Internet of vehicles data feature extraction method, device, equipment, storage medium and product

Abstract

The application discloses a kind of Internet of Vehicles data feature extraction method, device, equipment, storage medium and product, belong to Internet of Vehicles field, to accurately extract the data feature in Internet of Vehicles data.The method comprises: obtaining the extended detailed list XDR data and measurement report MR data generated by multiple terminal devices, and based on the preset pairing strategy, XDR data and MR data are paired and handled, to obtain multiple first paired data;Obtain Internet of Vehicles data, and cluster processing is carried out on Internet of Vehicles data, to determine multiple Internet of Vehicles data clusters and multiple Internet of Vehicles data corresponding to each Internet of Vehicles data cluster;The first paired data corresponding to Internet of Vehicles data in first paired data is determined as second paired data, to construct three-dimensional quality difference model based on multiple Internet of Vehicles data corresponding to each Internet of Vehicles data cluster and second paired data corresponding to each Internet of Vehicles data;Determine the multi-dimensional Internet of Vehicles data feature based on three-dimensional quality difference model and first paired data.

Description

Technical Field

[0001] This application belongs to the field of vehicle networking, and specifically relates to a method, device, equipment, storage medium and product for extracting vehicle networking data features. Background Technology

[0002] The new infrastructure of the Internet of Vehicles (IoV) breaks down the data boundaries between vehicles and people, cities, and transportation infrastructure, accelerating the connection between people, vehicles, roads, and the cloud, and bringing more mobile connections and data traffic demands. As the IoV becomes more sophisticated, wireless communication technologies such as LTE-V2X and 5G are deeply integrated with fiber optic networks, providing communication services with wide coverage, low latency, and high reliability. However, according to incomplete statistics, under the existing traffic technology system, the phenomenon of frequent traffic accidents remains prominent, with the number of road traffic accidents showing a generally increasing trend year by year, and even a worsening trend. This undoubtedly highlights the urgency of improving traffic safety management. In the IoV environment, numerous factors affect network performance, and these factors are often dynamic, real-time, and random. For example, vehicle speed, road conditions, weather conditions, and network signal strength are all intertwined and work together to affect the operation of the IoV system.

[0003] Therefore, in order to effectively cope with complex and ever-changing factors, a method is needed that can accurately extract data features from vehicle network data. Summary of the Invention

[0004] This application provides a method for extracting features from vehicle network data, which can accurately extract data features from vehicle network data.

[0005] In a first aspect, embodiments of this application provide a method for extracting features from vehicle-to-everything (V2X) data. The method includes: acquiring Extended Detail Record (XDR) data and Measurement Report (MR) data generated by various terminal devices, and performing pairing processing on the XDR data and MR data based on a preset pairing strategy to obtain multiple first paired data sets, wherein the terminal devices include V2X terminals; acquiring V2X data and performing clustering processing on the V2X data to determine multiple V2X data clusters and multiple V2X data sets corresponding to each V2X data cluster; determining the first paired data set corresponding to the V2X data set as second paired data sets, thereby constructing a three-dimensional quality difference model based on the multiple V2X data sets corresponding to each V2X data cluster and the second paired data sets corresponding to each V2X data set; and determining multi-dimensional V2X data features based on the three-dimensional quality difference model and the first paired data sets.

[0006] Secondly, embodiments of this application provide a vehicle-to-everything (V2X) data feature extraction device, comprising: a first acquisition module, configured to acquire Extended Detail Record (XDR) data and Measurement Report (MR) data generated by various terminal devices, and perform pairing processing on the XDR data and the MR data based on a preset pairing strategy to obtain multiple first paired data, wherein the terminal devices include a V2X terminal; a second acquisition module, configured to acquire V2X data, and perform clustering processing on the V2X data to determine multiple V2X data clusters and multiple V2X data corresponding to each V2X data cluster; a first determination module, configured to determine the first paired data corresponding to the V2X data in the first paired data as second paired data, so as to construct a three-dimensional quality difference model based on the multiple V2X data corresponding to each V2X data cluster and the second paired data corresponding to each V2X data; and a second determination module, configured to determine multi-dimensional V2X data features based on the three-dimensional quality difference model and the first paired data.

[0007] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0008] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0009] Fifthly, embodiments of this application provide a computer program product that, when executed by a processor, implements the steps of the method described in the first aspect.

[0010] In a sixth aspect, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0011] In this embodiment, the extended detailed data (XDR) and measurement report (MR) data generated by various terminal devices are first acquired, and the XDR and MR data are paired based on a preset pairing strategy to obtain multiple first paired data. The terminal devices include vehicle networking terminals. Then, vehicle networking data is acquired and clustered to determine multiple vehicle networking data clusters and multiple vehicle networking data corresponding to each cluster. Next, the first paired data corresponding to the vehicle networking data in the first paired data is determined as second paired data. A three-dimensional quality difference model is constructed based on the multiple vehicle networking data corresponding to each cluster and the second paired data corresponding to each cluster. Finally, multi-dimensional vehicle networking data features are determined based on the three-dimensional quality difference model and the first paired data, enabling rapid and accurate extraction of data features from the vehicle networking data. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a method for extracting features from vehicle network data provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the second method for extracting vehicle network data features provided in this application embodiment; Figure 3 This is a flowchart illustrating the third method for extracting vehicle network data features provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of a vehicle network data feature extraction device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of a vehicle network data feature extraction device provided in an embodiment of this application. Detailed Implementation

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

[0014] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0015] The method for extracting vehicle network data features provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0016] Figure 1 This illustration shows an embodiment of a vehicle-to-everything (V2X) data feature extraction method. This method can be executed by an electronic device, which may include a server and / or a terminal device, such as an in-vehicle terminal or a mobile terminal. In other words, the method can be executed by software or hardware installed in a V2X data feature extraction device, and includes the following steps: Step 102: Obtain extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices, and perform pairing processing on the XDR data and the MR data based on a preset pairing strategy to obtain multiple first pairing data.

[0017] The terminal equipment includes a vehicle networking terminal.

