A static user identification method, device, equipment, storage medium and product

Abstract

This invention discloses a static user identification method, apparatus, device, storage medium, and product. First, sampling point information of the user to be identified in the target scene is acquired. This sampling point information includes the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival. Then, the sampling point information is input into a scene classification model, which outputs the indoor / outdoor classification result of the user to be identified at a single moment. Finally, within a preset time period, if the proportion of the user's indoor / outdoor classification result being indoors reaches a preset percentage threshold, the user to be identified is considered a static user covered by an indoor cell in the target scene within the preset time period. This invention, through a scene classification model and a time period proportion determination mechanism, achieves accurate identification of indoor static users in a target scene. It solves the problems of traditional solutions being time-consuming and inefficient, having excessively high requirements for models and computing power, and failing to meet practical needs in terms of accuracy and efficiency. This improves the practicality and reliability of indoor / outdoor user separation and static user identification.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and more specifically, to a static user identification method, apparatus, device, storage medium, and product. Background Technology

[0002] With the development of mobile communication services, over 80% of these services occur indoors. Accurately separating indoor and outdoor users for network awareness assessment has become a crucial issue for the industry. Existing indoor user location and separation solutions mainly fall into three categories: First, by configuring the information mapping relationship between the target building and the base station, and using the signal strength difference calculated based on the signal penetration characteristics, location separation is achieved. Second, based on 4G technology and information from co-located base stations of next-generation mobile communication, a 4G fingerprint database is constructed, and the location of sampling points is determined using trilateration and backfilled into the target network data. Third, by determining the actual distance between the user and the base station using parameters such as Timing Advance (TA) and Horizontal Angle of Arrival (HAOA), and calculating the estimated distance using Reference Signal Receiving Power (RSRP), a confidence algorithm is configured to complete location separation.

[0003] However, existing solutions generally have limitations: the need to manually construct mapping relationships leads to time-consuming and inefficient processes, the requirements for the accuracy or computing power of the damage model are too high, and the overall recognition accuracy and efficiency are difficult to meet actual needs. Summary of the Invention

[0004] Based on this, the present invention provides a static user identification method, device, equipment, storage medium and product, which can obtain indoor and outdoor classification results through sampling point information and pre-trained models, and determine static users by combining time proportion, solving the problems of existing solutions that require a lot of manual operation, have high requirements and insufficient accuracy and efficiency, and providing reliable support for network perception evaluation.

[0005] To achieve the above objectives, embodiments of the present invention provide a static user identification method, comprising: Obtain sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes serving cell identifier, horizontal angle of arrival and vertical angle of arrival; The sampling point information is input into a pre-trained scene classification model, and the indoor / outdoor classification result of the user to be identified at a single moment is output. Within a preset time period, when the proportion of indoor classification results for the user to be identified at a single moment reaches a preset percentage threshold, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period.

[0006] To achieve the above objectives, embodiments of the present invention also provide a static user identification device, comprising: The sampling point information acquisition module is used to acquire sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival; The indoor / outdoor classification module is used to input the sampling point information into a pre-trained scene classification model and output the indoor / outdoor classification result of the user to be identified at a single moment. The static user identification module is used to determine that the user to be identified is a static user covered by the indoor cell of the target scene within the preset time period when the proportion of the indoor classification result of the user to be identified at a single moment reaches a preset proportion threshold.

[0007] To achieve the above objectives, embodiments of the present invention also provide a static user identification device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the static user identification method as described in any of the above embodiments.

[0008] To achieve the above objectives, embodiments of the present invention also provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the static user identification method as described in any of the above embodiments.

[0009] To achieve the above objectives, embodiments of the present invention also provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the static user identification method as described in any of the above embodiments.

