Regional green wave scheduling method and system based on traffic flow prediction, and storage medium

Abstract

The invention discloses a regional green wave scheduling method and system based on traffic flow prediction and a storage medium, and relates to the field of traffic control systems.The method comprises the steps that road vectors, traffic facility data and POI information in a region are extracted, and a road network topological structure is constructed; and road real-time flow is collected based on the road network topological structure, flow prediction is carried out by using the flow prediction model and taking the POI information as input, and road predicted flow is obtained. Data alignment is carried out according to the road real-time flow and the road predicted flow, and a dynamic adjacent weight matrix is constructed to reflect the main trunk-branch trunk priority change. And in combination with a road network topological structure, phase difference and time window adjustment optimization is carried out on the trunk and the branches, and a regional green wave scheduling strategy is obtained. The technical problems that in the prior art, a traffic scheduling system lacks dynamic response to real-time flow and predicted flow, a green wave band strategy cannot be effectively adjusted according to the change of the traffic flow, the traffic efficiency is low in the peak period, and congestion occurs are solved.

Description

Technical Field

[0001] The present invention relates to the field of traffic control systems, and in particular to a regional green wave scheduling method, system and storage medium based on traffic flow prediction. Background Art

[0002] With the continuous advancement of urbanization, the flow of urban traffic has increased, causing a large number of traffic congestion problems, especially during peak traffic hours, and the challenges of traffic management have become increasingly prominent. The traditional traffic light scheduling method is often fixed, lacking flexibility and adaptability, and cannot effectively respond to changes in traffic flow, resulting in low traffic efficiency and increased delays. In order to alleviate this problem, the adoption of green wave scheduling strategy has become an important means to improve the smoothness of urban traffic. The green wave scheduling strategy coordinates the traffic lights at intersections to ensure that a vehicle can continue to keep the green light when passing through multiple intersections, reducing the dwell time and improving the traffic efficiency. However, the traditional green wave scheduling method mainly relies on fixed time settings and lacks dynamic response to real-time traffic and predicted traffic changes.

[0003] Therefore, in the existing technology, the traffic dispatch system lacks dynamic response to real-time traffic and predicted traffic, and cannot effectively adjust the green wave belt strategy according to changes in traffic flow, resulting in low traffic efficiency during peak hours and technical problems such as congestion. Summary of the invention

[0004] This application solves the technical problem that the traffic dispatching system in the prior art lacks dynamic response to real-time traffic and predicted traffic, and cannot effectively adjust the green wave belt strategy according to changes in traffic flow, resulting in low traffic efficiency and congestion during peak hours by providing a regional green wave dispatching method, system and storage medium based on traffic flow prediction. The regional green wave dispatching method based on traffic flow prediction dynamically adjusts the priority, phase difference and green light time window of road signals by combining real-time traffic data collection with the predicted traffic model, thus realizing the optimized dispatching of regional green wave belts, significantly improving traffic fluency and reducing traffic congestion.

[0005] A first aspect of an embodiment of the present application provides a regional green wave scheduling method based on traffic flow prediction, the method comprising: extracting road vectors, traffic facility data and POI information in a region, and constructing a road network topology structure, wherein the road network topology structure comprises trunk roads and branch roads; collecting real-time road traffic based on the road network topology structure, and using a traffic prediction model to perform traffic prediction with POI information as input to obtain predicted road traffic, wherein the traffic prediction model is obtained by training and convergence using historical data; aligning data according to the real-time road traffic and the predicted road traffic, and constructing a dynamic adjacency weight matrix to reflect the priority changes of the trunk and branches; according to the priority changes of the trunk and branches combined with the road network topology structure, optimizing the phase difference and time window adjustment of the trunk and branches to obtain a regional green wave scheduling strategy.

[0006] In the implementation method, vector data of all roads in the area are extracted from map data, including road geometry, number of lanes, and turning restrictions; traffic facility data in the area are collected, including the location and attribute information of traffic lights, isolation guardrails, bus stops, crosswalks, and cameras; the name, category, coordinates, address data, service radius, peak hours, and event data of POI points in the area are obtained to establish the POI information; the road network is defined as a directed graph, and nodes and edges are extracted according to the positioning relationship based on the vector data of all roads, traffic facility data, and POI information to construct the road network topology structure.

[0007] In the implementation method, nodes and edges are extracted according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads, and the road network topology structure is constructed, including: identifying road intersections and confluence / divergence points, and defining nodes; defining the connection relationship between each node as an edge based on the road vector data, wherein the number of lanes, speed limit, and turning limit of each edge define the edge attributes; according to the road geometric connection relationship, the nodes, edges and edge attributes are connected to construct the road network topology structure, and attribute marking is performed according to the position positioning relationship of the traffic facility data and POI information.

[0008] In the implementation method, constructing the road network topology structure also includes: defining the determination rules of trunk roads and branch roads according to the number of lanes; determining the road grade of the lane number of each lane according to the determination rules, determining the road grade as a trunk lane or a branch lane, and marking the trunk roads and branch roads of the road network topology structure according to the road grade.

[0009] In the implementation method, data alignment is performed according to the real-time road traffic and the predicted road traffic, and a dynamic adjacency weight matrix is ​​constructed, including: performing time series alignment of traffic data according to the time relationship between the real-time road traffic and the predicted road traffic; setting the update frequency of the real-time road traffic and the predicted road traffic, configuring a sliding window based on the update frequency, and updating the traffic time series data according to the sliding window; and dynamically adjusting the adjacency weight matrix according to the updated traffic time series data.

