Navigation terminal screening method, system and equipment and computer storage medium
By combining unsupervised learning and anomaly rules in the navigation terminal screening method, the problems of intelligence, accuracy and adaptability of navigation terminal screening have been solved, achieving efficient screening and health scoring, and improving the accuracy and transparency of navigation services.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIAOXIN BEIDOU (HAINAN) TECH CO LTD
- Filing Date
- 2026-02-12
- Publication Date
- 2026-05-26
AI Technical Summary
Existing navigation terminal screening methods are insufficient in terms of intelligence, accuracy, interpretability, and adaptability, making it difficult to effectively identify abnormal data patterns and human cheating behavior in navigation terminals.
By combining unsupervised learning and anomaly rules, non-abnormal navigation terminals are selected and a health score is generated through feature extraction, model anomaly score generation, anomaly rule matching, and TOPSIS score generation.
It achieves high accuracy, transparency, and intelligence in navigation terminal screening, and is highly adaptable. It can effectively identify complex anomalies and human cheating behaviors, reducing human intervention.
Smart Images
Figure CN122087652A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of traffic data analysis and processing technology, and more specifically, to a navigation terminal screening method, system, device, and computer storage medium. Background Technology
[0002] Navigation systems are now widely used in numerous fields such as vehicle monitoring, logistics, and smart agriculture, with the number of connected navigation terminals reaching hundreds of millions. However, with the continuous expansion of this network, a large amount of invalid, drifting, duplicate reporting, and even forged location data is generated due to hardware failures, software defects, signal interference, or malicious cheating. This abnormal location data not only severely pollutes the navigation data source but also directly affects the accuracy and fairness of upper-level operations such as mileage statistics, behavior analysis, and safety monitoring based on location data. It may even lead to erroneous decisions and economic losses. To avoid this, it is necessary to screen and identify problematic terminals.
[0003] In the process of screening problematic terminals, rework rate assessment based on working time, terminal status prediction based on deep learning, or comprehensive assessment based on evaluation models can be used. However, the rework rate assessment based on working time calculates the rework rate by statistically analyzing the working time of navigation terminals and the number of historical failure samples, thereby assessing the overall reliability of navigation terminals. Although this method has a quantitative assessment basis, the assessment dimension is singular, making it difficult to accurately distinguish specific abnormal data patterns. It also lacks sensitivity to sporadic failures, intermittent anomalies, and software-level data cheating behaviors. The terminal status prediction method based on deep learning uses a pre-trained deep neural network model to score the status of navigation terminals and identify potential anomaly types. This method has certain advantages in complex pattern recognition, but the model has black-box characteristics, resulting in poor interpretability of the output results. At the same time, it highly relies on a large amount of high-quality labeled data, has limited generalization ability to unseen new cheating patterns, and has high model iteration and maintenance costs. The comprehensive evaluation method based on the evaluation model constructs an evaluation system that includes multi-dimensional quality indicators and sets weights based on expert experience to achieve a comprehensive quantitative evaluation of the quality of navigation terminals. Although this method considers the fusion of multiple indicators, the weight configuration is highly subjective, the static model is difficult to adapt to the dynamic changes in the behavior of navigation terminals, and it lacks the ability to continuously evolve and optimize in a closed loop.
[0004] In conclusion, improving the intelligence, accuracy, interpretability, and adaptability of navigation terminal screening is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] The purpose of this application is to provide a navigation terminal screening method, which can, to a certain extent, solve the technical problems of how to improve the intelligence, accuracy, interpretability, and adaptability of navigation terminal screening. This application also provides a navigation terminal screening system, electronic device, and computer-readable storage medium.
[0006] To achieve the above objectives, this application provides the following technical solution: A method for screening navigation terminals, comprising: Obtain travel data for a target number of navigation terminals; Feature extraction is performed on the trip data to obtain trip features; The travel features are processed through unsupervised learning to generate a model anomaly score for the navigation terminal; The trip data is matched according to the set exception rules to obtain the exception rule trigger result; Based on the model anomaly score and the anomaly rule triggering results, non-abnormal navigation terminals are selected as candidate terminals. Based on the aforementioned travel characteristics, generate TOPSIS scores for candidate terminals; Based on the TOPSIS score, a health score is generated for candidate terminals, and navigation services are processed based on the trip data and health score of the candidate terminals.
