AI-based urban traffic flow prediction and dynamic signal optimization method and system

Abstract

The invention discloses an AI-based urban traffic flow prediction and dynamic signal optimization method and system, and relates to the technical field of traffic management, and the method comprises the steps: obtaining real-time traffic data, historical flow data and external environment data of a target region, obtaining multi-source data, and constructing a space-time matrix corresponding to the target region according to the multi-source data; learning the space-time matrix by using a preset space-time fusion model, and outputting a prediction result of the traffic flow in a future preset time period through the space-time fusion model obtained by learning; wherein the space-time fusion model is a combined architecture of a graph convolutional network and a Transform network; and dynamically adjusting traffic signal control parameters in the target area through a multi-agent learning algorithm based on a prediction result. According to the method, closed-loop regulation and control of'data perception-prediction modeling-autonomous decision 'are formed, so that the limitation of spatial feature modeling staticization and time correlation analysis fragmentation of a traditional method is broken through, and the problem of regional imbalance caused by single-point optimization is effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of traffic management technology, and in particular to an AI-based urban traffic flow prediction and dynamic signal optimization method and system. Background Art

[0002] Amid the rapid development of smart cities, traditional traffic management models are struggling to cope with the systemic congestion caused by the surge in motor vehicle ownership. Existing traffic control systems generally employ a "detect-rule-response" mechanism. This single-dimensional data model and delayed decision-making have become key bottlenecks hindering the modernization of urban governance.

[0003] Existing traffic flow prediction models often independently process spatial topology and time series features, leading to two core issues: First, static road network modeling fails to reflect the dynamic attenuation of actual traffic efficiency (such as the chain congestion at related intersections caused by reduced vehicle speeds during rainy days); second, time series models such as LSTMs have long response delays to emergencies, making it difficult to meet real-time control requirements. Furthermore, traditional signal optimization uses a centralized computing architecture. When the number of intersections is large, the time required for collaborative decision-making increases exponentially, leading to high regional imbalance rates. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an AI-based urban traffic flow prediction and dynamic signal optimization method and system, aiming to solve the problems in the existing technology.

[0005] A first aspect of the present invention is to provide an AI-based urban traffic flow prediction and dynamic signal optimization method, the method comprising:

[0006] Acquire real-time traffic data, historical flow data, and external environment data of a target area to obtain multi-source data, and construct a spatiotemporal matrix corresponding to the target area based on the multi-source data;

[0007] The spatiotemporal matrix is ​​learned using a preset spatiotemporal fusion model, and the spatiotemporal fusion model obtained through learning outputs a traffic flow prediction result within a preset time period in the future; wherein the spatiotemporal fusion model is a joint architecture of a graph convolutional network and a Transformer network;

[0008] Based on the prediction results, traffic signal control parameters in the target area are dynamically adjusted through a multi-agent learning algorithm.

[0009] According to one aspect of the above technical solution, the steps of acquiring real-time traffic data, historical traffic data, and external environment data of a target area to obtain multi-source data and constructing a spatiotemporal matrix corresponding to the target area based on the multi-source data include:

[0010] Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data;

[0011] Extracting road network topology and time series data respectively according to the multi-source data;

[0012] Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data;

[0013] A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

[0014] According to one aspect of the above technical solution, the steps of learning the space-time matrix using a preset space-time fusion model and outputting a traffic flow prediction result for a preset future time period using the learned space-time fusion model include:

[0015] Using a preset spatiotemporal fusion model to learn the spatial weights and periodic features in the spatiotemporal matrix to obtain a learned spatiotemporal fusion model;

[0016] The spatiotemporal fusion model is used to output the traffic flow prediction results within a preset time period in the future.

[0017] According to one aspect of the above technical solution, the step of dynamically adjusting the traffic signal control parameters in the target area through a multi-agent learning algorithm based on the prediction results includes:

[0018] Based on the prediction results, output node features of each traffic control node in the target area through a multi-agent learning algorithm;

[0019] Control each traffic control node for centralized training and decentralized execution, output the phase / green light duration of each traffic control node, and evaluate the global value score of each traffic control node;

[0020] According to the global value score of the traffic control node, the phase / green light duration corresponding to the traffic control node is adjusted according to the time period characteristics.

[0021] According to one aspect of the above technical solution, the step of adjusting the phase / green light duration corresponding to the traffic control node according to the global value score of the traffic control node and the time period characteristics includes:

[0022] receiving a traffic prediction result corresponding to a target area from a cloud at a first preset time interval, so as to update a global value score of the traffic control node according to the traffic prediction result;

[0023] The control edge end regenerates a signal control scheme for each traffic control node every second preset time, wherein the signal control scheme includes a phase selection probability and a green light duration correction value.

[0024] According to one aspect of the above technical solution, the control edge terminal regenerates the signal control scheme for each traffic control node every second preset time, wherein the signal control scheme includes a phase selection probability and a green light duration correction value, including:

[0025] The control edge runs the Actor network every second preset time and regenerates the signal control plan for each traffic control node through the Actor network;

[0026] Among them, the Actor network outputs the action probability distribution by inputting the local state and neighbor state in the target area. The action probability distribution includes the phase selection probability and the green light duration correction value.

