Method and system for accurate highway toll collection based on analysis of vehicle models prone to misjudgment

Abstract

The invention relates to the technical field of highway toll collection management, and discloses a highway precise toll collection method and system based on analysis of vehicle models prone to misjudgment, and the method comprises the steps: calculating a feature vector, calculating a dispute index, and determining a vehicle model. The system corresponds to the method. According to the method and the device, the multi-source data of the vehicle is acquired, the feature vector is calculated, and the dispute index is calculated by combining the dynamic change of the time sequence analysis charging and dispute data, so that the misjudgment-prone vehicle is captured; vehicle information is updated to an error-prone vehicle model information base and shared, meanwhile, real-time high-speed data is collected to be associated with feature vectors, driving scenes and surrounding vehicle information are combined to determine the types of the error-prone vehicle models, finally, the final vehicle model is analyzed and determined based on dispute indexes, the vehicle model judgment accuracy is achieved, the charging accuracy and fairness are improved, and the vehicle model judgment efficiency is improved. And a reliable basis is provided for subsequent management and auditing, and the overall efficiency and quality of highway toll collection management are improved.

Description

Technical Field

[0001] The present application relates to the technical field of highway toll management, and specifically to a method and system for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment. Background Art

[0002] In the field of highway toll management, accurate identification of vehicle types is crucial to toll collection. At present, highways charge passenger vehicles according to types one to four. Among them, it is difficult to quickly identify the correct toll model of critical vehicles such as eight- and nine-seat passenger vehicles at the toll collection site because their vehicle characteristics are not obvious. This problem has caused many adverse consequences, such as frequent toll disputes. The difference in toll standards on different sections will also cause additional disputes, which greatly interferes with the traffic efficiency of highways. At the same time, handling these disputes increases operating and management costs and reduces the economic benefits and management efficiency of highway operations.

[0003] Although the existing technology has explored vehicle identification, it still has obvious shortcomings. For example, Chinese patent application number CN202411913545.1 discloses a highway contactless toll collection system and method based on Beidou positioning. The proposed highway contactless toll collection system and method based on Beidou positioning mainly focus on vehicle identity verification and passage path review to achieve contactless toll collection and abnormal behavior alarm. It is not specially designed for the identification of easily misjudged vehicle models, and it is difficult to accurately judge the toll model. For example, Chinese patent application number CN202411147156.2 discloses a vehicle identity automatic identification method and system based on multi-source data fusion. It does not emphasize the use of multi-source data to accurately analyze the easily misjudged vehicle models. The data comparison may only stay on the surface features, and lacks the dynamic and collaborative nature of information management. When faced with complex vehicle model data, it is difficult to accurately distinguish the easily misjudged vehicle models, and it is impossible to update information in a timely manner, which makes it difficult to ensure the accuracy and timeliness of toll management.

[0004] Therefore, a new technical solution for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment is urgently needed to solve these problems and achieve accurate highway toll collection. Summary of the invention

[0005] The purpose of this application is to provide a method and system for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment, so as to solve the technical problems raised in the above-mentioned background technology.

[0006] To achieve the above objectives, this application discloses the following technical solutions: In a first aspect, the present application discloses a method for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment, the method comprising the following steps: Collect multi-source data of the vehicle, fuse the multi-source data, and calculate feature vectors. The multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data, and electronic system configuration data; Combine time series analysis to analyze the dynamic changes of charging data and dispute data, calculate the dispute index of vehicles and use it to automatically capture vehicles that are prone to misjudgment; When the dispute index reaches the preset dispute index threshold, the vehicle information is updated to the preset misjudged vehicle model information library and shared; real-time high-speed data is collected and associated with the feature vector, the misjudged vehicle model information library is traversed, and the type of misjudged vehicle model is determined based on the driving scene and surrounding vehicle information, and the final vehicle model is determined based on the dispute index analysis.

[0007] Preferably, when calculating the feature vector by fusing multi-source data, the key items under each data category are clarified based on the preset data level, and the internal logical connection between the key items is analyzed; wherein the data level is used to characterize the importance of the multi-source data; Based on the actual situation of historical vehicles and corresponding charging dispute cases, the frequency of occurrence and the degree of mutual influence of each key item in different dispute situations are counted, and a corresponding weight is assigned to each key item based on the statistical results and the internal logical connection. A feature vector is generated based on the weight. The feature vector is used to reflect the key information related to the easily misjudged vehicle models in multi-source data.

[0008] Preferably, when calculating the dispute index of a vehicle by analyzing the dynamic changes of the charging data and the dispute data in combination with the time series, the charging data and the dispute data are arranged in chronological order to form a corresponding charging time series and dispute time series; Calculate the fluctuations in the types of vehicles and the amounts charged in the charging time series and the frequency and amount differences of disputes in the dispute time series, construct a dispute status classification standard, correspond the status of vehicles in different time periods to the corresponding levels in the dispute status classification standard, and calculate the dispute index based on the proportion of vehicles in each level. The dispute index is used to reflect the possibility of vehicle charging disputes.

[0009] Preferably, when real-time high-speed data is collected and associated with a feature vector, key content in the real-time high-speed data is extracted based on the key content in the feature vector to generate a real-time feature corresponding to the feature vector; The matching degree between the feature vector and the real-time feature is calculated, and the matching degree determination criteria are adjusted based on the current dispute index of the vehicle, and the adjustment is: When the dispute index is greater than or equal to a preset adjustment threshold, the matching degree judgment standard is increased, and when the dispute index is less than the adjustment threshold, the matching degree judgment standard is decreased.