[0007] Preferably, the step of processing the trip features through unsupervised learning to generate a model anomaly score for the navigation terminal includes: Detect whether all values of the travel feature meet a preset target threshold; If all values of the trip features satisfy the target threshold, then the corresponding navigation terminal is selected as a candidate terminal. If at least one value of the trip feature does not meet the target threshold, the trip feature is input into a pre-trained isolated forest model, and the model anomaly score of the navigation terminal output by the isolated forest model is received.
[0008] Preferably, the step of matching the trip data according to the set anomaly rules to obtain the anomaly rule trigger result includes: Obtain the set abnormal rules, which include repeated trajectory cheating rules, fixed position false driving rules, high frequency invalid data rules, signal pattern abnormal rules, and behavior logic contradiction rules; The trip data is matched according to the aforementioned anomaly rules to obtain the anomaly rule triggering result.
[0009] Preferably, the step of selecting non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the anomaly rule triggering result includes: Detect whether the abnormal score of the model is less than a preset threshold; If the abnormal score of the model is greater than or equal to the preset threshold, the corresponding navigation terminal is designated as an abnormal terminal. In response to the model's anomaly score being less than the preset threshold, the system checks whether the number of anomaly rule triggers the first rule and / or the number of triggers of the second rule is less than a set number. If the number of times the first rule and / or the second rule is triggered by the abnormal rule triggering result is greater than or equal to a set number, then the corresponding navigation terminal is designated as an abnormal terminal. If the first rule is not triggered and the number of times the second rule is triggered is less than a set number, then the corresponding navigation terminal is selected as a candidate terminal.
[0010] Preferably, generating the TOPSIS score for candidate terminals based on the travel characteristics includes: The travel characteristics of the candidate terminals are standardized to generate the feature weights of the candidate terminals; Based on the weight of the features, the entropy value of each feature in the travel features of the candidate terminal is generated; Based on the entropy value, the weight values of each feature in the travel features of the candidate terminal are generated; A weighted normalization matrix is constructed based on the weight values and the travel characteristics; Determine the positive and negative ideal solutions of the weighted canonical matrix; The TOPSIS score of the candidate terminal is generated based on the distance between the matrix value in the weighted normalization matrix and the positive ideal solution and the negative ideal solution.
[0011] Preferably, the step of extracting features from the trip data to obtain trip features includes: Feature extraction is performed on the trip data to obtain spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features; Spatiotemporal rationality, data integrity, data reliability, numerical validity, reporting behavior, and attachment behavior are used as itinerary characteristics.
[0012] Preferably, obtaining the travel data of the target number of navigation terminals includes: Obtain the location data streams of the target number of navigation terminals; The positioning data stream is preprocessed and the trajectory is reconstructed to obtain the travel data of the navigation terminal.
[0013] A navigation terminal screening system, comprising: The trip data acquisition module is used to acquire trip data for a target number of navigation terminals; The feature extraction module is used to extract features from the trip data to obtain trip features; An anomaly detection module is used to process the trip features through unsupervised learning to generate a model anomaly score for the navigation terminal; The anomaly rule matching module is used to match the trip data according to the set anomaly rules to obtain the anomaly rule triggering result; The candidate terminal filtering module is used to filter out non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the anomaly rule triggering result. The TOPSIS score generation module is used to generate TOPSIS scores for candidate terminals based on the travel characteristics. The normal terminal screening module is used to generate a health score for candidate terminals based on the TOPSIS score, and to process navigation services based on the travel data and health score of the candidate terminals.
[0014] An electronic device, comprising: Memory, used to store computer programs; A processor, used to implement the steps of any of the above-described navigation terminal screening methods when executing the computer program.
[0015] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the navigation terminal screening methods described above.
