Real-time traffic management system and method based on big data acquisition

Abstract

The invention discloses a real-time traffic management system and method based on big data acquisition, and relates to the technical field of real-time traffic management, quality evaluation is performed on acquired traffic data, and traffic data conforming to a quality evaluation rule is screened out as effective data of a key digital table; generating a traffic judgment threshold value of each road section by using the historical congestion characteristics, the historical accident characteristics, the historical unobstructed characteristics and the road static characteristics, and further judging the traffic state of each road section; providing an optimized path for the floating car according to the traffic state judgment result of each road section; the implementation effect after the floating car receives the optimized path is fed back to the key number table in real time, then changes of key numbers before and after the floating car receives the optimized path are compared, and a management strategy is adjusted according to the comparison result of the changes of the key numbers. The traffic state real-time evaluation, the dynamic path optimization and the closed-loop adaptive adjustment of the management strategy are realized, and the traffic efficiency and the resource utilization rate are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of real-time traffic management, and particularly to a real-time traffic management system and method based on big data collection. Background Art

[0002] With the acceleration of the urbanization process and the continuous increase in the number of motor vehicles, the problem of urban traffic congestion has become increasingly serious, bringing many inconveniences to people's travel and the development of cities. Traditional traffic management methods and technical means are difficult to meet the needs of modern traffic management. The development of big data technology provides new ideas and methods for traffic management. By collecting a large amount of traffic data and using advanced data analysis and processing technologies, the traffic conditions can be understood more comprehensively and accurately, and the optimal allocation of traffic resources and the intelligence of traffic management can be realized.

[0003] However, the existing traffic management systems lack an effective data quality assessment mechanism, resulting in the collected data having noise, errors or being incomplete, which affects the accuracy of subsequent analysis and decision-making. And only relying on a single indicator or a simple model to judge the traffic state, unable to comprehensively consider various factors such as historical data and road static characteristics, resulting in inaccurate and detailed judgment of the traffic state; the path planning of some existing systems does not fully combine the real-time traffic state and road segment association information, unable to provide the optimal driving path for vehicles, and easily causing waste of traffic resources and aggravation of congestion.

[0004] Therefore, in view of the above problems, there is an urgent need for a real-time traffic management system and method based on big data collection. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a real-time traffic management system and method based on big data collection, which solves the problems of the traditional traffic management system having a lag in response to real-time road conditions and inaccurate path planning.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A real-time traffic management system based on big data collection, including a data screening module, a traffic status evaluation module, an optimized route generation module, and a feedback adjustment module, wherein: The data screening module is used to perform quality evaluation on the collected traffic data based on a preset key digital table, screen out the traffic data that meets the quality evaluation rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; The traffic status evaluation module is used to extract the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics of each section in the key digital table, generate the traffic determination threshold of each section by using the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics, and then combine the traffic congestion index predicted by the multimodal traffic prediction model and the traffic determination threshold of the corresponding section to judge the traffic status of each section; The optimized route generation module is used to provide an optimized route for the floating vehicle according to the traffic status judgment result of each section, in combination with section association simulation and the target address of the floating vehicle; The feedback adjustment module is used to real-time feedback the implementation effect after the floating vehicle receives the optimized route to the key digital table, and then compare the changes in the key numbers before and after the floating vehicle receives the optimized route, and adjust the management strategy according to the comparison result of the key number changes.

[0007] Further, the data screening module is specifically analyzed as follows: Extract the traffic data quality characteristics of each traffic data collection time node, and the traffic data quality characteristics include data integrity, data timeliness, data relevance, and data consistency; Use logistic regression to construct a traffic data quality binary classification model, take the traffic data quality characteristics as the input of the traffic data quality binary classification model, and output the probability that the traffic data is valid data; Set the traffic data quality evaluation threshold for each section based on traffic scenario semantic understanding and traffic road type; Compare the output results of the traffic data quality binary classification model of each traffic data collection time node of each section with the traffic data quality evaluation threshold of the corresponding section respectively. When the output result of the traffic data quality binary classification model exceeds the traffic data quality evaluation threshold, mark the traffic data collected at the traffic data collection time node of this section as the valid data of the key digital table.

[0008] Further, the specific analysis of setting the traffic data quality assessment threshold for each section based on traffic scene semantic understanding and traffic road type is as follows: Set the initial traffic data quality assessment threshold based on the confusion matrix simulation of historical traffic data quality marked samples; Identify the scene complexity and weather severity based on traffic scene semantic understanding, and obtain the section importance score based on the traffic road type; Determine the data quality adjustment requirement value for each section by combining the scene complexity, weather severity, and section importance score, and then adjust the initial traffic data quality assessment threshold using the data quality adjustment requirement value for each section to obtain the traffic data quality assessment threshold for each section.

[0009] Further, the specific analysis of the traffic state assessment module is as follows: The traffic determination threshold specifically includes a congestion determination threshold and a smooth determination threshold; Obtain multi-modal traffic features, which specifically include weather features, road air features, peak state features, and holiday state features; Use the weather features, road air features, peak state features, and holiday state features as the input of the multi-modal traffic prediction model, and output the predicted section congestion index; Compare the section congestion index with the congestion determination threshold and the smooth determination threshold. When the section congestion index is greater than the congestion determination threshold, the section is initially in a congested state. When the section congestion index is less than the smooth determination threshold, the section is initially in a smooth state. When the section congestion index is between the congestion determination threshold and the smooth determination threshold, the section is initially in a slow-moving state.

