A truck model and axle count information reconciliation export transaction system

Abstract

This invention provides a truck model and axle count information comparison and export transaction system, relating to the field of intelligent transportation technology. It includes a multi-source data acquisition unit; an information processing and verification unit, communicatively connected to the multi-source data acquisition unit, for processing the multi-source data based on an artificial intelligence recognition model and constructing a dynamic digital twin model of the vehicle; a multi-dimensional cross-verification mechanism; and a hierarchical decision execution unit, communicatively connected to the information processing and verification unit, for executing hierarchical response operations based on the risk score. By introducing multi-source data fusion and artificial intelligence analysis, this invention significantly improves the accuracy of vehicle model and axle count verification, effectively reducing tolling errors caused by identification mistakes. Through the multi-dimensional cross-verification mechanism, the system achieves real-time comparison and risk warning of vehicle information at entrances and exits, enhancing the real-time identification and prevention capabilities for cheating behaviors such as modified vehicles and hidden axles.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to an export transaction system for comparing truck model and axle count information. Background Technology

[0002] The current highway truck export transaction system has technical deficiencies in vehicle type recognition, axle count verification, and fraud detection. In this system, traditional visual recognition technology is easily affected by occlusion and lighting conditions. For some specially modified vehicles (such as those with raised cargo boxes or hidden axles), the recognition accuracy is low, and the error rate may be as high as 15% to 20%. This means that in actual operation, inaccurate recognition may lead to large errors in the toll amount, causing trouble for users.

[0003] In terms of axle count verification, the system generally relies on data recorded by OBU (On-Board Unit) or CPC card (Composite Pass Card) or directly determines the number of axles of the truck through visual counting; however, this single data source verification method has a great risk and is prone to misjudgment of axle count, which can lead to billing disputes; at the same time, the existing system lacks the ability to cross-validate different data sources, which to some extent limits the improvement of recognition accuracy and precision. Traditional systems are also significantly inadequate in identifying cheating behavior. Currently, the system mainly relies on post-event tracing, which makes it difficult to capture the dynamic characteristics and driving behavior of vehicles in real time, such as cheating methods like "driving at the very end" and "obscuring license plates." This lag makes it difficult for the system to prevent these cheating behaviors in advance, which can easily lead to unfair situations in the actual charging process. Comprehensively improving the system's real-time analysis and prevention capabilities has become a key requirement to address the shortcomings of existing technologies. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by designing a truck model and axle count comparison export transaction system. This system overcomes the technical deficiencies of current highway truck export transaction systems in terms of model identification, axle count verification, and fraud detection, thereby improving the overall performance of the system and ensuring the accuracy and fairness of toll collection.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: An export transaction system for comparing truck model and axle count information includes: Multi-source data acquisition units are deployed at the entrance and exit lanes of toll stations to collect truck model outline data, axle number sensor data, vehicle identification images and driving trajectory data; The information processing and verification unit is communicatively connected to the multi-source data acquisition unit. It is used to process the multi-source data based on an artificial intelligence recognition model and construct a dynamic digital twin model of the vehicle; and to compare and score the vehicle information at the entrance and exit based on a multi-dimensional cross-verification mechanism. A tiered decision execution unit, which is communicatively connected to the information processing and verification unit, is used to execute tiered response operations based on the risk score. The distributed encrypted evidence storage unit is used to encrypt and store data related to system transactions and facilitate communication.

[0006] Furthermore, the multi-source data acquisition unit includes: The front-end perception subunit includes a fusion perception module for acquiring three-dimensional point cloud data of the vehicle model, a sensing module for dynamically detecting the number of axes, and an optical recognition module for capturing vehicle markings. The edge computing subunit is used to perform real-time preprocessing and feature extraction on the data collected by the front-end sensing subunit.

[0007] Furthermore, the multidimensional cross-verification mechanism includes: Spatial dimension verification is used to compare the consistency of the 3D models of vehicles at the entrance and exit. Time-based verification is used to associate the same vehicle at the entrance and exit. Physical dimension verification is used to compare the number of axles identified by the sensors with the number of axles recorded by the on-board electronic unit. Behavioral dimension verification is used to analyze vehicle behavior risks based on driving trajectory.

