Intelligent integrated highway control system based on digital twin
The intelligent highway management system addresses data processing and risk analysis issues by using digital twins and neural networks to integrate and analyze highway data, ensuring accurate and timely risk management and safety supervision.
Patent Information
- Application Number
- CN202411072286.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2044-08-06
AI Technical Summary
Existing highway management systems fail to effectively process and label data, lack data fusion and integration, and do not provide adequate risk analysis, leading to messy data, inaccurate analysis, and dangerous driving conditions.
An intelligent integrated highway management system based on digital twins that maps sensor and structure data with physical models, uses neural networks for data fusion, and performs real-time safety analysis to generate risk-level control data.
Ensures accurate data association with the physical world, provides real-time supervision, and enables detailed risk-level management, improving data processing efficiency, reducing accidents, and enhancing safety.
Smart Images
Figure CN118587897B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of highway control and management, and specifically to an intelligent integrated highway control and management system based on digital twins. Background Art
[0002] Highway management and control refers to a series of management and control measures on highways to ensure safe, smooth and efficient traffic.
[0003] The Chinese patent with publication number CN113112804A discloses a smart highway comprehensive management method, which mainly uses the setting of a drone aerial inspection system to take good photos along the highway, and can quickly upload the shooting information, and at the same time facilitate rapid operation to the designated location according to the command, so as to improve its emergency efficiency. By setting fixed-point monitoring and patrol systems along the way, key parts can be better monitored, and its mobility can be improved, so as to facilitate rapid police dispatch and improve processing efficiency and processing quality. By setting priorities, events can be divided into priorities, so as to facilitate priority processing of major situations and improve the safety and stability of transportation. Although the above patent solves the problem of highway control, the following problems still exist in actual operation:
[0004] 1. The highway data is not effectively processed and labeled, resulting in the data being too messy.
[0005] 2. There is no better data fusion and processing of different information on the highway, resulting in the inability to conduct more accurate data analysis.
[0006] 3. There is no more careful risk analysis and control of vehicles traveling on the highway, which leads to dangerous driving of vehicles on the highway. Summary of the invention
[0007] The purpose of the present invention is to provide an intelligent integrated highway management and control system based on digital twins, which maps sensor data and highway structure data with parameters in a physical model, can accurately associate actual data with model parameters, ensure the close correspondence between data and the actual physical world, and use a neural network model for data fusion, which can fully utilize the advantages of neural networks in processing complex data relationships. By matching the risk factors in the second highway model data with the degree of driving norms in the first highway model data, detailed management and control data can be generated, and the management and control data can be divided into risk levels, so that managers can take corresponding management and control measures according to different risk levels, which can solve the problems in the prior art.
[0008] To achieve the above object, the present invention provides the following technical solutions:
[0009] The highway intelligent integrated management and control system based on digital twin includes:
[0010] Highway data acquisition unit, used for:
[0011] The highway data is retrieved from the database, and the retrieved highway data is classified according to data attributes. After the classification is completed, a unique coding number is assigned. After the coding number is completed, the highway data is marked as the target highway data;
[0012] Highway data building block for:
[0013] Performing data preprocessing on the target highway data, performing data mapping fusion on the target highway data after the data preprocessing through a physical model, and marking the target highway data after the data mapping fusion as first highway data;
[0014] Real-time data acquisition unit for:
[0015] Acquire the real-time data on the highway, analyze the real-time data, determine the supervision scope of the real-time data according to the analysis results, and mark the determined real-time data as the second highway data;
[0016] Run the forecast data implementation unit to:
[0017] The first highway data and the second highway data are respectively modeled by a neural network model, and after the modeling is completed, the first highway model data and the second highway model data are obtained, and operation prediction is performed according to the first highway model data and the second highway model data, and the prediction result is sent to the terminal for control operation.
[0018] Preferably, the highway data acquisition unit comprises:
[0019] Highway data classification module for:
[0020] Highway data is divided into sensor data and highway structure data;
[0021] Among them, sensor data includes traffic flow sensors, vehicle speed sensors, ultrasonic sensors, radar sensors, fiber optic sensors, environmental monitoring sensors, smart lighting sensors, and vehicle GPS positioning sensors;
[0022] The high-speed structural data includes road network traffic data, video surveillance data, road network basic data and auxiliary data;
[0023] Sensor data and high-speed structure data are classified separately.
[0024] Preferably, the highway data acquisition unit further includes:
[0025] Classification data encoding and labeling module, used for:
[0026] The classified sensor data and high-speed structure data are uniquely coded and labeled through the database self-increment ID and timestamp;
[0027] Among them, the classified sensor data and high-speed structure data are stored in the database, and after the storage is completed, a unique ID is assigned through the database's self-incrementing ID;
[0028] Timestamping sensor data and high-speed structure data stored in a database and assigned unique IDs;
[0029] Generate time labels for sensor data and high-speed structure data that have been assigned unique IDs in the database according to timestamps;
[0030] Get the unique coding number of each sensor data and high-speed structure data according to the time label and the assigned unique ID;
[0031] The sensor data and highway structure data with unique coding labels are marked as target highway data.