[0010] In the implementation method, the adjacency weight matrix is ​​dynamically adjusted, including: configuring the static weight of each road according to the historical traffic flow of the main road and the branch road; dividing the real-time traffic flow of the road and the predicted traffic flow of the road by the maximum traffic flow of the road respectively, de-dimensionalizing and converting them into interval values ​​of [0,1] to obtain the real-time flow coefficient and the predicted flow coefficient; configuring weight coefficients for different traffic time periods, performing weighted calculation on the static weight, real-time flow coefficient, and predicted flow coefficient according to the weight coefficients, obtaining the adjacency weight of each road section, and constructing the adjacency weight matrix.

[0011] In the implementation method, according to the priority changes of the trunk-branch combined with the road network topology, the phase difference and time window of the trunk and branch are adjusted and optimized to obtain a regional green wave scheduling strategy, including: evaluating the convoy propagation speed on the road within a preset time period according to the real-time traffic flow of the road and the predicted traffic flow of the road; obtaining the priority of the trunk roads and branch roads from the dynamic adjacency weight matrix, and calculating the phase difference between adjacent intersecting roads according to the priority of the trunk roads and the branch roads and the convoy propagation speed; optimizing the green light time windows of the trunk and branches according to the traffic prediction results to ensure the continuity of the green wave band; obtaining the regional green wave scheduling strategy according to the phase difference and green light time window.

[0012] In the implementation method, obtaining the regional green wave scheduling strategy also includes: screening the target scheduling roads according to the dynamic adjacency weight matrix, the target scheduling roads are roads whose priorities meet the preset threshold; performing trunk-branch multi-intersection zoning based on the distribution of the target scheduling roads; performing execution status partition verification on the regional target green wave band through tracking the difference between the real-time road traffic and the predicted road traffic; when the regional target green wave band is not met, resetting the adjacency weight matrix to optimize and adjust the regional green wave scheduling strategy, wherein the regional green wave scheduling strategy optimization and adjustment includes dynamic adjustment of the phase difference and the green light time window; when the regional target green wave band is met, performing regional global green wave band evaluation and verification based on the real-time road traffic.

[0013] A second aspect of an embodiment of the present application provides a regional green wave scheduling system based on traffic flow prediction, the system comprising: a topology construction module, used to extract road vectors, traffic facility data and POI information in the region, and construct a road network topology structure, wherein the road network topology structure comprises trunk roads and branch roads; a road flow prediction module, used to collect real-time road flow based on the road network topology structure, and use a flow prediction model to perform flow prediction with POI information as input to obtain predicted road flow, wherein the flow prediction model is obtained by training and convergence using historical data; a weight matrix acquisition module, used to perform data alignment according to the real-time road flow and the predicted road flow, and construct a dynamic adjacency weight matrix to reflect the priority changes of the trunk-branch; a scheduling strategy acquisition module, used to adjust and optimize the phase difference and time window of the trunk and branch according to the priority changes of the trunk-branch in combination with the road network topology structure, and obtain a regional green wave scheduling strategy.

[0014] According to a third aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, and the computer program is used to execute the regional green wave scheduling method based on traffic flow prediction provided by the present application.

[0015] The regional green wave scheduling method, system and storage medium based on traffic flow prediction proposed in this application are proposed to construct a road network topology by extracting road vectors, traffic facility data and POI information in the region. Based on the road network topology, the real-time road traffic is collected, and the traffic prediction model is used to predict the traffic with POI information as input to obtain the predicted road traffic. Data alignment is performed according to the real-time road traffic and the predicted road traffic, and a dynamic adjacency weight matrix is ​​constructed to reflect the priority changes of the trunk-branch. In combination with the road network topology, the phase difference and time window adjustment and optimization of the trunk and branch are performed to obtain the regional green wave scheduling strategy. It solves the technical problem that the traffic scheduling system lacks dynamic response to real-time traffic and predicted traffic in the prior art, and cannot effectively adjust the green wave belt strategy according to the changes in traffic flow, resulting in low traffic efficiency and congestion during peak hours. The regional green wave scheduling method based on traffic flow prediction dynamically adjusts the priority, phase difference and green light time window of the road signal by combining the real-time collection of traffic data with the predicted traffic model, realizes the optimized scheduling of the regional green wave belt, significantly improves the traffic fluency, and reduces traffic congestion.

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0018] Figure 1 A flow chart of a regional green wave scheduling method based on traffic flow prediction provided in an embodiment of the present application; Figure 2 A schematic diagram of a process for constructing a road network topology structure using a regional green wave scheduling method based on traffic flow prediction provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of a regional green wave scheduling system based on traffic flow prediction provided in an embodiment of the present application.

[0019] Explanation of the reference numerals: topology structure building module 11, road flow prediction module 12, weight matrix acquisition module 13, scheduling strategy acquisition module 14. DETAILED DESCRIPTION

[0020] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0021] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0022] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0023] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a regional green wave scheduling method based on traffic flow prediction, the method comprising: Extract road vectors, traffic facility data and POI information in the area to construct a road network topology structure, wherein the road network topology structure includes trunk roads and branch roads; collect real-time road traffic based on the road network topology structure, and use a traffic prediction model to perform traffic prediction with POI information as input to obtain road predicted traffic, wherein the traffic prediction model is obtained by training and converging using historical data.

