A decision support system for urban traffic congestion management based on dynamic spatiotemporal graph neural network

Through the urban traffic congestion governance decision support system based on dynamic spatio-temporal graph neural network, the influence of vehicle volume and road length is introduced, and traffic flow prediction is optimized, and the problem of existing models ignore these factors is solved, achieving more accurate traffic management and more Efficient road use.

CN118736837BActive Publication Date: 2024-11-22CHONGQING DESIGN GRP CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411052185.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-22
Estimated Expiration
2044-08-01

AI Technical Summary

Technical Problem

The existing traffic model ignores the impact of different types of vehicles on road capacity and the cumulative effect of traffic flow on long-distance roads, resulting in inaccurate traffic forecasts and difficult to effectively alleviate urban traffic congestion.

Method used

The urban traffic congestion governance decision support system based on dynamic spatiotemporal graph neural network is adopted, and the influence of vehicle volume and road length is introduced through data collection and processing, spatiotemporal graph construction, traffic prediction and decision support modules are introduced to optimize the maximum traffic capacity and traffic flow prediction. And consider weather and holiday factors to provide refined traffic management decisions.

Benefits of technology

Improve the accuracy and reliability of traffic forecasts, support more efficient traffic management, reduce unexpected congestion, improve public travel experience, and improve road usage efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118736837B_ABST
    Figure CN118736837B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of urban traffic technology, and in particular, to a decision support system for urban traffic congestion management based on a dynamic spatiotemporal graph neural network. The system includes a data collection and processing unit for collecting multi-source real-time and historical traffic data; a spatiotemporal graph construction unit for defining nodes and edges based on the traffic data processed by the data collection and processing unit; a traffic prediction unit for predicting future traffic conditions based on real-time traffic data using a spatiotemporal graph; a congestion warning unit for issuing a warning to a traffic management center when a congestion condition is predicted; and a decision support unit for providing data support for urban traffic planning based on the predicted traffic conditions. The system not only improves the accuracy and reliability of traffic prediction, but also supports more efficient traffic management and safer road use, reduces unexpected congestion, and improves road use efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of urban traffic technology, and in particular to an urban traffic congestion management decision support system based on a dynamic spatiotemporal graph neural network. Background Art

[0002] Urban traffic congestion has become a common phenomenon in major cities around the world, especially during peak hours. Congestion can cause serious traffic delays, increase commuting time and costs, and the design and construction of urban roads often cannot keep up with the growth of vehicles. The limited road capacity cannot carry excessive traffic, especially at key traffic nodes and busy commercial areas. In response to these problems, city managers and traffic engineers have been seeking innovative solutions. The emergence of dynamic spatiotemporal graph neural network technology provides a new perspective and tool for solving urban traffic problems. Through real-time data collection, processing and prediction, combined with intelligent decision support, refined management and optimization of traffic flow can be achieved, thereby alleviating the above challenges and improving the overall performance of urban transportation systems. However, the traffic model in the existing technology ignores the impact of different types of vehicles on road capacity and the cumulative effect of traffic flow on long-distance roads. Therefore, a decision support system for urban traffic congestion management based on dynamic spatiotemporal graph neural network is designed. Summary of the invention

[0003] The purpose of the present invention is to provide an urban traffic congestion management decision support system based on a dynamic spatiotemporal graph neural network to solve the problem that the traffic model proposed in the above background technology ignores the impact of different types of vehicles on road capacity and the cumulative effect of traffic flow on long-distance roads.