[0010] Compared with existing technologies, the static user identification method, apparatus, device, storage medium, and product disclosed in this invention first acquire sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes serving cell identifier, horizontal angle of arrival, and vertical angle of arrival; then, the sampling point information is input into a pre-trained scene classification model, and the single-moment indoor / outdoor classification result of the user to be identified is output; finally, within a preset time period, when the proportion of the single-moment indoor / outdoor classification result of the user to be identified as indoor reaches a preset proportion threshold, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period. Therefore, this invention, through a scene classification model and a time period proportion determination mechanism, effectively achieves accurate identification of indoor static users in the target scene, solving the problems of traditional solutions being time-consuming and inefficient, having excessively high requirements for models and computing power, and having difficulty meeting practical needs in terms of identification accuracy and efficiency, thereby improving the practicality and reliability of indoor / outdoor user separation and static user identification. Attached Figure Description

[0011] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating a static user identification method according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating a static user identification method according to an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the acquisition of a residential area near a scenario-based electronic fence, provided by an embodiment of the present invention. Figure 4 This is a schematic diagram of latitude and longitude calculation of sampling points provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of indoor and outdoor sampling point markings provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the random forest principle provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the X-ray method provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of a static user identification device according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the structure of a static user identification device provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] See Figure 1 This is a flowchart illustrating a static user identification method according to an embodiment of the present invention. Specifically, the static user identification method includes steps S1 to S3: S1. Obtain sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes serving cell identifier, horizontal angle of arrival and vertical angle of arrival; S2. Input the sampling point information into the pre-trained scene classification model and output the single-moment indoor / outdoor classification result of the user to be identified; S3. Within a preset time period, when the proportion of indoor classification results for the user to be identified at a single moment reaches a preset proportion threshold, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period.

[0015] Specifically, sampling point information associated with the target scene is obtained, including the serving cell identifier (serving cell ID), horizontal angle of arrival, and vertical angle of arrival. For example, the sampling point information is shown in the table below:

[0016] After preprocessing and normalizing the serving cell ID, horizontal angle of arrival, and vertical angle of arrival data of the sampling points in the table above, these data are used as input to a pre-trained scene classification model. The model is then used to infer from the input data and output the predicted value corresponding to the scene, as shown in the table below:

[0017] Note: The values ​​in the predicted value column are the indoor / outdoor classification results at a single moment: 0 represents indoor and 1 represents outdoor.

[0018] Finally, statistical analysis is performed on all the single-moment indoor / outdoor classification results of the user to be identified within the preset time period. When the proportion of the single-moment indoor / outdoor classification results of the user to be identified as indoor reaches a preset proportion threshold, the user to be identified is determined to be a static user covered by the target scene indoor cell within the preset time period; otherwise, the user is not a static user covered by the target scene indoor cell within that time period.

[0019] Compared with existing technologies, the embodiments of the present invention effectively achieve accurate identification of indoor static users in target scenarios through scene classification models and time period ratio determination mechanisms. This solves the problems of traditional solutions being time-consuming and inefficient, having excessively high requirements for models and computing power, and having difficulty meeting actual needs in terms of identification accuracy and efficiency. As a result, the practicality and reliability of indoor and outdoor user separation and static user identification are improved.

[0020] In a preferred embodiment, step S3, which involves determining that the user to be identified is a static user within the target scenario's indoor cell coverage area during a preset time period when the proportion of the user's single-moment indoor / outdoor classification result being indoor reaches a preset percentage threshold, includes: Within a preset time period, the indoor / outdoor classification results of the user to be identified are statistically analyzed using a sliding window of a preset size at a single moment. When the number of indoor sampling points within the window reaches a first preset percentage threshold, the user to be identified is determined to be a scene user within the window; wherein, the indoor sampling point refers to the sampling point whose indoor / outdoor classification result is indoor at a single moment; When the number of user windows in a scene reaches a second preset percentage threshold within the preset time period, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period.

[0021] Specifically, based on the table data corresponding to the predicted values ​​mentioned above, firstly, the users in the occupied scenarios are identified. Then, the sampling point data of these users within the 5-minute period (this duration can be set according to the actual situation) are statistically analyzed, and the proportion of sampling points in a specific scenario is calculated. The sampling point proportion refers to the proportion of a user's sampling points in a specific scenario in the total sample. For a user in a certain scenario, its sampling point proportion can be calculated using the following formula: ; in, It is the percentage of sampling points. It represents the number of sampling points in the target scenario for that user within that 5-minute period. This is the total number of sampling points for the user within 5 minutes in the user sampling point data table.