[0032] Preferably, the data preprocessing process also includes filtering out repeated, missing and redundant related data in the highway data information, including:
[0033] Extracting duplicate data from the highway data information;
[0034] Extracting the data acquisition time of every two identical data in the repeated data;
[0035] The difference evaluation parameter of the repeated data is obtained according to the data acquisition time of each two identical data, wherein the difference evaluation parameter is obtained by the following formula:
[0036] ;
[0037] Where Q represents the difference evaluation parameter; Q0 represents the initial value of the parameter; T 01 and T 02 Respectively represent the data acquisition time of two identical data; T x Indicates the preset reference time difference value; N f The number of groups of two identical data whose difference in the time of acquisition of the two identical data is non-zero; N z Indicates the total number of data groups after two identical data in repeated data are grouped together;
[0038] When the difference evaluation parameter is lower than a preset difference threshold, the data corresponding to the difference evaluation parameter and obtained earlier than the previous one for every two identical data are deleted;
[0039] When the difference evaluation parameter is not lower than a preset difference threshold, it is determined whether every two identical data corresponding to the difference evaluation parameter should be retained.
[0040] Preferably, when the difference evaluation parameter is not lower than a preset difference threshold, it is determined whether all two identical data corresponding to the difference evaluation parameter are to be retained, including:
[0041] Extracting all difference evaluation parameters that are not lower than the difference threshold;
[0042] The compensation coefficient is set using all difference evaluation parameters that are not less than the difference threshold; wherein the compensation coefficient is obtained by the following formula:
[0043] ;
[0044] Wherein, λ represents the compensation coefficient; m represents the number of difference evaluation parameters whose difference evaluation parameters are not lower than the preset difference threshold; Q i represents the ith difference evaluation parameter; Q0 represents the difference threshold;
[0045] The difference threshold is adjusted using the compensation coefficient to obtain an adjusted difference evaluation parameter, wherein the adjusted difference evaluation parameter is obtained by the following formula:
[0046] ;
[0047] Among them, Q t represents the adjusted difference evaluation parameter; Q0 represents the difference threshold;
[0048] When the adjusted difference evaluation parameter is lower than the preset difference threshold, the data corresponding to the difference evaluation parameter and the data acquisition time of each two identical data are deleted;
[0049] When the adjusted difference evaluation parameter is not lower than the preset difference threshold, every two identical data judgments corresponding to the difference evaluation parameter are retained.
[0050] Preferably, the highway data construction unit includes:
[0051] Data preprocessing module, used to:
[0052] Preprocess the target highway data, which includes data cleaning, data integration and data standardization;
[0053] Data cleaning is to process missing values, outliers and duplicate values of target highway data;
[0054] Data integration is to integrate the target highway data after data cleaning to obtain a unified data set, and matching, conversion and redundancy are performed during the integration process;
[0055] Data standardization is to convert the target highway data after data integration into the same scale and format;
[0056] After data standardization, the preprocessed target highway data is obtained.
[0057] Preferably, the highway data construction unit further includes:
[0058] Data mapping fusion module for:
[0059] Retrieving the sensor data and highway structure data from the pre-processed target highway data respectively;
[0060] Retrieve the physical model from the database;
[0061] Mapping the sensor data and the high-speed structure data to the parameters in the physical model respectively, wherein the type of the sensor data and the characteristics of the high-speed structure data are used as the mapping data;
[0062] After the mapping of sensor data and high-speed structure data with parameters in the physical model is completed, data fusion is performed through a neural network model, wherein the neural network model is retrieved from a database;
[0063] The neural network model fuses the data to adjust the network parameters through the back propagation method and the gradient descent method. After the network parameters are adjusted, the sensor data and the high-speed structure data are fused through the forward propagation of the neural network model.
[0064] The sensor data and the highway structure data are fused to obtain the first highway data.
[0065] Preferably, the real-time data acquisition unit is further used for:
[0066] Real-time data on highways include real-time weather data and real-time traffic data;
[0067] Match the real-time meteorological data and real-time traffic data with each other through timestamp identification;
[0068] After the data correspondence, the real-time meteorological data and the real-time traffic data are associated using data mining methods;
[0069] Determine the safety range of current highway traffic flow based on real-time meteorological data;
[0070] Conduct safety analysis on the current safety range of high-speed traffic flow and real-time traffic data;
[0071] Determine whether the real-time traffic data is within the regulatory scope based on the analysis results;
[0072] Among them, when the real-time traffic data exceeds the safety range of the current high-speed traffic flow, the real-time traffic data is within the supervision range and marked as risk data;
[0073] When the real-time traffic data does not exceed the safety range of the current highway traffic volume, the real-time traffic data is not within the supervision range and is marked as safe data;
[0074] Safety data and risk data are uniformly labeled as second highway data.