[0004] To achieve the above objectives, the present invention aims to provide a decision support system for urban traffic congestion management based on a dynamic spatiotemporal graph neural network, comprising:

[0005] Data collection and processing unit: used to collect multi-source real-time and historical traffic data, and clean and integrate the real-time and historical traffic data;

[0006] Spatiotemporal graph construction unit: used to define nodes and edges according to the traffic data processed by the data collection and processing unit, assign attributes to each node and edge, generate a spatiotemporal graph of urban traffic, and regularly update the attributes of nodes and edges;

[0007] Traffic prediction unit: used to predict the traffic conditions in the future based on real-time traffic data using the spatiotemporal graph;

[0008] Among them, predicting the traffic conditions in the future includes predicting traffic flow and maximum capacity. In the process of predicting traffic flow and maximum capacity, the influence of vehicle volume and road length are introduced for optimization, which are as follows:

[0009] ;

[0010] Among them, is the maximum traffic capacity after optimization; is the proportion of the 𝑖th type of vehicles in the total flow; is the average volume of the 𝑖th type of vehicle; is the road length; is the correction factor for the effect of road length;

[0011] ;

[0012] in, The current traffic flow after optimization;

[0013] but

[0014] ;

[0015] in, The optimized congestion level; Congestion warning unit: used to send a warning to the traffic management center when a congestion situation is predicted;

[0016] Decision support unit: used to optimize the timing control of traffic lights, optimize the allocation of public transportation resources, optimize the driver's route selection according to the predicted traffic conditions, analyze historical congestion patterns, identify bottlenecks and accident-prone points, and provide data support for urban traffic planning;

[0017] Among them, the decision support unit includes a congestion prediction module, a traffic light adjustment module and a route planning module.

[0018] As a further improvement of the present technical solution, the traffic data in the data collection and processing unit includes traffic flow data, infrastructure data and time series data.

[0019] As a further improvement of the present technical solution, the nodes of the space-time graph in the space-time graph construction unit are intersections and road sections, and the attributes of the nodes include the location coordinates, number of lanes, maximum capacity, and average waiting time of the intersections and road sections; the edges of the space-time graph are a road, and the attributes of the edges include the length, direction, allowed vehicle types, and speed limit of a road.

[0020] As a further improvement of the present technical solution, the space-time graph construction unit can visualize the space-time graph and identify hot spots and time periods of traffic congestion.

[0021] As a further improvement of the technical solution, the congestion degree is obtained through the traffic flow and the maximum traffic capacity, and then the estimated travel time is obtained according to the congestion degree.

[0022] As a further improvement of the technical solution, the specific method of predicting the traffic flow and the maximum traffic capacity and obtaining the congestion degree through the traffic flow and the maximum traffic capacity is as follows:

[0023] ;

[0024] in, is the maximum traffic capacity; is the number of lanes; is the vehicle saturation flow rate; Green light time; is the signal cycle time;

[0025] ;

[0026] in, is the current traffic flow; As the basic traffic flow; is the growth rate; For time;

[0027] but

[0028] ;

[0029] in, The degree of congestion.

[0030] As a further improvement of the technical solution, the specific method for obtaining the estimated travel time according to the congestion level is as follows:

[0031] ;

[0032] in, is the estimated time of travel; is the average speed of the vehicle;

[0033] Introducing the influencing factors of weather and holidays into the above formula for optimization, the optimized expression is:

[0034] ;

[0035] in, is the estimated travel time after optimization; is the weather influence coefficient; is the holiday impact coefficient.

[0036] As a further improvement of the present technical solution, the congestion prediction module in the decision support unit is used to predict the degree of congestion using real-time and historical traffic data; the traffic light adjustment module is used to dynamically adjust the timing plan of traffic lights; the route planning module is used to provide drivers with alternative route suggestions to avoid congestion, and use map services to update navigation in real time.

[0037] As a further improvement of the present technical solution, the prediction result of the congestion prediction module includes the congestion location, duration and severity.