[0022] When the proportion of the sampling points is greater than the first preset proportion threshold (e.g., 85%), the user is considered a user in the 5-minute granularity scenario, and user fingerprint data is created; otherwise, the user is not a user in the 5-minute granularity scenario. The first preset proportion threshold is set based on expert experience.

[0023] Next, it determines whether the user is a static user within the target scenario's indoor community coverage area during the time period. The principle is as follows: Figure 2 As shown: Suppose a time series Each of them This represents the specific data value at the corresponding time point. In this embodiment, a value of size is defined. A sliding window that covers Continuous data points. For each time point The sliding window will cover from arrive Data points: ; Its core detection principle is: based on the condition that "all data points within the window are greater than a certain threshold". "As a condition for judgment, it is necessary to check the window." Do all data points satisfy the following conditions: ; Based on the sliding window detection logic described above, and combined with the static user identification method of this embodiment, continuous user fingerprint data is monitored and analyzed. This user fingerprint data comes from several (e.g., 6, the number is set according to the actual situation) indoor user sampling points in a specific scenario with an interval of 5 minutes. After analysis, the corresponding identification results are obtained. The specific identification logic is as follows: determine whether at least 5 of the 6 consecutive user fingerprint data points belong to the same target scenario. If so, determine that the user is a static user covered by the target scenario's indoor cell within the corresponding time period. At the same time, establish a scene static user identifier field, integrate the measurement time of each sampling point within the time period as the time partition of the user's belonging to this scenario, and update the above information to the table below. If the above conditions are not met, skip the user and continue monitoring the fingerprint data of the next user. Finally, output all static user data covered by the target scenario's indoor cell.

[0024] Optionally, the static user data for indoor cell coverage in the target scenario can be the static user data for indoor 5G cell coverage in the target scenario.

[0025] In a preferred embodiment, obtaining the sampling point information of the user to be identified in the target scene includes: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Using the data of the nearby communities of the scene's electronic fence, and associating it with the user sampling point data, user sampling point information of the nearby communities of the scene's electronic fence is obtained; wherein, the user sampling point information of the nearby communities of the scene's electronic fence includes sampling point information of at least one user to be identified.

[0026] For example, taking the scenario electronic fence data, cell engineering parameter data, and user sampling point data all in tabular form, and assuming the cell engineering parameter data is 5G cell engineering parameter data, the sampling point information is obtained as follows: 1. Obtain the scene electronic fence data table, 5G cell engineering parameter data table, and user sampling point information table from the network management system database.

[0027] 2. For example Figure 3 As shown, a circle with a radius r and a preset length (e.g., r=2km) is drawn with the electronic fence of the target scene as the center. The surrounding cells and their locations (e.g., latitude and longitude) are obtained by combining the 5G cell engineering parameter data table. This yields information on all cells within the circle and their locations, outputting a table of cells near the scene's electronic fence, containing fields such as {distance between site and scene, city, cell identifier, cell longitude, cell latitude, coverage type, scene name, and location information}. Figure 3 In the diagram, cells 1-5 represent the residential area, and the polygons within the circles represent the target scene.

[0028] 3. By associating the table of nearby residential areas with the user sampling point information table, the following table of user sampling points near the electronic fence is obtained:

[0029] In a preferred embodiment, the scene classification model is trained based on a training dataset, which contains several samples and their labels. Each sample includes the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampling point sample. The label is used to indicate whether the sample belongs to indoor or outdoor environments.

[0030] Specifically, the training process for the scene classification model is as follows: Obtain a training dataset; wherein the training dataset contains several samples and their labels, each sample including the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampling point sample, and the label is used to indicate whether the sample is indoors or outdoors; The scene classification model is obtained by training the model using the training sample dataset.

[0031] It is worth noting that the application of this model effectively saves the work of subsequent experts in reviewing, correcting and labeling the data, significantly reduces the reliance on expert experience, reduces labor costs, improves the efficiency of data processing, and automates the entire process.