[0010] Further, the specific acquisition method of the traffic determination threshold is as follows: The historical congestion features include congestion duration and minimum congestion speed, the historical accident features include accident impact range and secondary congestion probability, the historical smooth feature is the non-peak average speed, and the road static features include the number of lanes and speed limit value; Use the regression model with the historical congestion features and historical accident features as the input, and output the congestion determination threshold; Use the regression model with the historical smooth feature and road static features as the input, and output the smooth determination threshold.

[0011] Furthermore, the optimized path generation module is specifically analyzed as follows: summarize the traffic status judgment results of each road section to construct a traffic status map. The specific traffic status map is indexed by road section ID, marking the current traffic status of each road section. At the same time, establish a road section association matrix. The nodes in the road section association matrix are road section IDs, and the edges are the connection relationships between road sections. And assign corresponding weights to the edges based on the convenience of passing between road sections and the length of the distance; obtain the current position of the floating car in real time, and obtain the target address information of the floating car based on the vehicle navigation input information. Taking the current position of the floating car as the starting point and the target address as the end point, generate an initial path using the traffic status map and the road section association matrix; traverse each road section on the initial path, obtain the predicted road section congestion index of each road section, and take the average value of the predicted road section congestion indices of each road section as the congestion score of the initial path. Then feedback the initial path with the lowest congestion score to the floating car driver.

[0012] Furthermore, the feedback adjustment module is specifically analyzed as follows: set a feedback adjustment period, and then retrieve the implementation effect of the floating car after receiving the optimized path within the feedback adjustment period. The real-time effect is specifically the traffic efficiency of each road section; obtain the change amount of the traffic efficiency of each road section in the key table before and after the implementation of the optimized path, and then take the average value of the change amounts of the traffic efficiency of each road section as the effect evaluation value. Compare the effect evaluation value with the effect expected index. When the effect evaluation value is greater than or equal to the effect expected index, there is no need to trigger the management strategy adjustment mechanism; when the effect evaluation value is less than the effect expected index, trigger the management strategy adjustment mechanism. The management strategy adjustment includes increasing the promotion of the optimized path and optimizing and adjusting relevant parameters.

[0013] A real-time traffic management method based on big data collection, applying the above real-time traffic management system based on big data collection, includes the following steps: Step S1, perform quality evaluation on the collected traffic data based on a preset key digital table, screen out the traffic data that meets the quality evaluation rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; Step S2, extract the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics and road static characteristics of each road section in the key digital table, generate the traffic determination threshold of each road section using the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics and road static characteristics, and then combine the road section congestion index predicted by the multi-modal traffic prediction model and the traffic determination threshold of the corresponding road section to judge the traffic status of each road section; Step S3, according to the traffic status judgment results of each road section, combined with the road section association simulation and the floating car target address, provide an optimized path for the floating car; Step S4, feedback the implementation effect of the floating car after receiving the optimized path to the key digital table in real time, and then compare the changes in the key numbers before and after the floating car receives the optimized path, and adjust the management strategy according to the comparison results of the key number changes.

[0014] The present invention has the following beneficial effects:

[0015] The real-time traffic management system and method based on big data collection evaluates the quality of the original traffic data through a preset key digital table, only retains the valid data that conforms to the rules, ensures the reliability and accuracy of subsequent analysis, and classifies and stores the screened valid data into the key digital table in real time to form a dynamically updated data pool, providing a high-precision data basis for traffic state evaluation; comprehensively considering historical congestion, accidents, smoothness characteristics and road static characteristics, generates traffic determination thresholds for each road section, overcomes the limitations of single-characteristic evaluation, combines multi-modal traffic prediction models to predict the congestion index of road sections, and outputs the real-time traffic state after comparing with the determination threshold, improving the accuracy of evaluation; based on the real-time traffic state evaluation results, combines road section association simulation and floating vehicle target addresses to dynamically generate optimized paths, reducing travel time and fuel consumption; updates the key digital table in real time with the data after implementing the optimized path by floating vehicles, forming a closed loop of "path generation - effect feedback - strategy adjustment" to achieve the adaptive optimization of the traffic system; through real-time traffic state evaluation and path optimization, diverts vehicles on congested road sections in advance, reduces the overall congestion level, and improves the utilization rate of road resources.

[0016] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a structural diagram of a real-time traffic management system based on big data collection according to the present invention.

[0018] Figure 2 It is a flowchart of a real-time traffic management method based on big data collection according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The embodiments of the present application realize real-time traffic state evaluation, dynamic path optimization and closed-loop adaptive adjustment of management strategies through a real-time traffic management system and method based on big data collection, significantly improving traffic efficiency and resource utilization rate.

[0020] The overall idea of the embodiments of this application is as follows: Based on big data collection, the traffic data is quality-controlled through a data screening module, and the valid data is stored in a key digital table. Multiple features in the key digital table are used to generate traffic determination thresholds, and a multi-modal traffic prediction model is combined to accurately judge the traffic status of each section. An optimized path is planned for the vehicle based on the traffic status judgment result, section association simulation, and floating vehicle target address. The implementation effect of the optimized path for the floating vehicle is fed back to the key digital table, and the management strategy is adjusted by comparing the changes in the key numbers, forming a closed-loop real-time traffic management system from data collection, processing, analysis to decision execution and feedback optimization, so as to achieve the reasonable allocation of traffic resources and the dynamic optimization of traffic conditions.