[0018] The execution entity of the vehicle network data feature extraction method described in this application can be a vehicle network data feature extraction system, a vehicle network data feature extraction software, or other execution entities. This application embodiment will use a vehicle network data feature extraction system (hereinafter referred to as the feature extraction system) as an example for illustration.

[0019] The feature extraction system first acquires multiple Extended Detail Record (XDR) data and multiple Measurement Report (MR) data generated by various terminal devices. Then, based on a preset pairing strategy, it pairs the acquired XDR and MR data to obtain multiple pairing results, i.e., the first pairing data. The various terminal devices include vehicle-to-everything (V2X) terminals. These terminal devices are those capable of network interaction, such as mobile phones and V2X terminals. Furthermore, the various terminal devices refer to all network-interactive terminal devices currently in the network. Therefore, the XDR and MR data from these terminal devices also represent all XDR and MR data generated in the current network.

[0020] In other words, the feature extraction system first acquires XDR and MR data generated by terminals capable of network interaction, then pairs the acquired MR and XDR data to determine the MR and XDR data that are related to each other, and then merges the related XDR and MR data to obtain the merged result, which is the first pairing data.

[0021] Specifically, when pairing XDR and MR data, the feature extraction system can merge XDR and MR data with the same generation time and device code based on the data generation time and device code included in the XDR and MR data to obtain the merged result. In other words, the preset pairing strategy can be a pairing strategy based on the same data generation time and the same device code, or it can be other pairing strategies.

[0022] Furthermore, after acquiring XDR and MR data, the feature extraction system can perform data checks and cleaning operations on the XDR and MR data separately to ensure that the merged XDR and MR data are healthy data. For example, the feature extraction system identifies and removes invalid data (such as null values, outliers, and data with incorrect formats) and duplicate data from the XDR and MR data to ensure data accuracy and uniqueness, providing a clean data foundation for multi-dimensional massive data analysis; the feature system performs conversion and standardization processing according to a unified data format, converting the time format in the XDR and MR data to a unified "year-month-day hour:minute:second" format; and it encodes base station identifiers, converting base station numbers in different formats into a unified 10-digit code.

[0023] Step 104: Obtain vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

[0024] The feature extraction system acquires multiple vehicle network data points, and then performs clustering processing on these multiple vehicle network data points to determine multiple clustering results (i.e., multiple vehicle network data clusters) and the data included in each clustering result (i.e., multiple vehicle network data corresponding to each vehicle network data cluster).

[0025] In other words, the feature extraction system first acquires multiple pieces of vehicle-to-everything (V2X) data, and then performs clustering processing on these multiple pieces of V2X data to achieve the purpose of classifying the multiple pieces of V2X data. After classifying the multiple pieces of V2X data, the feature extraction system can identify multiple types of V2X data (i.e., multiple V2X data clusters) and the V2X data corresponding to each type (i.e., multiple V2X data corresponding to each V2X data cluster).

[0026] Specifically, when clustering multiple vehicle network data, the feature extraction system can use algorithms such as SparkKmeans clustering; when clustering vehicle network data according to clustering algorithms, it can perform clustering analysis based on multi-dimensional information such as location information and speed information in each vehicle network data.

[0027] Step 106: Determine the first pairing data corresponding to the vehicle network data in the first pairing data as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data.

[0028] After determining the first pairing data, the feature extraction system identifies the first pairing data that corresponds to the vehicle network data in the first pairing data, and determines the first pairing data corresponding to each vehicle network data as the second pairing data.

[0029] Specifically, the XDR and MR data acquired by the feature extraction system are data from multiple terminal devices, including vehicle-to-everything (V2X) terminals. Therefore, the XDR and MR data include data from V2X terminals. Thus, the first pairing data obtained by pairing the XDR and MR data also includes data from V2X terminals. The feature extraction system then determines the first pairing data corresponding to each V2X data from the first pairing data and identifies the corresponding first pairing data as the second pairing data.

[0030] After determining the second pairing data, the feature extraction system constructs a three-dimensional quality difference model based on multiple vehicle-to-everything (V2X) data corresponding to each V2X data cluster and the second pairing data corresponding to each V2X data cluster. In other words, the feature extraction system constructs a three-dimensional quality difference model based on multiple V2X data corresponding to each type.

[0031] Specifically, when constructing the three-dimensional quality difference model, the feature extraction system determines the signal information in multiple vehicle network data included in each vehicle network data cluster, such as the reference signal received power (RSRP) information and the signal-to-interference-plus-noise ratio (SINR) information. Then, based on the multiple signal information included in each vehicle network data cluster, the three-dimensional quality difference model is iteratively trained to obtain the pre-trained three-dimensional quality difference model.

[0032] Step 108: Determine multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data.

[0033] After determining the three-dimensional quality difference model, the feature extraction system determines multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first paired data. These multi-dimensional vehicle network data features include time dimension, location dimension, etc.

[0034] Specifically, the data included in the three-dimensional quality difference model is the signal feature data of the vehicle network data. The feature extraction system can determine the overall signal feature data of the vehicle network data through the three-dimensional quality difference model. The first pairing data includes not only terminal data in the signal dimension, but also information in the time, location and other dimensions. Therefore, by combining the three-dimensional quality difference model and the first pairing data, the feature extraction system can obtain the signal feature data of the vehicle network data in different time periods and locations in multiple dimensions.

[0035] More specifically, the feature extraction system leverages the MapReduce model in Hadoop to store the associated data on different nodes to complete cluster operation. It follows a three-stage approach: "reduce redundancy first, then implement parallelism, and finally enhance hardware support." Based on data scale, dimensionality, and accuracy requirements, it flexibly combines distributed, sampling, approximation, indexing, and GPU technologies to control costs while ultimately integrating the extracted data with other data into the data foundation. Through multi-dimensional modeling methods such as spatial and scenario analysis, it flexibly adjusts the required fields to the reporting center and outputs RSR GIS maps, SINR GIS maps, and low-quality GIS maps based on tile maps, presenting multi-dimensional information indicators.