[0024] Extract road vectors, traffic facility data and POI information in the area, wherein the road vector data is vector data of all roads in the area extracted from the map data, including the geometry of each road, the number of lanes and turning restrictions, etc. POI stands for "point of interest", which refers to a location with specific geographic coordinates on a map. The traffic facility data is obtained by collecting specific information of traffic facilities in the area, including the location and attribute information of traffic lights, isolation guardrails, bus stops, crosswalks, cameras, etc. The POI information is the name, category, address, etc. of important places such as shopping malls, schools, hospitals, etc. in the area. A road network topology structure is constructed based on the extracted information, and the road network topology structure includes trunk roads and branch roads. Further, based on the road network topology structure, real-time road traffic is collected, and traffic prediction is performed based on the constructed traffic prediction model with POI information as input, and the traffic of each road in the future period of time is predicted to obtain the predicted road traffic. The traffic prediction model is obtained by training and convergence using historical data. When constructing the traffic prediction model, the POI information recorded in the historical data and the traffic flow data of the road corresponding to the POI information in the future period of time are used. Furthermore, the POI information in the historical data is used as training data, and the traffic flow change data of the roads affected by the POI information in the future is used as supervision data to perform supervised training on the neural network model, so as to obtain the traffic flow change data of each affected road output by the model prediction, and further, the traffic flow data of each affected road is added to obtain the predicted road traffic flow.

[0025] like Figure 2 As shown, the method provided in the embodiment of the present application also includes: extracting vector data of all roads in the area from the map data, including road geometry, number of lanes, and turning restrictions; collecting traffic facility data in the area, including the location and attribute information of traffic lights, isolation guardrails, bus stops, crosswalks, and cameras; obtaining the name, category, coordinates, address data, service radius, peak hours, and event data of POI points in the area, and establishing the POI information; defining the road network as a directed graph, and extracting nodes and edges according to the positioning relationship based on the vector data of all roads, traffic facility data, and POI information, to construct the road network topology structure.

[0026] Constructing a road network topology structure includes: extracting vector data of all roads in the area from the geographic information system in the map data, including road geometry, number of lanes, and turning restrictions. The geometry of the road includes the starting point, end point, curve, intersection, etc. of the road. For each road, extract its lane number information. The number of lanes is crucial for traffic prediction and green wave scheduling, because roads with different numbers of lanes require different scheduling strategies. Some roads have turning restrictions, such as some intersections may only allow left or right turns. It is necessary to extract this information through vector data so that the impact of turning behavior on traffic flow can be considered in subsequent traffic prediction and scheduling. Collect traffic facility data in the area, including the location and attribute information of traffic lights, isolation guardrails, bus stops, crosswalks, and cameras. Obtain the name, category, coordinates, address data, service radius, peak hours, and event data of POI points in the area, and establish the POI information. The service radius is the area covered by the POI, and the event data is the event information that frequently exists in the POI object, such as school dismissal, concert performances, and other events that will affect traffic flow. Peak hours are peak hours with large traffic volume, such as morning peak hours and evening peak hours. Finally, after integrating all the above information, a road network topology is constructed through a directed graph. Nodes and edges are extracted according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads to construct the road network topology.

[0027] The method provided in the embodiment of the present application also includes: identifying road intersections, confluence / divergence points, and defining nodes; defining the connection relationship between each node as an edge based on the road vector data, wherein the number of lanes, speed limit, and turning limit of each edge define the edge attributes; connecting the nodes, edges, and edge attributes based on the road geometric connection relationship, constructing the road network topology structure, and marking the attributes according to the location positioning relationship of the traffic facility data and POI information.

[0028] Nodes and edges are extracted according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads, and the road network topology structure is constructed, including: identifying the road intersections, confluence / divergence points, and defining nodes. The road intersection is the location where two or more roads intersect with each other. Confluence is the confluence of multiple roads into one road, and the divergence point is the location where multiple roads branch out from one road. Road intersections and confluence / divergence points are defined as nodes. Subsequently, the connection relationship between nodes is defined as edges based on the vector data of the road. The number of lanes, speed limit, and turning restriction of each edge define the edge attributes. Exemplarily, a section of road from node X to node Y can be defined as edge XY; this section of edge XY has four lanes in both directions, a speed limit of 50 kilometers per hour, and a turning restriction at node Y, that is, left turn is prohibited. The above information will be explicitly recorded as the attributes of edge XY in the road network topology structure.

[0029] Furthermore, after the definition of nodes and edges is completed, based on the geometric connection relationship of the actual road, the above nodes and edges are connected according to the actual spatial position relationship of the road to form a complete, clear and orderly directed graph. And traffic facility data such as traffic lights, bus stops, crosswalks, cameras and POI information such as shopping malls, hospitals, schools, etc. are further attributed to the nodes and edges in the road network according to the spatial positioning relationship. Assuming that there is a shopping mall near node Z, and the intersection is equipped with traffic lights and crosswalks, these attribute information are marked when defining node Z so that the subsequent traffic prediction model can perform more accurate prediction analysis.

[0030] The method provided in the embodiment of the present application also includes: defining the determination rules of trunk roads and branch roads according to the number of lanes; determining the road grade of the lane number of each lane according to the determination rules, determining the road grade as a trunk lane or a branch lane, and marking the trunk roads and branch roads of the road network topology structure according to the road grade.