[0010] Preferably, when traversing the easily misjudged vehicle model information database, the vehicles in the easily misjudged vehicle model information database are sorted from large to small according to the dispute index, and starting from the vehicle with the largest dispute index, the vehicle information is sequentially associated with the real-time high-speed data and the number of matching items is counted. When the ratio of the number of matching items to the number of all features is greater than or equal to a preset matching degree threshold, the vehicle is determined to be an easily misjudged vehicle model and a corresponding easily misjudged vehicle model analysis is performed.

[0011] As a preferred method, the analysis of vehicle types that are prone to misjudgment includes: The types of vehicles that are easily misjudged are determined by combining the driving scene and the surrounding vehicle information, collecting the driving scene information and the surrounding vehicle information of the current vehicle, integrating them and comparing them with the characteristic performance of various types of vehicles in different scenarios in the said easily misjudged vehicle model information database, and combining the vehicle's own dispute index to comprehensively determine the candidate set of easily misjudged vehicle types, which includes at least two types of vehicles that are easily misjudged corresponding to the vehicle.

[0012] As a preferred method, when determining the final model based on the dispute index analysis, multiple attributes that affect the model judgment are listed, the actual occurrence frequency of each model and the charging dispute situation under the historical dispute index analysis are counted, and the corresponding weight is assigned to each attribute; For each candidate model, its performance value in each attribute is multiplied by the corresponding weight and the sum is calculated to get the comprehensive score. The model with the highest score is selected as the final model.

[0013] Preferably, when the real-time high-speed data of the vehicle exists and the dispute index is calculated, the vehicles in the same section of the highway at the same time period are defined as a set, and each vehicle in the set is used as a node to analyze the association relationship between the vehicles and construct an association network; Check the position of the vehicle in the associated network, count the dispute indexes of its neighboring nodes, and correct the dispute index of the vehicle based on the vehicle position and the dispute indexes of the neighboring nodes.

[0014] As a preferred method, after the final vehicle model is determined, the confidence interval of the toll vehicle model is adjusted based on the dispute index of the vehicle, and the determination process of the toll vehicle model, the applicable charging standards, and the reasons and specific ranges for the confidence interval adjustment are recorded and output; wherein, the confidence interval is adjusted as follows: When the dispute index is greater than a preset first confidence threshold, the confidence interval is expanded; When the dispute index is less than a preset second confidence threshold, the confidence interval is narrowed; wherein the second confidence threshold is less than the first confidence threshold.

[0015] In a second aspect, the present application discloses a system for accurate highway toll collection based on analysis of easily misjudged vehicle types, the system is applicable to the method for accurate highway toll collection based on analysis of easily misjudged vehicle types as described above, the system comprising a feature vector module, a dispute index module and a vehicle type determination module which are sequentially connected in communication; The feature vector module is used to collect multi-source data of the vehicle, fuse the multi-source data, and calculate the feature vector, the multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data and electronic system configuration data; The dispute index module is used to analyze the dynamic changes of charging data and dispute data in combination with time series, calculate the dispute index of the vehicle and automatically capture vehicles that are prone to misjudgment; The vehicle model determination module is used to update the vehicle information to a preset misjudged vehicle model information library and share it when the dispute index reaches a preset dispute index threshold; collect real-time high-speed data and associate it with feature vectors, traverse the misjudged vehicle model information library, determine the type of misjudged vehicle model in combination with the driving scene and surrounding vehicle information, and determine the final vehicle model based on dispute index analysis.

[0016] Beneficial effects: The method and system for accurate highway toll collection based on analysis of easily misjudged vehicle types of the present application utilizes multi-source data collected from vehicles and calculates feature vectors, and combines time series analysis to analyze the dynamic changes in toll and dispute data to calculate a dispute index, so as to capture easily misjudged vehicles; when the dispute index reaches a threshold, the vehicle information is updated to an easily misjudged vehicle type information database and shared, and real-time highway data is collected and associated with the feature vector, and the type of easily misjudged vehicle type is determined based on driving scenarios and surrounding vehicle information, and finally the final vehicle type is determined based on the dispute index analysis, thereby achieving accuracy in highway toll collection; the toll amount is calculated based on the final vehicle type and the output process is recorded, which not only improves the accuracy and fairness of the toll collection, but also provides a reliable basis for subsequent management and auditing, thereby improving the overall efficiency and quality of highway toll management. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flowchart of a method for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment provided in an embodiment of the present application; Figure 2 A structural block diagram of a system for precise highway toll collection based on analysis of vehicle types that are prone to misjudgment provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0020] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0021] The first aspect of this embodiment discloses Figure 1 A method for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment is shown, the method comprising the following steps: S1: Collect multi-source data of the vehicle, fuse the multi-source data, and calculate feature vectors. The multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data, and electronic system configuration data; S2: Combine time series analysis to analyze the dynamic changes of charging data and dispute data, calculate the dispute index of vehicles and use it to automatically capture vehicles that are prone to misjudgment; S3: When the dispute index reaches the preset dispute index threshold, the vehicle information is updated to the preset misjudged vehicle model information library and shared; real-time high-speed data is collected and associated with the feature vector, the misjudged vehicle model information library is traversed, and the type of misjudged vehicle model is determined based on the driving scene and surrounding vehicle information, and the final vehicle model is determined based on the dispute index analysis.