[0016] This application provides a navigation terminal screening method, which involves: acquiring travel data of a target number of navigation terminals; extracting features from the travel data to obtain travel features; processing the travel features through unsupervised learning to generate model anomaly scores for the navigation terminals; matching the travel data according to set anomaly rules to obtain anomaly rule triggering results; selecting non-anomaly navigation terminals as candidate terminals based on the model anomaly scores and anomaly rule triggering results; generating TOPSIS scores for the candidate terminals based on the travel features; and generating health scores for the candidate terminals based on the TOPSIS scores, thereby processing navigation services based on the travel data and health scores of the candidate terminals. In this application, travel features are extracted from the travel data of a target number of navigation terminals to obtain travel features that are adapted to the dynamic changes in the behavior of the navigation terminals. Then, unsupervised learning and anomaly rules are used to process the travel features, and non-abnormal candidate terminals are selected based on the processing results. Finally, TOPSIS scores are generated for the candidate terminals to produce a health score for normal terminals. This approach integrates unsupervised learning and anomaly rules to process travel features and select normal terminals, converting the TOPSIS scores of candidate terminals into health scores to guide navigation operations. This expands the dimensions of navigation terminal selection, resulting in high screening accuracy. The entire process is transparent and requires no human intervention, offering good interpretability and a high degree of intelligence. Furthermore, it is not limited by navigation terminals and has good adaptability. The navigation terminal screening system, electronic device, and computer-readable storage medium provided in this application also solve the corresponding technical problems. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0018] Figure 1 A flowchart illustrating a navigation terminal screening method provided in this application embodiment; Figure 2 Here is a complete architecture diagram of an exemplary navigation terminal screening scheme; Figure 3 This is a schematic diagram of the structure of a navigation terminal screening system provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application; Figure 5 This is another structural schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation
[0019] 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, and 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.
[0020] Please see Figure 1 , Figure 1 This is a flowchart of a navigation terminal screening method provided in an embodiment of this application.
[0021] This application provides a navigation terminal screening method, which may include the following steps: Step S101: Obtain the travel data of the target number of navigation terminals.
[0022] In practical applications, the travel data of a target number of navigation terminals can be obtained first. This travel data refers to the corresponding data after the navigation terminal applies the navigation system. The navigation system can be flexibly determined according to the application scenario. For example, the navigation system can be the Beidou navigation system, etc.
[0023] In an exemplary embodiment, during the process of acquiring the travel data of a target number of navigation terminals, the positioning data stream of the target number of navigation terminals can be acquired; the positioning data stream is preprocessed and reconstructed, the preprocessing including data cleaning, information parsing and enhancement, etc., and the trajectory reconstruction including trajectory generation and segmentation, dynamic and static behavior division, etc., to obtain the travel data of the navigation terminals.
[0024] In specific application scenarios, data cleaning can include removing invalid records such as those with out-of-range latitude and longitude, empty terminal numbers, and physically impossible speed and direction values, as well as deduplication; information parsing and enhancement includes accurately parsing key metadata such as ACC status, data retransmission identifier, and terminal model from the additional information fields of the original message; trajectory generation and segmentation includes sorting and aggregating positioning points according to terminal number and timestamp to form an initial trajectory, and then segmenting the trajectory according to preset time thresholds and spatial distance thresholds, such as time intervals exceeding 10 minutes or distances exceeding 10 kilometers; dynamic and static behavior division includes identifying stop points within each independent trip using configurable spatiotemporal threshold rules, such as continuous stays exceeding 20 minutes and activity ranges within 100 meters, and further dividing the trip into "dynamic driving segments" and "static stop segments".
[0025] Step S102: Extract features from the travel data to obtain travel features.
[0026] In practical applications, in order to facilitate the processing of trip data for screening navigation terminals, feature extraction can be performed on the trip data first to obtain trip features that reflect the characteristics of the data in the trip data. These trip features can then be used to screen navigation terminals. The type of trip features can be flexibly determined according to the application scenario.
[0027] In an exemplary embodiment, during the process of feature extraction from the trip data to obtain trip features, the trip data can be used to extract features such as spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features. Spatiotemporal rationality features may include jump point rate and out-of-line rate; data integrity features may include trajectory missing rate and attribute missing rate; data reliability features may include timeliness anomaly rate and timing disorder rate; numerical validity features may include dirty data rate; reporting behavior features may include retransmission rate; and attachment behavior features may include attachment loss rate and attachment timeliness anomaly rate. The determination method for each feature can be as shown in Table 1. The spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features are used as trip features. In this way, this application constructs a multi-dimensional, fine-grained trip feature index system, which can comprehensively characterize the navigation terminal status from multiple aspects such as spatiotemporal, data quality, and behavioral logic, providing a solid data foundation for subsequent accurate screening of navigation terminals.