[0075] Preferably, the operation prediction data implementation unit includes:
[0076] Highway data model building module for:
[0077] The characteristic data in the first highway data and the second highway data are retrieved respectively, wherein the characteristic data of the sensor data include road surface temperature, vehicle speed, and road surface slipperiness; the characteristic data of the highway structure data include road surface status and tunnel status; the characteristic data of the real-time meteorological data include wind speed, precipitation, and visibility; the characteristic data of the real-time traffic data include risk data and safety data;
[0078] Performing propagation training on the feature data of the first highway data and the second highway data respectively through a neural network model;
[0079] The propagation training is to initialize the weights of the characteristic data parameters of the first highway data and the second highway data. After the setting is completed, the parameter data is forward propagated through the convolution layer, the downsampling layer and the fully connected layer to obtain the output value;
[0080] When the error is greater than the expected value, the error is passed back to the network, and the errors of the fully connected layer, downsampling layer, and convolutional layer are obtained in turn;
[0081] Among them, the error of each layer is the total error of the network;
[0082] When the error is equal to or less than the expected value, the training is completed;
[0083] After the training is completed, the first highway model data and the second highway model data are obtained.
[0084] Preferably, the operation prediction data implementation unit further includes:
[0085] Run the prediction confirmation module to:
[0086] Determine the driving standard of vehicles on the current expressway according to the first highway model data;
[0087] Retrieve highway historical data from the database;
[0088] identifying risk factors in the second highway model data based on highway historical data;
[0089] The risk factors in the second highway model data are matched with the driving standardization degree of the first highway model data, and the control data is obtained after the risk matching is completed;
[0090] The control data is divided into risk levels, where the risk levels are divided into level one, level two and level three;
[0091] The control data whose risk level is not within the safe range is transmitted to the display terminal of the control vehicle through wireless signals for display, and control operation prompts are given while displaying;
[0092] Among them, the control operation is to send the vehicle's safe driving data to the display terminal.
[0093] Compared with the prior art, the present invention has the following beneficial effects:
[0094] 1. The digital twin-based intelligent integrated highway management and control system provided by the present invention maps sensor data and highway structure data with parameters in the physical model, can accurately associate actual data with model parameters, ensure the close correspondence between data and the actual physical world, and use neural network models for data fusion, which can fully utilize the advantages of neural networks in processing complex data relationships.
[0095] 2. The digital twin-based intelligent integrated highway management and control system provided by the present invention can determine whether the real-time traffic data is within the supervision scope through safety analysis, providing a powerful supervision tool for the traffic management department. It can accurately match data from different sources through timestamp identification for data correspondence, providing a solid foundation for subsequent data association and analysis.
[0096] 3. The digital twin-based intelligent integrated highway management and control system provided by the present invention continuously optimizes the model by comparing the error between the output value and the expected value. Accurate error control helps to ensure the accuracy and stability of the model. By constructing the first highway model data and the second highway model data, it can provide an in-depth understanding and prediction of the highway conditions. By matching the risk factors in the second highway model data with the driving standardization degree of the first highway model data, it can generate detailed management and control data and divide the management and control data into risk levels, so that managers can take corresponding management and control measures according to different risk levels. BRIEF DESCRIPTION OF THE DRAWINGS
[0097] Figure 1 This is a schematic diagram of the intelligent integrated control module for expressways of the present invention;
[0098] Figure 2 It is a schematic diagram of the intelligent comprehensive management and control method of highways according to the present invention. DETAILED DESCRIPTION
[0099] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0100] In order to solve the problem that the existing technology does not effectively process and label highway data, resulting in too much data clutter, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:
[0101] The highway intelligent integrated management and control system based on digital twin includes:
[0102] Highway data acquisition unit, used for:
[0103] The highway data is retrieved from the database, and the retrieved highway data is classified according to data attributes. After the classification is completed, a unique coding number is assigned. After the coding number is completed, the highway data is marked as the target highway data;
[0104] Highway data building block for:
[0105] Performing data preprocessing on the target highway data, performing data mapping fusion on the target highway data after the data preprocessing through a physical model, and marking the target highway data after the data mapping fusion as first highway data;
[0106] Real-time data acquisition unit for:
[0107] Acquire the real-time data on the highway, analyze the real-time data, determine the supervision scope of the real-time data according to the analysis results, and mark the determined real-time data as the second highway data;
[0108] Run the forecast data implementation unit to:
[0109] The first highway data and the second highway data are respectively modeled by a neural network model, and after the modeling is completed, the first highway model data and the second highway model data are obtained, and operation prediction is performed according to the first highway model data and the second highway model data, and the prediction result is sent to the terminal for control operation.