[0022] It should be noted that the basic data of the multi-source data in this embodiment includes at least vehicle information, the management data includes at least issuance data, the charging data includes at least transaction data, and the dispute data includes at least real-time vehicle passing data, image data, and audit data.

[0023] Based on the above, this embodiment uses multi-source data collected from vehicles, which covers data on foundation, management, charging, disputes, internal structural features, and electronic system configuration, to fuse these data and calculate feature vectors, and combine time series analysis to analyze the dynamic changes of charging and dispute data to calculate the dispute index to capture vehicles that are easily misjudged. When the dispute index reaches the threshold, the vehicle information is updated to the misjudged vehicle model information library and shared. At the same time, real-time high-speed data is collected and associated with the feature vector, and the type of vehicle that is easily misjudged is determined in combination with the driving scene and surrounding vehicle information. Finally, the final vehicle model is determined based on the dispute index analysis, achieving the accuracy of highway toll collection. Calculating the toll amount according to the final vehicle model and recording the output process not only improves the accuracy and fairness of the toll collection, but also provides a reliable basis for subsequent management and auditing, and improves the overall efficiency and quality of highway toll management.

[0024] Specifically, when calculating the feature vector by fusing multi-source data, the key items under each data category are identified based on the preset data level, and the internal logical connection between the key items is analyzed; wherein the data level is used to characterize the importance of multi-source data; Based on the actual situation of historical vehicles and corresponding charging dispute cases, the frequency of occurrence and degree of mutual influence of each key item in different dispute situations are counted, and a corresponding weight is assigned to each key item based on the statistical results and internal logical connections. Based on the weight, a feature vector is generated. The feature vector is used to reflect the key information related to the easily misjudged vehicle models in multi-source data.

[0025] As a preferred implementation of this embodiment, in the preset data level, multi-source data is divided into Multi-source data for key projects , the corresponding weight is , then the eigenvector In a specific example, the multi-source data contains three key items, namely, the number of seats in the vehicle 、Number of past fee disputes 、Complexity of electronic system configuration , and their weights are , , .like , , , then the eigenvector .

[0026] Based on the above, this embodiment uses the preset data level to clarify the key items, analyzes their internal logical connections, and based on historical cases, counts the frequency of occurrence and mutual influence of each item in different dispute situations, assigns weights to each key item to generate feature vectors, and realizes the effective integration of multi-source data. This enables the feature vector to accurately reflect the key information related to the easily misjudged vehicle model, providing a more targeted and accurate data basis for the subsequent calculation of the dispute index and analysis of the easily misjudged vehicle model.

[0027] Specifically, when calculating the dispute index of a vehicle by combining the dynamic changes of the charging data and the dispute data with the time series analysis, the charging data and the dispute data are arranged in chronological order to form corresponding charging time series and dispute time series; Calculate the fluctuations in the types of vehicles and the amounts charged in the charging time series, as well as the frequency and amount differences of disputes in the dispute time series, build a dispute status classification standard, correspond the status of vehicles in different time periods to the corresponding levels in the dispute status classification standard, and calculate the dispute index based on the proportion of vehicles in each level. The dispute index is used to reflect the possibility of vehicle charging disputes.

[0028] As a preferred implementation of this embodiment, the amount fluctuation in the collection charging time series is as follows: , the frequency of disputes in the dispute time series is The difference in amount is , the dispute status is divided into The vehicle status ratio in each level is , controversy index ,in, is the coefficient corresponding to each level. In a specific example, the amount fluctuation in the charging time series has 2 values: , , the frequency of disputes in the dispute time series , the amount difference The dispute status is divided into 3 levels, and the proportion of vehicles in each level is , , , corresponding coefficient , , , then the controversy index .

[0029] Based on the above, this embodiment arranges the charging data and dispute data in chronological order to form a time series, calculates the amount fluctuation, dispute frequency and amount difference, constructs the dispute status classification standard, and calculates the dispute index according to the status proportion of the vehicle in each level, thereby realizing a dynamic assessment of the possibility of vehicle charging disputes. This method fully considers the time dimension and dynamic changes of the data, so that the dispute index can more accurately reflect the actual charging dispute risk of the vehicle, and provides a strong basis for automatically capturing vehicles that are prone to misjudgment.

[0030] Specifically, when real-time high-speed data is collected and associated with a feature vector, key content in the real-time high-speed data is extracted based on the key content in the feature vector to generate a real-time feature corresponding to the feature vector; The matching degree between the feature vector and the real-time feature is calculated, and the matching degree judgment standard is adjusted based on the current dispute index of the vehicle. The adjustment is: When the dispute index is greater than or equal to the preset adjustment threshold, the matching degree judgment standard is increased, and when the dispute index is less than the adjustment threshold, the matching degree judgment standard is decreased.

[0031] As a preferred implementation of this embodiment, the matching degree between the feature vector and the real-time feature is calculated by any existing matching degree calculation method: , the preset adjustment threshold is The original matching criteria are , the adjusted matching criteria are When the controversy index hour, ;when hour, ,in, and is a preset adjustment amount. In a specific example, the original matching degree judgment standard , preset adjustment threshold , , When a vehicle's dispute index ,at this time, , then the adjusted matching criteria ; When another vehicle dispute index ,at this time, , then the adjusted matching criteria .