[0036] The vehicle-to-everything (V2X) data feature extraction method provided in this invention first acquires Extended Detail Record (XDR) data and Measurement Report (MR) data generated by various terminal devices, and performs pairing processing on the XDR data and MR data based on a preset pairing strategy to obtain multiple first paired data. The terminal devices include V2X terminals. Then, V2X data is acquired and clustered to determine multiple V2X data clusters and multiple V2X data corresponding to each V2X data cluster. Next, the first paired data corresponding to the V2X data in the first paired data is determined as second paired data. A three-dimensional quality difference model is constructed based on the multiple V2X data corresponding to each V2X data cluster and the second paired data corresponding to each V2X data. Finally, multi-dimensional V2X data features are determined based on the three-dimensional quality difference model and the first paired data, which can accurately extract data features from V2X data.

[0037] In one implementation, the step of clustering the vehicle network data to determine multiple vehicle network data clusters and multiple sets of vehicle network data corresponding to each vehicle network data cluster (step 104) can be performed via steps A1-A2: Step A1: Based on the GPS sub-data, IMSI sub-data, and discrete nickname ID sub-data in each of the vehicle network data, determine the vehicle network feature vector of each of the vehicle network data.

[0038] When performing clustering processing on vehicle network data, the feature extraction system first obtains the GPS sub-data, IMSI sub-data, and discrete nickname ID sub-data included in each vehicle network data, and then determines the feature vector (i.e., vehicle network feature vector) corresponding to each vehicle network data based on the GPS sub-data, IMSI sub-data, and ID sub-data.

[0039] Specifically, during cluster analysis, the feature extraction system integrates multi-dimensional feature parameters such as the vehicle's unique identifier (IMEI), geospatial coordinates (GPS), communication behavior feature vector (IMSI), and vehicle speed to obtain the vehicle network feature vector (e.g., unifying GPS, IMSI, and discrete IDs into a weighted average, and mapping GPS, IMSI, and discrete IDs to the same mixed space), which facilitates subsequent clustering of vehicle network data.

[0040] Step A2: Cluster the vehicle network feature vectors using a preset clustering algorithm to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

[0041] After determining the vehicle network feature vectors corresponding to each vehicle network data, the feature extraction system performs clustering processing on multiple vehicle network feature vectors using a preset clustering algorithm, thereby determining the clustering results (multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster). The preset clustering algorithm can be the Spark KMeans clustering algorithm.

[0042] Furthermore, the feature extraction system uses an improved k-means algorithm for clustering. The core of the improved k-means algorithm is no longer randomness, but rather the use of density + maximum distance K-means, making it a customized version for the Internet of Vehicles. It uses a hybrid calculation of "weighted average + vector median + mode" for different types of features (i.e., different Internet of Vehicles data clusters), which significantly reduces the impact of outliers. Based on the current cluster center, the sample data is relabeled and assigned to the corresponding cluster.

[0043] Specifically, during clustering, the feature extraction system initially selects K baseline cluster centers. Then, based on distance metrics within the feature space, it assigns each data sample point (vehicle network data or vehicle network feature vector) to the cluster corresponding to the nearest cluster center. After data point allocation, the feature extraction system recalculates the center parameters of each cluster and performs the data point allocation operation again. This process iterates until the variation in cluster centers reaches a preset convergence threshold, or the number of iterations reaches a set upper limit. Through this iterative optimization process, vehicle network users and devices with similar characteristics can be aggregated into the same cluster, thereby efficiently identifying and classifying user distribution and behavior patterns (e.g., smooth driving mode, constant acceleration mode, etc.) within the vehicle network area, and accurately extracting vehicle network data from the associated dataset.

[0044] The Spark Kmeans clustering algorithm provided in this application is improved in the following three aspects compared with the prior art: the distance function unifies GPS, IMSI, and discrete ID into a weighted and normalized mixed space; the initial center selection uses the Kmeans with the largest density and distance, making it a customized version for the Internet of Vehicles, reducing randomness and improving convergence speed; the center update strategy adopts a mixed calculation of "weighted average + vector median + mode" for different types of features, which significantly reduces the impact of outliers.

[0045] In one implementation, the step of constructing a three-dimensional quality difference model based on multiple vehicle network data corresponding to each vehicle network data cluster and the second paired data corresponding to each vehicle network data (step 106) can be performed via steps B1-B3: Step B1: Obtain the preset latency threshold and download speed threshold.

[0046] Before constructing a three-dimensional quality difference model based on multiple vehicle network data clusters, multiple vehicle network data corresponding to each vehicle network data cluster, and second paired data corresponding to each vehicle network data, the feature extraction system first needs to obtain various data required to construct the three-dimensional quality difference model.

[0047] Specifically, the feature extraction system first obtains the preset latency threshold and download speed threshold required to construct the three-dimensional poor quality model. The latency threshold and download speed threshold are preset by the operator. For example, the operator sets the latency threshold to 100ms and the download speed threshold to 10Mbps based on the vehicle network service perception test.

[0048] Step B2: Determine the Reference Signal Received Power (RSRP) data and the Signal-to-Interference-plus-Noise Ratio (SINR) data based on the vehicle network data and the second pairing data.

[0049] Before constructing the three-dimensional quality difference model, the feature extraction system also needs to determine the reference signal received power (RSRP) data and signal-to-interference-plus-noise ratio (SINR) data corresponding to each vehicle network data based on multiple vehicle network data and the second pairing data corresponding to each vehicle network data.

[0050] Step B3: Construct the three-dimensional quality difference model based on the latency threshold, the download rate threshold, the RSRP data, and the SINR data using a multinomial regression strategy.

[0051] After determining the latency threshold, download rate threshold, RSRP data, and SINR data, the feature extraction system constructs the three-dimensional quality poor model based on the acquired data (i.e., latency threshold, download rate threshold, RSRP data, and SINR data) using a multinomial regression strategy.

[0052] When constructing the three-dimensional quality defect model, the voltage level and SINR interact with the latency and rate, exhibiting a curvilinear relationship. Therefore, the feature extraction system uses methods such as multinomial regression to construct the three-dimensional quality defect model. When the voltage level is <-106dBm and SINR is <-3, the proportion of low-perception sampling points increases sharply. The R² evaluation of this model exceeds 0.82, ultimately forming a quality defect judgment basis and decision-making scheme that can directly guide network optimization and service scheduling.

