Road traffic dispersion and emergency command system construction method based on multi-source data fusion

Abstract

The invention relates to the technical field of traffic management, in particular to a road traffic dispersion and emergency command system construction method based on multi-source data fusion, and the method comprises the steps: collecting traffic flow sensor data, video monitoring data, meteorological data, social media user feedback data and the like through a multi-source data collection module; a data fusion and processing module is used, and data-level, information-level and decision-level fusion algorithms are adopted to carry out standardization processing, key information extraction and comprehensive evaluation and prediction on multi-source data. And constructing a traffic flow prediction model based on machine learning, such as a long and short-term memory network model, and accurately predicting the traffic flow. And according to the prediction result and the real-time traffic condition, a traffic dispersion strategy making module is used for adjusting signal lamp timing, issuing traffic guidance information and allocating traffic police resources. In the aspect of emergency command, multi-source data are integrated to evaluate event severity, deploy rescue resources and determine an optimal rescue route, and multi-department information sharing and cooperative work are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic management, and particularly to a method for constructing a traffic guidance and emergency command system based on multi-source data fusion when a traffic jam occurs due to a car accident on a highway. Background Art

[0002] In the highway traffic system, the problem of traffic jams caused by car accidents is becoming increasingly prominent. With the continuous growth of highway traffic volume and the high speed of vehicle travel, once a car accident occurs, it is extremely easy to cause serious traffic paralysis, which not only affects the travel efficiency of the public, but also may bring huge economic losses.

[0003] When the current highway traffic management system faces such situations, many deficiencies are exposed. In terms of car accident monitoring, relying on limited road monitoring cameras and driver alarms, it is difficult to achieve comprehensive and real-time monitoring. The camera coverage in mountainous and remote sections is insufficient, and the discovery of accidents often has a delay. Moreover, the driver alarm information is incomplete and inaccurate, resulting in the traffic management department being unable to quickly grasp the full picture of the accident.

[0004] In terms of traffic flow prediction, the traditional model has serious deficiencies in prediction accuracy when facing sudden car accident situations. The highway traffic flow is affected by various complex factors. After a car accident occurs, the traffic flow characteristics change sharply. The traditional model only relies on historical flow data and does not fully consider the actual situations such as road blockage, vehicle queuing, and rescue operations caused by the accident, making the prediction result deviate greatly from the actual situation and unable to provide a reliable basis for traffic guidance. For example, on a highway during peak hours, after a sudden car accident, the traditional model still predicts according to the normal flow trend without considering the impact of the accident on the traffic flow, resulting in subsequent guidance measures being lagged and unreasonable.

[0005] In the emergency command link, the information communication between departments is not smooth. The traffic police, fire, medical, and road rescue departments act independently and lack an efficient coordination mechanism. The information is transmitted untimely and inaccurately, resulting in unreasonable allocation of rescue resources and low rescue efficiency. For example, on the way to the accident scene, the fire department may encounter traffic congestion and delay the arrival time because it cannot obtain the real-time information of the traffic control around the accident by the traffic police; the medical department's understanding of the casualty situation at the accident scene lags behind, affecting the reasonable allocation of rescue resources.

[0006] Traffic diversion strategies lack pertinence and flexibility. Simple road closures and traffic diversions are often adopted without comprehensive consideration of the specific circumstances of the accident scene, such as the severity of the accident, the number of vehicles, the road capacity, and the surrounding road network structure. In a complex highway network, unreasonable diversion strategies are likely to cause traffic congestion to spread over a larger area, further exacerbating traffic chaos. For example, when a car accident occurs near an interchange, due to the lack of a comprehensive analysis of the surrounding road network and blind traffic diversion, vehicle flows converge on other sections, resulting in more serious traffic paralysis.

[0007] In summary, when the existing highway traffic management system responds to traffic jams caused by car accidents, it is difficult to achieve efficient traffic diversion and emergency command. There is an urgent need for an innovative multi-source data fusion system to enhance the ability to handle such emergencies. Summary of the Invention

[0008] Aiming at the problems in the existing technology, the present invention provides a method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion, which overcomes the defects of the existing system in accident monitoring, traffic flow prediction, emergency command coordination, and traffic diversion strategies. By integrating multi-source data, it realizes rapid response, accurate assessment, efficient diversion, and scientific emergency command for car accidents, minimizing the negative impact of accidents on highway traffic and ensuring the safe and unobstructed flow of the road.