[0110] Specifically, by classifying highway data according to sensors and highway structures through the highway data acquisition unit, data can be organized and managed more systematically, thereby improving the efficiency of data management; by mapping sensor data and highway structure data with parameters in the physical model through the highway data construction unit, actual data can be accurately associated with model parameters, ensuring a close correspondence between data and the actual physical world; by performing correlation analysis on real-time meteorological data and real-time traffic data through the real-time data acquisition unit, and judging the safety range of current highway traffic based on meteorological conditions, the traffic safety status of the highway can be evaluated in real time; by running the prediction data implementation unit, risk factors in the second highway model data are matched with the degree of driving norms in the first highway model data, and detailed management and control data can be generated.
[0111] Highway data acquisition unit, including:
[0112] Highway data classification module for:
[0113] Highway data is divided into sensor data and highway structure data;
[0114] Among them, sensor data includes traffic flow sensors, vehicle speed sensors, ultrasonic sensors, radar sensors, fiber optic sensors, environmental monitoring sensors, smart lighting sensors, and vehicle GPS positioning sensors;
[0115] The high-speed structural data includes road network traffic data, video surveillance data, road network basic data and auxiliary data;
[0116] Sensor data and high-speed structure data are classified separately.
[0117] Classification data encoding and labeling module, used for:
[0118] The classified sensor data and high-speed structure data are uniquely coded and labeled through the database self-increment ID and timestamp;
[0119] Among them, the classified sensor data and high-speed structure data are stored in the database, and after the storage is completed, a unique ID is assigned through the database's self-incrementing ID;
[0120] Timestamping sensor data and high-speed structure data stored in a database and assigned unique IDs;
[0121] Generate time labels for sensor data and high-speed structure data that have been assigned unique IDs in the database according to timestamps;
[0122] Get the unique coding number of each sensor data and high-speed structure data according to the time label and the assigned unique ID;
[0123] The sensor data and highway structure data with unique coding labels are marked as target highway data.
[0124] Specifically, by classifying highway data according to sensors and highway structures through the highway data classification module, data can be organized and managed more systematically, thereby improving the efficiency of data management. Different types of data can be specially stored, processed and analyzed according to their characteristics, making data management more efficient and orderly. Sensor data covers a variety of sensor types, from traffic flow, vehicle speed to environmental monitoring, etc. Each sensor data has its specific purpose and value. The classification data coding and labeling module combines the database's auto-increment ID and timestamp to ensure the unique coding label of each sensor data and highway structure data. The introduction of timestamps not only provides time dimension information for data, but also enhances the traceability of data. Through timestamps, data within a specific time period can be easily queried, and then the operating status or abnormal conditions of the highway within the time period can be analyzed. Through the automated coding and labeling process, the data management process is greatly simplified. There is no need to manually assign a unique identifier to each data item, which reduces the possibility of human error and improves the efficiency of data processing.
[0125] In order to solve the problem that the existing technology does not better integrate and process different information on the highway, thus making it impossible to conduct more accurate data analysis, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:
[0126] Highway data building unit, including:
[0127] Data preprocessing module, used to:
[0128] Preprocess the target highway data, which includes data cleaning, data integration and data standardization;
[0129] Data cleaning is to process missing values, outliers and duplicate values of target highway data;
[0130] Data integration is to integrate the target highway data after data cleaning to obtain a unified data set, and matching, conversion and redundancy are performed during the integration process;
[0131] Data standardization is to convert the target highway data after data integration into the same scale and format;
[0132] After data standardization, the preprocessed target highway data is obtained.
[0133] The data preprocessing process also includes filtering out duplicate, missing and redundant related data in highway data information, including:
[0134] Extracting duplicate data from the highway data information;
[0135] Extracting the data acquisition time of every two identical data in the repeated data;
[0136] The difference evaluation parameter of the repeated data is obtained according to the data acquisition time of each two identical data, wherein the difference evaluation parameter is obtained by the following formula:
[0137] ;
[0138] Where Q represents the difference evaluation parameter; Q0 represents the initial value of the parameter; T 01 and T 02 Respectively represent the data acquisition time of two identical data; T x Indicates the preset reference time difference value; N f The number of groups of two identical data whose difference in the time of acquisition of the two identical data is non-zero; N z Indicates the total number of data groups after two identical data in repeated data are grouped together;
[0139] When the difference evaluation parameter is lower than a preset difference threshold, the data corresponding to the difference evaluation parameter and obtained earlier than the previous one for every two identical data are deleted;
[0140] When the difference evaluation parameter is not lower than a preset difference threshold, it is determined whether every two identical data corresponding to the difference evaluation parameter should be retained.
[0141] The technical effect of the above technical solution is: since duplicate data is not only duplicate data generated by repeated data collection, it is very likely that the data is at the same value for a long time, therefore, blindly deleting duplicate data can easily cause data missing problems. Therefore, the above technical solution of this embodiment can effectively improve the identification of valid data in duplicate data, thereby reducing the probability of subsequent data missing and reducing the problem of duplicate data occupying space.