[0021] Please refer to Figure 1 , an embodiment of the present invention provides a technical solution: A real-time traffic management system based on big data collection, including a data screening module, a traffic status evaluation module, an optimized path generation module, and a feedback adjustment module, wherein: The data screening module is used to perform quality evaluation on the collected traffic data based on a preset key digital table, screen out the traffic data that meets the quality evaluation rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; The traffic status evaluation module is used to extract the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics of each section in the key digital table, generate traffic determination thresholds for each section using the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics, and then combine the section congestion index predicted by the multi-modal traffic prediction model and the traffic determination thresholds of the corresponding sections to judge the traffic status of each section; The optimized path generation module is used to provide an optimized path for the floating vehicle according to the traffic status judgment result of each section, combined with section association simulation and floating vehicle target address; The feedback adjustment module is used to feed back the implementation effect of the floating vehicle after receiving the optimized path to the key digital table in real time, and then compare the changes in the key numbers before and after the floating vehicle receives the optimized path, and adjust the management strategy according to the comparison result of the changes in the key numbers.

[0022] Specifically, the data screening module is specifically analyzed as follows: extracting the traffic data quality characteristics of each traffic data collection time node, where the traffic data quality characteristics include data integrity, data timeliness, data relevance, and data consistency; using logistic regression to construct a traffic data quality binary classification model, taking the traffic data quality characteristics as the input of the traffic data quality binary classification model, and outputting the probability that the traffic data is valid data; setting the traffic data quality evaluation threshold for each road section based on traffic scene semantic understanding and traffic road type; respectively comparing the output results of the traffic data quality binary classification model of each traffic data collection time node of each road section with the traffic data quality evaluation threshold of the corresponding road section. When the output result of the traffic data quality binary classification model exceeds the traffic data quality evaluation threshold, the traffic data collected at this traffic data collection time node of this road section is marked as valid data in the key digital table.

[0023] The specific analysis of setting the traffic data quality evaluation threshold for each road section based on traffic scene semantic understanding and traffic road type is as follows: setting the initial traffic data quality evaluation threshold based on the confusion matrix simulation of historical traffic data quality marked samples; identifying the scene complexity and weather severity based on traffic scene semantic understanding, and obtaining the road section importance score based on the traffic road type; combining the scene complexity, weather severity, and road section importance score to determine the data quality adjustment requirement value for each road section, and then using the data quality adjustment requirement value of each road section to adjust the initial traffic data quality evaluation threshold to obtain the traffic data quality evaluation threshold for each road section.

[0024] In this implementation plan, data integrity refers to the degree of no missing values in traffic data, reflecting the comprehensiveness of the data, which is obtained by counting the number of missing values in the data set and comparing the amount of data to be collected with the actual amount of data collected; data timeliness refers to the time interval from data collection to availability, reflecting the timeliness of the data, which is obtained by recording the data collection time and the time when it enters the system for processing and calculating the difference between the two; data relevance refers to the tightness of the logical relationship between different traffic data, reflecting the rationality of the data, which is identified by analyzing the correlation between different data variables, such as the relationship between vehicle speed and traffic flow, occupancy rate, etc., and measured using statistical indicators such as the Pearson correlation coefficient, with a value range of [-1, 1]. The closer the absolute value is to 1, the stronger the relevance; data consistency refers to the degree of consistency of the same data in different data sources or at different time points, reflecting the stability of the data, which is obtained by comparing the measured values of the same traffic parameter in different data sources or at different times, and calculating the difference between the measured values, such as the mean absolute error, mean square error, etc. The smaller the value, the better the consistency.

[0025] The specific steps to construct a binary classification model for traffic data quality using logistic regression are as follows: Collect historical traffic data containing data integrity, timeliness, relevance, and consistency features, as well as corresponding data quality labels (valid or invalid), and preprocess the data, such as filling missing values, handling outliers, and data standardization; Analyze the correlation between each feature and data quality, and select features with strong correlation as model inputs; Divide the dataset into a training set and a test set, usually in a ratio of 7:3 or 8:2, and use the training set data to train the logistic regression model, and optimize the model parameters through iteration; Use the test set data to evaluate the model performance, and calculate metrics such as accuracy, recall rate, and F1 value; Adjust the model parameters or features according to the evaluation results to improve the model performance. The model coefficients and error terms in the binary classification model for traffic data quality are obtained using the maximum likelihood estimation method, specifically determined by maximizing the log-likelihood function, and iterative solutions can be performed using optimization algorithms such as gradient descent.

[0026] An example of the specific expression of the binary classification model for traffic data quality is as follows: Let the input traffic data quality feature vector be , corresponding to data integrity, timeliness, relevance, and consistency respectively, then , where represents the probability that the traffic data is valid data, is the error term, are the model coefficients.

[0027] The specific steps to set the initial threshold for traffic data quality assessment based on the confusion matrix simulation of historical traffic data quality labeled samples are as follows: Prepare historical traffic data quality labeled samples, including true labels and model prediction probabilities; Set a series of different thresholds, compare the prediction probabilities with the thresholds to obtain classification results; For each threshold, calculate the four metrics of the confusion matrix: true positive (TP), false positive (FP), true negative (TN), and false negative (FN); Select a suitable threshold as the initial threshold according to business requirements and evaluation metrics (such as accuracy, recall rate, F1 value, etc.). The confusion matrix specifically includes: true positive (TP), the number of samples that are actually valid data and the model predicts as valid data; false positive (FP), the number of samples that are actually invalid data but the model predicts as valid data; true negative (TN), the number of samples that are actually invalid data and the model predicts as invalid data; false negative (FN), the number of samples that are actually valid data but the model predicts as invalid data.

