Bridge vehicle load monitoring system and method based on mobile phone positioning

Abstract

The invention provides a bridge vehicle load monitoring system and method based on mobile phone positioning. A data acquisition module acquires acceleration data and position information of a vehicle through a mobile phone SDK (Software Development Kit), namely a software development kit; the vehicle type recognition module analyzes the collected acceleration data and recognizes different types of vehicles; and the dynamic load calculation module calculates the dynamic load applied to the bridge when the vehicle passes through the bridge by using the acceleration data collected from the mobile phone SDK and combining the information of the vehicle. According to the invention, real-time and accurate monitoring of the bridge load is realized through sensor data acquisition, vehicle type identification and dynamic load calculation technologies, so that the efficiency and effectiveness of bridge safety management are improved, the cost and complexity of the monitoring system are reduced, and an economical, convenient and user-friendly solution is provided for maintenance and health management of the bridge.

Description

Technical Field

[0001] The invention provides a bridge vehicle load monitoring system based on mobile phone positioning and a method thereof, belonging to the technical field of bridge monitoring. Background Art

[0002] Existing bridge monitoring technologies often require complex hardware equipment and professional technicians to operate, which not only increases the monitoring cost, but also limits the real-time and extensiveness of monitoring. In addition, current technologies have certain limitations in vehicle type identification and load calculation, which affects the accuracy and practicality of monitoring data.

[0003] CN201410349146.7 discloses a random identification method for bridge deck moving vehicle loads. Gratings, strain and acceleration sensors are arranged on the bridge deck to obtain the speed of the moving vehicle and the dynamic strain and acceleration response of the bridge under the load of the moving vehicle; an accurate structural finite element model of the bridge structure is established, and the model is condensed so that the degree of freedom of the bridge model after condensation matches the dynamic strain signal measurement position; based on the measured dynamic strain response samples, the KL decomposition technology is used to obtain the deterministic components of the bridge's random dynamic displacement, velocity and acceleration; the direct inversion method and the Honginov regularization method are used to identify the deterministic components of the bridge deck random moving vehicle load; and the statistical characteristics of the random moving vehicle load are obtained: average time history and variance. However, this method has the following disadvantages:

[0004] 1. Cost and complexity: Arranging grating, strain and acceleration sensors on the bridge deck requires a lot of capital investment, especially on large bridges. The sensors need to be calibrated and maintained regularly after installation, which requires professional technicians to operate, increasing maintenance costs.

[0005] 2. Impact of environmental factors: The bridge environment is changeable, and the sensor may be affected by factors such as temperature, humidity, and corrosion, which may reduce its durability and accuracy. External environmental factors (such as wind and temperature changes) may interfere with the sensor data and affect the accuracy of the monitoring results.

[0006] 3. Data processing and analysis: Establishing a structural finite element model and performing model reduction, KL decomposition, load identification and other calculation processes require a lot of computing resources. The processing and analysis of dynamic strain response samples requires advanced signal processing technology and expertise, which increases the difficulty of data analysis.

[0007] 4. Real-time limitation: Due to the complexity of data processing and analysis, there may be a certain delay, which affects the effect of real-time monitoring. There may be response time issues when identifying random moving vehicle loads on the bridge deck, especially in high traffic conditions.

[0008] 5. Limitations: This method may be more suitable for specific bridge structures or specific monitoring purposes, and may not be applicable to all types of bridges. The accuracy of the finite element model depends on many factors, such as material properties, boundary conditions, etc. The error of the model may affect the accuracy of the monitoring results.

[0009] 6. Long-term stability: If exposed to harsh environments for a long time, the performance of the sensor may gradually decline, affecting the stability of long-term monitoring.

[0010] "Structural Monitoring Data Fusion Based on Visual and Inertial Sensing System" (Zhao Yanliang. Structural Monitoring Data Fusion Based on Visual and Inertial Sensing System [D]. Lanzhou University of Technology, 2023.) By installing sensors and acquisition equipment, the structure and function of the bridge are monitored in real time, and the monitoring data is analyzed and processed to ensure the safety and reliability of the bridge. However, this method has the following disadvantages: The data fusion of visual sensors and inertial sensors requires precise algorithms and calibration processes, which increases the complexity of the system. In practical applications, the algorithms in the fusion process may need to be continuously adjusted and optimized to ensure the accuracy of the data. In addition, although inertial sensors are integrated to reduce the requirements for light and environment, the visual system is still sensitive to changes in lighting. The performance of the visual system may be affected under strong light, shadows or changeable weather conditions. Finally, if the bridge structure is blocked during the monitoring process of the visual system, such as by vehicles, pedestrians or other obstacles, the accuracy of displacement monitoring will be affected.

