Highway dynamic speed limiting method based on multi-source data

Abstract

The invention discloses an expressway dynamic speed limit method based on multi-source data, and belongs to the technical field of expressway speed limit, and the method comprises the steps: obtaining accident data, constructing an accident risk degree prediction model, and generating an accident risk degree trend chart; acquiring real-time traffic flow data, constructing a traffic flow prediction model, and generating a traffic flow tendency chart; obtaining meteorological data, and generating a local meteorological state distribution diagram; performing space alignment and data fusion on the traffic flow tendency chart and the local meteorological state distribution diagram to generate a road section comprehensive risk index; performing time alignment on the accident risk degree trend chart and the comprehensive risk degree index of the road section to generate a real-time risk degree score of the current road section; smoothly adjusting the speed limit value by adopting a linear interpolation algorithm on the basis of the real-time risk degree score in combination with a preset speed limit adjustment rule; and comparing the current vehicle speed with the adjusted speed limit value, and early warning over-speed. According to the invention, accurate speed limit control based on multi-dimensional data is realized, and the road safety is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of highway speed limitation, and in particular relates to a dynamic speed limitation method for highways based on multi-source data. Background Art

[0002] In the dynamic speed limit system of highways, the fusion and real-time processing of multi-source data are key. However, in actual business scenarios, the diversity and heterogeneity of data sources have brought about technical contradictions. Accident data comes from historical records and real-time alarm systems, and contains information such as accident type, time and location of occurrence, but its temporal and spatial resolution is low, and it is difficult to accurately reflect the immediate degree of danger of the current road section. Meteorological data includes visibility, precipitation, temperature, etc., which are collected through meteorological stations and satellites, but meteorological stations are sparsely distributed, resulting in blind spots in data in some areas. Traffic flow data relies on cameras, radars, and induction coils to capture vehicle density and speed in real time, but equipment failure or signal interference may cause incomplete or distorted data.

[0003] These data sources differ significantly in time scale, spatial coverage and data format, and how to achieve efficient data alignment and fusion becomes a core issue. For example, accident data is updated less frequently, while traffic flow data is updated in seconds, and it is difficult to directly match the two in the time dimension. In addition, the spatial resolution of meteorological data is limited and cannot accurately reflect local micro-weather changes, while traffic flow data is limited to equipment installation points and is difficult to cover the entire highway.

[0004] Furthermore, there is a contradiction between the real-time requirements of the dynamic speed limit algorithm and the complexity of data processing. The system needs to complete data collection, cleaning, fusion and risk assessment in a very short time, while also considering data delay and loss. In addition, the dynamic adjustment of the speed limit value also needs to consider the driver's behavior and psychological factors. Too low or too fast speed limit changes may cause the driver to make misjudgments or operational errors, which in turn increases safety risks.

[0005] In summary, how to smoothly adjust the speed limit while ensuring the accuracy of risk assessment and avoiding adverse effects on driving experience and traffic flow is a technical problem that needs to be solved urgently. Summary of the invention

[0006] In order to solve the above technical problems, the present invention proposes a method for dynamic speed limit on expressways based on multi-source data to solve the above problems existing in the prior art.

[0007] To achieve the above object, the present invention provides a method for dynamic speed limit of a highway based on multi-source data, comprising the following steps:

[0008] Obtain accident data, construct an accident risk prediction model based on time series, and generate an accident risk trend chart;

[0009] Obtain real-time traffic flow data, construct a traffic flow prediction model, and generate a traffic flow trend chart;

[0010] Obtain meteorological data and generate a local meteorological status distribution map;

[0011] Align the traffic flow trend chart and the local meteorological status distribution map spatially and fuse the data to generate a comprehensive risk index for road segments;

[0012] Align the accident risk trend chart and the comprehensive risk index for road segments in time, and use a weighted fusion algorithm to generate a real-time risk score for the current road segment;

[0013] Based on the real-time risk score, combined with the preset speed limit adjustment rules, use a linear interpolation algorithm to smoothly adjust the speed limit value;

[0014] Compare the current vehicle speed with the adjusted speed limit value. If there is an overspeed behavior, give an alarm.