[0142] By extracting duplicate data from highway data, the technical solution is able to remove duplicate information, thereby reducing data redundancy. This helps improve data utilization and processing efficiency. The solution not only extracts duplicate data, but also pays special attention to the moment when each pair of duplicate data was obtained. This ensures that time factors are taken into account when processing data, especially when it comes to time-dependent information that fluctuates over time on highways.
[0143] The scheme calculates the difference evaluation parameter to quantify the difference between two sets of repeated data, which can effectively improve the recognition accuracy of highly similar (i.e. repeated) or long-term unchanged data.
[0144] Based on the calculated difference evaluation parameter, the solution dynamically decides how to handle the duplicate data. If the parameter is lower than the preset threshold, it means that the two sets of data are very similar, and the earlier acquired data will be deleted. On the contrary, if the parameter is higher than or equal to the threshold, it means that the two sets of data are significantly different, and they will both be retained. This adaptive processing method ensures the integrity and accuracy of the data.
[0145] By removing duplicate and highly similar data, the solution can significantly reduce storage space usage. This is very beneficial for systems that need to store large amounts of data for a long time, as it can help save storage costs and improve storage efficiency. Due to the reduction of duplicate and redundant data, the solution is able to complete tasks faster when processing data. This not only increases the speed of data processing, but also enables the system to respond to real-time changes or perform rapid data analysis more efficiently. At the same time, by accurately removing duplicate and highly similar data, the technical solution helps to improve the accuracy of data analysis.
[0146] Specifically, when the difference evaluation parameter is not lower than a preset difference threshold, it is determined whether all two identical data corresponding to the difference evaluation parameter are to be retained, including:
[0147] Extracting all difference evaluation parameters that are not lower than the difference threshold;
[0148] The compensation coefficient is set using all difference evaluation parameters that are not less than the difference threshold; wherein the compensation coefficient is obtained by the following formula:
[0149] ;
[0150] Wherein, λ represents the compensation coefficient; m represents the number of difference evaluation parameters whose difference evaluation parameters are not lower than the preset difference threshold; Q i represents the ith difference evaluation parameter; Q0 represents the difference threshold;
[0151] The difference threshold is adjusted using the compensation coefficient to obtain an adjusted difference evaluation parameter, wherein the adjusted difference evaluation parameter is obtained by the following formula:
[0152] ;
[0153] Among them, Q t represents the adjusted difference evaluation parameter; Q0 represents the difference threshold;
[0154] When the adjusted difference evaluation parameter is lower than the preset difference threshold, the data corresponding to the difference evaluation parameter and the data acquisition time of each two identical data are deleted;
[0155] When the adjusted difference evaluation parameter is not lower than the preset difference threshold, every two identical data judgments corresponding to the difference evaluation parameter are retained.
[0156] The technical effect of the above technical solution is: when the difference evaluation parameter is not lower than the preset threshold, the solution will determine whether to retain each pair of identical data. This mechanism ensures that even when there are certain differences in the data, data that is valuable for analysis can be retained. By using the compensation coefficient, the solution can dynamically adjust the difference threshold according to actual conditions. This adjustment is based on a comprehensive consideration of multiple difference evaluation parameters, thereby improving the accuracy of threshold setting. By introducing the compensation coefficient, the solution provides a flexible way to handle data that is close to the threshold. This processing method avoids overly strict or overly loose data deletion strategies, thereby retaining valuable information to the maximum extent while deleting redundant data.
[0157] Since the acquisition time of each pair of identical data is taken into account, the solution can more accurately decide whether to retain or delete data. This time sensitivity ensures the timeliness and relevance of data. By accurately determining whether to retain or delete data, the solution can further optimize the use of storage space. This helps to reduce storage costs and improve storage efficiency. Since the data processing mechanism in the solution can quickly respond to differential changes, it is very suitable for real-time data processing or application scenarios that require fast response. By accurately retaining and deleting data, the solution improves the accuracy of decisions based on this data. This makes the decisions of related systems more reliable.
[0158] Data mapping fusion module for:
[0159] Retrieving the sensor data and highway structure data from the pre-processed target highway data respectively;
[0160] Retrieve the physical model from the database;
[0161] Mapping the sensor data and the high-speed structure data to the parameters in the physical model respectively, wherein the type of the sensor data and the characteristics of the high-speed structure data are used as the mapping data;
[0162] After the mapping of sensor data and high-speed structure data with parameters in the physical model is completed, data fusion is performed through a neural network model, wherein the neural network model is retrieved from a database;
[0163] The neural network model fuses the data to adjust the network parameters through the back propagation method and the gradient descent method. After the network parameters are adjusted, the sensor data and the high-speed structure data are fused through the forward propagation of the neural network model.
[0164] The sensor data and the highway structure data are fused to obtain the first highway data.