[0011] In recent years, with the rapid development of smartphone technology, the built-in sensors and SDK (Software Development Kit) of mobile phones have provided a new possibility for bridge monitoring. Smartphones are generally equipped with high-precision sensors such as accelerometers and GPS modules, which can collect dynamic data and location information of vehicles in real time. However, how to effectively use this data to monitor and dynamically track bridge loads is still an unresolved problem. Therefore, there is an urgent need for a bridge monitoring system based on mobile SDK, which can use the portability and popularity of mobile phones to realize dynamic tracking of vehicle loads and provide real-time and accurate data support for bridge health assessment. Summary of the invention

[0012] The present invention provides a bridge vehicle load monitoring system and method based on mobile phone positioning to solve the following technical problems:

[0013] (1) Real-time dynamic monitoring: Existing technologies make it difficult to achieve real-time dynamic monitoring of bridge loads, resulting in delays in monitoring data and an inability to reflect the actual load conditions of the bridge in a timely manner.

[0014] (2) Monitoring cost and scalability: Traditional bridge monitoring systems are expensive and complex to install and maintain, which limits their application and expansion in a large number of bridges.

[0015] (3) Data accuracy and reliability: Existing monitoring methods may have errors in data collection and processing, affecting the accuracy and reliability of bridge load monitoring.

[0016] (4) Vehicle type identification and load estimation: How to accurately identify the vehicle type through data collected by the mobile phone SDK and estimate the vehicle load based on this is a challenge in the existing technology.

[0017] (5) Data integration and fusion: How to effectively integrate the acceleration, location and other information provided by the mobile SDK, and how to fuse it with other sensor data to improve monitoring accuracy.

[0018] (6) User privacy and security: When using mobile SDK for monitoring, it is necessary to consider the privacy protection of user data and the security of the system to prevent data leakage or malicious use.

[0019] The specific technical solution is:

[0020] A bridge vehicle load monitoring system based on mobile phone positioning, including a data acquisition module, a vehicle type identification module and a dynamic load calculation module;

[0021] The data acquisition module collects the acceleration data and position information of the vehicle through the mobile phone SDK (software development kit);

[0022] It includes the following submodules: acceleration data acquisition module, responsible for real-time acquisition of vehicle acceleration data; position positioning module, responsible for obtaining vehicle position information.

[0023] The vehicle type recognition module is used to analyze the collected acceleration data and identify different types of vehicles.

[0024] The dynamic load calculation module uses the acceleration data collected from the mobile phone SDK and combines the vehicle information to calculate the dynamic load imposed on the bridge when the vehicle passes through the bridge.

[0025] The data acquisition module of the present invention adopts the following design:

[0026] Low power design: To extend the battery life of mobile phones, the SDK uses low power mode when collecting data, and wakes up the sensor for data collection only when necessary.

[0027] Data encryption: In order to protect user privacy, the SDK uses encryption technology when collecting, transmitting and storing data to ensure data security.

[0028] User authorization: When the SDK is used for the first time, it will request user authorization to access accelerometer and GPS data to ensure compliance with relevant privacy protection regulations.

[0029] Compatibility testing: SDK needs to undergo extensive compatibility testing to ensure stable operation on mobile phones of different brands and models. Through the design and implementation of the above data acquisition module, the present invention can effectively collect acceleration data and position information for bridge monitoring, providing reliable data support for subsequent vehicle type identification, load estimation and bridge health status assessment.

[0030] The vehicle type identification module of the present invention has the following characteristics:

[0031] Feature extraction and selection: Extract key features from acceleration data, such as frequency components, peak values, time series patterns, etc. Select features that can effectively distinguish different vehicle models and optimize them.

[0032] Machine learning algorithms: Use support vector machine (SVM), random forest, neural network and other machine learning algorithms for vehicle type recognition. Select the best model parameters through cross-validation and other methods.

[0033] Online learning and updating: Design models that can learn online to adapt to new vehicle models and changing environmental conditions. Update models regularly to improve recognition accuracy and robustness.

[0034] Real-time and accuracy: Optimize algorithms to achieve fast and accurate vehicle model recognition and reduce latency. Improve the generalization ability of the model and reduce dependence on training data.