[0015] Optionally, the process of obtaining accident data, constructing an accident risk prediction model based on time series, and generating an accident risk trend includes:

[0016] Obtain the original fields of accident type, time point, and location name from the accident data, and construct a spatio-temporal distribution matrix; based on the spatio-temporal distribution matrix, calculate the accident distribution law, extract eigenvalue according to the accident distribution law, and determine the accident spatio-temporal distribution; based on the accident spatio-temporal distribution, combined with the time series analysis method, construct an accident risk prediction model; obtain the risk prediction value based on the accident risk prediction model, and generate an accident risk trend according to the risk prediction value.

[0017] Optionally, the process of obtaining real-time traffic flow data, constructing a traffic flow prediction model, and generating a traffic flow trend chart includes:

[0018] Obtain real-time traffic flow data and extract vehicle density and speed information; use a sliding window algorithm to smooth the vehicle density and speed to obtain the smoothed vehicle density value and speed value; according to the smoothed vehicle density value and speed value, combined with the time series analysis method, construct a traffic flow prediction model; calculate the traffic flow value for the future time period through the traffic flow prediction model and output the traffic flow trend chart.

[0019] Optionally, the process of obtaining meteorological data and generating a local meteorological status distribution map includes:

[0020] Extract three groups of data, namely visibility, precipitation, and temperature values, from meteorological data, and classify and store them according to the location information of meteorological stations; use interpolation algorithms to calculate data blind spots in sparse areas of meteorological stations and fill in missing data for visibility, precipitation, and temperature values; smooth the filled visibility, precipitation, and temperature values to obtain three groups of smoothed values; calculate the meteorological state of a local area based on the smoothed visibility, precipitation, and temperature values. If the meteorological state value exceeds the preset threshold range, determine that the meteorological state in this area is abnormal and generate abnormal identification information; combine spatial analysis methods to construct a meteorological state distribution map.

[0021] Optionally, the process of spatially aligning and data fusing the traffic flow trend map with the local meteorological state distribution map to generate the comprehensive risk index for road segments includes:

[0022] Obtain the traffic flow trend map and the local meteorological state distribution map, and use a geographic information system to perform spatial alignment on the two; input the spatially aligned data into a convolutional neural network model to extract the features of traffic flow data and the features of the meteorological state distribution map, and generate feature vectors; perform weighted fusion on the traffic flow feature vector and the meteorological state feature vector to obtain a comprehensive feature vector; calculate the comprehensive risk index for each road segment through a fully connected layer based on the comprehensive feature vector.

[0023] Optionally, the process of smoothly adjusting the speed limit value based on real-time risk scores and combining preset speed limit adjustment rules using a linear interpolation algorithm includes:

[0024] Obtain the real-time risk score of the current road segment, and based on the preset speed limit adjustment rules, use a linear interpolation algorithm to calculate the speed limit adjustment amplitude of the current road segment; combine the calculated speed limit adjustment amplitude with the current speed limit value to generate a smoothed new speed limit value; determine whether the new speed limit value exceeds the preset speed limit range. If it exceeds, adjust the speed limit value to the boundary value of the preset range; update the adjusted speed limit value to the road segment speed limit database.

[0025] Optionally, the process of comparing the current vehicle speed with the adjusted speed limit value and issuing a warning if there is a speeding behavior includes:

[0026] Use a comparative analysis algorithm to compare vehicle speed data with the adjusted speed limit value in real time to generate a speed difference; determine whether the speed difference exceeds the speed limit range according to a preset difference threshold. If it exceeds, determine it as a speeding behavior; generate a warning message for the speeding behavior, trigger a warning response mechanism based on the warning message, record the storage location and timestamp of the warning message to form a warning record library; generate a driving behavior analysis report based on the warning record library and vehicle speed data.

[0027] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory, where the processor executes the computer program to implement the steps of the method.

[0028] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method are implemented.

[0029] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method are implemented.