[0009] The technical solution adopted by the present invention to solve its technical problems is: a method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion, including the following modules: Multi-source data acquisition module Traffic flow sensors: High-precision traffic flow sensors are deployed at intervals along the highway. According to the formula "the number of vehicles passing through the cross-section of a certain section per unit time = sensor count / statistical time (the unit time is set to minutes)", traffic flow data is collected in real time. For example, if the sensor on a certain section counts 300 vehicles passing through every 5 minutes, then the traffic flow of this section in 5 minutes is 300 vehicles.

[0010] Video surveillance equipment: High-definition cameras along the highway are used to obtain real-time video images of the accident scene and its surroundings. With the help of advanced image recognition algorithms, according to the formula "vehicle speed = vehicle moving distance / video time interval", information such as vehicle speed, density, and vehicle type is extracted. For example, through image recognition, if a vehicle moves a distance of 75 meters in 3 seconds, its speed is 25 meters per second (converted to 90 kilometers per hour).

[0011] Accident scene sensors: After an accident occurs, smoke sensors, gas sensors (detecting the leakage of dangerous chemicals), vehicle deformation sensors, etc. are quickly deployed to obtain detailed information about the accident scene and evaluate the severity of the accident.

[0012] Social Media and User Feedback: Scrape accident-related information posted by users through social media platforms, such as photos, videos, and text descriptions. At the same time, receive accident reports submitted by drivers through traffic service hotlines or mobile applications to enrich the information sources.

[0013] Data Fusion and Processing Module Data-level Fusion: Standardize data in different formats and from different sources, unify the timestamps and data formats, and construct a unified storage structure. For example, align and integrate the data from traffic flow sensors, video surveillance, and accident scene sensors in chronological order to form a complete dataset.

[0014] Information-level Fusion: Calculate key traffic indicators based on the integrated data. For example, the traffic congestion index, with the formula: Congestion Index = (Actual Traffic Flow / Road Capacity) × (Average Vehicle Speed / Free Flow Speed). When the congestion index is greater than 0.7, the road section is determined to be congested. Combine the accident scene sensor data and video images to evaluate the severity of the accident, which is divided into three levels: minor, moderate, and severe.

[0015] Decision-level Fusion: Use artificial intelligence algorithms and expert systems to deeply analyze the fused data. Predict the impact range and duration of the accident on traffic flow, providing a decision-making basis for traffic guidance and emergency command. For example, through a machine learning model trained with historical accident data and real-time data, predict the possible congestion time and range of surrounding road sections under the current accident.

[0016] Traffic Flow Prediction Model Build a prediction model: Adopt deep learning algorithms, such as the Long Short-Term Memory (LSTM) model, and train it with multi-source data. The model inputs include historical traffic flow data, road conditions before the accident, real-time data at the accident scene, weather data, and surrounding road network structure data, etc. The prediction formula is , where is the predicted traffic flow at time t + 1, is the actual traffic flow at time t, is the data of other relevant factors at time t, and f is the prediction function of the LSTM model.

[0017] Model Optimization and Update: Continuously optimize the model parameters based on the real-time collected data to improve the prediction accuracy. Regularly update the model, incorporating new accident cases and traffic data to make the model adapt to traffic flow changes in different scenarios.

[0018] Traffic Guidance Strategy Formulation Module Dynamic adjustment of signal light timing: For the signal lights at highway entrances, exits and surrounding interchange ramps, according to traffic flow prediction and real-time road conditions, the green light duration in a certain direction is dynamically adjusted according to the formula "Green light duration in a certain direction = Total cycle duration × (Traffic flow in that direction / Total traffic flow in all directions)". For example, if an accident causes a significant increase in traffic flow on a certain section, the green light duration in that direction is increased to improve vehicle passing efficiency.

[0019] Precise traffic guidance: Through variable message signs along the highway, traffic broadcasts and mobile navigation applications, real-time road conditions and recommended detour routes are released to drivers. According to the vehicle's location and driving direction, a personalized guidance plan is provided to guide vehicles to avoid accident congestion areas.

[0020] Reasonable allocation of traffic police resources: According to the severity of the accident and the traffic congestion situation, traffic police are reasonably allocated to key sections for dredging. A traffic police allocation priority evaluation system is established, and the formula is: Allocation priority = Severity coefficient of the accident × Congestion index × Importance coefficient of the section. For example, for an important section with a serious accident and a high congestion index, traffic police are allocated preferentially.

[0021] Emergency command module Accident assessment and resource allocation: After an accident occurs, multi-source data is quickly integrated to comprehensively assess the accident. According to the severity, rescue resources such as fire trucks, ambulances, and road rescue vehicles are allocated. Using path planning algorithms, according to the formula "Best rescue route = argmin (Rescue time)", the best rescue route is determined, and the rescue time is calculated based on factors such as road conditions, vehicle speed, and traffic congestion.