[0053] Specifically, the feature extraction model uses a regression polynomial method to construct a three-dimensional quality difference model to accurately calibrate the threshold of the vehicle network quality difference model. While preserving the interpretability of the model, by introducing x² and x³ terms, it can accurately capture the inflection points of the "RSRP-SINR, Rate" and "RSRP-SINR, Timedelay" curves, reducing the fitting residual by more than 28%. Furthermore, cross-validation can avoid the overfitting risk brought by polynomials of order 4 and above.

[0054] The steps for constructing a three-dimensional poor quality model using the feature extraction system are as follows: First, batch import the original samples after fusing base station MR and XDR into memory. Second, use a third-order polynomial to capture the nonlinear curvature in the relationship between RSRP-SINR and Rate / Timedelay, maintaining a balance between model complexity and overfitting risk. Third, quantify the goodness of fit through cross-validation R², RMSE, and residual diagnosis. Fourth, visualize the curves and confidence intervals to intuitively demonstrate the fit reliability and provide a continuous mapping function for threshold calibration. Fifth, jointly solve the critical intersection point of RSRP < -106 dBm and SINR < -3 dB on the fitted curve, using a dual-threshold approach to precisely lock the terminal perception degradation boundary in the weak coverage-high interference coupling area, ultimately obtaining a pre-trained three-dimensional poor quality model.

[0055] In one implementation, the step of determining multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data (step 108) can be performed via steps C1-C2: Step C1: Determine the scenario information of the vehicle network data based on the first pairing data.

[0056] The scenario information includes data generation vendor information, data generation location information, and data generation time information.

[0057] After determining the three-dimensional quality difference model and the first pairing data, the feature extraction system determines the scene information of the vehicle network data based on the first pairing data. The scene information includes data manufacturer information, data generation location information, data generation time information, data professional type information, data domain information, and data application scenario domain information. Among them, the data manufacturer information refers to the manufacturer of the terminal equipment to which the vehicle network data belongs; the data generation location information can be the province, city, or community where the data was generated; the data generation time information can be the hour, minute, second, day, month, or year of data generation; the data professional type information can be northbound data, DPI-collected data, asset management data, or user data; the data domain information can be O domain information, B domain information, or M domain information; and the data application scenario domain information can be high-speed rail information, highway information, or low-altitude information.

[0058] Step C2: Determine the vehicle network data characteristics for various scenarios based on the scene information and the three-dimensional quality difference model.

[0059] After determining the various scenario information of the vehicle network data, the feature extraction system determines the vehicle network data features in each scenario based on various scenario information and the three-dimensional quality difference model, and then determines the vehicle network data features in multiple scenarios based on multiple scenario information.

[0060] Specifically, the feature extraction system can determine the target scene data features corresponding to the target scene based on the user's target scene feature display request, and then display them to the user. The target scene feature display request is used to request the feature extraction system to determine the vehicle-to-everything (V2X) data features under the target scene. The target scene is the scene to which the user expects the data features to belong, and the target scene data features are the V2X data features corresponding to the target scene. For example, when a user expects to obtain V2X data features within 15 minutes, the target scene is "within 15 minutes," and the target scene data features are the V2X data features within a preset area within 15 minutes. When a user expects to obtain V2X data features for manufacturer B, the target scene is "manufacturer B," and the target scene data features are the V2X data features of manufacturer B within a preset area. When a user expects to obtain V2X data features for Hebei Province, the target scene is "Hebei Province," and the target scene data features are the V2X data features for Hebei Province.

[0061] Figure 2 This is a flowchart illustrating a second method for extracting vehicle network data features according to an embodiment of this specification, as shown below. Figure 2 As shown, the schematic diagram includes: Step 202: Obtain extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices, and perform pairing processing on the XDR data and the MR data based on a preset pairing strategy to obtain multiple first pairing data.

[0062] The terminal equipment includes a vehicle networking terminal.

[0063] Step 204: Obtain multiple vehicle network data.

[0064] Step 206: Based on the GPS sub-data, IMSI sub-data, and discrete nickname ID sub-data in each of the vehicle network data, determine the vehicle network feature vector of each of the vehicle network data.

[0065] Step 208: Cluster the vehicle network feature vectors using a preset clustering algorithm to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

[0066] Step 210: Determine the first pairing data corresponding to the vehicle network data in the first pairing data as the second pairing data.

[0067] Step 212: Obtain the preset latency threshold and download speed threshold.

[0068] Step 214: Determine the Reference Signal Received Power (RSRP) data and the Signal-to-Interference-plus-Noise Ratio (SINR) data based on the vehicle network data and the second pairing data.

[0069] Step 216: Construct a three-dimensional quality difference model based on the latency threshold, the download rate threshold, the RSRP data, and the SINR data using a multinomial regression strategy.

[0070] Step 218: Determine the scenario information of the vehicle network data based on the first pairing data.

[0071] The scenario information includes data generation vendor information, data generation location information, and data generation time information.

[0072] Step 220: Determine the vehicle network data characteristics for various scenarios based on the scene information and the three-dimensional quality difference model.

[0073] In the embodiments described in the specification, multi-feature intelligent clustering is used: using IMEI, GPS, IMSI, speed, etc. as feature vectors, Spark KMEANs iterative clustering is run to achieve integrated clustering of "vehicle-person-device-number", which facilitates subsequent refined operations such as network slicing, business prioritization, and fault delimitation by "cluster"; through quality difference model analysis: on the associated high-reliability dataset, multinomial regression is used to establish three-dimensional quality difference models of "level-latency" and "SINR-rate", with an error ≤5%, directly outputting the three root causes of "weak coverage, high interference, and low experience" and optimization thresholds, forming a front-line "diagnosis-decision" manual; through distributed indicator output: relying on Hadoop / MapReduce, the association and indicator calculation are distributed to multiple nodes, which can be summarized at the hourly level according to any dimension such as "terminal-scenario-city", supporting real-time presentation through multiple channels such as large screens, work orders, and APP, and realizing minute-level instant response of "problem discovery-dispatch-closure".