[0030] Compared with the prior art, the present invention has the following advantages and technical effects:

[0031] The present invention discloses a highway dynamic speed limit method based on multi-source data. The method constructs an accident risk prediction model and a comprehensive road section risk index by analyzing accident data, real-time traffic flow, and meteorological information; the system aligns and fuses this information in space and time to generate a real-time road section risk score; based on this score, the present invention uses a linear interpolation algorithm to dynamically adjust the speed limit value and determines speeding behavior by comparing the actual speed of the vehicle; when a high-risk speeding situation occurs, the reinforcement learning algorithm can be used to optimize the speed limit strategy and further reduce the speed limit value to reduce potential safety hazards. Compared with the prior art, the present invention realizes precise speed limit control based on multi-dimensional data, can dynamically adjust the speed limit according to real-time road conditions, effectively improves road safety, and reduces the incidence of traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0033] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0035] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0036] Embodiment 1

[0037] AsFigure 1 As shown in Figure 1 , in this embodiment, a highway dynamic speed limit method based on multi-source data is provided, including the following steps:

[0038] Obtain accident data, construct an accident risk prediction model based on time series, and generate an accident risk trend chart;

[0039] Obtain real-time traffic flow data, construct a traffic flow prediction model, and generate a traffic flow trend chart;

[0040] Obtain meteorological data and generate a local meteorological state distribution map;

[0041] Spatially align and data-fuse the traffic flow trend chart and the local meteorological state distribution map to generate a comprehensive road section risk index;

[0042] Temporally align the accident risk trend chart with the comprehensive road section risk index, and use a weighted fusion algorithm to generate a real-time risk score for the current road section;

[0043] Based on the real-time risk score, in combination with a preset speed limit adjustment rule, use a linear interpolation algorithm to smoothly adjust the speed limit value;

[0044] Compare the current vehicle speed with the adjusted speed limit value. If there is a speeding behavior, give an alarm.

[0045] As an implementable method, it specifically includes the following steps:

[0046] S101. Extract the accident type, occurrence time, and location from the accident data, and in combination with the spatio-temporal distribution law of accident occurrences, construct an accident risk prediction model based on time series, and generate an accident risk trend.

[0047] Specifically, obtain original fields such as accident type, time point, and location name from the accident database. Construct a spatio-temporal distribution matrix for the accident type, time point, and location name, and calculate the accident distribution law. Extract eigenvalue according to the accident distribution law to determine the accident spatio-temporal distribution. Use the time series analysis method to construct an accident risk prediction model and set model parameters. Calculate the accident risk for a future time period through the accident risk prediction model to obtain a risk prediction value. Draw an accident trend chart based on the risk prediction value and output an accident risk trend chart.

[0048] Implementable. The accident database contains rich information, such as accident types, occurrence time and location, etc. By extracting these original fields, a spatio-temporal distribution matrix can be constructed to reveal the laws of accident occurrence. For example, in the traffic accident data of a certain city, it may be found that the accident frequency at the downtown crossroads is significantly increased during the morning and evening rush hours. The construction of the spatio-temporal distribution matrix lays the foundation for calculating the accident distribution law. The distribution law reflects the distribution characteristics of accidents in the time and space dimensions. Extracting eigenvalue from the distribution law helps to determine the spatio-temporal classification of accidents. Time series analysis methods play an important role in constructing accident prediction models. For example, using the autoregressive integrated moving average (ARIMA) model, the accident occurrence probability in a future period can be predicted based on historical data. The setting of model parameters directly affects the prediction accuracy and needs to be adjusted according to the actual situation. The accident risk level in the future period calculated by the prediction model can be used to draw a trend chart, and this visual display can intuitively reflect the change trend of accident risk.

[0049] S102. Obtain real-time traffic flow data, use the sliding window algorithm to smooth the vehicle density and speed, and combine with a preset flow threshold to determine whether the current traffic condition is abnormal.

[0050] Specifically, obtain real-time traffic flow data and extract vehicle density and speed information. Use the sliding window algorithm to smooth the vehicle density and speed to obtain the smoothed density value and speed value. Calculate the current traffic flow value according to the smoothed density value and speed value. Compare the current traffic flow value with the preset flow threshold to determine whether the traffic condition is abnormal. If the traffic condition is abnormal, generate an abnormal warning signal and record the time and location of the abnormality. Use the time series analysis method to construct a traffic flow prediction model and set model parameters. Calculate the traffic flow value in the future period through the prediction model and output a traffic flow trend chart.