[0165] Specifically, the missing values, outliers and duplicate values in the target highway data are processed by the data preprocessing module, thereby eliminating potential noise and errors and improving the accuracy and reliability of the data. The data integration process further ensures the integrity and consistency of the data. By matching, converting and eliminating redundant data, a more refined and unified data set can be obtained. Data standardization converts the data to the same scale and format, which is crucial for subsequent machine learning algorithms or model training. Standardization helps to avoid model bias caused by scale differences between different features, thereby improving the stability and prediction performance of the model. The sensor data and high-speed structure data are mapped to the parameters in the physical model by the data mapping fusion module, which can accurately associate the actual data with the model parameters and ensure the close correspondence between the data and the actual physical world. The use of the neural network model for data fusion can make full use of the advantages of the neural network in processing complex data relationships. The network parameters are adjusted by the back propagation method and the gradient descent method, making the fusion process more accurate and efficient, which helps to improve the accuracy and efficiency of data fusion. By retrieving the sensor data and high-speed structure data in the preprocessed target highway data and mapping them to the physical model, the complexity and redundancy of data processing are reduced, and the speed and efficiency of data processing are improved.
[0166] The real-time data acquisition unit is also used to:
[0167] Real-time data on highways include real-time weather data and real-time traffic data;
[0168] Match the real-time meteorological data and real-time traffic data with each other through timestamp identification;
[0169] After the data correspondence, the real-time meteorological data and the real-time traffic data are associated using data mining methods;
[0170] Determine the safety range of current highway traffic flow based on real-time meteorological data;
[0171] Conduct safety analysis on the current safety range of high-speed traffic flow and real-time traffic data;
[0172] Determine whether the real-time traffic data is within the regulatory scope based on the analysis results;
[0173] Among them, when the real-time traffic data exceeds the safety range of the current high-speed traffic flow, the real-time traffic data is within the supervision range and marked as risk data;
[0174] When the real-time traffic data does not exceed the safety range of the current highway traffic volume, the real-time traffic data is not within the supervision range and is marked as safe data;
[0175] Safety data and risk data are uniformly labeled as second highway data.
[0176] Specifically, real-time meteorological data and real-time traffic data on the highway are obtained in real time to ensure the timeliness and freshness of the data. Data correspondence through timestamp identification can accurately match data from different sources, providing a solid foundation for subsequent data association and analysis. This real-time and high efficiency enables the system to respond quickly to changes on the highway and provide effective information in a timely manner. By correlating and analyzing real-time meteorological data and real-time traffic data, and judging the safety range of the current highway traffic flow according to meteorological conditions, the traffic safety status of the highway can be evaluated in real time. When the real-time traffic data exceeds the safety range, it can be quickly identified and marked as risk data, thereby realizing real-time monitoring and early warning of risks. This function helps to reduce the occurrence of traffic accidents and improve road safety. Through safety analysis, it can determine whether the real-time traffic data is within the regulatory scope, providing a powerful regulatory tool for traffic management departments. At the same time, the second highway data marked as safety data and risk data can provide comprehensive information support for decision makers, helping them to formulate more reasonable and effective traffic management strategies.
[0177] In order to solve the problem of dangerous driving of vehicles on highways due to lack of more careful risk analysis and control in the existing technology, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:
[0178] Run the forecast data implementation unit, including:
[0179] Highway data model building module for:
[0180] The characteristic data in the first highway data and the second highway data are retrieved respectively, wherein the characteristic data of the sensor data include road surface temperature, vehicle speed, and road surface slipperiness; the characteristic data of the highway structure data include road surface status and tunnel status; the characteristic data of the real-time meteorological data include wind speed, precipitation, and visibility; the characteristic data of the real-time traffic data include risk data and safety data;
[0181] Performing propagation training on the feature data of the first highway data and the second highway data respectively through a neural network model;
[0182] The propagation training is to initialize the weights of the characteristic data parameters of the first highway data and the second highway data. After the setting is completed, the parameter data is forward propagated through the convolution layer, the downsampling layer and the fully connected layer to obtain the output value;
[0183] When the error is greater than the expected value, the error is passed back to the network, and the errors of the fully connected layer, downsampling layer, and convolutional layer are obtained in turn;
[0184] Among them, the error of each layer is the total error of the network;
[0185] When the error is equal to or less than the expected value, the training is completed;
[0186] After the training is completed, the first highway model data and the second highway model data are obtained.
[0187] Run the prediction confirmation module to:
[0188] Determine the driving standard of vehicles on the current expressway according to the first highway model data;
[0189] Retrieve highway historical data from the database;
[0190] identifying risk factors in the second highway model data based on highway historical data;
[0191] The risk factors in the second highway model data are matched with the driving standardization degree of the first highway model data, and the control data is obtained after the risk matching is completed;
[0192] The control data is divided into risk levels, where the risk levels are divided into level one, level two and level three;
[0193] The control data whose risk level is not within the safe range is transmitted to the display terminal of the control vehicle through wireless signals for display, and control operation prompts are given while displaying;
[0194] Among them, the control operation is to send the vehicle's safe driving data to the display terminal.