[0074] In one implementation, after acquiring the extended detail data (XDR) and measurement report (MR) data generated by various terminal devices (step 102), steps D1-D4 may also be performed: Step D1: Determine the number of the first cells based on the XDR data.

[0075] The first cell number refers to the number of base station cells included in the XDR data.

[0076] After acquiring XDR and MR data, the feature extraction system can perform integrity checks on the XDR and MR data to determine whether the acquired XDR and MR data are complete data generated by various terminal devices. Subsequent operations are only performed after ensuring that the XDR and MR data are complete data.

[0077] When determining whether the acquired XDR data and MR data are complete data generated by multiple terminal devices, the feature extraction system first determines the number of first cells based on the acquired XDR data, wherein the number of first cells is the number of base station cells included in the overall XDR data acquired by the feature extraction system.

[0078] Specifically, the XDR data acquired by the feature extraction system consists of XDR data from multiple terminal devices, which are terminal devices capable of network interaction. In other words, the acquired XDR data is all the XDR data generated in the current network. Therefore, the first cell number determined based on the XDR data is the number of base station cells involved in all the XDR data generated in the current network.

[0079] Step D2: Determine the number of second cells based on the MR data.

[0080] The second cell number refers to the number of base station cells included in the MR data.

[0081] When determining whether the acquired XDR data and MR data are complete data generated by multiple terminal devices, the feature extraction system will also determine the number of second cells based on the acquired MR data. The number of second cells is the number of base station cells included in the complete MR data acquired by the feature extraction system.

[0082] Specifically, the MR data acquired by the feature extraction system consists of XDR data from multiple terminal devices, which are terminal devices capable of network interaction. In other words, the acquired MR data is all the MR data generated in the current network. Therefore, the number of the second cell determined based on the MR data is the number of base station cells involved in all the MR data generated in the current network.

[0083] Step D3: Determine whether the XDR data and the MR data are complete based on the number of the first cell and the number of the second cell.

[0084] After determining the number of the first and second cells, the feature extraction system determines whether the XDR and MR data are complete based on these numbers. To verify the completeness of the XDR and MR data, the feature extraction system can perform two cell count checks. For example, the system first checks if the number of the first and second cells are the same for a first integrity verification. Then, after the first integrity verification passes, it checks if both the first and second cell counts are the existing cell counts in the current network for a second integrity verification.

[0085] Specifically, when determining the completeness of XDR and MR data based on the first and second cell counts, the feature extraction system obtains the actual number of cells in the current network. Then, it determines the completeness of the XDR data based on the actual number of cells in the current network and the first cell count, and determines the completeness of the MR data based on the actual number of cells in the current network and the second cell count. For example, the feature extraction system determines whether the actual number of cells in the current network and the first cell count are the same. If they are the same, the XDR data is considered complete; if they are different, the XDR data is considered incomplete. Similarly, it determines whether the actual number of cells in the current network and the second cell count are the same. If they are the same, the MR data is considered complete; if they are different, the MR data is considered incomplete.

[0086] More specifically, the XDR and MR data obtained by the feature extraction system are all the XDR and MR data generated in the current network. Therefore, theoretically, the number of base station cells designed in the XDR and MR data is the same as the number of base station cells in the current network. Thus, the feature extraction system can determine the integrity of the XDR and MR data based on the number of the first cell, the number of the second cell, and the actual number of cells in the current network.

[0087] In other words, the feature extraction system compares the number of cells obtained from XDR data and MR data. If the number of cells obtained from MR data equals the number obtained from XDR data, the first cell count verification passes. If the number of cells obtained from MR data is less than the number obtained from XDR data, the cause should be investigated on the MR data collection side, and the accuracy of the correlation between XDR and MR data should be improved after the issue is resolved. If the number of cells obtained from MR data is greater than the number obtained from XDR data, the cause should be investigated on the XDR data collection side. During the second cell count verification, the feature extraction system counts the number of cells in the XDR and MR data. The counted number of cells can be compared with the number of cells in the current network. If the counted number of cells is less than the number of cells in the current network, the cause should be investigated on the collection side, such as checking the base station configuration or whether there was file loss during data transmission.

[0088] Step D4: If so, the XDR data and the MR data are paired based on a preset pairing strategy to obtain multiple first paired data.

[0089] After confirming that both the XDR and MR data are complete data, the feature extraction system then pairs the XDR and MR data based on a preset pairing strategy to determine multiple first pairing data.

[0090] In one implementation, the XDR data and MR data are paired based on a preset pairing strategy to obtain multiple first paired data (step 102), and steps E1-E3 can be executed: Step E1: Based on the first preset rule, the XDR data and the MR data are paired to obtain the third paired data.

[0091] The first preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time is the same.

[0092] When pairing XDR data and MR data based on a preset pairing strategy, the preset pairing strategy can be a single pairing strategy or a multiple pairing strategy, such as a two-pairing strategy or a three-pairing strategy. That is, the feature extraction system can perform multiple pairings using multiple different pairing strategies to determine the first pairing data.

[0093] Specifically, the feature extraction system can determine the first pairing data through a two-stage pairing strategy. During the first pairing, the feature extraction system performs pairwise pairing processing on XDR data and MR data based on a first preset rule, and obtains the processing result, namely the third pairing data. The first preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time is the same.

[0094] In other words, during the first pairing (or data association), the feature extraction system combines data from the MR data showing the same time, location, and user for the same MME_UE_S1AP_ID with the XDR data for the same user. Ensuring the filtering fields are identical, for scenarios with equal site IDs and the same MME_UE_S1AP_ID, the start and end times of the MR are calculated to determine the difference, sorted by time difference, and the data with the smallest time difference is selected as the reliable data for the first association, i.e., the third pairing data.

[0095] Step E2: Based on the second preset rule, the XDR data and the MR data are paired to obtain the fourth paired data.

[0096] The second preset rule is that the base station identifiers are the same, the user identifiers are the same, and the data generation time interval is less than the first threshold.