[0038] As a further improvement of this technical solution, the specific method of the route planning module is as follows:

[0039] S51. Based on the space-time graph, use the Euclidean distance as the heuristic function:

[0040] ;

[0041] in, From the current node To the target node Heuristic estimation cost of ; and The target nodes of Coordinates and coordinate; and The current node of Coordinates and coordinate;

[0042] S52, calculating the path cost:

[0043] ;

[0044] in, From the starting point To the current node The actual With from To the target Heuristic estimate of cost The sum of From the starting point To the current node the actual cost of

[0045] ;

[0046] in, is the parent node of node 𝑛; From parent node to node Costs;

[0047] S53, finding the path with the lowest cost through iterative search;

[0048] S531, set open list and close list ,set up and A heuristic estimate of the distance from the starting point to the target;

[0049] S532. Select the one with the lowest Value Node , move 𝑛 from the open list to the closed list; for Every neighbor :if If already in the closed list, skip; otherwise, calculate and ;if Not in the open list, or through arrive Lower cost, newer The parent node is , and add or update it in the open list;

[0050] S533, if the open list is empty, no path is found; if the target node In the open list, find the lowest cost path.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. In this urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network, the influence of vehicle volume and road length is introduced into the formula of maximum capacity and current traffic flow for optimization. Considering vehicle volume can more accurately reflect the actual occupation of road space by different types of vehicles (such as large trucks, buses and private cars), which in turn affects the road's traffic efficiency. Considering road length helps to evaluate the accumulation and dissipation effects of traffic flow over long distances, which is important for predicting traffic bottlenecks and congestion points. This can not only improve the accuracy and reliability of traffic forecasts, but also support more efficient traffic management and safer road use.

[0053] 2. In this urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network, the travel time calculation takes into account the impact of weather and holidays, which improves the accuracy of traffic forecasts, helps to improve traffic management decisions, reduce unexpected congestion, improve road use efficiency, and ultimately enhance the public's travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is the overall flow chart of the present invention;

[0055] Figure 2 It is a system flow chart of the decision support unit in the present invention.

[0056] The meaning of each number in the figure is:

[0057] 1. Data collection and processing unit; 2. Space-time graph construction unit; 3. Traffic prediction unit; 4. Congestion warning unit; 5. Decision support unit; 51. Congestion prediction module; 52. Traffic light adjustment module; 53. Route planning module. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] See also Figure 1-2 As shown, a city traffic congestion management decision support system based on a dynamic spatiotemporal graph neural network is provided, comprising a data collection and processing unit 1, a spatiotemporal graph construction unit 2, a traffic prediction unit 3, a congestion warning unit 4 and a decision support unit 5;

[0060] The data collection and processing unit 1 is used to collect multi-source real-time and historical traffic data, and clean and merge the real-time and historical traffic data;

[0061] Data is obtained from multiple sources such as various sensors, cameras, GPS devices, social media, traffic incident reports, etc.; pressure sensors, infrared sensors, etc. are used to detect the number and speed of vehicles on the road; cameras are used to monitor road conditions, vehicle types and traffic violations; GPS devices obtain location data from vehicle navigation systems, mobile phone applications, etc. to understand vehicle distribution and movement trajectories; social media is used to collect information about traffic conditions posted by users, such as traffic accidents, construction warnings, etc.; traffic incident reports are real-time and planned traffic incident information obtained from official channels; data cleaning is then performed to remove noise and outliers, and data fusion is performed to ensure data consistency and integrity, providing a high-quality data foundation for subsequent analysis.

[0062] Among them, traffic data includes traffic flow data, infrastructure data and time series data; by real-time monitoring and prediction of traffic conditions, the system can take timely measures to reduce congestion and traffic accidents; by analyzing infrastructure data and traffic flow data, the system can reasonably allocate traffic management resources and improve resource utilization; through the analysis of time series data, the system can provide a scientific basis for urban traffic planning and policy formulation; by optimizing signal control and path selection, the system can improve road traffic efficiency, reduce travel time and cost, and enhance citizens' travel experience.