[0035] The dynamic load calculation module of the present invention has the following functions:

[0036] Dynamic model establishment: Establish the dynamic model of the vehicle, taking into account parameters such as the vehicle's mass and suspension system characteristics. Ensure that the model can adapt to different types of vehicles and driving conditions.

[0037] Acceleration and load conversion: Use Newton's second law (F=ma) to convert acceleration data into force. Combined with the vehicle's dynamic model, calculate the dynamic load of the vehicle in the vertical direction.

[0038] Data fusion and calibration: It may be necessary to fuse other data sources, such as vehicle speed, vehicle type, etc., to improve the accuracy of load calculations. The calculated dynamic loads can be calibrated and verified by comparing the results with traditional monitoring methods.

[0039] The present invention also provides a bridge vehicle load monitoring method based on mobile phone positioning, comprising the following steps:

[0040] Step 1: The data acquisition module collects the acceleration data and location information of the vehicle through the mobile phone SDK;

[0041] The acceleration data acquisition module collects the acceleration data of the vehicle in real time. First, the built-in three-axis accelerometer is accessed through the mobile SDK, and the accelerometer measures the acceleration changes of the mobile phone in the three vertical directions of X, Y, and Z. Secondly, the SDK will periodically read the value of the accelerometer and record the acceleration data of the vehicle, including static acceleration and dynamic acceleration. Finally, the data processing algorithm built into the SDK filters and calibrates the collected acceleration data. In addition, data preprocessing also includes denoising and removing zero drift of the data.

[0042] The location positioning module is responsible for obtaining the vehicle's location information. First, through the built-in GPS module of the mobile phone, the SDK obtains the vehicle's longitude, latitude and altitude information in real time. Secondly, the SDK uses positioning technology. Finally, the SDK ensures that the acceleration data and location information are synchronized in time.

[0043] Step 2: The vehicle type recognition module identifies different types of vehicles by analyzing the collected acceleration data.

[0044] First, the vehicle type recognition module extracts key features from the acceleration data. The key features include frequency components, peak values, and time series patterns, which are used to distinguish different vehicle types.

[0045] Secondly, the vehicle model recognition module uses machine learning algorithms to identify vehicle models.

[0046] In addition, the module uses optimized algorithms to process real-time data and adopts online learning mechanisms to adapt to new models and environmental changes.

[0047] Step 3: The dynamic load calculation module uses the acceleration data collected from the mobile SDK and the vehicle information to calculate the dynamic load imposed on the bridge when the vehicle passes through the bridge. Specifically, it includes:

[0048] Step 3.1, acceleration data analysis:

[0049] The dynamic load calculation module receives the acceleration data stream from the data acquisition module and analyzes the time series characteristics of the acceleration data, including the rate of change, direction and amplitude of acceleration.

[0050] The models used include:

[0051] (1) The single-degree-of-freedom vehicle model is a linear two-degree-of-freedom model, assuming that the vehicle moves only on the horizontal plane and only considers the two degrees of freedom of yaw and lateral motion;

[0052] The equation of motion for yaw, taking into account the linear relationship between lateral force and tire slip angle:

[0053]

[0054] Among them, I z is the moment of inertia of the vehicle about the vertical axis, ω r is the yaw rate, a and b are the distances from the center of mass to the front and rear axles, respectively, and F yf and F yr are the lateral forces on the front and rear axles respectively.

[0055] Lateral motion equation, assuming the vehicle has no longitudinal slip:

[0056]

[0057] Where m is the vehicle mass, V x is the speed of the vehicle along the x-axis (i.e. longitudinal direction), and β is the sideslip angle at the center of mass.

[0058] Lateral force calculation, using the linear tire model:

[0059] F yf =-C af a f

[0060] F yr =-C ar a r

[0061] Among them, C af and C ar are the cornering stiffness of the front and rear tires, a f and a r are the sideslip angles of the front and rear tires respectively. The sideslip angle is related to the vehicle's geometric parameters, yaw rate, and center of mass sideslip angle.

[0062] (2) Multi-degree-of-freedom vehicle model

[0063] Using the three-degree-of-freedom model, the longitudinal motion equation is added on the basis of the two-degree-of-freedom model:

[0064] Longitudinal motion equation:

[0065]

[0066] Among them, F x is the driving / braking force, F drag is the air resistance, F rolling is the rolling resistance, F grade is the slope resistance.

[0067] Step 3.2, dynamic load calculation:

[0068] Using Newton's second law F=ma, the acceleration data is converted into force; m is the mass of the vehicle and a is the acceleration. Combined with the vehicle's dynamic model, the dynamic load of the vehicle in the vertical direction, i.e. the load-bearing direction of the bridge, is calculated.