[0028] Table 1. Trip Feature Description Table
[0029] It should be noted that in the calculation of travel characteristics, the thresholds applied for speed, distance, etc. (such as the jump point speed threshold and the fly point distance threshold) can be dynamically adjusted based on the distribution of historical data. For example, the method of using the statistical percentile of historical data can be used for dynamic adjustment to enhance the universality of different navigation terminal types and usage scenarios and improve the evolution and updating capability of the method.
[0030] Step S103: Process the trip features through unsupervised learning to generate the model anomaly score of the navigation terminal.
[0031] In practical applications, after obtaining the travel features, considering that the behavior patterns of abnormal travel points are very different from those of normal travel points, they are usually more likely to be isolated and require shorter path lengths. Since unsupervised learning can screen outliers, the travel features can be processed through unsupervised learning to generate a model anomaly score for the navigation terminal, so as to detect whether the navigation terminal is an abnormal terminal based on the model anomaly score.
[0032] In an exemplary embodiment, during the process of generating model anomaly scores for navigation terminals by processing travel features through unsupervised learning, normal terminals can be initially screened based on travel features to reduce the amount of data processed subsequently and improve screening efficiency. For example, it can be detected whether all values of the travel features meet a preset target threshold, which is a threshold that characterizes the travel features as excellent features, such as missing rate and dirty data rate both <1%. In response to all values of the travel features meeting the target threshold, the corresponding navigation terminal is selected as a candidate terminal. In response to at least one value of the travel features not meeting the target threshold, the travel features are input into a pre-trained isolated forest model, for example, the travel features of each navigation terminal are aggregated into an evaluation matrix X. , aj represents the label of the statistical value in the trip features, 1-n represents the label of the navigation terminal participating in the statistics. The evaluation matrix is input into the isolated forest model, and the model anomaly score of the navigation terminal output by the isolated forest model is received. The model anomaly score can be a value between 0 and 1, and the closer the model anomaly score is to 1, the more "outlier" the navigation terminal is in the overall distribution, and the easier it is to be classified as an abnormal terminal.
[0033] Step S104: Match the trip data according to the set exception rules to obtain the exception rule trigger result.
[0034] In practical applications, navigation terminal screening can also be carried out by leveraging the knowledge of navigation experts and known, explicit cheating patterns. That is, the knowledge of navigation experts and known, explicit cheating patterns can be converted into set abnormal rules, and then the trip data can be matched according to the set abnormal rules to obtain the abnormal rule triggering results. In order to detect whether the navigation terminal is an abnormal terminal based on the abnormal rule triggering results.
[0035] In an exemplary embodiment, during the process of matching trip data according to set anomaly rules to obtain anomaly rule triggering results, the set anomaly rules can be obtained. These anomaly rules include rules for repeated trajectory cheating, rules for false driving at fixed locations, rules for high-frequency invalid data, rules for abnormal signal patterns, and rules for contradictory behavioral logic. These anomaly rules can be updated periodically or through user operation to improve the method's evolution and update capabilities. During the matching of trip data according to the anomaly rules to obtain anomaly rule triggering results, if the trip data detects that the same navigation terminal exhibits trips with highly consistent start / end coordinates, start / end times, and path shapes on different dates or between different navigation terminals, it can be determined that the navigation terminal has triggered a repeated trajectory cheating rule. If the trip data detects that the navigation terminal has been consistently... If data is reported within a very small geographical area (such as a parking lot), but the ACC status is "on" or the reported speed value is continuously non-zero, it can be determined that the navigation terminal has triggered a fixed-location false driving rule. If the navigation terminal reports data at an abnormally high frequency based on the trip data, but the proportion of invalid points such as jump points, flying points, and dirty data in the reported data exceeds a threshold, it can be determined that the navigation terminal has triggered a high-frequency invalid data rule. If the navigation terminal reports data with an unnatural and overly regular pattern in the time dimension based on the trip data, it can be determined that the navigation terminal has triggered a signal pattern abnormality rule. If there is a logical inconsistency between the multi-dimensional information reported by the navigation terminal based on the trip data, such as the ACC being off and continuously reporting non-zero speed, it can be determined that the navigation terminal has triggered a behavior logic contradiction rule.