[0008] Furthermore, the information processing and verification unit also includes an adaptive threshold adjustment module, which is used to dynamically adjust the judgment thresholds for verification of each dimension based on environmental and traffic flow context information.

[0009] Furthermore, the hierarchical decision execution unit is configured to execute at least one of the following responses: The warning response triggers an alarm and prompts for manual review. The intercept response automatically controls lane control equipment to guide vehicles into the verification area. The system will lock the response, coordinate with law enforcement systems, and restrict vehicle access.

[0010] Furthermore, the hierarchical decision execution unit also includes an augmented reality-assisted verification module, which is used to overlay and display vehicle entry information, historical data, and verification discrepancies to the staff terminal.

[0011] Furthermore, the system also includes a model continuous optimization unit, which comprises: The federated learning module is used to coordinate the participation of each node in the collaborative training of the artificial intelligence recognition model under the protection of data privacy. The simulation testing module is used to test and optimize the strategies of the information processing and verification unit and the hierarchical decision execution unit in a simulated environment.

[0012] Furthermore, the federated learning module employs privacy enhancement techniques to protect the privacy and security of local data at each node during model training.

[0013] Furthermore, the distributed encrypted evidence storage unit is implemented based on blockchain technology and is used to generate a unique and tamper-proof evidence storage record for each transaction.

[0014] Furthermore, the optical recognition module includes a license plate image restoration unit, which is used to enhance unclear license plate images and combine them with a character recognition model to achieve end-to-end vehicle identification recognition.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention significantly improves the accuracy of vehicle type and axle count verification by introducing multi-source data fusion and artificial intelligence analysis, effectively reducing tolling errors caused by identification mistakes. The system achieves real-time comparison and risk warning of vehicle information at the entrance and exit through a multi-dimensional cross-verification mechanism, enhancing the real-time identification and prevention capabilities of cheating behaviors such as modified vehicles and hidden axles. At the same time, the hierarchical response and encrypted evidence storage mechanism ensures the timeliness of handling and the credibility of transaction data, thereby improving the accuracy, fairness and efficiency of the toll collection process as a whole. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0018] The export transaction system for comparing truck model and axle count information disclosed in this invention, such as... Figure 1 As shown, it mainly consists of a multi-source data acquisition unit, an information processing and verification unit, a hierarchical decision execution unit, a distributed encryption and evidence storage unit, and a model continuous optimization unit: Among them, the multi-source data acquisition unit This unit is deployed at the entrance and exit lanes of the toll station and is responsible for collecting truck-related data, specifically including: Front-end perception subunit: Fusion Perception Module: Used to acquire 3D point cloud data of the vehicle model. Through multi-sensor fusion technology, it accurately captures the outline of the truck, providing basic data for the subsequent construction of a digital twin model.

[0019] Sensing module: Dynamically detects the number of axles. It uses high-precision sensors to obtain the number of axles of the truck in real time and accurately.

[0020] Optical recognition module: Captures vehicle identification, including a license plate image restoration unit, which can enhance unclear license plate images and combine with a character recognition model to achieve end-to-end vehicle identification recognition, ensuring accurate acquisition of vehicle identification information; at the same time, this module is also responsible for collecting vehicle trajectory data, recording the vehicle's driving trajectory in the toll station area through cameras and other equipment reasonably arranged in the lanes.

[0021] Edge computing subunit: performs real-time preprocessing and feature extraction on the data collected by the front-end sensing subunit, reducing the computational burden on subsequent information processing and verification units and improving the overall response speed of the system.

[0022] Information processing and verification unit This unit communicates with the multi-source data acquisition unit and mainly performs data processing, model building, and information verification, as detailed below: A vehicle type and axle number recognition model trained based on a federated learning framework: It coordinates the participation of each toll station node in federated learning training and secure aggregation and updates under local data encryption. While protecting the privacy and security of local data at each node, it makes full use of multi-source data to improve the accuracy and generalization ability of the model. The model is used to process multi-source data and construct a dynamic three-dimensional digital twin model of the vehicle, which can reflect the actual status of the vehicle in real time.