[0069] Step 3.3, data fusion and calibration:

[0070] First, multiple data sources are collected through the data acquisition module; second, a fusion model is established through machine learning algorithms to fuse information from different data sources; then, a large amount of real-time data is processed through optimization algorithms.

[0071] The present invention provides a bridge vehicle load monitoring system and method based on mobile phone positioning, which aims to achieve real-time and accurate monitoring of bridge loads through innovative sensor data acquisition, vehicle type identification and dynamic load calculation technology, thereby improving the efficiency and effectiveness of bridge safety management, while reducing the cost and complexity of the monitoring system, and providing an economical, convenient and user-friendly solution for bridge maintenance and health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 is a flow chart of the present invention;

[0073] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0074] like Figure 1 The flowchart shown and Figure 2 The structure shown.

[0075] This embodiment provides a bridge vehicle load monitoring system based on mobile phone positioning and a monitoring process thereof, including:

[0076] (I) Data acquisition module

[0077] Describes how the mobile SDK collects acceleration data and location information. The main function of the data collection module is to collect the acceleration data and location information of the vehicle through the mobile SDK (Software Development Kit). The module is designed to ensure the accuracy and real-time nature of the data while taking into account user privacy and system security.

[0078] 1. Acceleration data collection:

[0079] The acceleration data acquisition module is an important part of the present invention, which is responsible for real-time acquisition of the acceleration data of the vehicle. First, the built-in three-axis accelerometer is accessed through the mobile phone SDK. The accelerometer can measure the acceleration changes of the mobile phone in the three vertical directions of X, Y, and Z, providing a basis for subsequent data processing and analysis. Secondly, the SDK will read the value of the accelerometer regularly (for example, multiple times per second) and record the acceleration data of the vehicle, including static acceleration (such as gravity acceleration) and dynamic acceleration (such as acceleration changes during vehicle driving), to ensure the real-time and accuracy of data acquisition. Finally, in order to reduce noise and error, the data processing algorithm built into the SDK will filter and calibrate the collected acceleration data to ensure the accuracy of the data. In addition, data preprocessing also includes operations such as denoising and zero drift removal to further improve the quality of the data. Through these three steps, the acceleration data acquisition module can provide high-quality data support for subsequent vehicle model identification and dynamic load calculation, and provide strong protection for the safety monitoring of bridges.

[0080] 2. Location information collection:

[0081] The position positioning module is another key component in the present invention, which is responsible for obtaining the location information of the vehicle. First, through the built-in GPS module of the mobile phone, the SDK can obtain the longitude, latitude and altitude information of the vehicle in real time, and provide a location reference for bridge monitoring. Secondly, in order to improve the positioning accuracy, the SDK may adopt a variety of positioning technologies, such as GLONASS, Beidou or Wi-Fi positioning, and differential GPS technology to reduce errors and ensure the accuracy of the location information. Finally, the SDK ensures that the acceleration data and location information are synchronized in time to facilitate subsequent vehicle model identification and load estimation. Through these three steps, the GPS positioning module can provide accurate location information for bridge monitoring, provide strong support for real-time tracking of vehicles and accurate estimation of bridge loads, and provide strong protection for the safety management of bridges.

[0082] (II) Vehicle Type Identification Module

[0083] The vehicle type recognition module is one of the core technologies in this invention. Its function is to accurately identify different types of vehicles by analyzing the collected acceleration data. The design and implementation of this module involves multiple key steps and advanced technologies to ensure the accuracy and real-time performance of vehicle recognition.

[0084] First, the vehicle type recognition module needs to extract key features from the acceleration data. These features include frequency components, peak values, time series patterns, etc., which can effectively distinguish different vehicle types. Through these features, the module can identify the type of vehicle, such as trucks, cars, buses, etc.

[0085] Secondly, the vehicle model recognition module uses machine learning algorithms for vehicle model recognition. Support vector machine (SVM), random forest, neural network and other algorithms are widely used in vehicle model recognition tasks. Through methods such as cross-validation, the module selects the best model parameters to improve the accuracy of vehicle model recognition.

[0086] In addition, the vehicle model recognition module also needs to consider real-time performance and accuracy. In order to achieve fast and accurate vehicle model recognition, the module uses an optimized algorithm to process real-time data. At the same time, the module also designs an online learning mechanism to adapt to new vehicle models and environmental changes. In practical applications, the vehicle model recognition module needs to work in conjunction with other modules, such as the dynamic load calculation module and the data acquisition module. Data sharing and fusion between modules help improve the performance and efficiency of the entire system.