[0195] Specifically, the highway data model construction module can be used to specifically retrieve key features from various types of highway data. Targeted feature extraction helps to build a more accurate and effective highway data model. The neural network model is used to propagate and train feature data, which can give full play to the advantages of neural networks in processing complex and nonlinear relationships. Through steps such as initialization parameters, forward propagation and back propagation, the inherent laws and patterns between data can be learned, thereby building a more accurate and reliable highway data model. During the training process, the model is continuously optimized by comparing the error between the output value and the expected value. Accurate error control helps to ensure the accuracy and stability of the model. By constructing the first highway model data and the second highway model data, it can provide an in-depth understanding and prediction of highway conditions. By running the prediction confirmation module, the driving standardization of vehicles on the current highway can be judged in real time based on the first highway model data, and the risk factors in the second highway model data can be identified in combination with the highway historical data. This real-time risk assessment capability enables the system to quickly discover potential safety hazards and issue early warnings in a timely manner, thereby improving road driving safety. By matching the risk factors in the second highway model data with the driving standardization degree of the first highway model data, detailed control data can be generated, and the control data can be divided into risk levels, so that managers can take corresponding control measures according to different risk levels. The control data with risk levels that are not within the safe range are transmitted to the display terminal of the controlled vehicle via wireless signals for display, and control operation prompts are provided at the same time. This timely information feedback and control operation prompts help drivers to understand the safety status of the vehicle in a timely manner and take corresponding control measures to ensure driving safety.
[0196] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are 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 other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0197] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. The intelligent integrated highway management and control system based on digital twin is characterized by: include: Highway data acquisition unit, used for: The highway data is retrieved from the database, and the retrieved highway data is classified according to data attributes. After the classification is completed, a unique coding number is assigned. After the coding number is completed, the highway data is marked as the target highway data; Highway data building block for: Performing data preprocessing on the target highway data, performing data mapping fusion on the target highway data after the data preprocessing through a physical model, and marking the target highway data after the data mapping fusion as first highway data; Real-time data acquisition unit for: Acquire the real-time data on the highway, analyze the real-time data, determine the supervision scope of the real-time data according to the analysis results, and mark the determined real-time data as the second highway data; Run the forecast data implementation unit to: The first highway data and the second highway data are respectively constructed by a neural network model, and after the construction is completed, the first highway model data and the second highway model data are obtained, and operation prediction is performed according to the first highway model data and the second highway model data, and the prediction result is sent to the terminal for control operation; The highway data acquisition unit comprises: Highway data classification module for: Highway data is divided into sensor data and highway structure data; Among them, sensor data includes traffic flow sensors, vehicle speed sensors, ultrasonic sensors, radar sensors, fiber optic sensors, environmental monitoring sensors, smart lighting sensors, and vehicle GPS positioning sensors; The high-speed structural data includes road network traffic data, video surveillance data, road network basic data and auxiliary data; Classify sensor data and high-speed structure data separately; The highway data acquisition unit further includes: Classification data encoding and labeling module, used for: The classified sensor data and high-speed structure data are uniquely coded and labeled through the database self-increment ID and timestamp; Among them, the classified sensor data and high-speed structure data are stored in the database, and after the storage is completed, a unique ID is assigned through the database's self-incrementing ID; Timestamping sensor data and high-speed structure data stored in a database and assigned unique IDs; Generate time labels for sensor data and high-speed structure data that have been assigned unique IDs in the database according to timestamps; Get the unique coding number of each sensor data and high-speed structure data according to the time label and the assigned unique ID; The sensor data and highway structure data with unique coding labels are marked as target highway data; The highway data construction unit includes: Data preprocessing module, used to: Preprocess the target highway data, which includes data cleaning, data integration and data standardization; Data cleaning is to process missing values, outliers and duplicate values of target highway data; Data integration is to integrate the target highway data after data cleaning to obtain a unified data set, and matching, conversion and redundancy are performed during the integration process; Data standardization is to convert the target highway data after data integration into the same scale and format; After data standardization, the pre-processed target highway data is obtained; The data preprocessing process also includes filtering out duplicate, missing and redundant related data in highway data information, including: Extracting duplicate data from the highway data information; Extracting the data acquisition time of every two identical data in the repeated data; The difference evaluation parameter of the repeated data is obtained according to the data acquisition time of each two identical data, wherein the difference evaluation parameter is obtained by the following formula: ； Where Q represents the difference evaluation parameter; Q0 represents the initial value of the parameter; T 01 and T 02 Respectively represent the data acquisition time of two identical data; T x Indicates the preset reference time difference value; N f The number of groups of two identical data whose difference in the time of acquisition of the two identical data is non-zero; N z Indicates the total number of data groups after two identical data in repeated data are grouped together; When the difference evaluation parameter is lower than a preset difference threshold, the data corresponding to the difference evaluation parameter and obtained earlier than the previous one for every two identical data are deleted; When the difference evaluation parameter is not lower than the preset difference threshold, it is determined whether all two identical data corresponding to the difference evaluation parameter are to be retained; When the difference evaluation parameter is not lower than the preset difference threshold, it is determined whether all two identical data corresponding to the difference evaluation parameter are to be retained, including: Extracting all difference evaluation parameters that are not lower than the difference threshold; The compensation coefficient is set using all difference evaluation parameters that are not less than the difference threshold; wherein the compensation coefficient is obtained by the following formula: ； Wherein, λ represents the compensation coefficient; m represents the number of difference evaluation parameters whose difference evaluation parameters are not lower than the preset difference threshold; Q i represents the ith difference evaluation parameter; Q0 represents the difference threshold; The difference threshold is adjusted using the compensation coefficient to obtain an adjusted difference evaluation parameter, wherein the adjusted difference evaluation parameter is obtained by the following formula: ； Among them, Q t represents the adjusted difference evaluation parameter; Q0 represents the difference threshold; When the adjusted difference evaluation parameter is lower than the preset difference threshold, the data corresponding to the difference evaluation parameter and the data acquisition time of each two identical data are deleted; When the adjusted difference evaluation parameter is not lower than the preset difference threshold, every two identical data judgments corresponding to the difference evaluation parameter are retained.