[0063] The spatiotemporal graph construction unit 2 is used to define nodes and edges according to the traffic data processed by the data collection and processing unit 1, and assign attributes to each node and edge to generate a spatiotemporal graph of urban traffic, which can reflect the dynamic changes of the traffic network in time and space. For example, the spatiotemporal graph can show the changes in traffic flow of a certain road at different time periods of a day, and regularly update the attributes of nodes and edges to reflect the real-time traffic conditions, so as to ensure the accuracy and real-time performance of the spatiotemporal graph, for example, the traffic flow data is updated every 5 minutes;

[0064] The nodes of the spatiotemporal graph in the spatiotemporal graph construction unit 2 are intersections and road sections. The nodes represent the basic elements in the traffic network. The attributes of the nodes include the location coordinates, number of lanes, maximum capacity, and average waiting time of the intersections and road sections. The location coordinates can determine the exact location of the node in the geographic space, which is used for map matching and path planning. The number of lanes describes the number of lanes of the node (road section), which affects the capacity and traffic flow. The maximum capacity represents the maximum traffic flow that the node (road section) can accommodate under ideal conditions, which is used to evaluate the road load. The average waiting time records the average waiting time of vehicles at the node (intersection), reflecting the congestion of the node. The edge of the spatiotemporal graph is a road, which represents the connection relationship in the traffic network. The attributes of the edge include the length, direction, allowed vehicle type, and speed limit of a road. The length of the road indicates the actual length of the road, which is used to calculate the travel time and path planning; the direction indicates the direction of the road, one-way or two-way, which affects the management of traffic flow; the allowed vehicle type specifies which types of vehicles can pass through the road, such as motor vehicles, non-motor vehicles, buses, etc. The speed limit indicates the maximum speed limit of the road, which is used for safe driving and the enforcement of traffic regulations.

[0065] The traffic prediction unit 3 is used to use the time-space diagram to predict the traffic conditions in the future according to the real-time traffic data so as to respond in time;

[0066] Among them, predicting the traffic conditions in the future includes predicting the traffic flow and the maximum capacity, obtaining the congestion degree through the traffic flow and the maximum capacity, and then obtaining the estimated travel time according to the congestion degree;

[0067] Among them, the specific method of predicting traffic flow and maximum capacity and obtaining the congestion degree through traffic flow and maximum capacity is as follows:

[0068] ;

[0069] in, is the maximum traffic capacity; is the number of lanes; is the vehicle saturation flow rate; Green light time; is the signal cycle time;

[0070] ;

[0071] in, is the current traffic flow; As the basic traffic flow; is the growth rate; For time;

[0072] but

[0073] ;

[0074] in, The degree of congestion.

[0075] In the process of predicting traffic flow and maximum capacity, the influence of vehicle volume and road length are introduced for optimization, which are as follows:

[0076] ;

[0077] in, is the maximum traffic capacity after optimization; is the proportion of the 𝑖th type of vehicles in the total flow; is the average volume of the 𝑖th type of vehicle; is the road length; is the correction factor for the effect of road length;

[0078] The larger the vehicle size, the fewer vehicles will pass through the intersection per unit time, so the maximum capacity will be reduced accordingly. At the same time, the length of the road may affect the uniformity of traffic flow. Longer sections may cause unstable traffic flow and affect the capacity.

[0079] ;

[0080] in, The current traffic flow after optimization;

[0081] Vehicle volume affects the efficiency of road use, while road length may affect the distribution and accumulation of traffic flow;

[0082] but

[0083]

[0084] in, is the optimized congestion level.

[0085] This ratio represents the proportion of traffic flow to the maximum capacity. When this ratio is close to or exceeds 1, it means that the road is at or close to saturation and traffic congestion begins to occur.

[0086] Considering vehicle size can more accurately reflect the actual occupation of road space by different types of vehicles (such as large trucks, buses and private cars), which in turn affects the traffic efficiency of the road; different vehicle sizes mean that different types of vehicles have different impacts on traffic flow. For example, large vehicles may occupy more lane space, thereby reducing traffic capacity; considering road length helps to evaluate the accumulation and dissipation effects of traffic flow over long distances, which is important for predicting traffic bottlenecks and congestion points. In addition, traffic flow distribution on long-distance roads may be uneven, and some sections may be congested due to local events (such as accidents and construction). Considering road length can make predictions closer to actual traffic conditions;

[0087] Traditional traffic models may ignore the impact of different types of vehicles on road capacity and the cumulative effect of traffic flow on long-distance roads. By introducing the impact of vehicle volume and road length, the system can more finely predict traffic flow and maximum capacity, thereby providing more accurate traffic condition forecasts, which helps traffic managers better understand traffic patterns, predict bottlenecks and congestion points, and take corresponding preventive measures.