[0087] By analyzing the collected acceleration data, different types of vehicles can be accurately identified. By using machine learning algorithms and real-time processing technology, the module can achieve fast and accurate vehicle model identification. In addition, the module works in conjunction with other modules to help improve the performance and efficiency of the entire system.

[0088] (III) Dynamic load calculation module

[0089] The core task of the dynamic load calculation module is to use the acceleration data collected from the mobile SDK and other relevant information of the vehicle to calculate the dynamic load imposed on the bridge when the vehicle passes through the bridge. This process involves knowledge of physics, mechanics and signal processing.

[0090] 1. Acceleration data analysis:

[0091] The module receives the acceleration data stream from the data acquisition module and analyzes the time series characteristics of the acceleration data, including the rate of change, direction and amplitude of acceleration.

[0092] Vehicle dynamics model:

[0093] (1) Single-degree-of-freedom vehicle model, which is a linear two-degree-of-freedom model and is commonly used for lateral dynamics analysis.

[0094] The linear two-degree-of-freedom model of this embodiment includes yaw and lateral motion. This model assumes that the vehicle moves only in the horizontal plane and only considers the two degrees of freedom of yaw (rotation around the vertical axis) and lateral motion. The model ignores the pitch, roll, longitudinal and vertical motion of the vehicle.

[0095] Basic equation:

[0096] Yaw equation of motion (considering the linear relationship between lateral force and tire slip angle):

[0097]

[0098] Among them, I z is the moment of inertia of the vehicle about the vertical axis, ω r is the yaw rate, a and b are the distances from the center of mass to the front and rear axles, respectively, and F yf and F yr are the lateral forces on the front and rear axles respectively.

[0099] The equations of lateral motion, assuming no longitudinal slip of the vehicle:

[0100]

[0101] Where m is the vehicle mass, v x is the speed of the vehicle along the x-axis (i.e. longitudinal direction), and β is the sideslip angle at the center of mass.

[0102] Lateral force calculation, linear tire model:

[0103] F yf =-C af a f

[0104] F yr =-C ar a r

[0105] Among them, C af and C ar are the cornering stiffness of the front and rear tires, a f and a r are the sideslip angles of the front and rear tires respectively. The sideslip angle is related to the vehicle's geometric parameters, yaw rate, and center of mass sideslip angle.

[0106] (2) Multi-degree-of-freedom vehicle model

[0107] For more complex vehicle dynamics analysis, more degrees of freedom may need to be considered, such as vertical, pitch, roll, longitudinal, etc. Here, a simplified three-degree-of-freedom model (longitudinal, lateral, and yaw) is used as an example.

[0108] The three-degree-of-freedom model adds the longitudinal motion equation based on the two-degree-of-freedom model:

[0109] Longitudinal motion equation:

[0110]

[0111] Among them, F x is the driving / braking force, F drag is the air resistance, F rolling is the rolling resistance, F grade is the slope resistance.

[0112] This model is already much more complex than the single-degree-of-freedom model, and in practical applications more nonlinear factors may need to be considered, such as the nonlinear cornering characteristics of the tire, the elasticity of the suspension, etc.

[0113] 2. Dynamic load calculation:

[0114] Using Newton's second law (F=ma), the acceleration data is converted into force. Here "m" is the mass of the vehicle and "a" is the acceleration. Combined with the vehicle's dynamic model, the dynamic load of the vehicle in the vertical direction (i.e. the load-bearing direction of the bridge) is calculated.

[0115] 3. Data fusion and calibration:

[0116] First, multiple data sources are collected through the data acquisition module, such as acceleration data, location information, vehicle speed, etc.; secondly, a fusion model is established through machine learning algorithms to fuse information from different data sources to improve the overall data quality and accuracy; then, a large amount of real-time data is processed through optimization algorithms to ensure the speed and accuracy of dynamic load calculations.

Claims

1. A bridge vehicle load monitoring system based on mobile phone positioning, characterized in that: Including data acquisition module, vehicle type identification module and dynamic load calculation module; The data acquisition module collects the acceleration data and position information of the vehicle through the mobile phone SDK (software development kit); It includes the following submodules: acceleration data acquisition module, which is responsible for collecting the acceleration data of the vehicle in real time; Position positioning module, responsible for obtaining the vehicle's location information; The vehicle type recognition module is used to analyze the collected acceleration data and identify different types of vehicles; The dynamic load calculation module uses the acceleration data collected from the mobile phone SDK and combines the vehicle information to calculate the dynamic load imposed on the bridge when the vehicle passes through the bridge.