[0032] Based on the above, this embodiment uses the key content of real-time high-speed data extracted from the feature vector to generate real-time features, calculate the matching degree, and adjust the matching degree judgment standard based on the current dispute index of the vehicle, thereby realizing dynamic adjustment when collecting real-time high-speed data and associating it with the feature vector. The strictness of the association can be flexibly adjusted according to the dispute risk of the vehicle, which improves the accuracy of data association, avoids misjudgment caused by unified standards, and provides a more reliable data association basis for subsequent traversal of the easily misjudged vehicle model information library.

[0033] Specifically, when traversing the easily misjudged vehicle model information database, the vehicles in the easily misjudged vehicle model information database are sorted from large to small according to the dispute index. Starting from the vehicle with the largest dispute index, the vehicle information is associated with the real-time high-speed data and matched in sequence, and the number of matching items is counted. When the proportion of the number of matching items to the number of all features is greater than or equal to the preset matching degree threshold, the vehicle is determined to be an easily misjudged vehicle model and the corresponding easily misjudged vehicle model analysis is performed.

[0034] As a preferred implementation of this embodiment, the number of all features of vehicles in the easily misjudged vehicle model information database is , the number of matching items between vehicle information and real-time high-speed data is , the preset matching threshold is .when In a specific example, the number of all features of a vehicle in the easily misjudged vehicle information database is , preset matching threshold , the number of matches between the vehicle information and the real-time high-speed data , then at this time , and determined that the vehicle is a type that is easily misjudged.

[0035] Based on the above, this embodiment uses the method of sorting the vehicles in the easily misjudged vehicle model information database from large to small according to the dispute index, performing association matching in sequence, counting the number of matching items and comparing them with the preset matching degree threshold, thereby realizing efficient and accurate screening of easily misjudged vehicle models. Prioritizing vehicles with high dispute indexes reduces the blindness of traversal, improves the efficiency of finding easily misjudged vehicle models, and provides clear objects for subsequent analysis of easily misjudged vehicle models.

[0036] Specifically, the analysis of models that are prone to misjudgment includes: The types of vehicles that are easily misjudged are determined by combining the driving scene and surrounding vehicle information, collecting the driving scene information and surrounding vehicle information of the current vehicle, integrating them and comparing them with the characteristic performance of various types of vehicles in different scenarios in the easily misjudged vehicle model information database, and combining the vehicle's own dispute index to comprehensively determine the candidate set of easily misjudged vehicle models. The candidate set includes at least two types of vehicles that are easily misjudged as corresponding to the vehicle.

[0037] As a preferred implementation of this embodiment, based on any existing matching method, the matching score between the driving scene information and the vehicle model features in the information database is , the matching score between surrounding vehicle information and vehicle model features , the vehicle's own dispute index is , comprehensive score , , and is a preset weight coefficient, and a candidate set of easily misjudged vehicle types is determined based on the comprehensive score ranking. In a specific example, , , , the driving scene information matching score of a certain vehicle model , surrounding vehicle information matching score , Vehicle’s own controversy index , then the comprehensive score , based on the comprehensive score, the candidate set of vehicle types that are easy to misjudge is determined.

[0038] Based on the above, this embodiment collects driving scenes and surrounding vehicle information, compares it with the characteristic performance of various types of models in the easily misjudged model information database in different scenarios, and combines the vehicle's own dispute index for comprehensive judgment to achieve a more comprehensive and accurate determination of the candidate set of easily misjudged models. It fully considers the actual driving environment and the dispute situation of the vehicle itself, avoids the limitations of single factor judgment, and provides a more reasonable selection range for the subsequent determination of the final model.

[0039] Specifically, when determining the final model based on the dispute index analysis, multiple attributes that affect the model judgment are listed, the actual occurrence frequency of each model and the charging dispute situation under the historical dispute index analysis are counted, and the corresponding weight is assigned to each attribute; For each candidate model, its performance value in each attribute is multiplied by the corresponding weight and the sum is calculated to get the comprehensive score. The model with the highest score is selected as the final model.

[0040] As a preferred implementation of this embodiment, the information affecting the vehicle type judgment is collected. Attributes , the corresponding weight is , the performance value of a candidate model in each attribute is , comprehensive score , select the model with the highest comprehensive score as the final model. In a specific example, there are three attributes that affect the model judgment, namely, controversy index, driving scene matching degree, and surrounding vehicle similarity, and the weights are , , The performance values ​​of a candidate model in each attribute are , , , then the comprehensive score .

[0041] Based on the above, this embodiment uses multiple attributes that affect vehicle model judgment to list, statistically analyze historical data to assign weights to each attribute, calculate the comprehensive scores of candidate vehicle models, and select the vehicle model with the highest score as the final vehicle model, thereby achieving scientific decision-making based on multiple attributes. The impact of multiple factors on vehicle model judgment is comprehensively considered, the accuracy and reliability of the final vehicle model determination is improved, and the rationality of highway toll collection is ensured.