[0031] Constructing the road network topology structure also includes: after clarifying the basic structure and attributes of nodes and edges, it is necessary to define the determination rules of the main roads or branch roads of the road according to the number of lanes. Main road determination rules: Roads with four lanes or more in both directions are defined as main roads. Branch road determination rules: Roads with four lanes or less in both directions are defined as branch roads. According to the defined determination rules, the number of lanes of all edges, i.e. roads, in the road network structure are analyzed one by one, and the road grade is determined and classified separately, so as to determine whether the road grade is a main lane or a branch lane, and the main roads and branch roads are marked for the road network topology structure according to the road grade. Thus, a complete, accurate and orderly road network topology structure is formed. This sophisticated structure lays a solid data and structural foundation for the next step of real-time traffic data collection, traffic prediction and regional green wave traffic scheduling strategy based on dynamic adjustment, so that subsequent traffic scheduling can be more accurate, efficient and reliable.

[0032] Data alignment is performed based on the real-time road traffic and the predicted road traffic, and a dynamic adjacency weight matrix is ​​constructed to reflect the priority changes of the trunk and branches. According to the priority changes of the trunk and branches and combined with the road network topology, the phase difference and time window of the trunk and branches are adjusted and optimized to obtain a regional green wave scheduling strategy.

[0033] Data alignment is performed according to the real-time traffic volume of the road and the predicted traffic volume of the road, so that the real-time traffic volume data and the predicted traffic volume data are synchronized in time series, ensuring that the data have the same time dimension. Then, a dynamic adjacency weight matrix is ​​constructed to reflect the priority changes of the trunk-branch, and the adjacency weights of each road are recorded in the dynamic adjacency weight matrix. Finally, according to the priority changes of the trunk-branch combined with the road network topology, the phase difference of the trunk and branch is recalculated, and the green light time window is adjusted and optimized using the predicted traffic volume of the road to obtain the regional green wave scheduling strategy. It solves the technical problem that the traffic scheduling system in the prior art lacks dynamic response to real-time traffic and predicted traffic, and cannot effectively adjust the green wave belt strategy according to the changes in traffic flow, resulting in low traffic efficiency and congestion during peak hours. The regional green wave scheduling method based on traffic flow prediction dynamically adjusts the priority, phase difference and green light time window of the road signal by combining the real-time collection of traffic data with the predicted traffic model, realizes the optimized scheduling of the regional green wave belt, significantly improves the traffic fluency, and reduces traffic congestion.

[0034] The method provided in the embodiment of the present application also includes: aligning the traffic data time series according to the time relationship between the real-time road traffic and the predicted road traffic; setting the update frequency of the real-time road traffic and the predicted road traffic, configuring a sliding window based on the update frequency, and updating the traffic time series data according to the sliding window; and dynamically adjusting the adjacency weight matrix according to the updated traffic time series data.

[0035] According to the real-time traffic of the road and the predicted traffic of the road, data alignment is performed to construct a dynamic adjacency weight matrix, including: according to the time relationship between the real-time traffic of the road and the predicted traffic of the road, the traffic data is aligned in time series, and the real-time traffic refers to the road traffic data collected in real time by sensors, cameras and other equipment, which is usually collected every certain time interval, such as 1 minute or 5 minutes. The real-time traffic of the road and the predicted traffic of the road are aligned in time. Subsequently, the update frequency of the real-time traffic of the road and the predicted traffic of the road is set, such as updating the data every 5 minutes, or updating every 30 seconds. Further, the size of the sliding window is set according to the update frequency. For example, the update frequency is 5 minutes, and the size of the sliding window can be set to 30 minutes. The sliding window will gradually slide within a time range of 30 minutes to update the traffic data within the period. The current window is the real-time traffic data of the past 30 minutes. As time goes by, the sliding window will slide from the data of 00:00-00:30 to the data of 00:05-00:35, ensuring that each window contains the latest traffic information. Finally, according to the updated traffic time series data, the adjacency weight matrix is ​​dynamically adjusted.

[0036] The method provided in the embodiment of the present application also includes: configuring the static weight of each road according to the historical traffic flow of the main road and the branch road; dividing the real-time traffic flow of the road and the predicted traffic flow of the road by the maximum traffic flow of the road respectively, de-dimensionalizing and converting them into interval values ​​of [0,1] to obtain the real-time traffic coefficient and the predicted traffic coefficient; configuring weight coefficients for different traffic time periods, performing weighted calculation on the static weight, real-time traffic coefficient, and predicted traffic coefficient according to the weight coefficients, obtaining the adjacency weight of each road section, and constructing the adjacency weight matrix.