2. A bridge vehicle load monitoring method based on mobile phone positioning, characterized in that: The following steps are involved: Step 1: The data acquisition module collects the acceleration data and location information of the vehicle through the mobile phone SDK; Step 2: The vehicle type recognition module identifies different types of vehicles by analyzing the collected acceleration data; Step 3: The dynamic load calculation module uses the acceleration data collected from the mobile SDK and combines it with the vehicle information to calculate the dynamic load imposed on the bridge by the vehicle when passing through the bridge.

3. The bridge vehicle load monitoring method based on mobile phone positioning according to claim 2 is characterized in that: Step 1 specifically includes the following sub-steps: Step 1.1, the acceleration data acquisition module collects the acceleration data of the vehicle in real time; First, access the built-in three-axis accelerometer through the mobile phone SDK. The accelerometer measures the acceleration changes of the mobile phone in the three vertical directions of X, Y, and Z. Secondly, the SDK will periodically read the accelerometer value and record the vehicle's acceleration data, including static acceleration and dynamic acceleration; Finally, the SDK’s built-in data processing algorithm filters and calibrates the collected acceleration data; In addition, data preprocessing also includes denoising and removing zero drift from the data; Step 1.2, the location positioning module is responsible for obtaining the location information of the vehicle; First, through the built-in GPS module of the mobile phone, the SDK obtains the longitude, latitude and altitude information of the vehicle in real time; Secondly, the SDK uses positioning technology; Finally, the SDK ensures that the acceleration data and position information are synchronized in time.

4. The bridge vehicle load monitoring method based on mobile phone positioning according to claim 2 is characterized in that: Step 2 specifically includes the following sub-steps: First, the vehicle model recognition module extracts key features from the acceleration data; the key features include frequency components, peak values, and time series patterns, which are used to distinguish different vehicle models; Secondly, the vehicle model recognition module uses machine learning algorithms to identify vehicle models; Optimized algorithms are used to process real-time data; at the same time, online learning mechanisms are adopted to adapt to new vehicle models and environmental changes.

5. The bridge vehicle load monitoring method based on mobile phone positioning according to claim 2 is characterized in that: Step 3 specifically includes the following sub-steps: Step 3.1, acceleration data analysis: The dynamic load calculation module receives the acceleration data stream from the data acquisition module; analyzes the time series characteristics of the acceleration data, including the rate of change, direction and amplitude of acceleration; The models used include: (1) The single-degree-of-freedom vehicle model is a linear two-degree-of-freedom model, assuming that the vehicle moves only on the horizontal plane and only considers the two degrees of freedom of yaw and lateral motion; The equation of motion for yaw, taking into account the linear relationship between lateral force and tire slip angle: Among them, I z is the moment of inertia of the vehicle about the vertical axis, ω r is the yaw rate, a and b are the distances from the center of mass to the front and rear axles, respectively, and F yf and F yr are the lateral forces on the front and rear axles, respectively; Lateral motion equation, assuming the vehicle has no longitudinal slip: Where m is the vehicle mass, V x is the speed of the vehicle along the x-axis (i.e. longitudinal direction), β is the sideslip angle of the center of mass; Lateral force calculation, using the linear tire model: F yf =-C af a f F yr =-C ar a r Among them, C af and C ar are the cornering stiffness of the front and rear tires, a f and a r They are the sideslip angles of the front and rear tires respectively; the sideslip angles are related to the vehicle's geometric parameters, yaw rate, and center of mass sideslip angle; (2) Multi-degree-of-freedom vehicle model Using the three-degree-of-freedom model, the longitudinal motion equation is added on the basis of the two-degree-of-freedom model: Longitudinal motion equation: Among them, F x is the driving / braking force, F drag is the air resistance, F rolling is the rolling resistance, F grade is the slope resistance; Step 3.2, dynamic load calculation: Using Newton's second law F=ma, the acceleration data is converted into force; m is the mass of the vehicle, and a is the acceleration; combined with the vehicle's dynamic model, the dynamic load of the vehicle in the vertical direction, i.e. the load-bearing direction of the bridge, is calculated; Step 3.3, data fusion and calibration: First, multiple data sources are collected through the data collection module; Secondly, a fusion model is established through machine learning algorithms to integrate information from different data sources; then, a large amount of real-time data is processed through optimization algorithms.

Images