[0023] Multidimensional cross-verification mechanism (four-dimensional comparison mechanism): Spatial dimension verification: Calculate the intersection-over-union (IoU) ratio between the 3D vehicle models at the entrance and exit, and compare the consistency between the 3D vehicle models at the entrance and exit to determine whether there are any abnormal changes in the spatial morphology of the vehicle.

[0024] Time-based verification: By using cross-camera re-identification (ReID) technology to match the vehicle image features at the entrance and exit, the same vehicle at the entrance and exit is associated to ensure accurate comparison of information for the same vehicle.

[0025] Physical dimension verification: Compare the number of axles identified by the sensors with the number of axles recorded by the on-board OBU (On-Board Unit) or CPC card to verify the accuracy of the axle count information.

[0026] Behavioral dimension verification: Based on the characteristics of vehicle driving trajectory, a machine learning model is used to calculate the behavioral risk score and analyze whether there are any abnormalities in vehicle behavior, such as frequent lane changes or speeding.

[0027] The adaptive threshold adjustment module dynamically adjusts the verification thresholds for each dimension of the four-dimensional comparison mechanism based on contextual features such as weather, traffic flow, and vehicle type distribution, using a machine learning model. For example, under adverse weather conditions, the spatial dimension verification threshold is appropriately relaxed; when traffic flow is high, the temporal dimension verification threshold is adjusted to improve processing efficiency. A risk score is generated by integrating multi-dimensional information, providing a basis for subsequent decision-making.

[0028] Hierarchical decision-making execution unit This unit communicates with the information processing and verification unit and performs tiered response operations based on the risk score, specifically including: Early warning response: If the risk score reaches the first-level threshold, an audible and visual alarm will be triggered, prompting the toll collector to conduct a manual review using the AR device. The AR device can overlay the vehicle's entry information, historical data, and verification discrepancies onto the staff's terminal, facilitating quick and accurate review by the staff.

[0029] Interception Response: If the risk score reaches the level 2 threshold, the roadblock equipment will be automatically controlled to guide the vehicle into the verification area for secondary verification, ensuring that suspicious vehicles are further inspected.

[0030] Lockdown Response: If the risk score reaches the Level 3 threshold, the system will be activated to record the vehicle's trajectory and restrict its passage to prevent the offending vehicle from continuing to drive and causing greater harm.

[0031] Distributed encrypted evidence storage unit This unit, based on blockchain technology, is used to encrypt and store transaction data in the system; it generates a unique and tamper-proof record for each transaction, ensuring the authenticity and integrity of the transaction data and providing a reliable basis for subsequent auditing and traceability.

[0032] Model Continuous Optimization Unit This unit includes: Federated Learning Module: Used to coordinate the collaborative training of AI recognition models by various nodes under the protection of data privacy, continuously optimize the performance of vehicle type and axle number recognition models, and improve the accuracy and reliability of the system; it adopts privacy enhancement technology to protect the privacy and security of local data of each node during model training.

[0033] Simulation testing module: Tests and optimizes the strategies of information processing and verification units and hierarchical decision execution units in a simulated environment. By simulating various real-world scenarios, it identifies and resolves potential problems in advance, ensuring the stability and effectiveness of the system in actual operation.

[0034] A method for comparing truck model and axle count information based on digital twins and federated learning is applied to the aforementioned export transaction system. The specific steps are as follows: Data collection steps: At the entrance and exit lanes of the toll station, use a multi-source data acquisition unit to collect 3D point cloud data of the truck's vehicle profile, axle number sensor data, license plate image and vehicle trajectory data.

[0035] Processing and verification steps: The vehicle type and axle number recognition model trained based on the federated learning framework processes the collected data and constructs a dynamic three-dimensional digital twin model of the vehicle; and based on the preset four-dimensional comparison mechanism (spatial, temporal, physical and behavioral dimensions), cross-verifies the vehicle information at the entrance and exit and generates a risk score; at the same time, according to the contextual features of weather, traffic flow and vehicle type distribution, the verification threshold of each dimension in the four-dimensional comparison mechanism is dynamically adjusted using a machine learning model.