[0051] Implementable, the real-time traffic flow monitoring system evaluates road conditions by obtaining vehicle density and speed information. Taking a main road in Beijing as an example, during the morning rush hour, the system collects data once a minute. The original data may fluctuate. For example, in one minute, the vehicle density is 80 vehicles per kilometer and the speed is 40 kilometers per hour. In the next minute, it may suddenly change to 90 vehicles per kilometer and 35 kilometers per hour. To eliminate the impact of such short-term fluctuations, a sliding window algorithm is used for smoothing. Assuming a 5-minute window is set, then each calculation will consider the data of the most recent 5 minutes. The smoothed density value obtained in this way may be 85 vehicles per kilometer, and the speed value is 38 kilometers per hour, which can better reflect the real traffic conditions. Based on the smoothed data, the current traffic flow value can be calculated. A simple method is to multiply the density by the speed to get the number of vehicles passing through a certain point per hour. In this example, the flow value is 85×38 = 3230 vehicles per hour. This value is compared with the preset flow threshold to determine whether the traffic is abnormal. Assuming the normal flow threshold for this section is 3000 vehicles per hour, then the current situation is determined to be abnormal. Once an abnormality is detected, the system will generate a warning signal. Record the time when the abnormality occurred (such as 8:30 am on June 22, 2024) and the location (such as a certain section in Chaoyang District, Beijing). This information can be used for subsequent traffic management and diversion. To predict future traffic conditions, the system uses time series analysis methods to build a prediction model. For example, an autoregressive moving average (ARMA) model can be used. This model considers the trends and periodicities of historical data and can accurately predict short-term traffic flow changes. By setting appropriate model parameters, such as the number of autoregressive terms and the number of moving average terms, the prediction accuracy can be improved. Using the built model, the system can predict the traffic flow in the next few hours. For example, the prediction results show that in the next two hours, the traffic flow will gradually drop to about 2800 vehicles per hour. Such prediction results can be plotted into a trend chart to visually display the change trend of traffic flow and provide decision-making support for traffic management departments. It not only realizes the real-time monitoring and early warning of traffic conditions, but also provides forward-looking guidance for traffic management through the prediction function. It can help detect traffic abnormalities in a timely manner, prevent congestion from forming, and improve road traffic efficiency. In addition, the long-term accumulated data can also be used for deeper traffic governance work such as optimizing traffic signal timing and planning road reconstruction.

[0052] S103. Extract visibility, precipitation, and temperature information from meteorological data, and use interpolation algorithms to fill the data blind spots in sparse areas of meteorological stations to generate a local meteorological state distribution map.

[0053] Specifically, three sets of data, namely visibility, precipitation, and temperature values, are extracted from meteorological data and classified and stored according to the location information of meteorological stations. An interpolation algorithm is used to calculate the data blind spots in sparse areas of meteorological stations and fill in the missing data of visibility, precipitation, and temperature values. The filled visibility, precipitation, and temperature values are smoothed to obtain three sets of smoothed values. Based on the smoothed visibility, precipitation, and temperature values, the meteorological state values of local areas are calculated. If the meteorological state value exceeds the preset threshold range, it is determined that the meteorological state in this area is abnormal, and abnormal identification information is generated. A spatial analysis method is adopted to construct a meteorological state distribution map, and the abnormal identification information is marked on the distribution map. The meteorological state trend of the local area is output through the distribution map, and future meteorological changes are predicted in combination with time series.

[0054] Implementable, meteorological data processing is crucial for accurate weather prediction and analysis. The raw data obtained from meteorological stations contains key information such as visibility, precipitation, and temperature. For data blind spots, interpolation algorithms can effectively fill in the missing data. For example, in the area between two adjacent meteorological stations, the inverse distance weighted method (IDW) can be used for interpolation. Data smoothing helps to eliminate short-term fluctuations and highlight long-term trends. This processing can effectively reduce the influence of outliers and improve prediction accuracy. The calculation of local meteorological state values usually takes into account multiple factors. For example, a simple meteorological index can be defined: MI = V + (100 - P) + (25 - |T - 25|), where V is visibility (km), P is precipitation (mm), and T is temperature (°C). If the smoothed data in a certain area is: visibility 10 km, precipitation 5 mm, and temperature 28 °C, then MI = 10 + 95 + 22 = 127. When setting the threshold to judge the abnormal state, the local climate characteristics need to be considered. The construction of the meteorological state distribution map can use the contour method. The city is divided into grids, and each grid is assigned the interpolated MI value. Connect the points with the same MI value to form contours. Abnormal areas can be marked with special colors. This visualization method intuitively shows the spatial distribution of the meteorological state. Combining time series analysis, future meteorological changes can be predicted.