[0097] After determining the third pairing data, the feature extraction system performs a second data pairing. During the second data pairing, the feature extraction system can pair the XDR data and MR data based on the second preset rule to obtain the fourth pairing data, or it can pair the remaining XDR data and remaining MR data based on the second preset rule to obtain the fourth pairing data. The second preset rule is that the base station identifiers are the same, the identifiers are the same, and the data generation time interval is less than a first threshold, such as 20 minutes. The remaining XDR data is the XDR data that has not undergone the first pairing, and the remaining MR data is the MR data that has not undergone the first pairing. In other words, the remaining XDR data is the XDR data that remains after the first pairing, and the remaining MR data is the MR data that remains after the first pairing.

[0098] Specifically, during the second pairing, the feature extraction system can first identify the remaining unpaired XDR and MR data after the first pairing in all XDR and MR data, and then pair the remaining XDR and MR data based on the second pairing rule to obtain the fourth pairing data.

[0099] In other words, the second data association involves the feature extraction system re-associating unassociated data based on the same field, MISIDN. The system then retrieves data from 20 minutes prior to the current time period and 20 minutes forward, performing search matching, merging, and deduplication.

[0100] Step E3: Determine the first pairing data based on the third pairing data and the fourth pairing data.

[0101] After determining the third and fourth paired data, the feature extraction system uses the sum of the third and fourth paired data as the first paired data. After determining the first paired data, the feature extraction system can store the determined first paired data in a preset database for subsequent operations.

[0102] Specifically, when the feature extraction system performs multiple pairings, it can store the paired data obtained from each pairing into a preset database after each pairing is completed. Then, after completing the last pairing and storing the paired data into the preset database, it can deduplicate all the paired data in the database to obtain the first paired data.

[0103] Figure 3 This is a flowchart illustrating a third method for extracting vehicle network data features according to an embodiment of this specification, as shown below. Figure 3 As shown, the schematic diagram includes: Step 302: Obtain extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices.

[0104] The terminal equipment includes a vehicle networking terminal.

[0105] Step 304: Determine the number of the first cells based on the XDR data.

[0106] The first cell number refers to the number of base station cells included in the XDR data.

[0107] Step 306: Determine the number of second cells based on the MR data.

[0108] The second cell number refers to the number of base station cells included in the MR data.

[0109] Step 308: Determine whether the XDR data and the MR data are complete based on the first number of cells and the second number of cells.

[0110] Step 310: If so, the XDR data and the MR data are paired according to the first preset rule to obtain the third paired data.

[0111] The first preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time is the same.

[0112] Step 312: Based on the second preset rule, the XDR data and the MR data are paired to obtain the fourth paired data.

[0113] The second preset rule is that the base station identifiers are the same, the user identifiers are the same, and the data generation time interval is less than the first threshold.

[0114] Step 314: Determine the first pairing data based on the third pairing data and the fourth pairing data.

[0115] Step 316: Obtain vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

[0116] Step 318: Determine the first pairing data corresponding to the vehicle network data in the first pairing data as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data.

[0117] Step 320: Determine multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data.

[0118] In the embodiments described in the specification, by verifying the number of cells, it can be ensured that subsequent association and modeling are built on a "clean and reliable" data foundation, providing a solid data basis for subsequent data extraction and analysis. By adopting a "two-stage association" strategy: first, precise matching is performed using "same time-location-user", and then fuzzy filling is performed 20 minutes before and after MISDN, improving the matching rate of MR and XDR from the local optimum of the first association to the global optimum of the second association, ensuring the integrity of user-time-space three-dimensional trajectory data.

[0119] It should be understood that the training and prediction processes of the AI ​​models involved in the various embodiments of this specification all adhere to multiple legal and compliant principles, including legal data sources, compliant data content, compliant data governance, compliant training objectives and schemes, compliant training processes, compliant training environments and tools, and compliant ethical verification of training results, and comply with the requirements of Article 5 of the Patent Law. Among them: Data source legitimacy: All datasets used for AI model training were obtained through legal means, covering three categories: publicly authorized data, data authorized by partners, and self-collected compliant data. Publicly authorized data comes from compliant data sources following open-source licenses such as Apache 2.0, with complete copyright attribution and authorization scope clearly marked, and no unauthorized open-source code or data reuse. Data authorized by partners has been subject to formal data usage agreements, clearly defining the scope, duration, and confidentiality obligations, and possessing a complete authorization chain. For self-collected data involving personal information, strict informed consent procedures have been followed, and anonymization processes (including but not limited to field masking, feature anonymization, and differential privacy technology applications) have been implemented to remove personally identifiable information, fully complying with the requirements of relevant laws and regulations such as the "Interim Measures for the Administration of Generative Artificial Intelligence Services" and the "Personal Information Protection Law."

[0120] Data content compliance: The AI ​​model's dataset undergoes multiple screenings and cleaning processes to remove all content that may violate social morality or harm public interests. It contains no obscene, pornographic, violent, discriminatory, or information that endangers national or public safety, nor does it involve the illegal acquisition or use of genetic resources. For data in sensitive fields (such as healthcare and finance), an additional privacy-preserving computation module (including federated learning and secure multi-party computation technologies) ensures that the data is "usable but not visible," avoiding compliance risks during the original data transmission process and ensuring that the data application scenarios and uses comply with public order and good morals and industry regulatory requirements.

[0121] Data governance norms: A complete data traceability system is established during the AI ​​model training process to automatically record the source, collection time, annotation process, cleaning rules, and permission allocation of training data, generating traceable compliance reports to ensure that the data is verifiable throughout its entire lifecycle. The dataset annotation process for AI models is completed by a professional human R&D team, clearly defining the proportion of human creative contributions and avoiding reliance on AI-generated data that has not undergone substantial human modification, thus meeting the examination requirements for "human main contributions" in AI patent applications.

[0122] Training objectives and plans are compliant: The AI ​​model training objective focuses on training scenarios for three-dimensional poor-quality models. The training scheme and final output results do not violate any mandatory provisions of laws and administrative regulations, do not harm the public interest or the legitimate rights and interests of others, and do not pose any potential risks of being used for illegal activities, infringing on privacy, or disrupting public safety. It strictly adheres to the ethical principle of "intelligent for good".

