter and a receiver. The transceiver 930 may further include antennas, and the number of antennas may be one or more.

[0209] Optionally, the positioning device 900 may specifically be the first network function in the embodiments of this application, and the positioning device 900 may implement the corresponding processes implemented by the first network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0210] Optionally, the positioning device 900 may specifically be a second network function in the embodiments of this application, and the positioning device 900 may implement the corresponding processes implemented by the second network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0211] Optionally, the positioning device 900 may specifically be a third network function in the embodiments of this application, and the positioning device 900 may implement the corresponding processes implemented by the third network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0212] Figure 10 This is a schematic structural diagram of the chip according to an embodiment of this application. Figure 10 The chip 1000 shown includes a processor 1010, which can call and run computer programs from memory to implement the methods in the embodiments of this application.

[0213] Optionally, such as Figure 10 As shown, chip 1000 may further include memory 1020. Processor 1010 can retrieve and run computer programs from memory 1020 to implement the methods described in this embodiment.

[0214] The memory 1020 can be a separate device independent of the processor 1010, or it can be integrated into the processor 1010.

[0215] Optionally, the chip 1000 may also include an input interface 1030. The processor 1010 can control the input interface 1030 to communicate with other devices or chips, specifically, to acquire information or data sent by other devices or chips.

[0216] Optionally, the chip 1000 may also include an output interface 1040. The processor 1010 can control the output interface 1040 to communicate with other devices or chips, specifically, to output information or data to other devices or chips.

[0217] Optionally, the chip can be applied to the first network function in the embodiments of this application, and the chip can implement the corresponding processes implemented by the first network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0218] Optionally, the chip can be applied to the second network function in the embodiments of this application, and the chip can implement the corresponding processes implemented by the second network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0219] Optionally, the chip can be applied to the third network function in the embodiments of this application, and the chip can implement the corresponding processes implemented by the third network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0220] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0221] It should be understood that the processor in the embodiments of this application may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0222] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0223] It should be understood that the above-described memory is exemplary and not a limiting description. For example, the memory in the embodiments of this application may also be static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM), etc. That is to say, the memory in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0224] This application also provides a computer program product, including a computer program.

[0225] Optionally, the computer program product can be applied to the first network function in the embodiments of this application, and when the computer program is executed by the processor, it implements the corresponding processes implemented by the first network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0226] Optionally, the computer program product can be applied to the second network function in the embodiments of this application, and when the computer program is executed by the processor, it implements the corresponding processes implemented by the second network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0227] Optionally, the computer program product can be applied to the third network function in the embodiments of this application, and when the computer program is executed by the processor, it implements the corresponding processes implemented by the third network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0228] This application also provides a computer-readable storage medium for storing computer programs.

[0229] Optionally, the computer-readable storage medium can be applied to the first network function in the embodiments of this application, and the computer program causes the computer to execute the corresponding processes implemented by the first network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0230] Optionally, the computer-readable storage medium can be applied to the second network function in the embodiments of this application, and the computer program causes the computer to execute the corresponding processes implemented by the second network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0231] Optionally, the computer-readable storage medium can be applied to the third network function in the embodiments of this application, and the computer program causes the computer to execute the corresponding processes implemented by the third network function in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0232] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0233] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0234] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0235] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0236] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0237] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0238] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A positioning method, characterized in that, Applied to a first network function, the method includes: Receive first information sent by the service consumer, the first information being used to request subscription to the location analysis service; Based on the first information, a location analysis is performed using the first network model to obtain a location result, which is then sent to the service consumer.

2. The method according to claim 1, characterized in that, The method further includes: Send a second message to the second network function, the second message being used to subscribe to the training of the first network model; Receive the trained first network model sent by the second network function.

3. The method according to claim 2, characterized in that, The localization analysis based on the first network model includes: Localization analysis is performed based on the first trained network model.

4. The method according to any one of claims 1 to 3, characterized in that, The first information includes service identification information and location result format information.

5. The method according to claim 4, characterized in that, The location result format information is used to indicate the format of the location result, and the format of the location result includes the location data and the confidence interval of the location data.

6. The method according to any one of claims 1 to 3, characterized in that, The first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers; The localization analysis based on the first network model includes: Multiple features are extracted from data of multiple dimensions through the multiple feature extraction layers; The multiple features are fused through the feature fusion layer to obtain fused features; The fused features are analyzed by the logistic regression layer to obtain the location data and the confidence interval of the location data.

7. A positioning method, characterized in that, Applied to a second network function, the method includes: Receive second information sent by the first network function, the second information being used to request subscription to the training of the first network model; The first network model is trained using the dataset, and the trained first network model is sent to the first network function.