2. The highway intelligent integrated management and control system based on digital twin according to claim 1 is characterized by: The highway data construction unit further includes: Data mapping fusion module for: Retrieving the sensor data and highway structure data from the pre-processed target highway data respectively; Retrieve the physical model from the database; Mapping the sensor data and the high-speed structure data to the parameters in the physical model respectively, wherein the type of the sensor data and the characteristics of the high-speed structure data are used as the mapping data; After the mapping of sensor data and high-speed structure data with parameters in the physical model is completed, data fusion is performed through a neural network model, wherein the neural network model is retrieved from a database; The neural network model fuses the data to adjust the network parameters through the back propagation method and the gradient descent method. After the network parameters are adjusted, the sensor data and the high-speed structure data are fused through the forward propagation of the neural network model. The sensor data and the highway structure data are fused to obtain the first highway data.
3. The highway intelligent integrated management and control system based on digital twin according to claim 2 is characterized by: The real-time data acquisition unit is further used for: Real-time data on highways include real-time weather data and real-time traffic data; Match the real-time meteorological data and real-time traffic data with each other through timestamp identification; After the data correspondence, the real-time meteorological data and the real-time traffic data are associated using data mining methods; Determine the safety range of current highway traffic flow based on real-time meteorological data; Conduct safety analysis on the current safety range of high-speed traffic flow and real-time traffic data; Determine whether the real-time traffic data is within the regulatory scope based on the analysis results; Among them, when the real-time traffic data exceeds the safety range of the current high-speed traffic flow, the real-time traffic data is within the supervision range and marked as risk data; When the real-time traffic data does not exceed the safety range of the current highway traffic volume, the real-time traffic data is not within the supervision range and is marked as safe data; Safety data and risk data are uniformly labeled as second highway data.
4. The highway intelligent integrated management and control system based on digital twin according to claim 3 is characterized by: The operation prediction data implementation unit comprises: Highway data model building module for: The characteristic data in the first highway data and the second highway data are retrieved respectively, wherein the characteristic data of the sensor data include road surface temperature, vehicle speed, and road surface slipperiness; the characteristic data of the highway structure data include road surface status and tunnel status; the characteristic data of the real-time meteorological data include wind speed, precipitation, and visibility; the characteristic data of the real-time traffic data include risk data and safety data; Performing propagation training on the feature data of the first highway data and the second highway data respectively through a neural network model; The propagation training is to initialize the weights of the characteristic data parameters of the first highway data and the second highway data. After the setting is completed, the parameter data is forward propagated through the convolution layer, the downsampling layer and the fully connected layer to obtain the output value; When the error is greater than the expected value, the error is passed back to the network, and the errors of the fully connected layer, downsampling layer, and convolutional layer are obtained in turn; Among them, the error of each layer is the total error of the network; When the error is equal to or less than the expected value, the training is completed; After the training is completed, the first highway model data and the second highway model data are obtained.
5. The highway intelligent integrated management and control system based on digital twin according to claim 4 is characterized by: The operation prediction data implementation unit further includes: Run the prediction confirmation module to: Determine the driving standard of vehicles on the current expressway according to the first highway model data; Retrieve highway historical data from the database; identifying risk factors in the second highway model data based on highway historical data; The risk factors in the second highway model data are matched with the driving standardization degree of the first highway model data, and the control data is obtained after the risk matching is completed; The control data is divided into risk levels, where the risk levels are divided into level one, level two and level three; The control data whose risk level is not within the safe range is transmitted to the display terminal of the control vehicle through wireless signals for display, and control operation prompts are given while displaying; Among them, the control operation is to send the vehicle's safe driving data to the display terminal.