[0022] Multi-department collaborative command: Build an information sharing and collaborative command platform for multiple departments such as traffic police, fire departments, medical departments, and road rescue. Each department can obtain accident information in real time through the platform and jointly formulate rescue plans to improve rescue efficiency. For example, the fire department can obtain traffic control information from the traffic police through the platform during the journey and adjust the driving route; the medical department can prepare medical resources in advance based on the casualty feedback at the accident scene.

[0023] Monitoring and adjustment during the rescue process: Through on-site video monitoring and the mobile devices of rescue personnel, the rescue process is monitored in real time. According to the actual situation, the rescue plan and traffic dredging strategy are adjusted in a timely manner. For example, if the accident scene is complex and the rescue difficulty increases, additional rescue forces are dispatched in a timely manner and the scope of traffic control is adjusted.

[0024] Advantages of the present invention: (1) The method for constructing a highway traffic dredging and emergency command system based on multi-source data fusion described in the present invention improves the accident monitoring and response speed: Multi-source data collection makes the monitoring of highway accidents more comprehensive and timely. Combining image recognition and social media data, information can be obtained within minutes after an accident occurs. Compared with traditional methods, the response time is shortened by more than 50%, saving precious time for rescue and traffic dredging.

[0025] (2) The method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion according to the present invention improves the accuracy of traffic flow prediction: The traffic flow prediction model based on multi-source data fusion fully considers the impact of accidents on traffic flow, and the prediction accuracy is improved by more than 30% compared with traditional models. It can accurately predict the traffic congestion range and duration caused by accidents, providing strong support for the formulation of scientific guidance strategies.

[0026] (3) The method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion according to the present invention optimizes the traffic guidance effect: According to real-time road conditions and traffic flow predictions, it dynamically adjusts signal timings, precisely guides, and reasonably allocates traffic police resources, effectively alleviating congestion on sections around accidents. In practical applications, after implementing the guidance strategy of the present invention on a certain accident section, the average vehicle passing speed has increased by more than 20 km / h, and the congestion time has been shortened by more than 40%.

[0027] (4) The method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion according to the present invention enhances the efficiency of emergency command coordination: The multi-department collaborative command platform realizes information sharing and collaborative decision-making, and the communication efficiency between departments has been greatly improved. During the rescue process, each department responds quickly and cooperates with each other. The rescue time has been shortened by more than 30% compared with the past, improving the success rate of accident rescue and reducing the impact range and duration of accidents on traffic. Description of the Drawings

[0028] The present invention will be further described below with reference to the drawings and embodiments.

[0029] Figure 1 It is the system architecture diagram of the method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion provided by the present invention; Figure 2 It is the data flow diagram of the method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion provided by the present invention; Figure 3 It is the model schematic diagram of the method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion provided by the present invention, taking the LSTM traffic flow prediction model as an example. Detailed Embodiments

[0030] In order to make the technical means, creative features, achieved purposes, and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0031] As Figures 1-3 shown, the method for constructing a highway traffic guidance and emergency command system based on multi-source data fusion according to the present invention includes the following steps: Multi-source data collection steps: Use traffic flow sensors to collect traffic flow data for each road section. The deployment interval of traffic flow sensors does not exceed 500 meters to ensure that the collected data can accurately reflect the traffic flow changes of the road section. The calculation formula is the number of vehicles passing through the cross-section of a certain road section per unit time = sensor count / statistical time; Obtain real-time road images and video data through video surveillance equipment, and use image recognition technology to extract information such as vehicle speed and density. The vehicle speed calculation formula is vehicle speed = vehicle moving distance / video time interval; Collect weather data provided by the meteorological department; Obtain user feedback data from social media platforms; Data fusion and processing steps: Adopt a data fusion algorithm for data-level fusion, standardize data in different formats and sources, and unify the storage format; Conduct information-level fusion, calculate the traffic congestion index based on data such as traffic flow, vehicle speed, and density. The congestion index = (actual traffic flow / road capacity) × (average vehicle speed / free flow speed); Conduct decision-level fusion, combine expert systems and artificial intelligence technologies to comprehensively evaluate and predict the traffic conditions. For the unstructured text data obtained from social media, use the named entity recognition technology in natural language processing to extract key information such as accident location, time, and type to improve data availability; Steps for constructing a traffic flow prediction model: Construct a traffic flow prediction model based on machine learning, such as a long short-term memory network model. The input includes multi-source data such as historical traffic flow data, real-time road conditions information, and weather data. It uses transfer learning technology to use the model parameters trained in other similar regions as initialization parameters and fine-tune them in combination with local data to shorten the model training time and improve the prediction efficiency. The prediction formula is where is the predicted traffic flow at time t + 1, is the actual traffic flow at time t, is the data of other relevant factors at time t, and f is the prediction function of the LSTM model; Steps for formulating traffic guidance strategies: According to the traffic flow prediction results and real-time traffic conditions, formulate traffic guidance strategies, including adjusting signal timing. The green light duration in a certain direction = total cycle duration × (traffic flow in that direction / sum of traffic flows in all directions); Release traffic guidance information; Allocate traffic police resources. The allocation priority = congestion index × road importance coefficient. When releasing traffic guidance information, consider the characteristics of different travel modes and provide exclusive optimized detour routes for various vehicles; Emergency command steps: When an emergency occurs, integrate multi-source data, evaluate the severity of the event, formulate an emergency rescue plan, including deploying resources such as rescue vehicles, personnel, and supplies, use path planning algorithms to determine the best rescue route, the best rescue route = argmin, monitor the rescue process in real time and adjust the rescue plan, achieve information sharing and collaborative work among various departments, establish an emergency resource reserve database, and update the quantity, location, and available status of fire, medical, and rescue vehicle resources in real time to provide an accurate basis for resource allocation.