[0028] The scene complexity refers to the complexity of the traffic scene, such as the number of intersections, lane changes, traffic flow fluctuations, etc. Scene information can be obtained through traffic surveillance video analysis, sensor data statistics, etc. The scene complexity can be divided into several levels, such as simple, medium, and complex, and assigned values of 1, 2, and 3 respectively; It can also be represented by continuous numerical values, which are obtained by weighted calculation of relevant indicators.

[0029] The severity of weather refers to the degree of impact of weather conditions on traffic, such as rainfall, snowfall, heavy fog, etc. Weather data, including weather type, precipitation, visibility, etc., is obtained from the meteorological department. A quantitative standard is formulated according to the weather type and relevant parameters. For example, a clear day is assigned a value of 0, light rain is assigned a value of 1, heavy rain is assigned a value of 2, etc.

[0030] The importance score of a road segment refers to the importance of the road segment in the traffic network, such as arterial roads, secondary arterial roads, branch roads, etc. It is evaluated according to factors such as road grade, traffic flow, and surrounding facilities. The importance of road segments can be divided into different levels and corresponding scores can be assigned. It is also possible to obtain a continuous score value through multi-index comprehensive evaluation.

[0031] An example of the specific calculation formula for determining the data quality adjustment requirement value for each road segment by combining the scene complexity, the severity of weather, and the importance score of the road segment is as follows: , where is the weight coefficient, and , represents the data quality adjustment requirement value, represents the scene complexity, represents the severity of weather, represents the importance score of the road segment.

[0032] Using the data quality adjustment requirement value for each road segment to adjust the initial threshold of traffic data quality assessment, an example of the specific calculation formula for obtaining the traffic data quality assessment threshold for each road segment is as follows: , where represents the traffic data quality assessment threshold, represents the initial threshold of traffic data quality assessment, represents the adjustment coefficient, which is used to control the influence degree of the data quality adjustment requirement value on the initial threshold. The value of

[0033] can be adjusted and optimized according to the actual situation. For example, a suitable value can be determined through experiments or experience, so that the adjusted threshold can better adapt to the characteristics and requirements of different road segments.

[0034] Specifically, the traffic status evaluation module is specifically analyzed as follows: The traffic determination thresholds specifically include a congestion determination threshold and a smooth determination threshold; multi-modal traffic features are obtained, and the multi-modal traffic features specifically include weather features, road air features, peak status features, and holiday status features; the weather features, road air features, peak status features, and holiday status features are used as inputs to a multi-modal traffic prediction model, and the predicted road congestion index is output; the road congestion index is compared with the congestion determination threshold and the smooth determination threshold. When the road congestion index is greater than the congestion determination threshold, the road is initially in a congested state. When the road congestion index is less than the smooth determination threshold, the road is initially in a smooth state. When the road congestion index is between the congestion determination threshold and the smooth determination threshold, the road is initially in a slow-moving state.

[0035] The specific method for obtaining the traffic determination thresholds is as follows: The historical congestion features include congestion duration and minimum congestion speed. The historical accident features include accident impact range and secondary congestion probability. The historical smoothness features are the non-peak average speed. The road static features include the number of lanes and speed limit values. The regression model is used with the historical congestion features and historical accident features as inputs to output the congestion determination threshold. The regression model is used with the historical smoothness features and road static features as inputs to output the smooth determination threshold.

[0036] In this implementation plan, the weather features include temperature, humidity, wind speed, wind direction, precipitation conditions (such as rain, snow, fog, etc.), light intensity, etc. Relevant data is obtained through devices such as weather stations, meteorological satellites, and roadside weather sensors. Temperature can be represented by specific numerical values, such as degrees Celsius. Humidity is represented by a percentage. Wind speed is represented in meters per second. Wind direction can be represented by an angle or a specific direction (such as north wind, south wind, etc.). Precipitation conditions can be represented by precipitation amount (millimeters), or in binary to indicate whether there is precipitation (0 means no, 1 means yes). Light intensity is represented in lux.

[0037] The road air features include the Air Quality Index (AQI) and pollutant concentrations (such as particulate matter concentration, carbon monoxide concentration, sulfur dioxide concentration, etc.). Data is collected through devices such as air quality monitoring stations and roadside air sensors. The Air Quality Index is a comprehensive value calculated according to certain standards based on different pollutant concentrations. Pollutant concentrations are represented in specific chemical substance concentration units, such as micrograms per cubic meter.

[0038] Peak state features include the time periods of the morning rush hour and the evening rush hour, as well as the traffic flow and pedestrian flow during these periods. The traffic flow is counted through traffic flow monitoring devices (such as cameras, geomagnetic sensors, etc.), and the pedestrian flow is counted through pedestrian flow monitoring devices (such as infrared sensors, video analysis, etc.). The peak time periods are determined based on historical data and actual monitoring. The morning rush hour and the evening rush hour can be represented by specific time intervals, such as the morning rush hour from 7:00 - 9:00, and the traffic flow and pedestrian flow are represented by specific quantities, such as the number of vehicles and the number of people passing through per hour.