[0036] Step S105: Based on the model anomaly score and the anomaly rule triggering results, select non-abnormal navigation terminals as candidate terminals.
[0037] In practical applications, after obtaining the model anomaly score and the result of the anomaly rule triggering, if the model anomaly score and the result of the anomaly rule triggering determine that the navigation terminal is abnormal, then the navigation terminal must be an abnormal terminal. However, if the model anomaly score and the result of the anomaly rule triggering determine that the navigation terminal is not abnormal, then the navigation terminal is not necessarily a normal terminal and further screening is required. Therefore, it is necessary to comprehensively select non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the result of the anomaly rule triggering. At this time, the selected candidate terminals can be considered as normal navigation terminals.
[0038] In an exemplary embodiment, during the process of selecting non-abnormal navigation terminals as candidate terminals based on model anomaly scores and anomaly rule triggering results, the following steps can be taken: First, check if the model anomaly score is less than a preset threshold. If the model anomaly score is greater than or equal to the preset threshold, the corresponding navigation terminal is designated as an abnormal terminal. If the model anomaly score is less than the preset threshold, check if the number of times the first rule and / or the second rule are triggered is less than a set number. The first rule is considered more dangerous than the second rule; for example, the first rule might be a high-risk rule, and the second rule a warning rule. If the number of times the first rule and / or the second rule are triggered is greater than or equal to the set number, the corresponding navigation terminal is designated as an abnormal terminal. If the first rule is not triggered and the number of times the second rule is triggered is less than the set number, the corresponding navigation terminal is designated as a candidate terminal. This approach not only efficiently detects known simple anomalies (such as data exceeding limits) but also uncovers complex cheating and hidden faults involving multiple combinations of anomaly features that traditional methods cannot identify. This overcomes the problem of difficulty in distinguishing "normal" terminals due to abnormal data interference, thus improving the accuracy of navigation terminal screening.
[0039] Step S106: Generate the TOPSIS score of the candidate terminal based on the travel characteristics.
[0040] Step S107: Generate a health score for candidate terminals based on the TOPSIS score, and process navigation services based on the trip data and health score of the candidate terminals.
[0041] In practical applications, after identifying candidate terminals, further screening can be conducted based on the similarity between travel features. This similarity can be represented by the TOPSIS score. Specifically, a TOPSIS score is generated for each candidate terminal based on its travel features, and then a health score is generated based on that score. For example, for candidate terminals with a TOPSIS score higher than or equal to a set threshold, a health score greater than the first set threshold is generated, with higher TOPSIS scores resulting in higher health scores, indicating a high-quality terminal. Conversely, for candidate terminals with a TOPSIS score lower than the set threshold, a health score less than or equal to the first set threshold is generated, with lower TOPSIS scores resulting in lower health scores, indicating a low-quality terminal. It should be noted that after identifying normal terminals, navigation services can be processed based on the travel data and health scores of the candidate terminals. For example, the travel data from the candidate terminal with the highest health score can be used to update the navigation data source for mileage statistics, behavior analysis, and safety monitoring. Furthermore, the health scores of candidate terminals can also serve as background information for subsequent terminal analysis, such as for evaluating terminal suppliers. This application does not impose specific limitations on these applications.