[0032] Specifically, the system has an automatic inspection function, regularly detects the status of traffic flow sensors and video surveillance equipment hardware facilities, and when a device failure is found, it automatically issues an alarm and records the fault information to ensure the continuity of data collection.

[0033] Specifically, during the information sharing process among various departments, an encrypted transmission protocol is adopted to ensure the security and integrity of data during transmission and prevent information leakage.

[0034] Specifically, the system sets up a user feedback and evaluation module to collect the evaluation opinions of traffic participants on traffic guidance strategies and emergency command effects, which are used to continuously optimize the functions and strategies of the system.

[0035] Appendix Figure 1 Among them, the multi-source data collection module collects various traffic-related data and transmits it to the data fusion and processing module for integration and analysis. The processed data is used on the one hand to build a traffic flow prediction model, and on the other hand to provide a decision-making basis for the traffic guidance strategy formulation module and the emergency command module. The traffic guidance strategy and emergency command plan are implemented through the traffic facility execution unit, and the traffic status is fed back to the multi-source data collection module to form a closed loop. At the same time, the data generated and used by each module is stored in the data storage center for easy data query, call, and historical analysis.

[0036] Specifically, the multi-source data acquisition module: This module is responsible for collecting various types of data related to highway traffic and is the foundation for the operation of the system. Among them, traffic flow sensors are distributed on each section of the highway, and traffic flow data is collected at set time intervals (such as every 5 minutes). According to the formula "the number of vehicles passing through the cross-section of a certain section per unit time = sensor count / statistical time", the real-time traffic flow of each section can be accurately obtained. The video surveillance equipment uses high-definition cameras to monitor the road comprehensively. With the help of image recognition technology, information such as vehicle speed, density, and vehicle type can be extracted. The vehicle speed calculation formula is "vehicle speed = vehicle moving distance / video time interval". The meteorological department interface accesses meteorological data in real time, such as temperature, precipitation, wind speed, etc. These meteorological factors have a significant impact on traffic conditions. The social media platform interface collects information such as traffic accident reports, road congestion descriptions, accident scene photos and videos posted by users through data scraping technology, broadening the data acquisition channels and enabling the system to comprehensively grasp traffic dynamics more fully.

[0037] Data fusion and processing module: This module receives various types of data transmitted from the multi-source data acquisition module and performs a series of processing. First, data standardization processing is carried out to convert data in different formats and from different sources into a unified format and unify the timestamps for subsequent fusion. Then, data-level fusion is carried out to align and integrate the standardized data according to the time series to form a complete data set. In the information-level fusion stage, key traffic indicators are calculated based on the integrated data, such as the traffic congestion index. The calculation formula is "congestion index = (actual traffic flow / road capacity) × (average vehicle speed / free flow vehicle speed)", and based on this, the congestion status of the section is judged. At the same time, combined with accident scene sensor data and video images, the severity of the accident is evaluated, which is divided into three levels: minor, general, and serious. Finally, decision-level fusion is carried out using artificial intelligence algorithms and expert systems to comprehensively evaluate and predict the traffic conditions, providing reliable data support for subsequent modules.