[0039] Holiday state features include whether it is a holiday and the type of holiday (such as legal holidays, weekends, etc.). The holidays are determined based on calendar information and relevant government regulations. The type of holiday can be recorded in a database or system and can be represented in binary (0 represents non-holiday, 1 represents holiday). The type of holiday can be represented by a numerical code, such as 1 representing a legal holiday and 2 representing a weekend, etc.

[0040] The construction method of the multimodal traffic prediction model is as follows: A deep learning model is adopted, such as a combination of a convolutional neural network (CNN) and a recurrent neural network (RNN). First, the CNN is used to extract image or spatial-related features such as weather features and road air features, and then the RNN is used to process time series data such as peak state features and holiday state features. Finally, the results of the two are fused and predicted through a fully connected layer. An example of the expression of the multimodal traffic prediction model is: Assume that the input weather feature is , the road air feature is , the peak state feature is , the holiday state feature is , the parameters of the model are , then the expression of the model can be represented as , where is the predicted congestion index of the road section, is the mapping function of the model.

[0041] The model parameters are obtained by training the model with a large amount of historical data, and optimization algorithms (such as stochastic gradient descent, Adagrad, etc.) are used to adjust the model to minimize the error between the predicted result of the model and the actual congestion index of the road section.

[0042] The congestion duration refers to the continuous time during which the road section is in a congested state. The traffic flow, vehicle speed, etc. of the road section are recorded through traffic flow monitoring devices. When the vehicle speed is lower than a certain threshold and lasts for a certain period of time, it is determined to be in a congested state, and thus the congestion duration is counted and represented in hours or minutes.

[0043] The minimum congestion speed represents the lowest speed at which vehicles travel on a road section in a congested state. From the vehicle speed data obtained by traffic flow monitoring devices, the minimum speed value is found during congested periods and is expressed in kilometers per hour.

[0044] The accident impact range represents the length of the road or the area affected by a traffic accident after the accident occurs. The location of the accident and the affected road section range are determined through traffic accident reports, on-site investigation records, traffic monitoring videos, etc. The road length is expressed in meters or kilometers, or the area range is expressed by the number of lanes or intersections affected by the accident.

[0045] The probability of secondary congestion represents the likelihood of subsequent traffic congestion after a traffic accident occurs. Based on historical accident data, the ratio of the number of times secondary congestion occurs to the total number of accidents in similar accident situations is statistically calculated as an estimated value of the probability of secondary congestion, and the probability is expressed as a value between 0 and 1.

[0046] The non-peak average speed represents the average speed at which vehicles travel on a road section during non-peak hours. The vehicle speed data during non-peak hours is obtained through traffic flow monitoring devices, and the non-peak average speed is calculated by taking the average value and is expressed in kilometers per hour.

[0047] The number of lanes represents the number of lanes available for vehicles to travel on a road section, which is obtained through road design drawings, on-site investigation, or Geographic Information System (GIS) data; the speed limit value represents the maximum speed limit for vehicles on a road section, which is obtained through regulations of the traffic management department, road signs, or relevant database queries.

[0048] An example of the congestion determination threshold calculation formula is: Assume historical congestion characteristics , where is the congestion duration, is the minimum congestion speed, and historical accident characteristics , where is the accident impact range, is the probability of secondary congestion, and the congestion determination threshold is: , and the regression coefficient is estimated through training data, is the error term.

[0049] An example of the smooth determination threshold calculation formula is: Assume historical smooth characteristics (non-peak average speed), and road static characteristics , where is the number of lanes, is the speed limit value, then the smooth determination threshold is: , and the regression coefficient is estimated through training data , is the error term.

[0050] The regression coefficient and error term are fitted to historical data using methods such as the least squares method, and are estimated by solving the minimum value of the objective function so that the sum of squared errors between the model's predicted value and the true value is minimized. In actual calculations, statistical software or related functions in the machine learning library can be used to estimate the regression coefficient.

[0051] Taking into account a variety of factors to evaluate traffic conditions can more comprehensively and accurately reflect the actual traffic conditions on a road section, provide a more precise basis for traffic management, help formulate more reasonable traffic management strategies, improve traffic operation efficiency, and reduce congestion and accidents.

[0052] Specifically, the optimization path generation module is specifically analyzed as follows: the traffic status judgment results of each road section are summarized to construct a traffic status map. The specific traffic status map uses the road section ID as the index, marks the current traffic status of each road section, and establishes a road section association matrix. The nodes in the road section association matrix are the road section IDs, and the edges are the connection relationships between road sections. The edges are given corresponding weights based on the convenience and distance of travel between road sections; the current position of the floating vehicle is obtained in real time, and the target address information of the floating vehicle is obtained based on the vehicle navigation input information. The current position of the floating vehicle is used as the starting point and the target address is used as the end point. The initial path is generated using the traffic status map and the road section association matrix; each road section on the initial path is traversed to obtain the predicted road section congestion index of each road section, and the average of the predicted road section congestion index of each road section is taken as the congestion score of the initial path, and then the initial path with the lowest congestion score is fed back to the floating vehicle driver.