[0042] Specifically, when there is real-time high-speed data of vehicles and the dispute index is calculated, the vehicles in the same section of the highway at the same time period are defined as a set, and each vehicle in the set is used as a node to analyze the association relationship between vehicles and build an association network; Check the position of the vehicle in the associated network, count the dispute index of its neighboring nodes, and correct the dispute index of the vehicle based on the vehicle position and the dispute index of the neighboring nodes.

[0043] As a preferred implementation of this embodiment, the original dispute index of the vehicle is calculated as follows: , the position coefficient of the vehicle in the association network is , the average dispute index of neighbor nodes is , the modified controversy index In a specific example, the vehicle original dispute index , the position coefficient of the vehicle in the association network , the average dispute index of neighbor nodes , then the modified controversy index .

[0044] Based on the above, this embodiment defines the vehicles in the same section of the highway at the same time as a set, constructs an association network, and corrects the dispute index of the vehicle according to the position of the vehicle in the network and the dispute index of the neighboring nodes, thereby further optimizing the dispute index. The association relationship between vehicles and overall environmental factors are taken into account, so that the dispute index can better reflect the actual possibility of toll disputes of vehicles in actual traffic, providing a more accurate basis for subsequent vehicle model analysis and toll decision-making.

[0045] Specifically, after the final vehicle model is determined, the confidence interval of the toll vehicle model is adjusted based on the dispute index of the vehicle, and the determination process of the toll vehicle model, the applicable charging standards, and the reasons and specific ranges for the confidence interval adjustment are recorded and output; among which, the adjustment of the confidence interval is: When the dispute index is greater than a preset first confidence threshold, the confidence interval is expanded; When the dispute index is less than a preset second confidence threshold, the confidence interval is narrowed; wherein the second confidence threshold is less than the first confidence threshold.

[0046] As a preferred implementation of this embodiment, the confidence interval of the original charging vehicle type is , the first confidence threshold is preset as , the second confidence threshold is ,and , the expansion factor is , the reduction factor is When the controversy index When , the new confidence interval is ;when When , the new confidence interval is In a specific example, the confidence interval corresponding to the original toll model is , the first confidence threshold , the second confidence threshold , expansion factor , reduction factor . A vehicle controversy index ,at this time , then the new confidence interval is ; Another vehicle controversy index ,at this time , then the new confidence interval is .

[0047] Based on the above, this embodiment uses the confidence interval of the toll model to be adjusted according to the dispute index of the vehicle after the final model is determined, and records the relevant process and information, thereby realizing dynamic adjustment and transparent management of the toll model. The toll range can be flexibly adjusted according to the dispute risk of the vehicle, which not only ensures the rationality of the charge, but also reduces the occurrence of toll disputes. At the same time, the recorded information provides convenience for subsequent audits and inquiries.

[0048] In the specific application of this embodiment, based on the basic trusted data source of vehicle data, the trusted information source can be but not limited to the vehicle information database of the Provincial Transportation Administration Bureau, the provincial ETC issuance data, etc., based on this, the accurate verification of the vehicle toll model is completed in the system background, and the easy-to-misjudge model library is successfully constructed. This authoritative and reliable data source has laid a solid foundation for subsequent accurate charging. On this basis, with the help of the existing toll system network, the cloud-based easy-to-misjudge model library is distributed to the lane system end of the whole province in real time. The front-end toll system relies on these real-time easy-to-misjudge model library data to more accurately determine and identify the toll model, thereby realizing accurate charging, effectively reducing charging errors, and improving operating efficiency. The specific implementation method of this application is: 1. The provincial center analyzes, calculates and forms a database of license plates and vehicle models to be verified. The provincial center forms a database of license plates and vehicle models to be verified based on vehicle transaction records every day using existing data. The existing data includes basic data, historical transaction record data, historical image data, etc. The basic data includes but is not limited to the information registered by the vehicle management office and the issuance data of ETC products. 2. The road section verifies the license plate and vehicle model database one by one to form an accurate vehicle model database; the provincial center sends the license plate and vehicle model database to be verified to the road section unit through the existing audit system, and the road section unit verifies the data through the real-time vehicle passing situation and the logs, pictures and other data of the vehicle's driving, and forms a vehicle model database after verification; 3. The provincial center will add the vehicle model database to the key focus list and use the original data to distribute it to every toll lane system in the province; 4. The road section uses the key focus list to accurately determine the type of vehicle, and the lane toll collection system can achieve accurate charging without adding any hardware equipment and software.

[0049] In a simple example of this embodiment, there is a passenger car that often travels on the highway network of a certain province. A series of steps are used to analyze whether the vehicle is a misjudged vehicle based on the change of the dispute index, and to determine the final toll-related matters. Specifically: Data collection and feature vector generation: 1. Multi-source data collection: Basic data: The vehicle registration shows that its brand is X, model is Y, the approved number of passengers is 9, and the length is 5.5 meters, etc.

[0050] Management data: The vehicle's annual inspection record is normal. There was a minor violation record due to overloading in the past, but it was handled in a timely manner.

[0051] Toll data: In the past six months, the vehicle has traveled 20 times on different sections of the road, with the toll amount fluctuating between 50 yuan and 150 yuan each time. On three occasions, the toll amount was significantly higher or lower than the average toll for the same type of vehicles on the same section of the road.

[0052] Disputed data: We have received two inquiries from car owners regarding the judgment of the toll model, reflecting that they believe the toll model may be wrong.