[0055] S104. Spatially align the traffic flow data with the meteorological state distribution map, and use a convolutional neural network to fuse the two sets of data to generate a comprehensive road section hazard index.

[0056] Specifically, traffic flow data and a meteorological status distribution map are obtained, and the two are spatially aligned using a geographic information system to ensure the consistency of the data in the spatial dimension. A convolutional neural network model is used to construct a fusion network for the traffic flow data and the meteorological status distribution map, and the aligned data is input. In the convolutional neural network model, the features of the traffic flow data and the features of the meteorological status distribution map are extracted to generate feature vectors. The traffic flow feature vector and the meteorological status feature vector are weighted and fused to obtain a comprehensive feature vector. According to the comprehensive feature vector, the comprehensive risk index of each road section is calculated through a fully connected layer. If the comprehensive risk index of a road section exceeds a preset threshold, the road section is determined to be a high-risk road section, and high-risk road section identification information is generated. The high-risk road section identification information is marked on the traffic map to form a high-risk road section distribution map, and the risk trend data of the road section is output.

[0057] Feasible. Obtaining traffic flow data and meteorological state distribution maps is the basis for fusion analysis. Traffic flow data usually includes indicators such as vehicle quantity, speed, and density, which can be collected through road sensors, cameras, or mobile devices. The meteorological state distribution map covers information such as temperature, precipitation, and visibility, and is generated by a network of meteorological stations. To ensure the consistency of these two types of data in the spatial dimension, it is necessary to use a geographic information system for spatial alignment. For example, matching the collection points of traffic flow data with the grid cells of the meteorological state distribution map so that each road section can correspond to the corresponding meteorological state. The convolutional neural network model is the key to realizing the fusion of traffic flow and meteorological state. This model can effectively extract spatial features and is suitable for processing data with geographical distribution characteristics. In the network structure, parallel convolutional layers can be designed to process traffic flow and meteorological state data respectively. For example, for traffic flow, 1x1 convolution can be used to extract the vehicle flow density feature; for meteorological state, 3x3 convolution can be used to capture local weather patterns. In this way, the network can learn the spatial correlation between traffic and meteorology at the same time. The weighted fusion of feature vectors is the core step in generating comprehensive features. An attention mechanism can be used to dynamically adjust the weights of different features. For example, in rainy days, meteorological features may obtain higher weights because precipitation has a greater impact on traffic safety. In this way, the model can flexibly fuse features according to specific situations and improve the prediction accuracy. The calculation of the comprehensive risk index needs to consider multiple factors. In addition to traffic flow and meteorological state, information such as road conditions and historical accident data can also be introduced. For example, for a mountain road, even if the current weather is good, due to its characteristics of many sharp turns and large slopes, it may be given a high risk index. Through the fully connected layer, the model can learn the complex relationships between these factors and thus obtain a more accurate risk assessment. The identification and visualization of high-risk road sections are the final outputs. Dynamic thresholds can be set and adaptively adjusted according to the characteristics of different time periods and regions. For example, during the morning and evening rush hours, the threshold may need to be increased accordingly. When marking high-risk road sections on the traffic map, different colors can be used to represent the levels of danger, such as red for extremely dangerous and yellow for moderately dangerous, etc. This intuitive display method helps traffic management departments quickly identify problem areas and take corresponding traffic control measures in a timely manner. Through this fusion analysis method, an accurate assessment and prediction of the safety status of road sections can be achieved. In the long run, this analysis can also provide a basis for road planning and renovation, thereby fundamentally improving traffic safety levels.

[0058] S105. Align the accident risk trend and the comprehensive risk index of the road section in time, and use a weighted fusion algorithm to calculate the real-time risk score of the current road section.

[0059] Specifically, a time alignment algorithm is used to align the accident risk degree trend with the comprehensive road section risk degree index in the time dimension to generate time-aligned data. According to the time-aligned data, a weighted fusion algorithm is used to calculate the real-time risk degree score of the current road section. If the real-time risk degree score exceeds the preset threshold, high-risk road section identification information is generated, and the high-risk road section identification information is marked on the traffic map to form a high-risk road section distribution map.