[0032] In a preferred embodiment, the training dataset is obtained in the following manner: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Based on the data of nearby communities in the scene's electronic fence, the location of each community is determined. Combined with the geometric surface information of the target scene, indoor communities covering the target scene are identified as community samples. Data matching the serving cell identifier with the cell sample is selected from the user sampling point data and used as sampling point samples; and the location information of the sampling point samples is determined by combining the antenna position of the cell sample, the horizontal angle of arrival, the vertical angle of arrival and the distance from the antenna in the user sampling point data, and the antenna azimuth angle in the cell engineering parameter data. Based on the location information of the sampling point samples and the data of the nearby communities of the scene's electronic fence, the sampling point samples are identified as either indoors or outdoors, and the identification result is used as the label of the sampling point samples; A training dataset is constructed based on the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampled points, as well as the labels of the sampled points.

[0033] For example, the training dataset can be obtained as follows: 1. Obtain from the network management system database: Scene electronic fence data table, containing fields such as {scene name, center longitude, center latitude, scene location information (latitude and longitude set)}; 5G cell engineering parameter data table, containing fields such as {city, cell identifier, cell name, longitude, latitude, coverage type (indoor or outdoor), antenna azimuth}; User sampling point information table, containing fields such as {measurement time, city OID, serving cell ID, horizontal angle of arrival (HAOA), vertical angle of arrival (VAOA), TA (distance from antenna), scene name, location information}. Among them, OID stands for Object Identifier; HAOA stands for Horizontal Angle of Arrival, a key parameter characterizing the horizontal angle of incidence of a signal; VAOA stands for Vertical Angle of Arrival, a key parameter characterizing the vertical angle of incidence of a signal; and TA stands for Timing Advance, which physically reflects the distance between the user terminal and the base station antenna and is one of the core parameters for user positioning.

[0034] 2. Draw a circle with a preset radius centered on the center point of the electronic fence of the target scene. Combine the 5G cell engineering parameter data table to obtain the surrounding cells and their latitude and longitude. Obtain all cells and their location information within the circle. Output a table of cells near the scene's electronic fence, including the fields {distance between site and scene, city, cell identifier, cell longitude, cell latitude, coverage type, scene name, location information}.

[0035] 3. Based on the ray method, indoor cells covering the target scene are identified according to the longitude and latitude of the cell and the geometric information of the specific scene.

[0036] 4. Based on the indoor cells covering the target scene identified in step 3, and combined with the serving cell ID in the user sampling point information table from step 1, filter out the user sampling point information occupying these indoor cells. Further, using the horizontal angle of arrival (HAOA), vertical angle of arrival (VAOA), and TA (distance from the antenna) from the user sampling point information table, and combining the antenna latitude and longitude of the cell site with the antenna azimuth angle from the 5G cell engineering parameter data table, calculate the latitude and longitude of the user sampling point.

[0037] See Figure 4 As shown, the latitude and longitude of the sampling points The calculation method is as follows: ; In the formula, HAOA is a field in the user sampling point information table; ; ; ; in, It is the total azimuth angle of the antenna (mechanical azimuth angle + electronic azimuth angle). It is the antenna's latitude and longitude. and It is a constant related to the Earth's radius.

[0038] Furthermore, by using the latitude and longitude coordinate transformation formula, the two-dimensional coordinates of each user sampling point can be obtained. coordinate: ; .

[0039] 5. Based on the ray casting method, and combining the latitude and longitude information of the user sampling points calculated in step 4, as well as the indoor cell covering the target scene generated in step 3, the spatial location of each user sampling point is determined to identify whether it is located inside or outside the target scene, and then... Figure 5The labels shown are for easy differentiation in subsequent analysis. It is important to note that due to the complexity and diversity of the data, the automatic identification process may contain some errors or uncertainties. Therefore, it is necessary to carefully check and correct the data annotation results using expert experience and expertise. This step is crucial because experts can identify and correct potential errors in the automatic annotation process based on the actual situation and their professional knowledge, ensuring the accuracy and reliability of the data. This process may require a significant amount of work, but it is a key link in ensuring the quality of the overall analysis results.