[0037] Dynamically adjusting the adjacency weight matrix includes: before constructing the dynamic adjacency weight matrix, configuring a static weight for each road, wherein the static weight is set according to the attributes of the road, i.e., a main road or a branch road, and the historical traffic volume. The weight of the main road will be higher than that of the branch road, and the weight of the road with a large traffic volume will also be higher than that of the road with a small traffic volume, because the main road carries more traffic volume and is usually a road with a higher priority in traffic scheduling. Subsequently, for each road, its maximum flow is obtained, and then the real-time flow and the predicted flow are respectively divided by the maximum flow of the road, so as to convert the real-time flow and the predicted flow into dimensionless coefficients with interval values ​​of [0,1], and obtain the real-time flow coefficient and the predicted flow coefficient. For example, if the maximum flow of a road is 1,000 vehicles per hour, and the current real-time flow is 500 vehicles per hour, the real-time flow coefficient is 500 / 1000 = 0.5. If the predicted flow of the road is 700 vehicles per hour, the predicted flow coefficient is 700 / 1000 = 0.7. Finally, configure the weight coefficients of different traffic time periods. The weight coefficients of the traffic time periods are weight configuration parameters of the static weights, real-time flow coefficients, and predicted flow coefficients at different times. The specific weight coefficients can be set based on the actual time period. For example, the predicted traffic flow in the early morning has less impact on the overall traffic flow, so the weight coefficient of the predicted traffic flow can be reduced. For example, the predicted traffic flow in the school area during the school dismissal period has a greater impact on the overall traffic flow, so the weight coefficient of the predicted traffic flow can be increased. The static weights, real-time flow coefficients, and predicted flow coefficients are weighted and calculated by the weight coefficients to obtain the adjacent weights of each road section. The adjacent weights are used to measure the comprehensive carrying capacity and priority of the traffic in the corresponding road. The adjacent weight matrix is ​​constructed according to the adjacent weights of each road section.

[0038] The method provided in the embodiment of the present application also includes: evaluating the convoy propagation speed in the road within a preset time period based on the real-time traffic flow of the road and the predicted traffic flow of the road; obtaining the priority of the main roads and branch roads from the dynamic adjacency weight matrix, and calculating the phase difference between adjacent intersecting roads based on the priority of the main roads and branch roads and the convoy propagation speed; optimizing the green light time windows of the main roads and branch roads based on the traffic prediction results to ensure the continuity of the green wave band; and obtaining the regional green wave scheduling strategy based on the phase difference and the green light time window.

[0039] According to the change of the priority of the trunk-branch combined with the road network topology, the phase difference and time window adjustment optimization of the trunk and branch are performed to obtain the regional green wave scheduling strategy, including: according to the real-time traffic flow of the road and the predicted traffic flow of the road, the propagation speed of the convoy in the road within the preset time period is evaluated, and the convoy propagation speed is based on the current traffic flow data obtained by sensors, cameras and other equipment and the future traffic flow data calculated by the traffic prediction model. Further, the average speed of the convoy passing through each road section within the preset time period is calculated, and then the convoy propagation speed is obtained according to the average speed. When the convoy propagation speed is determined, the priority of the trunk road and the branch road is obtained from the dynamic adjacency weight matrix, and the priority is the adjacency weight of the corresponding road in the dynamic adjacency weight matrix. Then, according to the priority of the trunk road and the branch road and the convoy propagation speed, the phase difference between adjacent intersections is calculated. The phase difference refers to the time difference between the switching of the signal lights of two adjacent intersections, that is, the time difference between the green light of a certain intersection and the green light of the next intersection. Reasonable phase difference can ensure that the green light is continuous when the vehicle passes through multiple intersections, avoiding red light interference, thereby improving traffic efficiency. The formula for calculating the phase difference is the ratio of the road length to the propagation speed of the convoy minus the calculated result of the green light time of the previous intersection multiplied by 1 plus the difference in road priority at the traffic light intersection. Furthermore, based on the flow prediction results, the green light time windows of the trunk and branches are optimized to ensure the continuity of the green wave band. The green light time window is set accordingly based on the flow prediction results. The larger the flow prediction result, the longer the green light time window of the corresponding road is to meet the driving needs of high-flow vehicles. Specifically, a corresponding table of traffic flow and green light time window of each lane can be set, and the green light time window can be determined based on the corresponding table. The green light time window is the duration of the green light at the road intersection. Finally, based on the phase difference and green light time window, the regional green wave scheduling strategy is obtained.

[0040] The method provided in the embodiment of the present application also includes: screening target scheduling roads according to the dynamic adjacency weight matrix, the target scheduling roads are roads whose priorities meet a preset threshold; performing trunk-branch multi-intersection zoning based on the distribution of the target scheduling roads; performing execution status zoning verification on the regional target green wave band through tracking the difference between the real-time road traffic and the predicted road traffic, and when the regional target green wave band is not met, resetting the adjacency weight matrix to optimize and adjust the regional green wave scheduling strategy, wherein the regional green wave scheduling strategy optimization and adjustment includes dynamic adjustment of the phase difference and the green light time window; when the regional target green wave band is met, performing regional global green wave band evaluation and verification based on the real-time road traffic.

[0041] Obtaining the regional green wave scheduling strategy also includes: screening the target scheduling roads according to the dynamic adjacency weight matrix, the target scheduling roads are roads whose priorities meet the preset threshold, that is, the priority parameter is greater than or equal to the preset priority threshold. These roads are the focus of green wave belt scheduling, ensuring that the traffic flow of these roads can be optimized. Based on the distribution of the target scheduling roads, trunk-branch multi-intersection partitioning is performed, and the trunk-branch multi-intersection partitioning is: the target scheduling road is divided into several relatively independent optimization control areas based on the relationship between the trunk road and the branch road, with the intersection as the node.