[0088] Furthermore, the specific method for obtaining the estimated travel time according to the congestion level is as follows:

[0089] ;

[0090] in, is the estimated time of travel; is the average speed of the vehicle;

[0091] Introducing the influencing factors of weather and holidays into the above formula for optimization, the optimized expression is:

[0092] ;

[0093] in, is the estimated travel time after optimization; is the weather influence coefficient; is the holiday impact coefficient;

[0094] The weather impact factor is used to indicate the effect of weather conditions on the speed of travel. For example, rainy and snowy weather will reduce the friction coefficient of the road, forcing drivers to slow down to maintain safety, thereby reducing the average speed. In addition, reduced visibility may also cause drivers to reduce speed; if the weather is bad, It may be greater than 1, indicating that the weather reduces the speed. The holiday impact coefficient is used to indicate the impact of holidays on the speed. During holidays, as many people travel for vacation or attend social events, the traffic volume on some sections may increase, while other sections may decrease. It may also be greater than 1, indicating that the traffic speed is reduced during holidays;

[0095] This ensures that the congestion level will not exceed 1 (complete congestion) nor be less than 0 (no congestion at all). The result may become very large or very small due to extreme weather or holiday effects, so it is necessary to limit its range to between 0 and 1. In practical applications, the weather influence coefficient and holiday impact coefficient It needs to be set based on historical data and expert experience.

[0096] By introducing the influence of weather and holidays, the formula can more accurately reflect the actual traffic conditions. In bad weather or holidays, the estimated travel time will be adjusted according to these factors, providing drivers with more reliable time estimates, while also providing traffic management departments with more accurate data to take appropriate measures to alleviate congestion.

[0097] The congestion warning unit 4 is used to send a warning to the traffic management center when a congestion situation is predicted;

[0098] When congestion is predicted, an early warning is automatically issued to the traffic management center. For example, when a road section is predicted to be severely congested in 10 minutes, an early warning is immediately issued, and a congestion alert is issued to relevant departments and the public to remind them to pay attention to the congested road section, for example, through mobile phone apps, social media and other channels. The emergency response plan is then launched to coordinate resources to alleviate congestion, for example, traffic police are dispatched to the scene to direct traffic and adjust the timing of signal lights. At the same time, travel recommendations for the public are updated to guide the reasonable choice of travel routes and methods, for example, drivers are advised to bypass congested sections and choose public transportation.

[0099] The decision support unit 5 is used to optimize the timing control of traffic lights, optimize the allocation of public transportation resources, optimize the driver's path selection according to the predicted traffic conditions, and analyze the historical congestion pattern, identify bottlenecks and accident-prone points, and provide data support for urban traffic planning;

[0100] Among them, the decision support unit 5 includes a congestion prediction module 51, a traffic light adjustment module 52 and a route planning module 53; the congestion prediction module 51 is used to predict the degree of congestion using real-time and historical traffic data; the traffic light adjustment module 52 is used to dynamically adjust the timing plan of traffic lights; the route planning module 53 is used to provide drivers with alternative route suggestions to avoid congestion, and use map services to update navigation in real time.