8. The method according to claim 7, characterized in that, The method further includes: Send a third message to a third network function, the third message being used to request subscription to the dataset, and receive the dataset sent by the third network function.

9. The method according to claim 7, characterized in that, The first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

10. The method according to claim 7, characterized in that, The dataset includes data and the corresponding labels for the data; Training the first network model using the dataset includes: The first network model is trained based on the first loss function, the data, and the labels; The first loss function is determined based on the first information, which is sent by the service consumer to the first network function. The first information is used to request subscription to the location analysis service.

11. The claim according to claim 10, characterized in that, The first information includes service identification information and location result format information.

12. The method according to claim 11, characterized in that, The location result format information is used to indicate the format of the location result, and the format of the location result includes the location data and the confidence interval of the location data.

13. The method according to claim 8, characterized in that, Before receiving the dataset sent by the third network function, the method further includes: Receive fourth information sent by the third network function, the fourth information being used to inform the second network function to obtain the dataset; The fifth information is sent to the third network function based on the fourth information, and the fifth information is used to obtain the dataset from the third network function.

14. The method according to any one of claims 7 to 13, characterized in that, The dataset includes radar data, sensing base station data, and image data, as well as the location information of the terminals corresponding to the radar data, the sensing base station data, and the image data, respectively; wherein, the location information includes longitude information, latitude information, and altitude information.

15. A positioning method, characterized in that, Applied to a third network function, the method includes: Receive third information sent by the second network function, the third information being used to request subscription to a dataset, the dataset being used to train a first network model, and the first network model being used for localization analysis; Based on the third information, determine whether the dataset needs to be collected, and obtain the determination result.

16. The method according to claim 15, characterized in that, When the determination result indicates that data collection is required, the method further includes: Send a sixth message to the data source, the sixth message being used to request subscription to the dataset; Receive the seventh message sent by the data source, the seventh message including the dataset.

17. The method according to claim 16, characterized in that, The method further includes: Send a fourth message to the second network function, the fourth message being used to inform the second network function to obtain the dataset; The second network function receives a fifth message, which is used by the second network function to obtain the dataset from the third network function. The dataset is sent to the second network function.

18. The method according to claim 15, characterized in that, When the determination result indicates that collection is not required, the method further includes: Send a fourth message to the second network function, the fourth message being used to inform the second network function to obtain the dataset; The second network function receives a fifth message sent by the second network function, the fifth message being used by the second network function to obtain the dataset from the third network function; The dataset is sent to the second network function.

19. The method according to any one of claims 15 to 18, characterized in that, After receiving the third information sent by the second network function, the method further includes: Obtain user authorization information from the user data management dataset.

20. The method according to any one of claims 15 to 18, characterized in that, The dataset includes radar data, sensing base station data, and image data, as well as the location information of the terminals corresponding to the radar data, the sensing base station data, and the image data, respectively; wherein, the location information includes longitude information, latitude information, and altitude information.

21. The method according to any one of claims 15 to 18, characterized in that, The first network model includes multiple feature extraction layers, feature fusion layers, and logistic regression layers.

22. A positioning device, characterized in that, The device, applied to a first network function, includes: The first communication unit is used to receive first information sent by a service consumer, the first information being used to request subscription to the location analysis service; The first processing unit is used to perform localization analysis based on the first information and a first network model to obtain the localization result. The first communication unit is used to send the positioning result to the service consumer.

23. A positioning device, characterized in that, For use in a second network function, the device includes: The second communication unit is used to receive second information sent by the first network function, the second information being used to request subscription to the training of the first network model; The second processing unit is used to train the first network model using the dataset; The second communication unit is used to send the trained first network model to the first network function.

24. A positioning device, characterized in that, The device, applied to a third network function, includes: The third communication unit is used to receive third information sent by the second network function, the third information being used to request subscription to a dataset; the dataset is used to train a first network model, the first network model being used for localization analysis; The third processing unit is used to determine whether the dataset needs to be collected based on the third information, and to obtain the determination result.

25. A positioning device, characterized in that, include: A processor and a memory for storing a computer program, the processor for calling and running the computer program stored in the memory to perform the method as described in any one of claims 1 to 6, or the method as described in any one of claims 7 to 14, or the method as described in any one of claims 15 to 21.

26. A computer program product, characterized in that, include: A computer program that, when executed by a processor, implements the method according to any one of claims 1 to 6, or the method according to any one of claims 7 to 14, or the method according to any one of claims 15 to 21.

27. A computer-readable storage medium, characterized in that, Used to store a computer program that causes a computer to perform the method as described in any one of claims 1 to 6, or the method as described in any one of claims 7 to 14, or the method as described in any one of claims 15 to 21.

Images