[0053] Internal structure characteristic data: The seat layout in the car is relatively flexible and can be adjusted according to needs.

[0054] Electronic system configuration data: equipped with advanced on-board navigation and vehicle monitoring systems, etc.

[0055] 2. Fusion of multi-source data to calculate feature vectors: According to the preset data level, the system sorts out the key items under each data category, such as the number of seats and vehicle length in the basic data, the violation record in the management data, and the fluctuation of the toll amount in the toll data. Analyze the internal logical connection between these key items. For example, when the number of seats is close to the critical value (such as 8-9 seats) and the toll amount fluctuates greatly, the vehicle is more likely to have a toll dispute. Based on the actual situation of a large number of vehicles in the past and the corresponding toll dispute cases, the frequency of occurrence and degree of mutual influence of each key item in different dispute situations are counted, and each key item is assigned a corresponding weight. Assume that the weight of the number of seats is 0.3, the weight of the violation record is 0.1, the weight of the toll amount fluctuation is 0.4, and the weight of the dispute data is 0.2. After weighted summation, the feature vector of the vehicle is generated for subsequent analysis.

[0056] Dispute index calculation, combining time series analysis of the dynamic changes in charging data and dispute data to calculate the dispute index of the vehicle: 1. Data collation and sequence formation: Arrange the toll data in chronological order to form a toll time series, and also collate the dispute data in the corresponding chronological order to form a dispute time series. For example, the toll time series shows that the toll amount is generally high during holidays and relatively stable on weekdays; the dispute data time series shows that the frequency of disputes has increased after the toll policy of certain sections of the road has been changed.

[0057] 2. Analyze data characteristics and classify levels: Calculate the fluctuations of the charging amount in the charging time series, mark the time periods with large fluctuations and the corresponding amount changes, and count the frequency of disputes in the dispute time series and the difference in the amount involved in each dispute. Based on these factors, establish a standard for classifying dispute status, which is divided into three dispute levels: low, medium, and high. For example, when the charging amount fluctuates within 10% and the frequency of disputes is less than 1 time / month, it is identified as a low dispute level; when the fluctuation is between 10%-30% and the frequency of disputes is 1-3 times / month, it is identified as a medium dispute level; when the fluctuation exceeds 30% and the frequency of disputes is more than 3 times / month, it is identified as a high dispute level.

[0058] 3. Calculate the dispute index: Observe the proportion of the vehicle in different dispute levels in each time period in the past six months. Assume that the time in the low dispute level accounts for 30%, the medium dispute level accounts for 40%, and the high dispute level accounts for 30%. The corresponding level coefficients are set to 1, 2, and 3. By calculating (30%×1+40%×2+30%×3), the dispute index of the vehicle is 2.2, indicating that the vehicle has a certain possibility of toll disputes.

[0059] Update the information base and associate real-time data based on the dispute index: 1. Information database update: The system preset dispute index threshold is 1.5. Since the dispute index of the vehicle is 2.2, which is greater than the threshold, the vehicle information is updated to the preset misjudged vehicle model information database and shared, so that the relevant toll stations and management departments can obtain the vehicle information in a timely manner.

[0060] 2. Real-time data association: When the vehicle enters the highway again, its real-time highway data is collected, such as the vehicle's current appearance features (consistent with registration), driving speed (normal driving speed range), lane position (driving lane) and other key contents in real time, and the matching degree of the two is calculated based on the corresponding feature description in the previously generated feature vector. Assuming that the initial matching degree judgment standard is 0.8, since the vehicle dispute index of 2.2 is greater than the preset adjustment threshold (assuming it is 1.8), the matching degree judgment standard is improved, requiring that the matching degree of each key content and the feature vector reaches 0.9 before it is considered to be associated.

[0061] Traverse the easily misjudged vehicle model information database and determine the type of vehicle model that is easily misjudged: 1. Database traversal: When traversing the database of vehicles that are prone to misjudgment, the system has sorted all vehicles in the database from large to small according to the dispute index. Starting from the vehicle with the highest dispute index, the vehicle information is taken out in turn for correlation matching with the real-time high-speed data of the vehicle, and the number of matching items is counted. Assuming that when the vehicle is matched with the vehicle information in the database, the number of all features is 10, the number of matching items reaches 7, and the preset matching degree threshold is 0.6 (i.e. 6 items), the vehicle is determined to be a possible misjudgment vehicle model and enters the misjudgment vehicle model analysis process.

[0062] 2. Determine the types of vehicles that are easily misjudged: Collect information about the driving scene in which the vehicle is currently located, such as the situation that the vehicle is currently on a mountain highway section, it is the peak tourist season, and the surrounding vehicles are mostly small passenger cars and tourist buses. At the same time, collect information about surrounding vehicles such as the type, number, and speed of surrounding vehicles. After integrating this information, compare it with the characteristic performance of various types of vehicles in the easily misjudged vehicle model information database in similar scenarios, combined with the vehicle's own dispute index of 2.2. After comprehensive judgment, it is found that the vehicle is easily misjudged as an eight-seater passenger car or a seven-seater passenger car in the current scenario, forming a candidate set of easily misjudged vehicle types.