[0123] Training process compliance: A closed-loop training framework is adopted to ensure compliance and controllability of the training process. The specific process is as follows: First, training samples are obtained through compliant data sources. After the aforementioned data cleaning and desensitization, they are input into the neural network model to generate preliminary training results. Second, an expert system is introduced to verify the preliminary results. Based on preset rules and human expert experience, the feasibility of the results is evaluated, and outputs that may pose ethical risks or compliance hazards are corrected (such as removing decision-making logic that violates public order and good morals, and adjusting model parameters that do not comply with safety regulations). Finally, the loss function weights are dynamically optimized based on expert system feedback to strengthen the model's learning of compliant results, avoid overfitting errors or non-compliant labels, and form a closed-loop control of "data input - model training - expert verification - parameter optimization - result feedback" to ensure that the entire training process complies with A5 ethical review requirements.

[0124] Training environment and tool compliance: AI model training is implemented using nationally licensed chips and a compliant training platform. All open-source frameworks and components used in the training process have obtained their corresponding licenses, and copyright statements and patent citation information are fully retained, with no instances of infringement or reuse. The training environment is built using virtual devices (containers / virtual machines) with fixed random seeds and initial parameter configurations to ensure the reproducibility of the training process. Furthermore, through access control and operation log recording, risks such as data leakage and parameter tampering during training are prevented, ensuring the security and compliance of the training process.

[0125] Training results ethical verification compliance: After the model is trained, it undergoes additional third-party ethical compliance assessment and algorithm filing review to verify that the model output does not violate social morality or harm public interests. For potentially sensitive scenarios (such as public services and intelligent decision-making), a special result verification mechanism is established to ensure that the model always complies with Article 5 of the Patent Law and relevant laws and regulations in practical applications.

[0126] In summary, the data and training process used in the AI ​​model of this specification strictly comply with the relevant provisions of Article 5 of the Patent Law and the Patent Examination Guidelines (2023 Edition), and there are no violations of laws, social ethics, public interests, or illegal use of genetic resources. It fully meets the compliance requirements for patent authorization.

[0127] It should be noted that the vehicle network data feature extraction method provided in this application embodiment can be executed by a vehicle network data feature extraction device, or a control module in the vehicle network data feature extraction device for executing the vehicle network data feature extraction method. This application embodiment uses the execution of the vehicle network data feature extraction method by a vehicle network data feature extraction device as an example to illustrate the vehicle network data feature extraction device provided in this application embodiment.

[0128] Figure 4 This is a schematic diagram of the structure of a vehicle network data feature extraction device according to an embodiment of the present invention. Figure 4 As shown, the vehicle network data feature extraction device includes: a first acquisition module 402, a second acquisition module 404, a first determination module 406, and a second determination module 408.

[0129] The first acquisition module 402 is used to acquire extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices, and to perform pairing processing on the XDR data and the MR data based on a preset pairing strategy to obtain multiple first pairing data. The terminal devices include vehicle networking terminals. The second acquisition module 404 is used to acquire vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster. The first determining module 406 is used to determine the first pairing data corresponding to the vehicle network data in the first pairing data as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data. The second determining module 408 is used to determine multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data.

[0130] The vehicle network data feature extraction in this application embodiment can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This application embodiment does not impose specific limitations.

[0131] The vehicle network data feature extraction device in this application embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit it.

[0132] The vehicle network data feature extraction device provided in this application embodiment can achieve... Figures 1 to 3 The various processes implemented in the method embodiments are not described in detail here to avoid repetition.

[0133] Based on the same technical concept, embodiments of this application also provide an electronic device for performing the above-described vehicle network data feature extraction method. Figure 5 This is a schematic diagram of the structure of an electronic device to implement various embodiments of this application. The electronic device can vary significantly due to differences in configuration or performance, and may include a processor 502, a communications interface 504, a memory 506, and a communication bus 508. The processor 502, communications interface 504, and memory 506 communicate with each other via the communication bus 508. The processor 502 can call a computer program stored in the memory 506 and executable on the processor 502 to perform the following steps: The system acquires Extended Detail Data (XDR) and Measurement Report (MR) data generated by various terminal devices, and performs pairing processing on the XDR data and MR data based on a preset pairing strategy to obtain multiple first pairing data. The terminal devices include vehicle networking terminals. Acquire vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster; The first pairing data corresponding to the vehicle network data in the first pairing data is determined as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data. Based on the three-dimensional quality difference model and the first pairing data, multi-dimensional vehicle network data features are determined.

[0134] In one implementation, the step of clustering the vehicle network data to determine multiple vehicle network data clusters and multiple sets of vehicle network data corresponding to each vehicle network data cluster includes: Based on the GPS sub-data, IMSI sub-data, and discrete nickname ID sub-data in each of the vehicle network data, determine the vehicle network feature vector of each of the vehicle network data; The vehicle network feature vectors are clustered using a preset clustering algorithm to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

[0135] In one implementation, constructing a three-dimensional quality difference model based on multiple vehicle-to-everything (V2X) data corresponding to each of the V2X data clusters and the second paired data corresponding to each of the V2X data clusters includes: Obtain the preset latency threshold and download speed threshold; Based on the vehicle network data and the second pairing data, the reference signal received power (RSRP) data and the signal-to-interference-plus-noise ratio (SINR) data are determined. The three-dimensional quality difference model is constructed using a multinomial regression strategy based on the latency threshold, the download rate threshold, the RSRP data, and the SINR data.

[0136] In one implementation, determining multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data includes: Based on the first pairing data, the scenario information of the vehicle network data is determined, and the scenario information includes data generation manufacturer information, data generation location information, and data generation time information. Based on the scene information and the three-dimensional quality difference model, the characteristics of vehicle network data in various scenarios are determined.