[0042] In an exemplary embodiment, the process of generating TOPSIS scores for candidate terminals based on travel characteristics can be combined with the entropy weight method. The complete process can be as follows: Figure 2 As shown, the travel features of candidate terminals are first standardized to generate the feature weights of candidate terminals. Assume the first... j under the feature number i The characteristic proportion of each navigation terminal is expressed as: ,but Based on the feature weight, generate the entropy value of each feature in the travel features of the candidate terminal. e j , , k =1 / ln n k is to ensure The adjustment coefficient; based on the entropy value, the weight value ω of each feature in the travel characteristics of the candidate terminal is generated. j , Among them, the smaller the entropy value, the greater the information content of the trip feature, and the higher its weight value; a weighted normalization matrix is constructed based on the weight value and the trip feature, for example, by first standardizing the trip feature to obtain... Then, a weighted normalization matrix V is constructed based on the weight values, where Determine the positive and negative ideal solutions of the weighted normalization matrix, and the positive ideal solution Z. + This represents the set of optimal values for each indicator within the weighted normalization matrix. Negative ideal solution Z - This represents the set of worst values for each indicator in the weighted normalization matrix. The positive and negative ideal solutions can be obtained by maxing out ( v ij The corresponding benefit-type indicators and min( v ij The corresponding cost indicators are determined; based on the distance between the matrix values in the weighted normative matrix and the positive and negative ideal solutions, the TOPSIS scores of the candidate terminals are generated.
[0043] In specific application scenarios, during the process of generating the TOPSIS score of a candidate terminal based on the distances between the matrix values in the weighted normalization matrix and the positive and negative ideal solutions, the distances from the travel features of the candidate terminal to the positive ideal solutions can be generated. and the distance to the negative ideal solution , , Then based on distance and distance Generate TOPSIS score S i , , S i The value is between 0 and 1, with a larger value indicating a better health status for the candidate terminal.
[0044] This application provides a navigation terminal screening method, which involves: acquiring travel data of a target number of navigation terminals; extracting features from the travel data to obtain travel features; processing the travel features through unsupervised learning to generate model anomaly scores for the navigation terminals; matching the travel data according to set anomaly rules to obtain anomaly rule triggering results; selecting non-anomaly navigation terminals as candidate terminals based on the model anomaly scores and anomaly rule triggering results; generating TOPSIS scores for the candidate terminals based on the travel features; and generating health scores for the candidate terminals based on the TOPSIS scores, thereby processing navigation services based on the travel data and health scores of the candidate terminals. In this application, travel features are extracted from the travel data of a target number of navigation terminals to obtain travel features, which are adapted to the dynamic changes in the behavior of the navigation terminals. Then, the travel features are processed through unsupervised learning and anomaly rules, and non-abnormal candidate terminals are selected based on the processing results. Finally, the TOPSIS scores of the candidate terminals are generated based on the travel data to generate a health score for normal terminals. This application realizes the comprehensive processing of travel features using unsupervised learning and anomaly rules to select normal terminals, and converts the TOPSIS scores of candidate terminals into health scores to guide the processing of navigation services. This expands the selection dimensions of navigation terminals, has high screening accuracy, and the entire process is open and transparent without human intervention. It has good interpretability and a high degree of intelligence. In addition, it is not limited by navigation terminals and has good adaptability.
[0045] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of a navigation terminal screening system provided in an embodiment of this application.
[0046] This application provides a navigation terminal screening system, which may include: The trip data acquisition module 101 is used to acquire trip data of a target number of navigation terminals; Feature extraction module 102 is used to extract features from travel data to obtain travel features; The anomaly detection module 103 is used to process the travel features through unsupervised learning and generate a model anomaly score for the navigation terminal. The exception rule matching module 104 is used to match the trip data according to the set exception rules to obtain the exception rule triggering result; The candidate terminal filtering module 105 is used to filter out non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the result of anomaly rule triggering. The TOPSIS score generation module 106 is used to generate TOPSIS scores for candidate terminals based on travel characteristics. The normal terminal filtering module 107 is used to filter out normal terminals from candidate terminals based on TOPSIS scores, so as to process navigation services based on the trip data of normal terminals.
[0047] This application provides a navigation terminal screening system, in which an anomaly detection module is specifically used to: detect whether all values of the travel feature meet a preset target threshold; in response to all values of the travel feature meeting the target threshold, select the corresponding navigation terminal as a candidate terminal; in response to at least one value of the travel feature not meeting the target threshold, input the travel feature into a pre-trained isolated forest model and receive the model anomaly score of the navigation terminal output by the isolated forest model.
[0048] This application provides a navigation terminal screening system, in which an abnormal rule matching module is specifically used to: acquire set abnormal rules, including repeated trajectory cheating rules, fixed location false driving rules, high frequency invalid data rules, signal mode abnormal rules, and behavior logic contradiction rules; and match the trip data according to the abnormal rules to obtain the abnormal rule triggering result.