[0040] After carefully checking and correcting the data annotation results using expert experience, the following training dataset is output; 70% of it is used for model training and 30% for model evaluation and validation.

[0041]

[0042] 6. Use the data in the table above as the training data for the model. In order to solve the classification problem of distinguishing whether user data belongs to indoor or outdoor, given that random forest has high accuracy, strong anti-overfitting ability and good robustness, and can effectively handle high-dimensional data, this implementation method chooses random forest as the classification algorithm. It is worth noting that the specific algorithm is not limited to random forest algorithm, but can also be other algorithms.

[0043] like Figure 6 As shown, the principle of random forest is: Random forests consist of multiple decision trees. When determining whether a sample belongs to a certain category, each decision tree makes a judgment and outputs the classification. The final classification result is determined by voting among the trees. Decision trees use the Gini index to measure the impurity of nodes to distinguish sample categories. The Gini index is calculated as follows: ; In the formula, Category in the current node The smaller the Gini index, the purer the nodes.

[0044] The following table is output after prediction by the random forest: Table 2: Random Forest Predictions

[0045] Note: Values ​​in the predicted value column: 0 represents indoor, 1 represents outdoor.

[0046] In a preferred embodiment, determining the location of each cell based on the cell data near the scene's electronic fence, and identifying indoor cells covering the target scene using the geometric information of the target scene as cell samples, includes: The location of each community is determined based on the data of communities near the electronic fence in the described scenario. Using the location of the aforementioned cell as the starting point, a ray with a preset slope is emitted; When the number of intersections between the ray and the edge of the geometric surface information of the target scene is odd, the cell is determined to be an indoor cell covering the target scene and is used as a cell sample.

[0047] Specifically, the ray casting method is used to determine the indoor cells covering the target scene. The ray casting method is mainly used to determine the positional relationship between points and polygons, and its principle is as follows: Figure 7 As shown: Using the latitude and longitude of the residential area as the starting point of the ray, emit a ray with a slope of... A ray intersects a polygon at an odd number of points; if the number of intersections is even, the cell is inside the polygon. The points where the ray intersects each side of the polygon are: The formula for calculating the coordinates is as follows: ; In the formula, , The coordinates of the ray's origin. , , , Let these be the coordinates of the two endpoints of the polygonal line segment. The slope of the ray. Let the slope of the line segment be denoted as . ,in , , Should meet or Otherwise, the point will not be counted in the number of intersections.

[0048] After processing using the ray-mapping method, the indoor coverage cells are marked, among which... Figure 7 Cell1, cell2, and cell3 in the text refer to cells within the scene's electronic fence (i.e., indoor cells covering the target scene).

[0049] To make the technical solution of the static user identification method clearer and easier to understand, the application architecture and functional implementation process are explained in detail below with specific examples: 1. Architecture Introduction The static user identification method is implemented based on a static user identification architecture, which consists of four core functional units: a data acquisition unit, a data preprocessing unit, an indoor 5G sampling point identification unit, and an indoor 5G static user identification unit. The functions of each unit are defined as follows: Data Acquisition Unit: The data acquisition unit is responsible for filtering and extracting data tables that match the application requirements of this architecture from the network management system database. Specifically, this includes scene electronic fence data tables, 5G cell engineering parameter data tables, and user sampling point information tables, providing raw data support for subsequent data processing and model inference throughout the entire process.

[0050] Data preprocessing unit: Performs in-depth processing on the collected data, accurately extracts data information that matches the characteristics of users covered by the community by associating scenario-level and user-level table data, and lays an accurate and reliable data foundation for the subsequent user identification process.

[0051] Indoor 5G sampling point identification unit: Receives data processed by the data preprocessing unit, uses machine learning algorithms to train the model, and constructs a scene classification model that can automatically identify the indoor and outdoor coverage attributes of sampling points, providing key identification basis for subsequent static user identification.

[0052] Indoor 5G Static User Identification Unit: Taking the sampling point classification results output by the indoor 5G sampling point identification unit as input, it refines and distinguishes user attributes based on the preset static user judgment logic, and finally accurately identifies static users within the indoor cell coverage area, providing important data reference for the optimization of network resource allocation strategies, that is, providing data support for network optimization, resource allocation, user experience improvement and business decisions.