[0101] The prediction result of the congestion prediction module 51 includes the congestion location, duration and severity;

[0102] Among them, knowing the location of congestion can inform relevant departments (such as traffic management centers and emergency service departments) in advance about the places where congestion may occur, so as to deploy resources in advance and take measures to alleviate congestion; then release congestion warning information to the public, inform drivers through broadcasting, electronic bulletin boards, mobile applications, etc., to help them adjust their travel plans; at the same time, long-term tracking of congestion locations can help urban planners and engineers understand traffic bottlenecks and optimize road design and transportation infrastructure construction;

[0103] Knowing the duration, drivers can decide whether to take a detour or postpone their trip based on the estimated duration of the congestion to avoid being stuck in a long traffic jam. Logistics companies and the transportation industry can adjust the timetable for cargo delivery and reduce the cost losses caused by traffic delays. Traffic management departments can reasonably arrange the working hours of police and maintenance personnel to ensure that there are enough people on duty during peak congestion periods.

[0104] High severity means that immediate attention and processing are required, which helps traffic management centers determine which congested areas should be prioritized for action; severe traffic congestion may increase the risk of traffic accidents, so understanding the severity can help relevant departments strengthen monitoring and safety measures; for drivers, understanding the severity of congestion helps determine whether they need to find alternative routes and estimate the extra time to reach their destination.

[0105] The comprehensive application of this information can greatly improve the operating efficiency of the road network, reduce commuting time, improve the driving experience, reduce the risk of traffic accidents, and have a positive impact on environmental protection because it reduces pollution caused by idling emissions. In addition, for urban planners and traffic policy makers, long-term congestion data analysis is an important basis for improving traffic planning and infrastructure investment.

[0106] The specific method of the route planning module 53 is as follows:

[0107] S51. Based on the space-time graph, use the Euclidean distance as the heuristic function:

[0108] ;

[0109] in, From the current node To the target node Heuristic estimation cost of ; and The target nodes of Coordinates and coordinate; and The current node of Coordinates and coordinate;

[0110] S52, calculating the path cost:

[0111] ;

[0112] in, From the starting point To the current node The actual With from To the target Heuristic estimate of cost The sum of From the starting point To the current node the actual cost of

[0113] ;

[0114] in, is the parent node of node 𝑛, that is, the directly previous node in the currently known optimal path; From parent node to node The cost of an edge is usually determined by the edge weight;

[0115] S53, finding the path with the lowest cost through iterative search;

[0116] S531, set open list and close list ,set up and The open list stores nodes that have yet to be evaluated, that is, nodes that have not been fully explored but may be part of the lowest-cost path. The closed list stores evaluated nodes, that is, nodes that have been removed from the open list and all their neighbors have been evaluated.

[0117] S532. Select the one with the lowest Value Node , move 𝑛 from the open list to the closed list; for Every neighbor :if If already in the closed list, skip; otherwise, calculate and ;if Not in the open list, or through arrive Lower cost, newer The parent node is , and add or update it in the open list;

[0118] S533, if the open list is empty, no path is found; if the target node In the open list, find the lowest cost path.

[0119] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A decision support system for urban traffic congestion management based on dynamic spatiotemporal graph neural network, characterized by: include: Data collection and processing unit (1): used to collect multi-source real-time and historical traffic data, and clean and integrate the real-time and historical traffic data; The spatiotemporal graph construction unit (2) is used to define nodes and edges according to the traffic data processed by the data collection and processing unit (1), assign attributes to each node and edge, generate a spatiotemporal graph of urban traffic, and regularly update the attributes of the nodes and edges; Traffic prediction unit (3): used to predict the traffic conditions in the future based on the real-time traffic data using the spatiotemporal graph; Among them, predicting the traffic conditions in the future includes predicting traffic flow and maximum capacity. In the process of predicting traffic flow and maximum capacity, the influence of vehicle volume and road length are introduced for optimization, which are as follows: ; Among them, is the maximum traffic capacity after optimization; is the proportion of the 𝑖th type of vehicles in the total flow; is the average volume of vehicles of type 𝑖; is the road length; is the correction factor for the effect of road length; ; in, The current traffic flow after optimization; but ; in, is the optimized congestion level; Congestion warning unit (4): used to issue a warning to the traffic management center when a congestion situation is predicted; Decision support unit (5): used to optimize the timing control of traffic lights, optimize the allocation of public transportation resources, optimize the driver's route selection according to the predicted traffic conditions, and analyze historical congestion patterns to identify bottlenecks and accident-prone points, providing data support for urban traffic planning; The decision support unit (5) includes a congestion prediction module (51), a traffic light adjustment module (52) and a route planning module (53).