[0049] This application provides a navigation terminal screening system. The candidate terminal screening module is specifically used for: detecting whether a model anomaly score is less than a preset threshold; if the model anomaly score is greater than or equal to the preset threshold, then designating the corresponding navigation terminal as an anomaly terminal; if the model anomaly score is less than the preset threshold, then detecting whether the number of times the first rule and / or the second rule are triggered by the anomaly rule triggering results is less than a set number; if the number of times the first rule and / or the second rule are triggered by the anomaly rule triggering results is greater than or equal to a set number, then designating the corresponding navigation terminal as an anomaly terminal; if the first rule is not triggered and the number of times the second rule is triggered is less than a set number, then designating the corresponding navigation terminal as a candidate terminal.
[0050] This application provides a navigation terminal screening system, wherein the TOPSIS score generation module may include: The standardization processing unit is used to standardize the travel characteristics of candidate terminals and generate the feature weights of candidate terminals. The entropy generation unit is used to generate the entropy value of each feature in the travel features of the candidate terminal according to the feature weight. The weight value generation unit is used to generate weight values for each feature in the travel features of the candidate terminal based on the entropy value. A matrix construction unit, used to construct a weighted normalized matrix based on weight values and travel characteristics; Ideal solution determination unit, used to determine the positive and negative ideal solutions of the weighted canonical matrix; The TOPSIS score generation unit is used to generate TOPSIS scores for candidate terminals based on the distances between matrix values in the weighted normalization matrix and the positive and negative ideal solutions.
[0051] This application provides a navigation terminal screening system, in which the feature extraction module is specifically used to: extract features from the trip data to obtain spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features; and use the spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features as trip features.
[0052] This application provides a navigation terminal screening system, in which the trip data acquisition module is specifically used to: acquire the positioning data streams of a target number of navigation terminals; preprocess and reconstruct the trajectory of the positioning data streams to obtain the trip data of the navigation terminals.
[0053] This application also provides an electronic device and a computer-readable storage medium, both of which have the corresponding effects of the navigation terminal screening method provided in the embodiments of this application. Please refer to... Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0054] An electronic device provided in this application includes a memory 201 and a processor 202. The memory 201 stores a computer program, and when the processor 202 executes the computer program, it implements the steps of the navigation terminal screening method described in any of the above embodiments.
[0055] Please see Figure 5 Another electronic device provided in this application embodiment may further include: an input port 203 connected to the processor 202 for transmitting commands input from the outside to the processor 202; a display unit 204 connected to the processor 202 for displaying the processing results of the processor 202 to the outside; and a communication module 205 connected to the processor 202 for enabling communication between the electronic device and the outside. The display unit 204 may be a display panel, a laser scanning display, etc.; the communication method adopted by the communication module 205 includes, but is not limited to, Mobile High-Definition Link (MHL), Universal Serial Bus (USB), High-Definition Multimedia Interface (HDMI), wireless connection: Wireless Fidelity (WiFi), Bluetooth communication technology, Bluetooth Low Energy communication technology, and communication technology based on IEEE 802.11s.
[0056] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of the navigation terminal screening method described in any of the above embodiments.
[0057] The computer-readable storage media involved in this application include random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs (compact disc read-only memory), or any other form of storage media known in the art.
[0058] This application provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the navigation terminal screening method described in any of the above embodiments.
[0059] For descriptions of relevant parts in the navigation terminal screening system, electronic device, and computer-readable storage medium provided in this application's embodiments, please refer to the detailed description of the corresponding parts in the navigation terminal screening method provided in this application's embodiments; they will not be repeated here. Furthermore, parts of the technical solutions provided in this application that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.
[0060] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 said element.
[0061] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for screening navigation terminals, characterized in that, include: Obtain travel data for a target number of navigation terminals; Feature extraction is performed on the trip data to obtain trip features; The travel features are processed through unsupervised learning to generate a model anomaly score for the navigation terminal; The trip data is matched according to the set exception rules to obtain the exception rule trigger result; Based on the model anomaly score and the anomaly rule triggering results, non-abnormal navigation terminals are selected as candidate terminals. Based on the aforementioned travel characteristics, generate TOPSIS scores for candidate terminals; Based on the TOPSIS score, a health score is generated for candidate terminals, and navigation services are processed based on the trip data and health score of the candidate terminals.