[0053] In this implementation plan, the specific method of assigning corresponding weights to edges based on the convenience and distance of travel between road segments is analyzed as follows: Consider factors such as the connection method between road segments, intersection types, and traffic control. For example, for directly connected road segments without traffic signal restrictions at intersections, the travel convenience is high, while for road segments connected through multiple complex intersections with frequent traffic control, the travel convenience is low. For road segments with high travel convenience, the weight of the edge is relatively low, meaning that vehicles can more easily travel between these road segments; shorter road segment distances should account for a smaller proportion in weight calculation because a short distance may mean a relatively short travel time. Longer road segment distances should be assigned larger weights to reflect the more time and cost that vehicles may spend traveling on this road segment. A specific example is as follows: Suppose there are three road segments A, B, and C. A is directly connected to B, and the intersection is a simple crossroads without traffic lights. A is connected to C through an intersection with traffic lights and frequent congestion. At the same time, the distance from A to B is 1 kilometer, and the distance from A to C is 2 kilometers. Then, when assigning weights, the edge weight between A and B may be set to 0.3 (considering both high travel convenience and short distance), and the edge weight between A and C may be set to 0.7 (considering low travel convenience and long distance). When generating a path, giving priority to connecting road segments with high travel convenience and relatively short distances is more in line with the path selection preferences in actual traffic.

[0054] By constructing a traffic state map and a road segment association matrix, the real-time traffic state and road segment connection relationship can be comprehensively considered to generate a relatively reasonable initial path for floating cars, which helps improve travel efficiency and reduce congestion time; calculating the congestion score of the initial path and selecting the path with the lowest score to feedback to the driver makes the path selection more targeted and accurate, and can better meet the driver's need to quickly reach the destination.

[0055] Specifically, the feedback adjustment module is analyzed as follows: Set a feedback adjustment period, and then retrieve the implementation effect of the floating car after receiving the optimized path within the feedback adjustment period. The real-time effect is specifically the traffic efficiency of each road segment; obtain the change amount of the traffic efficiency of each road segment in the key table before and after the implementation of the optimized path, and then take the average value of the change amounts of the traffic efficiency of each road segment as the effect evaluation value. Compare the effect evaluation value with the effect expected index. When the effect evaluation value is greater than or equal to the effect expected index, there is no need to trigger the management strategy adjustment mechanism; when the effect evaluation value is less than the effect expected index, trigger the management strategy adjustment mechanism. The management strategy adjustment includes increasing the promotion of the optimized path and optimizing and adjusting relevant parameters.

[0056] In this implementation plan, the feedback adjustment period is set as follows: Analyze historical traffic flow data to determine the time periods with significant traffic flow changes. For example, during the morning rush hour (7:00 - 9:00) and evening rush hour (17:00 - 19:00) on weekdays, the traffic flow is large and complex, and a shorter feedback adjustment period can be set, such as 15 - 30 minutes. During off-peak hours, the traffic flow is relatively stable, and the feedback adjustment period can be appropriately extended, such as 1 - 2 hours. If traffic accidents frequently occur in a certain area recently, in order to timely grasp the impact of accidents on traffic and the adjustment effect of management strategies, the feedback adjustment period should be shortened. If the traffic conditions in this area are stable for a long time, the period can be appropriately extended. At the same time, when there are special situations such as major events and construction, the traffic management department's demand for real-time control of traffic conditions increases. At this time, the feedback adjustment period is shortened to quickly respond and adjust strategies. Under normal circumstances, a relatively long period can be adopted.

[0057] The traffic efficiency of each road section reflects the ratio of the number of vehicles passing through the road section per unit time to the theoretical maximum traffic capacity of the road section, which reflects the actual utilization efficiency and traffic smoothness of the road section. For example, if 1000 vehicles actually pass through a certain road section per hour and its theoretical maximum traffic capacity is 1500 vehicles / hour, then the traffic efficiency of this road section is 0.67. Detect the passing number and time interval of vehicles on the road section through geomagnetic sensors, microwave sensors, etc., combine with the road section length, calculate the average vehicle speed, and then estimate the traffic efficiency according to the relationship model between speed and flow; it is also possible to collect the driving trajectories and time information of a large number of floating cars on the road section, count the number of floating cars passing through this road section during a certain time period, and the average driving time of these vehicles on this road section, and compare with the free flow driving time of the road section to estimate the traffic efficiency; or use the cameras installed on the road, identify the passing number and type of vehicles through image recognition technology, analyze the driving trajectories and speeds of vehicles, and thus calculate the traffic efficiency.

[0058] The method for setting the expected effect indicators is as follows: Analyze the improvement of traffic efficiency after taking similar optimization measures under similar traffic conditions in the past (such as the same season, weekday / weekend, weather conditions, etc.), and set the expected effect indicators based on this. For example, if the traffic efficiency has been increased by an average of 15% after taking a similar optimization path strategy during spring weekdays in the past, then the expected effect indicators for this time can be set between 10% - 20%. If the goal of the traffic management department is to reduce the congestion index by 20% in a specific area, according to the relationship between the congestion index and traffic efficiency, convert the corresponding traffic efficiency improvement goal, and use this as the expected effect indicator. For example, after analysis and calculation, to achieve a 20% reduction in the congestion index, the traffic efficiency needs to be increased by 18%, then 18% is used as the expected effect indicator.

[0059] The specific analysis of the management strategy adjustment is as follows: For promoting the optimized path by adding and optimizing it, through multiple channels such as traffic radio, mobile navigation applications, official websites of traffic management departments, and social media accounts, push the optimized path information to more drivers to improve the awareness rate and utilization rate of the optimized path. For example, it is possible to cooperate with mainstream mobile navigation applications to set the optimized path as one of the default recommended routes, and regularly broadcast the optimized path information during the peak-hour programs of traffic radio. It is also possible to provide personalized optimized path recommendations for different drivers according to their travel habits, historical driving data, etc. For example, for drivers who often travel during the morning and evening rush hours on weekdays, give priority to recommending paths optimized for peak hours; for drivers with higher requirements for driving speed, recommend optimized paths with a relatively higher average vehicle speed.