[0053] 2. Architecture and Functional Overview The functionality of the architecture described in this embodiment is implemented in two core stages: scene classification model training and data inference classification. The specific process is as follows: (1) Scene classification model training stage: The historical data collected by the data acquisition unit is first cleaned by the data preprocessing unit, and then the indoor and outdoor attributes of the data are labeled in combination with expert experience to construct a training dataset; the scene classification model is trained and verified based on the dataset, and finally the trained model is deployed to the indoor 5G sampling point identification unit. For example, 1) First, obtain three types of core raw data: scene electronic fence data table, 5G cell engineering parameter data table, and user sampling point information table, to provide a data foundation for subsequent processes; 2) Combine the scene electronic fence data table and the 5G cell engineering parameter data table to filter out cells near the scene electronic fence; 3) Use the ray casting method to map the nearby cells selected in the previous step and mark the cells located within the scene electronic fence; 4) Associate the cells within the scene electronic fence marked in the previous step with the user sampling point information table to obtain the sampling point information table of all users near the scene; 5) Use the ray casting method again to label the user sampling points obtained in the previous step with their "indoor / outdoor" attributes, providing labeled sample data for subsequent model training; 6) Use the user sampling point data with "indoor / outdoor attribute labels" obtained in the previous step to train the random forest algorithm, and finally build a scene classification model to complete the entire training process.

[0054] (2) Data Inference and Classification Stage: The raw data collected by the data acquisition unit is first cleaned by the data preprocessing unit. Then it is input into the indoor 5G sampling point identification unit, which uses the deployed scene classification model to infer the indoor and outdoor attributes of the sampling points. Finally, the inference results are input into the indoor 5G static user identification unit, which uses the preset static user determination rules to identify users and finally outputs the static user data of indoor 5G cell coverage.

[0055] Compared with existing technologies, the static user identification method provided in this invention effectively achieves accurate identification of indoor static users in target scenarios through scene classification models and time period ratio determination mechanisms. It solves the problems of traditional solutions being time-consuming and inefficient, having excessively high requirements for models and computing power, and having difficulty meeting actual needs in terms of identification accuracy and efficiency. This improves the practicality and reliability of indoor and outdoor user separation and static user identification.

[0056] See Figure 8 , Figure 8 This is a schematic diagram of a static user identification device provided in an embodiment of the present invention. The static user identification device 20 includes: The sampling point information acquisition module 21 is used to acquire sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes the serving cell identifier, horizontal angle of arrival and vertical angle of arrival; The indoor / outdoor classification module 22 is used to input the sampling point information into a pre-trained scene classification model and output the indoor / outdoor classification result of the user to be identified at a single moment. The static user identification module 23 is used to determine that the user to be identified is a static user covered by the indoor community of the target scene within the preset time period when the proportion of the indoor and outdoor classification results of the user to be identified at a single moment reaches a preset proportion threshold.

[0057] In one embodiment, the static user identification module 23 is specifically used for: Within a preset time period, the indoor / outdoor classification results of the user to be identified are statistically analyzed using a sliding window of a preset size at a single moment. When the number of indoor sampling points within the window reaches a first preset percentage threshold, the user to be identified is determined to be a scene user within the window; wherein, the indoor sampling point refers to the sampling point whose indoor / outdoor classification result is indoor at a single moment; When the number of user windows in a scene reaches a second preset percentage threshold within the preset time period, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period.

[0058] In one embodiment, the sampling point information acquisition module 21 is specifically used for: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Using the data of the nearby communities of the scene's electronic fence, and associating it with the user sampling point data, user sampling point information of the nearby communities of the scene's electronic fence is obtained; wherein, the user sampling point information of the nearby communities of the scene's electronic fence includes sampling point information of at least one user to be identified.

[0059] In one embodiment, the apparatus further includes a model training module for: training the scene classification model based on a training dataset; wherein the training dataset contains several samples and their labels, each sample including the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampling point sample, and the label is used to indicate whether the sample belongs to indoor or outdoor environments.