2. The method according to claim 1, characterized in that, The step of processing the trip features through unsupervised learning to generate a model anomaly score for the navigation terminal includes: Detect whether all values of the travel feature meet a preset target threshold; If all values of the trip features satisfy the target threshold, then the corresponding navigation terminal is selected as a candidate terminal. If at least one value of the trip feature does not meet the target threshold, the trip feature is input into a pre-trained isolated forest model, and the model anomaly score of the navigation terminal output by the isolated forest model is received.
3. The method according to claim 1, characterized in that, The process of matching the trip data according to the set anomaly rules to obtain the anomaly rule trigger result includes: Obtain the set abnormal rules, which include repeated trajectory cheating rules, fixed position false driving rules, high frequency invalid data rules, signal pattern abnormal rules, and behavior logic contradiction rules; The trip data is matched according to the aforementioned anomaly rules to obtain the anomaly rule triggering result.
4. The method according to claim 1, characterized in that, The step of selecting non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the anomaly rule triggering results includes: Detect whether the abnormal score of the model is less than a preset threshold; If the abnormal score of the model is greater than or equal to the preset threshold, the corresponding navigation terminal is designated as an abnormal terminal. In response to the model's anomaly score being less than the preset threshold, the system detects whether the number of anomaly rule triggers the first rule and / or the number of triggers of the second rule is less than a set number. If the number of times the first rule and / or the second rule is triggered by the abnormal rule triggering result is greater than or equal to a set number, then the corresponding navigation terminal is designated as an abnormal terminal. If the first rule is not triggered and the number of times the second rule is triggered is less than a set number, then the corresponding navigation terminal is selected as a candidate terminal.
5. The method according to claim 1, characterized in that, The step of generating TOPSIS scores for candidate terminals based on the travel characteristics includes: The travel characteristics of the candidate terminals are standardized to generate the feature weights of the candidate terminals; Based on the weight of the features, the entropy value of each feature in the travel features of the candidate terminal is generated; Based on the entropy value, the weight values of each feature in the travel features of the candidate terminal are generated; A weighted normalization matrix is constructed based on the weight values and the travel characteristics; Determine the positive and negative ideal solutions of the weighted canonical matrix; The TOPSIS score of the candidate terminal is generated based on the distance between the matrix value in the weighted normalization matrix and the positive ideal solution and the negative ideal solution.
6. The method according to any one of claims 1 to 5, characterized in that, The step of extracting features from the trip data to obtain trip features includes: Feature extraction is performed on the trip data to obtain spatiotemporal rationality features, data integrity features, data reliability features, numerical validity features, reporting behavior features, and attachment behavior features; Spatiotemporal rationality, data integrity, data reliability, numerical validity, reporting behavior, and attachment behavior are used as itinerary characteristics.
7. The method according to claim 1, characterized in that, The acquisition of the travel data of the target number of navigation terminals includes: Obtain the location data streams of the target number of navigation terminals; The positioning data stream is preprocessed and the trajectory is reconstructed to obtain the travel data of the navigation terminal.
8. A navigation terminal screening system, characterized in that, include: The trip data acquisition module is used to acquire trip data for a target number of navigation terminals; The feature extraction module is used to extract features from the trip data to obtain trip features; An anomaly detection module is used to process the trip features through unsupervised learning to generate a model anomaly score for the navigation terminal; The anomaly rule matching module is used to match the trip data according to the set anomaly rules to obtain the anomaly rule triggering result; The candidate terminal filtering module is used to filter out non-abnormal navigation terminals as candidate terminals based on the model anomaly score and the anomaly rule triggering result. The TOPSIS score generation module is used to generate TOPSIS scores for candidate terminals based on the travel characteristics. The normal terminal screening module is used to generate a health score for candidate terminals based on the TOPSIS score, and to process navigation services based on the travel data and health score of the candidate terminals.
9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the navigation terminal screening method as described in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the navigation terminal screening method as described in any one of claims 1 to 7.