[0060] Implementable, the time alignment algorithm can adopt the dynamic time warping (DTW) technology to match the historical accident data with the current road conditions. This can discover potential dangerous patterns, such as a certain road section having frequent accidents during a specific time period. The calculation of the real-time risk degree score needs to comprehensively consider various factors. Different weights can be set, such as the current road conditions accounting for 40%, the historical accident trend accounting for 30%, and the weather conditions accounting for 30%. This flexible weight allocation method enables the system to adapt to the characteristics of different regions. The visualization of high-risk road sections is crucial for reminding drivers and traffic management departments. Different colors can be used to mark the risk levels on the map, such as green indicating safety, yellow indicating medium risk, and red indicating high risk. At the same time, a pop-up window can be added to display detailed information, such as the current traffic flow, weather conditions, etc., to help users better understand the causes of danger.

[0061] S106. According to the real-time risk degree score, combined with the preset speed limit adjustment rules, a linear interpolation algorithm is used to smoothly adjust the speed limit value to avoid the speed limit change being too fast and affecting driving behavior.

[0062] Specifically, obtain the real-time risk degree score of the current road section and read the preset speed limit adjustment rules. According to the speed limit adjustment rules, use a linear interpolation algorithm to calculate the speed limit adjustment range of the current road section. Combine the calculated speed limit adjustment range with the current speed limit value to generate a smoothed new speed limit value. Determine whether the new speed limit value exceeds the preset speed limit range. If it exceeds, adjust the speed limit value to the boundary value of the preset range. Update the adjusted speed limit value to the road section speed limit database to form a speed limit update record. According to the speed limit update record, generate speed limit change trend data for subsequent analysis. Combine the speed limit change trend data with the driving behavior data to evaluate the impact degree of the speed limit adjustment on driving behavior.

[0063] Implementable, real-time risk assessment is a key indicator for road safety management, which comprehensively considers various factors such as traffic flow, meteorological conditions, and accident history. For example, the real-time risk assessment score of a certain highway section may rise from 65 points in normal times to 85 points on a rainy day. At this time, corresponding speed limit measures need to be taken. The speed limit adjustment rules are formulated based on a large amount of historical data and expert experience, and usually include the speed limit adjustment ranges corresponding to multiple risk levels. For example, when the risk level is between 75 and 85, the speed limit is reduced by 10%; when the risk level is between 85 and 95, the speed limit is reduced by 20%. These rules ensure the scientificity and consistency of speed limit adjustment. The linear interpolation algorithm plays an important role in calculating the specific speed limit adjustment range. Suppose the current risk level of a certain road section is 80 points, which is in the range of 75 - 85 points, corresponding to a speed limit adjustment range of 10% - 20%. Through linear interpolation, it can be calculated that the speed limit should be reduced by 15% in this case. This method makes the speed limit adjustment smoother and avoids potential safety hazards caused by sudden changes. The generation of the new speed limit value needs to consider the current speed limit. If the original speed limit is 120 km / h and it is reduced by 15%, the new speed limit is 102 km / h. This progressive adjustment can effectively reduce risks without causing significant fluctuations in traffic flow. The preset speed limit range is an important mechanism to ensure that the adjusted speed limit remains within a reasonable range. For example, the minimum speed limit on a highway may be set at 60 km / h. If the calculated new speed limit is lower than this value, it will be automatically adjusted to 60 km / h to maintain basic traffic efficiency. The generation and storage of speed limit update records are crucial for the traceability and subsequent analysis of the system. Each record may include information such as the road section ID, the speed limit values before and after adjustment, and the adjustment time. These data form the basis for analyzing the trend of speed limit changes. The speed limit change trend data can reveal some valuable patterns. For example, a certain road section often needs to reduce the speed limit during the rush hour every evening, which may imply that there is room for optimizing traffic organization during this period. Another example is that a certain road section frequently adjusts the speed limit during the rainy season, indicating that the drainage facilities may need to be improved. Combining the speed limit change trend with driving behavior data can evaluate the actual effect of speed limit adjustment. For example, by analyzing the changes in the actual driving speeds of vehicles, it can be known whether drivers comply with the new speed limit. If it is found that most vehicles are still speeding, the rationality of the speed limit value may need to be re-evaluated. This dynamic speed limit adjustment system can not only improve road safety but also optimize traffic flow, reduce congestion and accident rates. Through continuous data collection and analysis, the system can continuously improve itself and provide strong support for intelligent traffic management.