[0137] In one implementation, after acquiring the Extended Detail Data (XDR) and Measurement Report (MR) data generated by various terminal devices, the method further includes: The number of first cells is determined based on the XDR data, where the number of first cells is the number of base station cells included in the XDR data; The number of second cells is determined based on the MR data, where the number of second cells is the number of base station cells included in the MR data; The completeness of the XDR data and the MR data is determined based on the number of the first cell and the number of the second cell. If so, the XDR data and the MR data are paired based on a preset pairing strategy to obtain multiple first paired data.

[0138] In one implementation, the XDR data and MR data are paired based on a preset pairing strategy to obtain multiple first paired data, including: The XDR data and the MR data are paired based on a first preset rule to obtain third paired data. The first preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time is the same. The XDR data and the MR data are paired based on the second preset rule to obtain the fourth pairing data. The second preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time interval is less than the first threshold. The first pairing data is determined based on the third pairing data and the fourth pairing data.

[0139] The specific execution steps can be found in the various steps of the above-described embodiment of the vehicle network data feature extraction method, and can achieve the same technical effect. To avoid repetition, they will not be repeated here.

[0140] It should be noted that the electronic devices in the embodiments of this application include: servers, terminals, or other devices besides terminals.

[0141] The above electronic device structure does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or arrange them differently. For example, an input unit may include a Graphics Processing Unit (GPU) and a microphone, and a display unit may use a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar display panels. User input units include at least one of a touch panel and other input devices. A touch panel is also called a touchscreen. Other input devices may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be elaborated further here.

[0142] Memory can be used to store software programs and various data. Memory can primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area can store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, memory can include volatile memory or non-volatile memory, or both. 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. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0143] The processor may include one or more processing units; optionally, the processor integrates an application processor and a modem processor, wherein the application processor mainly handles operations related to the operating system, user interface, and applications, while the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into the processor.

[0144] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described vehicle network data feature extraction method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0145] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0146] This application also provides a computer program product. When the computer program product is executed by a processor, it implements the various processes of the above-described vehicle network data feature extraction method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0147] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described vehicle network data feature extraction method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0148] 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.

[0149] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0150] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0151] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A vehicle-to-everything data feature extraction method, characterized in that, include: The system acquires Extended Detail Data (XDR) and Measurement Report (MR) data generated by various terminal devices, and performs pairing processing on the XDR data and MR data based on a preset pairing strategy to obtain multiple first pairing data. The terminal devices include vehicle networking terminals. Acquire vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster; The first pairing data corresponding to the vehicle network data in the first pairing data is determined as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data. Based on the three-dimensional quality difference model and the first pairing data, multi-dimensional vehicle network data features are determined.

2. The method of claim 1, wherein, The step of clustering the vehicle network data to determine multiple vehicle network data clusters and multiple sets of vehicle network data corresponding to each vehicle network data cluster includes: Based on the GPS sub-data, IMSI sub-data, and discrete nickname ID sub-data in each of the vehicle network data, determine the vehicle network feature vector of each of the vehicle network data; The vehicle network feature vectors are clustered using a preset clustering algorithm to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster.

3. The method of claim 1, wherein, The construction of a three-dimensional quality difference model based on multiple vehicle-to-everything (V2X) data corresponding to each of the V2X data clusters and the second paired data corresponding to each of the V2X data clusters includes: Obtain the preset latency threshold and download speed threshold; Based on the vehicle network data and the second pairing data, the reference signal received power (RSRP) data and the signal-to-interference-plus-noise ratio (SINR) data are determined. The three-dimensional quality difference model is constructed using a multinomial regression strategy based on the latency threshold, the download rate threshold, the RSRP data, and the SINR data.

4. The method of claim 1, wherein, The determination of multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data includes: Based on the first pairing data, the scenario information of the vehicle network data is determined, and the scenario information includes data generation manufacturer information, data generation location information, and data generation time information. Based on the scene information and the three-dimensional quality difference model, the characteristics of vehicle network data in various scenarios are determined.

5. The method according to claim 1, characterized in that, After acquiring the extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices, the method further includes: The number of first cells is determined based on the XDR data, where the number of first cells is the number of base station cells included in the XDR data; The number of second cells is determined based on the MR data, where the number of second cells is the number of base station cells included in the MR data; The completeness of the XDR data and the MR data is determined based on the number of the first cell and the number of the second cell. If so, the XDR data and the MR data are paired based on a preset pairing strategy to obtain multiple first paired data.

6. The method according to claim 1, characterized in that, The XDR data and MR data are paired based on a preset pairing strategy to obtain multiple first paired data, including: The XDR data and the MR data are paired based on a first preset rule to obtain third paired data. The first preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time is the same. The XDR data and the MR data are paired based on the second preset rule to obtain the fourth pairing data. The second preset rule is that the base station identifier is the same, the user identifier is the same, and the data generation time interval is less than the first threshold. The first pairing data is determined based on the third pairing data and the fourth pairing data.

7. A device for extracting features from vehicle network data, characterized in that, include: The first acquisition module is used to acquire extended detailed list (XDR) data and measurement report (MR) data generated by various terminal devices, and to perform pairing processing on the XDR data and the MR data based on a preset pairing strategy to obtain multiple first pairing data. The terminal devices include vehicle networking terminals. The second acquisition module is used to acquire vehicle network data and perform clustering processing on the vehicle network data to determine multiple vehicle network data clusters and multiple vehicle network data corresponding to each vehicle network data cluster. The first determining module is used to determine the first pairing data corresponding to the vehicle network data in the first pairing data as the second pairing data, so as to construct a three-dimensional quality difference model based on the multiple vehicle network data corresponding to each vehicle network data cluster and the second pairing data corresponding to each vehicle network data. The second determining module is used to determine multi-dimensional vehicle network data features based on the three-dimensional quality difference model and the first pairing data.

8. An electronic device, characterized in that, The device includes: Processor; and A memory configured to store computer-executable instructions configured to be executed by the processor, the executable instructions including steps for performing the vehicle-to-everything (V2X) data feature extraction method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium is used to store computer-executable instructions that cause the computer to perform the vehicle network data feature extraction method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The method includes a computer program that, when executed by a processor, implements the vehicle network data feature extraction method according to any one of claims 1 to 6.

Images