[0060] Regarding optimizing and adjusting relevant parameters, adjust the parameters related to traffic conditions, road section weights, time costs, etc. in the path planning algorithm. For example, increase the weight of congested road sections to make the path planning algorithm more inclined to avoid congested road sections. It is also possible to adjust the floating car data collection parameters, and according to the feedback effect, optimize parameters such as the frequency and scope of floating car data collection. For example, in areas where the effect of the optimized path is not good, increase the floating car data collection frequency to more accurately grasp the traffic conditions, expand the data collection scope, include data of more surrounding road sections, and provide more comprehensive information for path planning.

[0061] By regularly retrieving the implementation effect of the floating car optimized path, flexibly adjust the management strategy based on the effect evaluation, so that traffic management can closely follow the changes in the actual traffic conditions, achieve dynamic optimization, and improve the overall operation efficiency of the traffic system. Quantify the change in traffic efficiency into an effect evaluation value and compare it with the expected indicators, which can accurately judge whether the current management strategy has achieved the expected effect, avoid blind adjustment, and improve the pertinence and effectiveness of the management strategy. When the effect evaluation is good, do not blindly trigger the management strategy adjustment to save human and material resources; when the effect is not good, adjust the strategy in a targeted manner to ensure that resources are invested in the links that really need improvement and achieve reasonable resource allocation.

[0062] Please refer to Figure 2, A real-time traffic management method based on big data collection, applying the above-mentioned real-time traffic management system based on big data collection, includes the following steps: Step S1, conduct quality assessment on the collected traffic data based on a preset key digital table, screen out the traffic data that meets the quality assessment rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; Step S2, extract the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics of each road section in the key digital table, generate the traffic determination threshold of each road section by using the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics, and then combine the congestion index predicted by the multi-modal traffic prediction model and the traffic determination threshold of the corresponding road section to judge the traffic state of each road section; Step S3, according to the traffic state judgment results of each road section, combine the road section association simulation and the floating car target address to provide an optimized path for the floating car; Step S4, feedback the implementation effect after the floating car receives the optimized path to the key digital table in real time, and then compare the changes in the key numbers before and after the floating car receives the optimized path, and adjust the management strategy according to the comparison results of the changes in the key numbers.

[0063] In summary, the present application has at least the following effects:

[0064] Ensure the effectiveness and reliability of the data in the key digital table through quality assessment, and provide an accurate data basis for subsequent analysis and decision-making; comprehensively consider various characteristics to generate traffic determination thresholds, and combine with the multi-modal traffic prediction model, which can more accurately judge the traffic state of each road section, and help to detect traffic congestion and abnormal situations in a timely manner; provide an optimized path for the floating car according to the traffic state, which can effectively improve the vehicle passing efficiency, reduce congestion, lower energy consumption and exhaust emissions, and enhance the overall operation efficiency of urban traffic; through real-time feedback and comparative analysis, the management strategy can be dynamically adjusted according to the actual effect, enabling the traffic management system to have self-adaptability and the ability to continuously optimize, so as to better cope with complex and changeable traffic conditions.

[0065] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods and systems. Therefore, the present invention can adopt the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0066] The present invention is described with reference to the flowcharts and structural diagrams of methods and systems according to embodiments of the present invention. It should be understood that each process and module combination in the flowcharts and structural diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in one process Figure 1 one process or multiple processes and structures Figure 1 or multiple modules.

[0067] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one process Figure 1 one process or multiple processes and structures Figure 1 or multiple modules.

[0068] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and structures Figure 1 or multiple modules.

[0069] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed as including the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0070] Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. A real-time traffic management system based on big data collection, characterized in that, It includes a data screening module, a traffic status evaluation module, an optimized path generation module, and a feedback adjustment module, where: The data screening module is used to perform quality evaluation on the collected traffic data based on a preset key digital table, screen out the traffic data that meets the quality evaluation rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; The traffic status evaluation module is used to extract the historical congestion characteristics, historical accident characteristics, historical smooth characteristics, and road static characteristics of each section in the key digital table, generate the traffic determination threshold of each section by using the historical congestion characteristics, historical accident characteristics, historical smooth characteristics, and road static characteristics, and then combine the traffic congestion index predicted by the multi-modal traffic prediction model and the traffic determination threshold of the corresponding section to judge the traffic status of each section; The optimized path generation module is used to provide an optimized path for the floating vehicle according to the traffic status judgment result of each section, in combination with section association simulation and the target address of the floating vehicle; The feedback adjustment module is used to timely feedback the implementation effect after the floating vehicle receives the optimized path to the key digital table, and then compare the changes in the key numbers before and after the floating vehicle receives the optimized path, and adjust the management strategy according to the comparison result of the key number changes.

2. The real-time traffic management system based on big data collection according to claim 1, wherein The data screening module is specifically analyzed as follows: Extract the traffic data quality characteristics of each traffic data collection time node, and the traffic data quality characteristics include data integrity, data timeliness, data relevance, and data consistency; Use logistic regression to construct a traffic data quality binary classification model, take the traffic data quality characteristics as the input of the traffic data quality binary classification model, and output the probability that the traffic data is valid data; Set the traffic data quality evaluation threshold for each section based on traffic scenario semantic understanding and traffic road type; Compare the output results of the traffic data quality binary classification model at each traffic data collection time node of each section with the traffic data quality evaluation threshold of the corresponding section respectively. When the output result of the traffic data quality binary classification model exceeds the traffic data quality evaluation threshold, mark the traffic data collected at this traffic data collection time node of this section as the valid data of the key digital table.