[0064] S107. By comparing the vehicle speed with the adjusted speed limit value, determine whether there is speeding behavior, generate a warning message and push it to the dynamic speed limit system in real time.

[0065] Specifically, obtain the real-time monitored vehicle speed data and store it in the information repository. At the same time, extract the adjusted speed limit value of the current section from the dynamic speed limit system. Use a comparative analysis algorithm to compare the vehicle speed data with the adjusted speed limit value in real time to generate a speed difference. Determine whether the speed difference exceeds the speed limit range according to the preset difference threshold. If it exceeds, it is judged as a speeding behavior. Generate a warning message for the speeding behavior, and use an information generation rule to integrate the speeding time, location, and speed difference into the warning message. Send the warning message to the dynamic speed limit system in real time through the preset push logic to trigger the warning response mechanism. Record the storage location and timestamp of the warning message in the dynamic speed limit system to form a warning record repository. Generate a driving behavior analysis report based on the warning record repository and vehicle speed data for subsequent optimization of the speed limit adjustment rule.

[0066] Implementable. The core of the dynamic speed limit system lies in real-time monitoring and analysis of vehicle speed data and comparison with the adjusted speed limit value. This process begins with obtaining real-time vehicle speed data, which is collected through sensors installed on the road or in-vehicle devices. This data is stored in a dedicated information repository, providing the basis for subsequent analysis. At the same time, the system extracts the latest speed limit value for the current section from the dynamic speed limit database. This speed limit value is the result of dynamic adjustment based on factors such as road conditions and weather. The comparison and analysis algorithm is the key to achieving speed control. It compares the actual vehicle speed with the speed limit value in real time and calculates the speed difference. For example, if a vehicle is traveling at 95 km / h on a section with a speed limit of 80 km / h, the system will calculate a speed difference of 15 km / h. This difference is compared with a preset threshold, which may be set at 10 km / h. In this example, the vehicle is speeding at 15 km / h, exceeding the threshold, and is thus determined to be speeding. For speeding behavior, the system generates warning information. The generation of warning information follows specific rules and includes key information such as the time of speeding, location, and speed difference. For example, a warning message may be: "At 14:30 on June 22, 2024, at the 2536 km mark of the Beijing-Hong Kong-Macao Expressway, the vehicle with license plate number Jing A12345 is speeding by 15 km / h". Such detailed information is helpful for subsequent analysis work. After the warning information is generated, the system sends it to the control center of the dynamic speed limit system through the preset push logic. This process may involve various communication technologies, such as 4G / 5G networks or dedicated traffic management networks. The timeliness of the push is crucial for effective traffic management, allowing managers to respond quickly, such as adjusting speed limit signs or notifying traffic police for on-site management. The system also stores the warning information in the warning record repository, including the storage location and timestamp. These records are not only used for immediate traffic management but also provide valuable data for long-term traffic behavior analysis. By analyzing this data, high-incidence speeding sections or periods can be identified, providing a basis for traffic management departments to formulate more targeted management measures. Finally, the system generates a driving behavior analysis report based on the warning record repository and vehicle speed data. This report may include multi-dimensional analysis, such as the speeding frequency at different times and the speeding trends of different vehicle types. For example, the report may show that trucks have a higher speeding frequency during early morning hours, providing a direction for formulating targeted management measures. These analysis results are not only used to optimize speed limit adjustment rules but also to guide traffic safety education and road design improvements. Through this series of data collection, analysis, and feedback mechanisms, the dynamic speed limit system continuously optimizes its operation strategy, improving road safety and traffic efficiency. This intelligent traffic management method represents the development direction of future traffic systems, combining big data analysis with real-time control, and contributing to a safer and more efficient road traffic environment.

[0067] Embodiment 2

[0068] This embodiment also provides a computer device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the method.

[0069] Embodiment Three

[0070] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0071] Embodiment Four

[0072] This embodiment also provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the method are implemented.