[0060] In one embodiment, the apparatus further includes a sample acquisition module, used for: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Based on the data of nearby communities in the scene's electronic fence, the location of each community is determined. Combined with the geometric surface information of the target scene, indoor communities covering the target scene are identified as community samples. Data matching the serving cell identifier with the cell sample is selected from the user sampling point data and used as sampling point samples; and the location information of the sampling point samples is determined by combining the antenna position of the cell sample, the horizontal angle of arrival, the vertical angle of arrival and the distance from the antenna in the user sampling point data, and the antenna azimuth angle in the cell engineering parameter data. Based on the location information of the sampling point samples and the data of the nearby communities of the scene's electronic fence, the sampling point samples are identified as either indoors or outdoors, and the identification result is used as the label of the sampling point samples; A training dataset is constructed based on the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampled points, as well as the labels of the sampled points.

[0061] In one implementation, determining the location of each cell based on cell data near the scene's electronic fence, and identifying indoor cells covering the target scene using geometric information of the target scene as cell samples, includes: The location of each community is determined based on the data of communities near the electronic fence in the described scenario. Using the location of the aforementioned cell as the starting point, a ray with a preset slope is emitted; When the number of intersections between the ray and the edge of the geometric surface information of the target scene is odd, the cell is determined to be an indoor cell covering the target scene and is used as a cell sample.

[0062] It is worth noting that the specific working process of the static user identification device can be referred to the working process of the static user identification method described in the above embodiments, and will not be repeated here.

[0063] Compared with existing technologies, the static user identification device disclosed in this invention first acquires sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes serving cell identifier, horizontal angle of arrival, and vertical angle of arrival; then, the sampling point information is input into a pre-trained scene classification model, and the single-moment indoor / outdoor classification result of the user to be identified is output; finally, within a preset time period, when the proportion of the single-moment indoor / outdoor classification result of the user to be identified as indoor reaches a preset proportion threshold, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period. Therefore, this invention, through a scene classification model and a time period proportion determination mechanism, effectively achieves accurate identification of indoor static users in the target scene, solving the problems of traditional solutions being time-consuming and inefficient, having excessively high requirements for models and computing power, and having difficulty meeting actual needs in terms of identification accuracy and efficiency, thereby improving the practicality and reliability of indoor / outdoor user separation and static user identification.

[0064] See Figure 9 , Figure 9 This is a schematic diagram of the structure of a static user identification device 30 provided in an embodiment of the present invention. The static user identification device 30 includes a processor 31, a memory 32, and a computer program stored in the memory 32 and configured to be executed by the processor 31. When the processor 31 executes the computer program, it implements the steps as described in the above-described static user identification method embodiment, for example... Figure 1 The steps S1 to S3 described above; or, when the processor 31 executes the computer program, it implements the functions of each module in the above-described device embodiments.

[0065] For example, the computer program can be divided into one or more modules, which are stored in the memory 32 and executed by the processor 31 to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the static user identification device. For example, the computer program can be divided into multiple modules, and the specific working process of each module can be referred to the working process of the static user identification device described in the above embodiments, which will not be repeated here.

[0066] The static user identification device can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The static user identification device may include, but is not limited to, a processor 31 and a memory 32. Those skilled in the art will understand that the static user identification device may also include input / output devices, network access devices, buses, etc.

[0067] The processor 31 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 31 is the control center of the static user identification device, connecting all parts of the static user identification device via various interfaces and lines.

[0068] The memory 32 can be used to store the computer programs and / or modules. The processor 31 implements various functions of the static user identification device by running or executing the computer programs and / or modules stored in the memory 32 and calling the data stored in the memory 32. The memory 32 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory 32 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0069] If the modules integrated into the static user identification device 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, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by the processor 31, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0070] This invention also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the static user identification method as described in any of the above embodiments.