[0042] Furthermore, by comparing the tracking difference between the real-time traffic flow and the predicted traffic flow, the actual implementation of the target green wave band can be judged. The tracking difference is the difference between the real-time traffic value and the predicted traffic value, which is used to evaluate the effectiveness of the current traffic control strategy. For example, the predicted traffic shows that the "Central Avenue" should have 1,200 vehicles per hour during the morning peak, while the real-time traffic shows 1,400 vehicles per hour, resulting in a difference of 200 vehicles, indicating that the actual traffic pressure exceeds the prediction and the current signal strategy needs to be optimized and adjusted. According to the tracking difference, the execution status of the regional target green wave band is partitioned and verified to determine whether the tracking difference is less than or equal to the preset tracking difference value. If it is less than, it means that the traffic pressure is small and can meet the regional target green wave band. On the contrary, the traffic pressure is large and cannot meet the regional target green wave band. When the regional target green wave band is not met, the latest real-time traffic data is collected again, the output of the traffic prediction model is updated, and the dynamic adjacency weight matrix is ​​reset, and the road priority is recalculated based on the new data. For example, during the school dismissal period, the measured and predicted traffic volume of the branch road is close to saturation. The dynamic adjacency weight matrix after data update is dynamically increased to 0.6, which is significantly higher than the priority of the main road. At this time, the corresponding green light time window is obtained based on the traffic volume data, and the main green wave bandwidth is compressed: the green wave bandwidth of the main road is adjusted from 60 seconds to 57 seconds according to the traffic volume data, and the traffic volume of the main road and the branch road is balanced. According to the phase difference and green light time window, the regional green wave scheduling strategy is obtained to ensure that vehicles on the branch road have priority. The green wave scheduling strategy is a linkage adjustment strategy set according to the acquired phase difference combined with the green light time window of the corresponding intersection road. If the phase difference is 50 seconds, the green light start time interval of the corresponding adjacent road should be 50 seconds to ensure that vehicles can pass through the adjacent intersection quickly and efficiently. The phase difference and green light time window of the roads in the region are dynamically adjusted to optimize the execution effect of the green wave band. Among them, the optimization and adjustment of the regional green wave scheduling strategy includes the dynamic adjustment of the phase difference and the green light time window. When the target green wave band of the region is met, it means that the current green wave scheduling strategy is effective and the existing scheme is kept stable. The regional global green wave band evaluation and verification is carried out based on the real-time road traffic, that is, real-time monitoring of traffic flow data and cyclic execution of state partition verification.

[0043] The technical solution provided by the embodiment of the present invention constructs a road network topology structure by extracting road vectors, traffic facility data and POI information in the region, wherein the road network topology structure includes trunk roads and branch roads; collects real-time road traffic based on the road network topology structure, and uses a traffic prediction model to predict traffic with POI information as input to obtain predicted road traffic, wherein the traffic prediction model is obtained by training and convergence using historical data; performs data alignment based on the real-time road traffic and the predicted road traffic, constructs a dynamic adjacency weight matrix to reflect the priority changes of the trunk-branch; performs phase difference and time window adjustment optimization on the trunk and branch based on the priority changes of the trunk-branch combined with the road network topology structure, and obtains a regional green wave scheduling strategy. This solves the technical problem that the traffic scheduling system in the prior art lacks dynamic response to real-time traffic and predicted traffic, and cannot effectively adjust the green wave belt strategy according to changes in traffic flow, resulting in low traffic efficiency and congestion during peak hours. The regional green wave scheduling method based on traffic flow prediction dynamically adjusts the priority, phase difference and green light time window of road signals by combining real-time traffic data collection with the predicted traffic model, thereby achieving optimal scheduling of regional green wave bands, significantly improving traffic smoothness and reducing traffic congestion.

[0044] Embodiment 2 is based on the same inventive concept as the regional green wave scheduling method based on traffic flow prediction in the above embodiment. Figure 3 As shown, the present application provides a regional green wave scheduling system based on traffic flow prediction, and the system and method embodiments in the embodiments of the present application are based on the same inventive concept. Among them, the system includes: a topology structure construction module 11, which is used to extract road vectors, traffic facility data and POI information in the region, and construct a road network topology structure, wherein the road network topology structure includes trunk roads and branch roads; a road flow prediction module 12, which is used to collect real-time road flow based on the road network topology structure, and use the flow prediction model to use POI information as input to perform flow prediction, and obtain road predicted flow, wherein the flow prediction model is obtained by training and convergence using historical data; a weight matrix acquisition module 13, which is used to align data according to the real-time road flow and the predicted road flow, and construct a dynamic adjacency weight matrix to reflect the priority changes of the trunk-branch; a scheduling strategy acquisition module 14, which is used to adjust and optimize the phase difference and time window of the trunk and branch according to the priority changes of the trunk-branch combined with the road network topology structure, and obtain a regional green wave scheduling strategy.

[0045] The specific configuration of the topology construction module 11 will be described in detail below. The topology construction module 11 may further include: extracting vector data of all roads in the area from the map data, including road geometry, number of lanes, and turning restrictions; collecting traffic facility data in the area, including the location and attribute information of traffic lights, isolation guardrails, bus stops, crosswalks, and cameras; obtaining the name, category, coordinates, address data, service radius, peak hours, and event data of POI points in the area, and establishing the POI information; defining the road network as a directed graph, extracting nodes and edges according to the positioning relationship based on the vector data of all roads, traffic facility data, and POI information, and constructing the road network topology structure.