[0073] The above are only the preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for dynamic speed limit on highways based on multi-source data, characterized in that: The following steps are involved: Obtain accident data, build an accident risk prediction model based on time series, and generate an accident risk trend chart; Obtain real-time traffic flow data, build a traffic flow prediction model, and generate a traffic flow trend chart; Obtain meteorological data and generate a local meteorological status distribution map; Spatial alignment and data fusion of traffic flow trend map and local meteorological status distribution map are performed to generate comprehensive danger index of road section; The accident risk trend graph is time-aligned with the comprehensive risk index of the road section, and a weighted fusion algorithm is used to generate a real-time risk score for the current road section; Based on the real-time hazard score and the preset speed limit adjustment rules, the speed limit value is smoothly adjusted using a linear interpolation algorithm; Compare the current vehicle speed with the adjusted speed limit value, and issue a warning if speeding occurs.

2. The method according to claim 1, characterized in that The process of obtaining accident data, building an accident risk prediction model based on time series, and generating accident risk trends includes: The original fields of accident type, time point and place name are obtained from the accident data to construct a spatiotemporal distribution matrix; based on the spatiotemporal distribution matrix, the accident distribution law is calculated, characteristic values ​​are extracted according to the accident distribution law, and the spatiotemporal distribution of the accident is determined; based on the spatiotemporal distribution of the accident and in combination with the time series analysis method, an accident hazard prediction model is constructed; based on the accident hazard prediction model, a hazard prediction value is obtained, and an accident hazard trend is generated according to the hazard prediction value.

3. The method according to claim 1, characterized in that The process of obtaining real-time traffic flow data, building a traffic flow prediction model, and generating a traffic flow trend chart includes: Acquire real-time traffic flow data and extract vehicle density and speed information; use sliding window algorithm to smooth vehicle density and speed to obtain smoothed vehicle density and speed values; build a traffic flow prediction model based on the smoothed vehicle density and speed values ​​combined with time series analysis method; calculate the traffic flow value of future time period through the traffic flow prediction model and output the traffic flow trend chart.

4. The method according to claim 1, characterized in that: The process of obtaining meteorological data and generating a local meteorological status distribution map includes: Three groups of data, namely visibility, precipitation and temperature, are extracted from meteorological data and classified and stored according to the location information of meteorological stations. Interpolation algorithm is used to calculate the data blind spots in sparse areas of meteorological stations to fill in the missing data of visibility, precipitation and temperature values. The filled visibility, precipitation and temperature values ​​are smoothed to obtain three groups of smoothed values. The meteorological status of the local area is calculated based on the smoothed visibility, precipitation and temperature values. If the meteorological status value exceeds the preset threshold range, the meteorological status of the area is judged to be abnormal and abnormal identification information is generated. Combined with spatial analysis methods, a meteorological status distribution map is constructed.

5. The method according to claim 1, characterized in that The process of spatially aligning and fusing the traffic flow trend map with the local meteorological status distribution map to generate the comprehensive danger index of the road section includes: Obtain the traffic flow trend map and the local meteorological status distribution map, and use the geographic information system to spatially align the two; input the spatially aligned data into the convolutional neural network model, extract the characteristics of the traffic flow data and the characteristics of the meteorological status distribution map, and generate a feature vector; perform weighted fusion of the traffic flow feature vector and the meteorological status feature vector to obtain a comprehensive feature vector; based on the comprehensive feature vector, calculate the comprehensive hazard index of each road section through the fully connected layer.

6. The method according to claim 1, characterized in that Based on the real-time hazard score and the preset speed limit adjustment rules, the process of smoothly adjusting the speed limit value using a linear interpolation algorithm includes: Obtain the real-time hazard score of the current road section, and calculate the speed limit adjustment range of the current road section based on the pre-set speed limit adjustment rules using a linear interpolation algorithm; combine the calculated speed limit adjustment range with the current speed limit value to generate a smoothed new speed limit value; determine whether the new speed limit value exceeds the preset speed limit range, and if so, adjust the speed limit value to the preset range boundary value; and update the adjusted speed limit value to the road section speed limit database.

7. The method according to claim 1, characterized in that Compare the current vehicle speed with the adjusted speed limit value. If there is speeding, the warning process includes: A comparative analysis algorithm is used to compare the vehicle speed data with the adjusted speed limit value in real time to generate a speed difference; a preset difference threshold is used to determine whether the speed difference exceeds the speed limit range, and if so, it is determined to be speeding; warning information is generated for speeding behavior, and a warning response mechanism is triggered based on the warning information, and the storage location and timestamp of the warning information are recorded to form a warning record library; a driving behavior analysis report is generated based on the warning record library and vehicle speed data.

8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images