[0038] Traffic flow prediction model: It is constructed using deep learning algorithms, such as the long short-term memory network (LSTM) model. The inputs of this model include historical traffic flow data, real-time road conditions information (obtained from the data fusion and processing module), weather data, and other relevant factor data (such as accident information, road construction conditions, etc.). The prediction formula is , where is the predicted traffic flow at time t + 1, is the actual traffic flow at time t, is the other relevant factor data at time t, and f is the prediction function of the LSTM model. Through learning and training on a large amount of historical data and real-time data, the model can capture the changing patterns of traffic flow and the complex correlation relationships between various factors, accurately predict future traffic flow, and provide key prediction basis for traffic guidance and emergency command.

[0039] Traffic Diversion Strategy Formulation Module: Based on the results output by the traffic flow prediction model and the real-time traffic condition information provided by the data fusion and processing module, formulate targeted traffic diversion strategies. For example, for signal timing adjustment, according to the proportion of traffic flow in different directions, use the formula "green light duration of a certain direction = total cycle duration × (traffic flow of this direction / sum of traffic flows in all directions)" to dynamically extend the green light duration of the direction with a large traffic flow, improving the road traffic efficiency. Utilize the traffic guidance system to broadcast real-time traffic conditions and recommended detour routes to drivers through variable message signs along highways, traffic radio, and mobile navigation applications, guiding vehicles to avoid congested sections. Meanwhile, according to the severity of accidents and traffic congestion conditions, combined with the formula "dispatching priority = accident severity coefficient × congestion index × section importance coefficient", rationally allocate traffic police resources to key sections for traffic diversion.

[0040] Emergency Command Module: When emergencies such as car accidents occur on highways, quickly integrate multi-source data and comprehensively evaluate the events. According to the severity of the accidents, allocate corresponding rescue resources, such as fire trucks, medical ambulances, road rescue vehicles, etc. Use path planning algorithms, combined with the accident location, distribution of surrounding rescue resources, and real-time traffic condition information, to determine the best rescue route according to the formula "best rescue route = argmin (rescue time)", ensuring that rescue resources can quickly reach the accident scene. Establish an information sharing and collaborative command platform for multiple departments including traffic police, fire, medical, and road rescue. Each department can obtain accident information in real time through the platform and jointly formulate rescue plans to improve rescue efficiency. During the rescue process, through on-site video monitoring and the mobile devices of rescue personnel, monitor the rescue progress in real time and adjust the rescue plan and traffic diversion strategy in a timely manner according to the actual situation.

[0041] Traffic Facility Execution Unit: Responsible for executing the instructions issued by the traffic diversion strategy formulation module and the emergency command module. In terms of traffic diversion, control the signal lights along the highway and switch the signal light status according to the adjusted timing plan; display real-time traffic conditions and traffic guidance information through variable message signs; guide traffic police to designated sections for on-site traffic diversion. In emergency command, dispatch fire trucks, medical ambulances, road rescue vehicles, etc. to the accident scene according to the planned best rescue route to ensure the orderly progress of rescue work.

[0042] Traffic Status Feedback: Through traffic flow sensors, video monitoring devices, and other traffic monitoring means, collect traffic status data after the traffic facility execution unit implements traffic diversion and emergency command measures in real time. For example, information such as changes in traffic flow, vehicle speed recovery, and alleviation of congested sections. Transmit this feedback data back to the multi-source data collection module, enabling the system to continuously optimize the data collection, analysis, and decision-making processes according to the changes in the actual traffic status, forming a closed-loop dynamic optimization system.

[0043] Data storage center: It is used to store the data generated and used by each module during operation. The raw data collected by the multi-source data acquisition module, the intermediate data processed by the data fusion and processing module, the training data and prediction results of the traffic flow prediction model, the decision-making data of the traffic guidance strategy formulation module and the emergency command module, etc. are all stored here. The data storage center provides data query and call services for each module, facilitating the modules to obtain historical data for comparative analysis at any time, and providing data support for the continuous optimization and improvement of the system. At the same time, the data storage center has data backup and security protection functions to ensure the integrity and security of the data.

[0044] Appendix Figure 2 Among them, the data collected by traffic flow sensors, video surveillance devices, meteorological department interfaces, and social media platform interfaces are aggregated and then sequentially undergo data standardization processing, data-level fusion, information-level fusion, and decision-level fusion. The fused data is used for the training and prediction of the traffic flow prediction model, as well as the formulation of traffic guidance strategies and emergency command plans. After implementing traffic guidance and emergency command measures, traffic status feedback data is collected and re-entered into the data aggregation link to achieve the recycling of data and the continuous optimization of the system.