[0046] The specific configuration of the topology structure building module 11 will be described in detail below. The topology structure building module 11 may further include: extracting nodes and edges according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads, and constructing the road network topology structure including: identifying road intersections, confluence / divergence points, and defining nodes; defining the connection relationship between nodes as edges according to the road vector data, wherein the number of lanes, speed limit, and turning limit of each edge define edge attributes; connecting nodes, edges, and edge attributes according to the road geometric connection relationship, constructing the road network topology structure, and marking attributes according to the location positioning relationship of the traffic facility data and POI information.

[0047] The specific configuration of the topology construction module 11 will be described in detail below. The topology construction module 11 may further include: constructing the road network topology structure, and also include: defining a determination rule for trunk roads and branch roads according to the number of lanes; determining the road grade of the lane number of each lane according to the determination rule, determining the road grade as a trunk lane or a branch lane, and marking the trunk roads and branch roads of the road network topology structure according to the road grade.

[0048] The specific configuration of the weight matrix acquisition module 13 will be described in detail below. The weight matrix acquisition module 13 further includes: performing data alignment according to the real-time road traffic and the predicted road traffic, and constructing a dynamic adjacency weight matrix, including: performing flow data time series alignment according to the time relationship between the real-time road traffic and the predicted road traffic; setting the update frequency of the real-time road traffic and the predicted road traffic, configuring a sliding window based on the update frequency, and updating the flow time series data according to the sliding window; and dynamically adjusting the adjacency weight matrix according to the updated flow time series data.

[0049] The specific configuration of the weight matrix acquisition module 13 will be described in detail below. The weight matrix acquisition module 13 further includes: dynamically adjusting the adjacent weight matrix, including: configuring the static weight of each road according to the historical traffic volume of the trunk road and the branch road; dividing the real-time traffic volume of the road and the predicted traffic volume of the road by the maximum traffic volume of the road, de-dimensionalizing and converting them into interval values ​​of [0,1] to obtain the real-time flow coefficient and the predicted flow coefficient; configuring the weight coefficients of different traffic time periods, performing weighted calculation on the static weight, real-time flow coefficient, and predicted flow coefficient according to the weight coefficient, obtaining the adjacent weight of each road section, and constructing the adjacent weight matrix.

[0050] The specific configuration of the scheduling strategy acquisition module 14 will be described in detail below. The scheduling strategy acquisition module 14 may further include: according to the priority change of the trunk-branch combined with the road network topology, the phase difference and time window adjustment optimization of the trunk and branch are performed to obtain the regional green wave scheduling strategy, including: according to the real-time traffic flow of the road and the predicted traffic flow of the road, the convoy propagation speed in the road within the preset time period is evaluated; the priority of the trunk road and the branch road is obtained from the dynamic adjacency weight matrix, and the phase difference between adjacent intersection roads is calculated according to the priority of the trunk road and the branch road and the convoy propagation speed; according to the traffic prediction results, the green light time window of the trunk and the branch is optimized to ensure the continuity of the green wave band; according to the phase difference and the green light time window, the regional green wave scheduling strategy is obtained.

[0051] The specific configuration of the scheduling strategy acquisition module 14 will be described in detail below. The scheduling strategy acquisition module 14 may further include: obtaining a regional green wave scheduling strategy, and also include: screening a target scheduling road according to the dynamic adjacency weight matrix, wherein the target scheduling road is a road whose priority meets a preset threshold; performing trunk-branch multi-crossing partitioning based on the distribution of the target scheduling road; performing execution status partitioning verification on the regional target green wave band through the tracking difference between the real-time road traffic and the predicted road traffic, and when the regional target green wave band is not met, resetting the adjacency weight matrix to optimize and adjust the regional green wave scheduling strategy, wherein the regional green wave scheduling strategy optimization and adjustment includes dynamic adjustment of the phase difference and the green light time window; when the regional target green wave band is met, performing regional global green wave band evaluation verification based on the real-time road traffic.

[0052] The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0053] Embodiment 3, based on the same inventive concept as the regional green wave scheduling method based on traffic flow prediction in the above-mentioned embodiment, this embodiment provides a computer-readable storage medium, which can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the regional green wave scheduling method based on traffic flow prediction in the embodiment of the present application. The processor executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory, that is, realizing the above-mentioned regional green wave scheduling method based on traffic flow prediction.

[0054] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. Regional green wave scheduling method based on traffic flow prediction, characterized in that: include: Extract road vectors, traffic facility data and POI information in the area, and construct a road network topology structure, which includes trunk roads and branch roads; Based on the road network topology structure, real-time road traffic is collected, and traffic prediction is performed using a traffic prediction model with POI information as input to obtain predicted road traffic, wherein the traffic prediction model is obtained by training and converging using historical data; Performing data alignment based on the real-time traffic volume of the road and the predicted traffic volume of the road, constructing a dynamic adjacency weight matrix to reflect the priority changes of the trunk and branches; According to the priority changes of the trunk and branches combined with the road network topology, the phase difference and time window adjustment and optimization of the trunk and branches are performed to obtain a regional green wave scheduling strategy.