[0036] Decision-making and execution steps: Implement tiered responses based on risk scores, including warning, interception, or locking actions; specifically: if the risk score reaches the first-level threshold, trigger an audible and visual alarm and prompt the toll collector to conduct manual verification via AR equipment; if the risk score reaches the second-level threshold, automatically control the road barrier equipment to guide the vehicle into the verification area for secondary verification; if the risk score reaches the third-level threshold, link with the law enforcement system to record the vehicle's trajectory and restrict its passage permissions.

[0037] Data storage steps: Transaction process data is encrypted and stored using blockchain technology to ensure the authenticity and integrity of the data.

[0038] Through the above system architecture and method, this invention achieves accurate comparison of truck model and axle number information, effectively improving the efficiency and security of export transactions, while ensuring data privacy and security.

[0039] It should 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 process, method, article, or apparatus.

[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A truck model and axle count comparison export transaction system, characterized in that, include: Multi-source data acquisition units are deployed at the entrance and exit lanes of toll stations to collect truck model outline data, axle number sensor data, vehicle identification images and driving trajectory data; The information processing and verification unit is communicatively connected to the multi-source data acquisition unit and is used to process the multi-source data based on an artificial intelligence recognition model and construct a dynamic digital twin model of the vehicle. And based on a multi-dimensional cross-verification mechanism, vehicle information at the entrance and exit is compared and risk-scored; A tiered decision execution unit, which is communicatively connected to the information processing and verification unit, is used to execute tiered response operations based on the risk score. The distributed encrypted evidence storage unit is used to encrypt and store data related to system transactions and facilitate communication.

2. The system according to claim 1, characterized in that, The multi-source data acquisition unit includes: The front-end perception subunit includes a fusion perception module for acquiring three-dimensional point cloud data of the vehicle model, a sensing module for dynamically detecting the number of axes, and an optical recognition module for capturing vehicle markings. The edge computing subunit is used to perform real-time preprocessing and feature extraction on the data collected by the front-end sensing subunit.

3. The system according to claim 1, characterized in that, The multidimensional cross-verification mechanism includes: Spatial dimension verification is used to compare the consistency of the 3D models of vehicles at the entrance and exit. Time-based verification is used to associate the same vehicle at the entrance and exit. Physical dimension verification is used to compare the number of axles identified by the sensors with the number of axles recorded by the on-board electronic unit. Behavioral dimension verification is used to analyze vehicle behavior risks based on driving trajectory.

4. The system according to claim 3, characterized in that, The information processing and verification unit also includes an adaptive threshold adjustment module, which is used to dynamically adjust the judgment thresholds for verification of each dimension based on environmental and traffic flow context information.

5. The system according to claim 1, characterized in that, The hierarchical decision execution unit is configured to execute at least one of the following responses: The warning response triggers an alarm and prompts for manual review. The interception response automatically controls lane control equipment to guide vehicles into the verification area; The system will lock the response, coordinate with law enforcement systems, and restrict vehicle access.

6. The system according to claim 5, characterized in that, The hierarchical decision execution unit also includes an augmented reality-assisted verification module, which is used to overlay and display vehicle entry information, historical data, and verification discrepancies to the staff terminal.

7. The system according to claim 1, characterized in that, The system also includes a model continuous optimization unit, which comprises: The federated learning module is used to coordinate the participation of each node in the collaborative training of the artificial intelligence recognition model under the protection of data privacy. The simulation testing module is used to test and optimize the strategies of the information processing and verification unit and the hierarchical decision execution unit in a simulated environment.

8. The system according to claim 7, characterized in that, The federated learning module employs privacy enhancement techniques to protect the privacy and security of local data at each node during model training.

9. The system according to claim 1, characterized in that, The distributed encrypted evidence storage unit is implemented based on blockchain technology and is used to generate a unique and tamper-proof evidence storage record for each transaction.

10. The system according to claim 2, characterized in that, The optical recognition module includes a license plate image restoration unit, which is used to enhance unclear license plate images and combine them with a character recognition model to achieve end-to-end vehicle identification recognition.

Images