[0045] Appendix Figure 3 Among them, historical traffic flow data, real-time road condition information, weather data, and other relevant factor data are used as inputs and enter the long short-term memory (LSTM) model. The LSTM model captures the traffic flow change rules and the correlation relationships between various factors through learning and analyzing these multi-source data, and thus outputs traffic flow prediction results, providing key prediction basis for traffic guidance and emergency command.

[0046] Example 1: Traffic guidance and emergency command for a regular car accident Accident monitoring and data collection: On a certain expressway section, traffic flow sensors continuously monitor the changes in traffic volume, and video surveillance devices continuously capture road pictures. When a two-vehicle rear-end collision occurs, the video surveillance system quickly detects the accident through image recognition technology and automatically triggers an alarm. At the same time, surrounding drivers upload accident photos and location information through the mobile traffic application. The traffic management department obtains the accident information within 1 minute and immediately activates the emergency response mechanism.

[0047] Data Fusion and Processing: The multi-source data acquisition module transmits traffic flow sensor data, video surveillance image data, and driver feedback data to the data fusion and processing module. This module standardizes the data, unifies the time format, and calculates the traffic congestion index. After calculation, the congestion index of the accident section rises to 0.8 within 10 minutes after the accident, determining that this section is in a congested state. At the same time, combining video images and driver descriptions, the accident is evaluated as an ordinary accident.

[0048] Traffic Flow Prediction: The traffic flow prediction model is trained and predicted using multi-source data. The model predicts that in the current accident situation, traffic congestion will occur within 2 kilometers of the accident section and its surrounding areas, and the congestion duration is expected to be 2 hours.

[0049] Traffic Diversion and Emergency Command: According to the traffic flow prediction results and accident assessment, the traffic diversion strategy formulation module quickly formulates a plan. First, adjust the signal timing of the entrances and exits around the accident section, increasing the green light duration in the accident direction from 30 seconds to 45 seconds to improve the vehicle evacuation speed. Second, through variable message signs and mobile navigation applications, release accident information and detour routes to drivers to guide vehicles to detour. At the same time, the traffic management department quickly deploys 2 traffic police officers to the accident scene for traffic diversion according to the deployment priority formula. The emergency command module coordinates the joint work of the traffic police, fire department, and medical department. After receiving the notice, the fire department and medical department quickly rush to the accident scene along the best rescue route. During the rescue process, each department communicates in real time through the joint command platform to ensure the smooth progress of the rescue work. After 1 hour of rescue and traffic diversion, the accident vehicles are cleared and the road returns to normal traffic.

[0050] Example 2: Response to Car Accidents in Bad Weather Meteorological Data Acquisition and Accident Occurrence: On a certain highway in winter, the meteorological department issued a heavy snow weather warning in advance. After the traffic management system obtains the meteorological data, it adjusts the traffic flow prediction model to reduce the road traffic capacity. In heavy snow weather, a multi-vehicle chain-reaction accident occurred on a certain section. The traffic flow sensor detected a sharp decrease in traffic volume, and some of the video surveillance equipment's images were blurred due to heavy snow, but the accident situation was still initially judged through image recognition technology. At the same time, there were a large number of discussions about this accident and photos and videos uploaded by users on social media.

[0051] Data Fusion and Accident Assessment: The data fusion and processing module integrates meteorological data, traffic flow data, video surveillance data, and social media data. Considering the impact of bad weather on traffic, recalculate the traffic congestion index. Due to the slippery road surface caused by heavy snow and the reduced vehicle driving speed, the congestion index of the accident section quickly rises to 0.9. Combining multi-source data, the accident is evaluated as a serious accident and there may be casualties.

[0052] Traffic Flow Prediction and Diversion Strategy: The traffic flow prediction model predicts that the accident will cause traffic congestion within a 5-kilometer radius around the accident site, and the congestion duration is expected to exceed 3 hours. Based on the prediction results, the traffic diversion strategy formulation module takes a series of measures. On the one hand, variable message signs are set up at multiple locations along the highway to release accident information and severe weather warnings, reminding drivers to slow down or choose alternative routes. On the other hand, some entrances and exits around the accident section are closed to reduce the number of vehicles entering the congested area. At the same time, 5 traffic policemen are deployed to the accident site and key sections around it for traffic diversion, strengthening the guidance and control of vehicles.