3. A real-time traffic management system based on big data collection according to claim 2, characterized in that, The specific analysis of setting the traffic data quality evaluation threshold for each section based on traffic scenario semantic understanding and traffic road type is as follows: Based on the confusion matrix simulation of historical traffic data quality marked samples, set the initial traffic data quality evaluation threshold; Identify the scenario complexity and weather severity based on traffic scenario semantic understanding, and obtain the section importance score based on the traffic road type; Combine the scenario complexity, weather severity, and section importance score to determine the data quality adjustment requirement value of each section, and then use the data quality adjustment requirement value of each section to adjust the initial traffic data quality evaluation threshold to obtain the traffic data quality evaluation threshold of each section.

4. A real-time traffic management system based on big data collection according to claim 1, characterized in that, The traffic status evaluation module is specifically analyzed as follows: The traffic determination threshold specifically includes a congestion determination threshold and a smooth determination threshold; Obtain multi-modal traffic characteristics, and the multi-modal traffic characteristics specifically include weather characteristics, road air characteristics, peak state characteristics, and holiday state characteristics; Taking weather characteristics, road air characteristics, peak status characteristics, and holiday status characteristics as the inputs of the multimodal traffic prediction model, the predicted road congestion index is output; Comparing the road congestion index with the congestion determination threshold and the smooth determination threshold. When the road congestion index is greater than the congestion determination threshold, the road is initially in a congested state. When the road congestion index is less than the smooth determination threshold, the road is initially in a smooth state. When the road congestion index is between the congestion determination threshold and the smooth determination threshold, the road is initially in a slow-moving state.

5. A real-time traffic management system based on big data collection according to claim 4, characterized in that, The specific acquisition method of the traffic determination threshold is as follows: The historical congestion characteristics include congestion duration and minimum congestion speed. The historical accident characteristics include accident impact range and secondary congestion probability. The historical smoothness characteristics are non-peak average speed. The road static characteristics include the number of lanes and speed limit value; Using a regression model with historical congestion characteristics and historical accident characteristics as inputs, the congestion determination threshold is output; Using a regression model with historical smoothness characteristics and road static characteristics as inputs, the smooth determination threshold is output.

6. A real-time traffic management system based on big data collection according to claim 1, characterized in that, The optimization path generation module is specifically analyzed as follows: Summarize the traffic state judgment results of each road section to construct a traffic state map. The specific traffic state map is indexed by road section ID, marking the current traffic state of each road section. At the same time, establish a road section association matrix. The nodes in the road section association matrix are road section IDs, and the edges are the connection relationships between road sections. And corresponding weights are assigned to the edges based on the convenience of passing between road sections and the length of the distance; Obtain the current position of the floating car in real time, and obtain the target address information of the floating car based on the vehicle navigation input information. Taking the current position of the floating car as the starting point and the target address as the end point, use the traffic state map and the road section association matrix to generate an initial path; Traverse each road section on the initial path, obtain the predicted road congestion index of each road section, and take the average value of the predicted road congestion index of each road section as the congestion score of the initial path. Then, feedback the initial path with the lowest congestion score to the floating car driver.

7. A real-time traffic management system based on big data collection according to claim 1, characterized in that, The feedback adjustment module is specifically analyzed as follows: Set the feedback adjustment period, and then retrieve the implementation effect of the floating car after receiving the optimized path within the feedback adjustment period. The real-time effect is specifically the traffic efficiency of each road section; Obtain the change amount of the traffic efficiency of each road section in the key table before and after the implementation of the optimized path, and then take the average value of the change amount of the traffic efficiency of each road section as the effect evaluation value. Compare the effect evaluation value with the effect expected index. When the effect evaluation value is greater than or equal to the effect expected index, there is no need to trigger the management strategy adjustment mechanism; When the effect evaluation value is less than the effect expected index, trigger the management strategy adjustment mechanism. The management strategy adjustment includes increasing the promotion of the optimized path and optimizing and adjusting relevant parameters.

8. A real-time traffic management method based on big data collection, which applies the real-time traffic management system based on big data collection described in any one of claims 1-7, characterized in that Including the following steps: Step S1, based on a preset key digital table, conduct a quality assessment on the collected traffic data, screen out the traffic data that meets the quality assessment rules as the valid data of the key digital table, and then organize the valid data into the key digital table for classified storage; Step S2: Extract the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics of each road section in the key digital table, generate the traffic determination thresholds for each road section using the historical congestion characteristics, historical accident characteristics, historical smoothness characteristics, and road static characteristics, and then combine the congestion index predicted by the multi-modal traffic prediction model for each road section and the traffic determination thresholds for the corresponding road sections to determine the traffic status of each road section; Step S3: According to the traffic status judgment results of each road section, combined with the road section association simulation and the floating car target address, provide an optimized path for the floating car; Step S4: Real-time feedback the implementation effect after the floating car receives the optimized path to the key digital table, and then compare the changes in the key digits before and after the floating car receives the optimized path, and adjust the management strategy according to the comparison results of the changes in the key digits.

Images