2. According to claim 1, a decision support system for urban traffic congestion management based on dynamic spatiotemporal graph neural network is characterized by: The traffic data in the data collection and processing unit (1) includes traffic flow data, infrastructure data and time series data.

3. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 2 is characterized by: The nodes of the space-time graph in the space-time graph construction unit (2) are intersections and road sections, and the attributes of the nodes include the location coordinates, number of lanes, maximum traffic capacity, and average waiting time of the intersections and road sections; the edges of the space-time graph are roads, and the attributes of the edges include the length, direction, allowed vehicle types, and speed limit of a road.

4. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 3 is characterized by: The space-time graph construction unit (2) is capable of visualizing the space-time graph and identifying hot spots and time periods of traffic congestion.

5. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 4 is characterized by: The congestion degree is obtained through the traffic flow and the maximum traffic capacity, and then the estimated travel time is obtained according to the congestion degree.

6. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 5 is characterized by: The specific method of predicting the traffic flow and the maximum capacity and obtaining the congestion degree through the traffic flow and the maximum capacity is as follows: ; in, is the maximum traffic capacity; is the number of lanes; is the vehicle saturation flow rate; Green light time; is the signal cycle time; ; in, is the current traffic flow; As the basic traffic flow; is the growth rate; For time; but ; in, The degree of congestion.

7. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 6 is characterized by: The specific method for obtaining the estimated travel time according to the congestion level is as follows: ; in, is the estimated time of travel; is the average speed of the vehicle; Introducing the influencing factors of weather and holidays into the above formula for optimization, the optimized expression is: ; in, is the estimated travel time after optimization; is the weather influence coefficient; is the holiday impact coefficient.

8. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 7 is characterized by: The congestion prediction module (51) in the decision support unit (5) is used to predict the degree of congestion using real-time and historical traffic data; the signal light adjustment module (52) is used to dynamically adjust the timing plan of the traffic light; and the route planning module (53) is used to provide the driver with alternative route suggestions to avoid congestion, and to update the navigation in real time using a map service.

9. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 8 is characterized by: The prediction result of the congestion prediction module (51) includes the congestion location, duration and severity.

10. The urban traffic congestion management decision support system based on dynamic spatiotemporal graph neural network according to claim 9 is characterized by: The specific method of the route planning module (53) is as follows: S51. Based on the space-time graph, use the Euclidean distance as the heuristic function: ; in, From the current node To the target node Heuristic estimation cost of ; and The target nodes of Coordinates and coordinate; and The current node of Coordinates and coordinate; S52, calculating the path cost: ; in, From the starting point To the current node The actual With from To the target Heuristic estimate of cost The sum of From the starting point To the current node the actual cost of ; in, is the parent node of node 𝑛; From parent node to node Costs; S53, finding the path with the lowest cost through iterative search; S531, set open list and close list ,set up and A heuristic estimate of the distance from the starting point to the target; S532. Select the one with the lowest Value Node , move 𝑛 from the open list to the closed list; for Every neighbor :if If already in the closed list, skip; otherwise, calculate and ;if Not in the open list, or through arrive Lower cost, newer The parent node is , and add or update it in the open list; S533, if the open list is empty, no path is found; if the target node In the open list, find the lowest cost path.

Citation Information

Patent Citations

  • Urban road traffic flow state estimation method and system based on block chain

    CN117671961A

  • Intelligent bus scheduling method and system

    CN117671992A

  • Urban road carbon neutralization estimation analysis method

    CN117854283A

  • GIS-based big data real-time acquisition device and method for smart city construction

    CN118396224A

  • Large-scale real-time traffic flow prediction method based on fuzzy logic and deep LSTM

    US20210209939A1