2. The regional green wave scheduling method based on traffic flow prediction according to claim 1 is characterized in that: Construct the road network topology, including: Extract vector data of all roads in the area from the map data, including road geometry, number of lanes, and turn restrictions; Collect data on traffic facilities in the area, including the location and attribute information of traffic lights, isolation barriers, bus stops, crosswalks, and cameras; Obtain the name, category, coordinates, address data, service radius, peak hours, and event data of POI points in the area, and establish the POI information; The road network is defined as a directed graph, and nodes and edges are extracted according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads to construct the road network topology structure.

3. The regional green wave scheduling method based on traffic flow prediction according to claim 2 is characterized in that: Extracting nodes and edges according to the positioning relationship based on the vector data, traffic facility data, and POI information of all roads, and constructing the road network topology structure includes: Identify road intersections, confluence / divergence points, and define nodes; Based on the road vector data, the connection relationship between nodes is defined as the edge, where the number of lanes, speed limit, and turn limit of each edge define the edge attributes; According to the geometric connection relationship of the roads, the nodes, edges and edge attributes are connected to construct the road network topology structure, and the attributes are marked according to the location positioning relationship of the traffic facility data and POI information.

4. The regional green wave scheduling method based on traffic flow prediction according to claim 3 is characterized in that: Constructing the road network topology structure also includes: Define the rules for determining trunk and branch roads based on the number of lanes; The lane number of each lane is determined by road grade according to the determination rule, the road grade is determined to be a trunk lane or a branch lane, and the road network topology structure is marked as a trunk road and a branch road according to the road grade.

5. The regional green wave scheduling method based on traffic flow prediction according to claim 1 is characterized in that: Data alignment is performed according to the real-time traffic volume of the road and the predicted traffic volume of the road to construct a dynamic adjacency weight matrix, including: Performing time series alignment of traffic data according to the time relationship between the real-time traffic of the road and the predicted traffic of the road; Setting an update frequency of the road real-time traffic and the road predicted traffic, configuring a sliding window based on the update frequency, and updating the traffic time series data according to the sliding window; The adjacency weight matrix is ​​dynamically adjusted according to the updated traffic timing data.

6. The regional green wave scheduling method based on traffic flow prediction according to claim 5 is characterized in that: Dynamically adjust the adjacency weight matrix, including: According to the historical traffic volume of the trunk roads and branch roads, a static weight of each road is configured; The real-time flow rate of the road and the predicted flow rate of the road are divided by the maximum flow rate of the road respectively, and the values ​​are converted into interval values ​​of [0,1] without dimensioning to obtain a real-time flow rate coefficient and a predicted flow rate coefficient; The weight coefficients of different traffic time periods are configured, and the static weight, real-time flow coefficient, and predicted flow coefficient are weightedly calculated according to the weight coefficients to obtain the adjacency weight of each road section and construct the adjacency weight matrix.

7. The regional green wave scheduling method based on traffic flow prediction according to claim 6 is characterized in that: According to the trunk-branch priority change and the road network topology, the trunk and branch are adjusted and optimized for phase difference and time window to obtain a regional green wave scheduling strategy, including: According to the real-time traffic volume of the road and the predicted traffic volume of the road, evaluating the propagation speed of the convoy on the road within a preset time period; Obtaining the priorities of the trunk roads and the branch roads from the dynamic adjacency weight matrix, and calculating the phase difference between adjacent intersection roads according to the priorities of the trunk roads and the branch roads and the propagation speed of the fleet; According to the traffic forecast results, optimize the green light time window of the trunk and branch to ensure the continuity of the green wave band; The regional green wave scheduling strategy is obtained according to the phase difference and the green light time window.

8. The regional green wave scheduling method based on traffic flow prediction according to claim 7 is characterized in that: Obtain regional green wave scheduling strategy, including: According to the dynamic adjacency weight matrix, a target scheduling road is selected, wherein the target scheduling road is a road whose priority meets a preset threshold; Based on the distribution of the target dispatching roads, multiple intersection partitions of trunks and branches are performed; By tracking the difference between the real-time road traffic and the predicted road traffic, the regional target green wave band is verified for execution status partitioning. When the regional target green wave band is not met, the adjacent weight matrix is ​​reset to optimize and adjust the regional green wave scheduling strategy, where the regional green wave scheduling strategy optimization adjustment includes dynamic adjustment of the phase difference and the green light time window; When the regional target green wave band is met, the regional global green wave band evaluation and verification is performed based on the real-time road traffic.

9. Regional green wave dispatching system based on traffic flow prediction, characterized by: The system is used to implement the regional green wave scheduling method based on traffic flow prediction according to any one of claims 1 to 8, and the system includes: A topological structure construction module is used to extract road vectors, traffic facility data and POI information in the region and construct a road network topological structure, wherein the road network topological structure includes trunk roads and branch roads; A road traffic prediction module is used to collect real-time road traffic based on the road network topology structure, and use a traffic prediction model to perform traffic prediction with POI information as input to obtain predicted road traffic, wherein the traffic prediction model is obtained by training and converging with historical data; A weight matrix acquisition module is used to align data according to the real-time traffic volume of the road and the predicted traffic volume of the road, and to construct a dynamic adjacency weight matrix to reflect the priority changes of the trunk and the branch. The scheduling strategy acquisition module is used to optimize the phase difference and time window adjustment of the trunk and branches according to the priority changes of the trunk and branches combined with the road network topology to obtain the regional green wave scheduling strategy.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the regional green wave scheduling method based on traffic flow prediction as described in any one of claims 1-8 is implemented.

Images