[0063] Determine the final model based on the controversy index: List multiple attributes that affect vehicle model judgment, including dispute index, driving scene matching (determined by comparing the overlap between the current vehicle driving scene and the standard scene feature of the corresponding vehicle model), and surrounding vehicle similarity (comparing the similarity between surrounding vehicles and the common surrounding vehicle distribution of the corresponding vehicle model). Reviewing the actual frequency of each vehicle model in similar driving scenes and surrounding vehicle distribution in the past and the corresponding toll disputes, assign corresponding weights to the above attributes. Assume that the dispute index weight is 0.4, the driving scene matching weight is 0.3, and the surrounding vehicle similarity weight is 0.3. For the two candidate models of eight-seat passenger cars and seven-seat passenger cars, calculate their comprehensive scores respectively. Assuming that the dispute index matching degree of the eight-seat passenger car is 0.8 (compared with the current vehicle dispute index), the driving scene matching degree is 0.7 (the degree of conformity with the mountain tourism peak season scene), and the surrounding vehicle similarity is 0.6 (compared with the surrounding vehicle distribution), its comprehensive score is calculated as (0.4×0.8+0.3×0.7+0.3×0.6)=0.71. The dispute index match of the seven-seat bus is 0.6, the driving scene match is 0.8, the surrounding vehicle similarity is 0.7, and the comprehensive score is (0.4×0.6+0.3×0.8+0.3×0.7)=0.7. Comparing the comprehensive scores of the two candidate models, the eight-seat bus scored higher, so the final model was determined to be an eight-seat bus.

[0064] Through such a complete analysis process of vehicles that are prone to misjudgment based on changes in the dispute index, highway management departments can more accurately manage vehicle tolls, reduce toll disputes, and improve operational efficiency.

[0065] The second aspect of this embodiment discloses Figure 2 A system for accurate highway toll collection based on analysis of easily misjudged vehicle types is shown, the system is applicable to the method for accurate highway toll collection based on analysis of easily misjudged vehicle types as described above, the system comprises a feature vector module, a dispute index module, a vehicle type determination module and a toll collection module which are sequentially connected in communication; The feature vector module is used to collect multi-source data of the vehicle, fuse the multi-source data, and calculate the feature vector. The multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data, and electronic system configuration data. The dispute index module is used to analyze the dynamic changes of charging data and dispute data in combination with time series, calculate the dispute index of vehicles and automatically capture vehicles that are prone to misjudgment; The vehicle model determination module is used to update the vehicle information to the preset misjudged vehicle model information library and share it when the dispute index reaches the preset dispute index threshold; collect real-time high-speed data and associate it with the feature vector, traverse the misjudged vehicle model information library, determine the type of misjudged vehicle model based on the driving scene and surrounding vehicle information, and determine the final vehicle model based on the dispute index analysis; The charging module is used to calculate the charging amount according to the final vehicle model and record the output determination process and calculation process.

[0066] It should be noted that the system for precise highway toll collection based on analysis of vehicle types that are prone to misjudgment in this embodiment corresponds to the aforementioned method for precise highway toll collection based on analysis of vehicle types that are prone to misjudgment in this embodiment. Therefore, the contents not specifically described in the system for precise highway toll collection based on analysis of vehicle types that are prone to misjudgment in this embodiment may refer to but are not limited to the functional definition, working principle and technical effects, etc., and the records in the aforementioned method for precise highway toll collection based on analysis of vehicle types that are prone to misjudgment in this embodiment will not be repeated in this text.

[0067] In summary, the method and system for accurate highway toll collection based on analysis of easily misjudged vehicle types in this embodiment utilizes multi-source data of collected vehicles and calculates feature vectors, and combines time series analysis to analyze the dynamic changes of toll and dispute data to calculate the dispute index, so as to capture easily misjudged vehicles; when the dispute index reaches a threshold, the vehicle information is updated to the easily misjudged vehicle type information library and shared, and at the same time, real-time high-speed data is collected and associated with the feature vector, and the type of vehicle type that is easily misjudged is determined in combination with the driving scene and surrounding vehicle information, and finally the final vehicle type is determined based on the dispute index analysis, thereby achieving accuracy in highway toll collection; the toll amount is calculated according to the final vehicle type and the output process is recorded, which not only improves the accuracy and fairness of the toll collection, but also provides a reliable basis for subsequent management and auditing, thereby improving the overall efficiency and quality of highway toll management.

[0068] In the embodiments provided in the present application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to implement the functions described herein, or a combination thereof. For software implementation, part or all of the processes of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The storage medium can be any available medium that can be accessed by a computer. Computer-readable storage media can include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer.

[0069] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment, characterized in that: The method comprises the following steps: Collect multi-source data of the vehicle, fuse the multi-source data, and calculate feature vectors. The multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data, and electronic system configuration data; Combine time series analysis to analyze the dynamic changes of charging data and dispute data, calculate the dispute index of vehicles and use it to automatically capture vehicles that are prone to misjudgment; When the dispute index reaches the preset dispute index threshold, the vehicle information is updated to the preset misjudged vehicle model information library and shared; real-time high-speed data is collected and associated with the feature vector, the misjudged vehicle model information library is traversed, and the type of misjudged vehicle model is determined based on the driving scene and surrounding vehicle information, and the final vehicle model is determined based on the dispute index analysis.

2. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 1 is characterized in that: When calculating feature vectors by fusing multi-source data, key items under each data category are identified based on preset data levels, and the internal logical connections between the key items are analyzed; wherein the data levels are used to characterize the importance of multi-source data; Based on the actual situation of historical vehicles and corresponding charging dispute cases, the frequency of occurrence and the degree of mutual influence of each key item in different dispute situations are counted, and a corresponding weight is assigned to each key item based on the statistical results and the internal logical connection. A feature vector is generated based on the weight. The feature vector is used to reflect the key information related to the easily misjudged vehicle models in multi-source data.

3. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 2 is characterized in that: When calculating the dispute index of a vehicle by analyzing the dynamic changes of the charging data and the dispute data in combination with the time series, the charging data and the dispute data are arranged in chronological order to form a corresponding charging time series and dispute time series; Calculate the fluctuations in the types of vehicles and the amounts charged in the charging time series and the frequency and amount differences of disputes in the dispute time series, construct a dispute status classification standard, correspond the status of vehicles in different time periods to the corresponding levels in the dispute status classification standard, and calculate the dispute index based on the proportion of vehicles in each level. The dispute index is used to reflect the possibility of vehicle charging disputes.

4. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 1 is characterized in that: When real-time high-speed data is collected and associated with a feature vector, key content in the real-time high-speed data is extracted based on the key content in the feature vector to generate a real-time feature corresponding to the feature vector; The matching degree between the feature vector and the real-time feature is calculated, and the matching degree determination criteria are adjusted based on the current dispute index of the vehicle, and the adjustment is: When the dispute index is greater than or equal to a preset adjustment threshold, the matching degree judgment standard is increased, and when the dispute index is less than the adjustment threshold, the matching degree judgment standard is decreased.

5. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 4 is characterized in that: When traversing the easily misjudged vehicle model information database, the vehicles in the easily misjudged vehicle model information database are sorted from large to small according to the dispute index. Starting from the vehicle with the largest dispute index, the vehicle information is sequentially associated with the real-time high-speed data and the number of matching items is counted. When the proportion of the number of matching items to the number of all features is greater than or equal to a preset matching degree threshold, the vehicle is determined to be an easily misjudged vehicle model and the corresponding easily misjudged vehicle model analysis is performed.

6. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 5 is characterized in that: Analysis of models that are prone to misjudgment includes: The types of vehicles that are easily misjudged are determined by combining the driving scene and the surrounding vehicle information, collecting the driving scene information and the surrounding vehicle information of the current vehicle, integrating them and comparing them with the characteristic performance of various types of vehicles in different scenarios in the said easily misjudged vehicle model information database, and combining the vehicle's own dispute index to comprehensively determine the candidate set of easily misjudged vehicle types, which includes at least two types of vehicles that are easily misjudged corresponding to the vehicle.

7. The method for accurate highway toll collection based on analysis of easily misjudged vehicle types according to claim 6 is characterized in that: When determining the final model based on the dispute index analysis, multiple attributes that affect the model judgment are listed, the actual frequency of occurrence of each model and the charging disputes under the historical dispute index analysis are counted, and the corresponding weight is assigned to each attribute; For each candidate model, its performance value in each attribute is multiplied by the corresponding weight and the sum is calculated to get the comprehensive score. The model with the highest score is selected as the final model.

8. The method for accurate highway toll collection based on analysis of easily misjudged vehicle types according to claim 3 is characterized in that: When the vehicle has the real-time high-speed data and the dispute index is calculated, the vehicles in the same section of the highway at the same time period are defined as a set, and each vehicle in the set is used as a node to analyze the association relationship between the vehicles and build an association network; Check the position of the vehicle in the associated network, count the dispute indexes of its neighboring nodes, and correct the dispute index of the vehicle based on the vehicle position and the dispute indexes of the neighboring nodes.

9. The method for accurate highway toll collection based on analysis of vehicle types prone to misjudgment according to claim 7, characterized in that: After the final vehicle model is determined, the confidence interval of the toll vehicle model is adjusted based on the dispute index of the vehicle, and the determination process of the toll vehicle model, the applicable toll standards, and the reasons and specific ranges for the confidence interval adjustment are recorded and output; among which, the adjustment of the confidence interval is: When the dispute index is greater than a preset first confidence threshold, the confidence interval is expanded; When the dispute index is less than a preset second confidence threshold, the confidence interval is narrowed; wherein the second confidence threshold is less than the first confidence threshold.

10. A system for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment, the system being applicable to the method for accurate highway toll collection based on analysis of vehicle types that are prone to misjudgment as described in any one of claims 1 to 9, characterized in that: The system includes a feature vector module, a dispute index module and a vehicle type determination module which are sequentially connected in communication; The feature vector module is used to collect multi-source data of the vehicle, fuse the multi-source data, and calculate the feature vector, the multi-source data includes basic data, management data, charging data, dispute data, internal structure feature data and electronic system configuration data; The dispute index module is used to analyze the dynamic changes of charging data and dispute data in combination with time series, calculate the dispute index of the vehicle and automatically capture vehicles that are prone to misjudgment; The vehicle model determination module is used to update the vehicle information to a preset misjudged vehicle model information database and share it when the dispute index reaches a preset dispute index threshold; Collect real-time high-speed data and associate it with feature vectors, traverse the easily misjudged vehicle model information database, determine the type of vehicle model that is easily misjudged based on driving scenarios and surrounding vehicle information, and determine the final vehicle model based on dispute index analysis.

Images