[0053] Emergency Rescue Coordination: The emergency command module quickly coordinates the traffic police, fire department, medical department, and road rescue department. The fire department rushes to the scene with professional snow removal and rescue equipment. The medical department dispatches multiple ambulances and prepares medical resources to deal with injuries such as hypothermia and frostbite. Each department shares information in real time through the collaborative command platform and adjusts the rescue plan according to the accident scene situation. During the rescue process, due to the continuous heavy snow, the road is severely snow-covered, increasing the rescue difficulty. The traffic management department promptly dispatches snow removal vehicles to conduct snow removal operations on the accident site and surrounding roads to ensure the passage of rescue vehicles and evacuation vehicles. After more than 2 hours of intense rescue, the accident vehicles are cleared, and the road gradually resumes traffic.

[0054] Example 3: Traffic Diversion and Emergency Command for a Major Car Accident Major Accident Monitoring and Information Collection: A major car accident involving more than 10 vehicles occurred near an interchange on a certain highway. The traffic flow sensor detected a sharp drop in the traffic flow on this section, and the video surveillance equipment captured a serious chaotic scene at the accident site. The accident information spread rapidly on social media, attracting wide attention. The traffic management department obtained the accident information through multiple channels within 5 minutes and immediately launched the highest-level emergency response.

[0055] Multi-source Data Fusion and Evaluation: The data fusion and processing module deeply fuses traffic flow data, video surveillance image data, social media data, and accident site sensor data (subsequently deployed). The calculated congestion index of the accident section and its surrounding areas is as high as 1.2, which is determined to be severely congested. Combining the on-site video and sensor data, the accident is evaluated as an extremely serious accident, possibly involving a large number of casualties and the risk of dangerous chemical leakage.

[0056] Traffic Flow Prediction and Diversion Plan Formulation: The traffic flow prediction model forecasts that the accident will cause traffic congestion within a 10-kilometer radius around the accident site, and the congestion duration is expected to exceed 5 hours. The traffic diversion strategy formulation module has developed a detailed diversion plan. First, a comprehensive analysis of the road network around the accident is conducted, and multiple detour routes are released to drivers through traffic radio and mobile navigation applications to guide vehicles to avoid the accident area. Second, traffic control is implemented at multiple entrances and exits around the accident to restrict vehicles from entering the congested section. Meanwhile, 10 traffic police officers and multiple police cars are deployed to the accident site and important nodes in the vicinity for traffic diversion and order maintenance. In addition, multiple temporary traffic diversion points are set up to reasonably guide vehicle diversion and relieve the congestion pressure.

[0057] Emergency Rescue Coordination and Resource Allocation: The emergency command module quickly organizes multi-department collaborative operations among traffic police, fire, medical, environmental protection and other departments. The fire department dispatches multiple fire trucks and professional rescue teams to carry out fire extinguishing and rescue work at the accident site. The medical department sends a large number of ambulances and medical staff to set up a temporary medical treatment point at the scene to provide emergency treatment for the injured. The environmental protection department conducts environmental monitoring at the accident site to prevent pollution to the environment caused by the leakage of dangerous chemicals. Each department communicates in real time through the collaborative command platform and adjusts the rescue and diversion plans according to the progress of the rescue and the traffic conditions. After more than 4 hours of arduous efforts, the accident site has been effectively cleared and the road has gradually resumed traffic.

[0058] Example 4: Response to a Car Accident during Highway Construction Construction Information and Accident Occurrence: Road surface construction is underway on a certain section of the highway, and corresponding traffic warning signs and guiding facilities are set up in the construction area. Near the construction area, a car accident occurs due to a driver avoiding a construction vehicle. The traffic flow sensors detect abnormal changes in the traffic flow in the construction area and its vicinity, and the video surveillance equipment captures the accident scene. At the same time, the construction personnel report the accident situation to the traffic management department in a timely manner.

[0059] Data Fusion and Traffic Condition Assessment: The data fusion and processing module fuses and processes the construction information, traffic flow data, video surveillance data, and the data reported by the construction personnel. Considering the impact of construction on the road traffic capacity, the traffic congestion index is recalculated. After the accident, the congestion index in the construction area and its surrounding sections has risen to 0.85, which is determined to be in a congested state. Combining multi-source data, the accident is evaluated as a general accident, but due to the special situation in the construction area, it may bring certain difficulties to traffic diversion and rescue work.

[0060] Traffic Flow Prediction and Traffic Diversion Strategy Adjustment: The traffic flow prediction model predicts that the accident will cause traffic congestion within a 3-kilometer radius of the construction area and its surrounding areas, and the duration of the congestion is expected to be 2.5 hours. The traffic diversion strategy formulation module adjusts the traffic diversion strategy based on the prediction results and the construction situation. On the one hand, more traffic warning signs are set up in front of and behind the construction area to remind drivers to pay attention to the accident and construction situation and slow down. On the other hand, coordinate with the construction unit to temporarily adjust the construction progress and operation mode to provide more passing space for rescue vehicles and evacuation vehicles. At the same time, through the traffic guidance system, guide vehicles to avoid the construction area and the accident congestion section. In addition, deploy 3 traffic policemen to the accident site and surrounding roads for traffic diversion to maintain traffic order.