[0071] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A static user identification method, characterized in that, include: Obtain sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes serving cell identifier, horizontal angle of arrival and vertical angle of arrival; The sampling point information is input into a pre-trained scene classification model, and the indoor / outdoor classification result of the user to be identified at a single moment is output. Within a preset time period, when the proportion of indoor classification results for the user to be identified at a single moment reaches a preset percentage threshold, the user to be identified is determined to be a static user covered by the indoor cell of the target scenario within the preset time period.

2. The static user identification method as described in claim 1, characterized in that, The step of determining that the user to be identified is a static user within the target scene's indoor cell coverage area during a preset time period when the proportion of indoor classification results for the user to be identified at a single moment reaches a preset percentage threshold includes: Within a preset time period, the indoor / outdoor classification results of the user to be identified are statistically analyzed using a sliding window of a preset size at a single moment. When the number of indoor sampling points within the window reaches a first preset percentage threshold, the user to be identified is determined to be a scene user within the window; wherein, the indoor sampling point refers to the sampling point whose indoor / outdoor classification result is indoor at a single moment; When the number of user windows in a scene reaches a second preset percentage threshold within the preset time period, the user to be identified is determined to be a static user covered by the indoor cell of the target scene within the preset time period.

3. The static user identification method as described in claim 1, characterized in that, The step of obtaining the sampling point information of the user to be identified in the target scene includes: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Using the data of the nearby communities of the scene's electronic fence, and associating it with the user sampling point data, user sampling point information of the nearby communities of the scene's electronic fence is obtained; wherein, the user sampling point information of the nearby communities of the scene's electronic fence includes sampling point information of at least one user to be identified.

4. The static user identification method as described in claim 1, characterized in that, The scene classification model is trained based on a training dataset, which contains several samples and their labels. Each sample includes the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampling point sample. The label is used to indicate whether the sample belongs to indoor or outdoor environments.

5. The static user identification method as described in claim 4, characterized in that, The training dataset was obtained in the following way: Acquire scene electronic fence data, community engineering parameter data, and user sampling point data; Combining the scene's electronic fence data and the community's engineering parameter data, and using the center point of the electronic fence in the target scene as a reference, data on the communities near the scene's electronic fence are generated. Based on the data of nearby communities in the scene's electronic fence, the location of each community is determined. Combined with the geometric surface information of the target scene, indoor communities covering the target scene are identified as community samples. Data matching the serving cell identifier with the cell sample is selected from the user sampling point data and used as sampling point samples; and the location information of the sampling point samples is determined by combining the antenna position of the cell sample, the horizontal angle of arrival, the vertical angle of arrival and the distance from the antenna in the user sampling point data, and the antenna azimuth angle in the cell engineering parameter data. Based on the location information of the sampling point samples and the data of the nearby communities of the scene's electronic fence, the sampling point samples are identified as either indoors or outdoors, and the identification result is used as the label of the sampling point samples; A training dataset is constructed based on the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival of the sampled points, as well as the labels of the sampled points.

6. The static user identification method as described in claim 5, characterized in that, The process of determining the location of each cell based on the cell data near the electronic fence of the scene, and identifying indoor cells covering the target scene using the geometric surface information of the target scene as cell samples, includes: The location of each community is determined based on the data of communities near the electronic fence in the described scenario. Using the location of the aforementioned cell as the starting point, a ray with a preset slope is emitted; When the number of intersections between the ray and the edge of the geometric surface information of the target scene is odd, the cell is determined to be an indoor cell covering the target scene and is used as a cell sample.

7. A static user identification device, characterized in that, include: The sampling point information acquisition module is used to acquire sampling point information of the user to be identified in the target scene; wherein, the sampling point information includes the serving cell identifier, horizontal angle of arrival, and vertical angle of arrival; The indoor / outdoor classification module is used to input the sampling point information into a pre-trained scene classification model and output the indoor / outdoor classification result of the user to be identified at a single moment. The static user identification module is used to determine that the user to be identified is a static user covered by the indoor cell of the target scene within the preset time period when the proportion of the indoor classification result of the user to be identified at a single moment reaches a preset proportion threshold.

8. A static user identification device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the static user identification method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the static user identification method as described in any one of claims 1 to 6.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the static user identification method as described in any one of claims 1 to 6.

Images