[0061] Emergency Rescue Coordination: The emergency command module quickly coordinates the traffic police, fire department, and medical department. The fire department and the medical department quickly rush to the accident site to rescue and treat the accident vehicles and the injured. The traffic police department conducts traffic control at the accident site and surrounding roads to ensure the smooth progress of the rescue work. Each department shares information in real time through the collaborative command platform and adjusts the rescue and traffic diversion measures according to the accident site situation. After more than one hour of efforts, the accident vehicles are cleared, the construction area resumes normal construction, and the road resumes traffic.

[0062] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of protection required by the present invention. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion, characterized in that: The following steps are involved: Multi-source data collection steps: Use traffic flow sensors to collect traffic flow data on each road section. The calculation formula is the number of vehicles passing through a cross section of a road section per unit time = sensor count / statistical time; The real-time images and video data of the road are obtained through video monitoring equipment, and the vehicle speed and density information are extracted using image recognition technology. The vehicle speed calculation formula is vehicle speed = vehicle moving distance / video time interval; Collect weather data provided by the meteorological department; Obtain user feedback data from social media platforms; Data fusion and processing steps: Use data fusion algorithm to perform data-level fusion, standardize data of different formats and sources, and unify the storage format; perform information-level fusion, calculate the traffic congestion index based on traffic volume, vehicle speed and density data, and the congestion index = (actual traffic volume / road capacity) × (average speed / free flow speed); perform decision-level fusion, combine expert systems and artificial intelligence technology, and conduct comprehensive evaluation and prediction of traffic conditions; Traffic flow prediction model construction steps: Build a traffic flow prediction model based on machine learning, such as a long short-term memory network model, and input multi-source data including historical traffic flow data, real-time road conditions information, and weather data. The prediction formula is ,in is the predicted traffic flow at time t+1, is the actual traffic flow at time t, is other relevant factor data at time t, and f is the prediction function of the LSTM model; Traffic diversion strategy formulation steps: According to the traffic flow prediction results and real-time traffic conditions, formulate traffic diversion strategies, including adjusting the timing of signal lights. The green light duration in a certain direction = total cycle duration × (traffic flow in that direction / total traffic flow in all directions); Release traffic guidance information; deploy traffic police resources, deployment priority = congestion index × road section importance coefficient; Emergency command steps: When an emergency occurs, integrate multi-source data, assess the severity of the incident, and formulate an emergency rescue plan, including the deployment of rescue vehicles, personnel and material resources, and use the path planning algorithm to determine the best rescue route, the best rescue route = argmin, monitor the rescue process in real time and adjust the rescue plan to achieve information sharing and collaborative work among departments.

2. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: The traffic flow sensors are deployed at intervals of no more than 500 meters to ensure that the collected data can accurately reflect changes in traffic flow on the road section.

3. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: In the data fusion and processing steps, the named entity recognition technology in natural language processing is used to extract key information such as accident location, time, and type from the unstructured text data obtained from social media, thereby improving data availability.

4. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: In the step of building the traffic flow prediction model, transfer learning technology is used to take the model parameters trained in other similar areas as initialization parameters, and fine-tune them in combination with local data to shorten the model training time and improve the prediction efficiency.

5. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: In the process of formulating traffic diversion strategies, when issuing traffic guidance information, the characteristics of different travel modes are taken into consideration to provide exclusive optimized detour routes for various types of vehicles.

6. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: During the emergency command process, an emergency resource reserve database is established to update the quantity, location and availability of fire, medical and rescue vehicle resources in real time to provide an accurate basis for resource allocation.

7. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: The system has an automatic inspection function, which regularly checks the status of traffic flow sensors and video surveillance equipment hardware facilities. When equipment failure is detected, it automatically issues an alarm and records the failure information to ensure the continuity of data collection.

8. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1 is characterized in that: In the process of information sharing among departments, encrypted transmission protocols are used to ensure the security and integrity of data during transmission and prevent information leakage.

9. The method for constructing a highway traffic diversion and emergency command system based on multi-source data fusion according to claim 1, characterized in that: The system sets up a user feedback evaluation module to collect traffic participants' evaluation opinions on traffic diversion strategies and emergency command effects, which is used to continuously optimize the system's functions and strategies.

Images