[0027] A second aspect of the present invention is to provide an AI-based urban traffic flow prediction and dynamic signal optimization system, which is applied to the method described in the above technical solution, and the system includes:

[0028] A data acquisition module is used to acquire real-time traffic data, historical flow data and external environment data of the target area to obtain multi-source data, and to construct a spatiotemporal matrix corresponding to the target area based on the multi-source data;

[0029] A prediction output module is configured to learn the spatiotemporal matrix using a preset spatiotemporal fusion model, and output a traffic flow prediction result for a preset future time period using the spatiotemporal fusion model obtained through learning; wherein the spatiotemporal fusion model is a combined architecture of a graph convolutional network and a Transformer network;

[0030] A dynamic optimization module is used to dynamically adjust the traffic signal control parameters in the target area based on the prediction results through a multi-agent learning algorithm.

[0031] According to one aspect of the above technical solution, the data acquisition module is used to:

[0032] Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data;

[0033] Extracting road network topology and time series data respectively according to the multi-source data;

[0034] Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data;

[0035] A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

[0036] A third aspect of the present invention is to provide a readable storage medium having a computer program stored thereon, which implements the method described in the above technical solution when executed by a processor.

[0037] The fourth aspect of the present invention is to provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above technical solution when executing the computer program.

[0038] Compared with the existing technology, the AI-based urban traffic flow prediction and dynamic signal optimization method and system shown in this embodiment has the following beneficial effects:

[0039] The present invention combines graph convolutional networks to extract dynamic spatial dependencies of road networks and Transformer to analyze multi-scale temporal features to achieve accurate traffic flow prediction. On this basis, a multi-agent collaborative control architecture with centralized training and distributed execution is adopted to enable each intersection to optimize signal parameters in real time based on local status and global reward signals, forming a closed-loop regulation of "data perception-predictive modeling-autonomous decision-making", thereby breaking through the limitations of static spatial feature modeling and fragmented temporal correlation analysis of traditional methods, effectively solving the regional imbalance problem caused by single-point optimization, and at the same time reducing decision-making delays through edge computing, achieving the triple technical effects of improved prediction accuracy, enhanced real-time control, and optimized regional traffic efficiency, and systematically improving urban traffic congestion problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments with reference to the accompanying drawings, in which:

[0041] Figure 1 Schematic diagram of a flow chart of an AI-based urban traffic flow prediction and dynamic signal optimization method according to one embodiment of the present invention;

[0042] Figure 2 This is a structural block diagram of an AI-based urban traffic flow prediction and dynamic signal optimization system shown in one embodiment of the present invention. DETAILED DESCRIPTION

[0043] To make the objectives, features, and advantages of the present invention more readily apparent, the following detailed description of specific embodiments of the present invention is provided in conjunction with the accompanying drawings. The accompanying drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0045] Example 1

[0046] See also Figure 1 The first aspect of the present invention provides an AI-based urban traffic flow prediction and dynamic signal optimization method. The method shown in this embodiment includes steps S10 to S30:

[0047] Step S10 , acquiring real-time traffic data, historical traffic data, and external environment data of the target area to obtain multi-source data, and constructing a spatiotemporal matrix corresponding to the target area based on the multi-source data.

[0048] In this embodiment, the target area is the area where traffic flow prediction and dynamic signal optimization are required. It is usually a key area within the city, which usually has urban main roads and various supporting facilities. Traffic flow changes dynamically with time periods. For example, areas where shopping complexes and schools are located are the target areas shown in this embodiment.

[0049] The multi-source data includes:

[0050] Real-time traffic data: traffic flow and speed information from cameras, geomagnetic sensors, and vehicle GPS;

[0051] External environmental data: weather conditions, holiday signs, emergencies, and road construction plans;

[0052] Historical traffic data: records of intersection traffic volume and signal phases stored in the form of time slices.

[0053] Specifically, in this embodiment, by obtaining real-time traffic data, historical flow data and external environment data of the target area, the real-time traffic data, historical flow data and external environment data are integrated to obtain multi-source data, and then a space-time matrix corresponding to the target area is constructed based on the multi-source data. The constructed space-time matrix is ​​used to express the correspondence between traffic flow and time and space in the target area.

[0054] More specifically, the real-time traffic data of the target area is obtained based on cameras, geomagnetic sensors, and vehicle-mounted GPS, and usually includes the real-time average speed of vehicles in the target area and the average waiting time of vehicles at intersections. The historical traffic data obtained includes the average speed of vehicles in the past and the average waiting time of vehicles at intersections. The external environment data mainly includes the supporting conditions of the target area and records the spatial information within the target area. The above data are then fused to obtain multi-source data, and based on this, a space-time matrix is ​​constructed to express the dual correspondence between traffic flow and time and space in the target area.

[0055] Step S20: Using a preset space-time fusion model to learn the space-time matrix, the space-time fusion model obtained through learning outputs a prediction result of traffic flow in a preset time period in the future.

[0056] Among them, the spatiotemporal fusion model is a joint architecture of graph convolutional network and Transformer network.

[0057] Specifically, after obtaining real-time traffic data, historical flow data and external environment data of the target area to obtain multi-source data and construct a space-time matrix, the preset space-time fusion model will be used to learn the pre-constructed space-time matrix. The space-time fusion model obtained through learning will output the predicted results of the traffic flow in the target area within the preset time period in the future.

[0058] In this embodiment, the steps of learning the space-time matrix using a preset space-time fusion model and outputting a traffic flow prediction result within a preset time period in the future using the learned space-time fusion model include:

[0059] Using a preset spatiotemporal fusion model to learn the spatial weights and periodic features in the spatiotemporal matrix to obtain a learned spatiotemporal fusion model;

[0060] The spatiotemporal fusion model is used to output the traffic flow prediction results within a preset time period in the future.

[0061] More specifically, in this embodiment, a preset space-time fusion model is used to learn the spatial weights and periodic features in the space-time matrix to obtain a learned space-time fusion model, so that the learned space-time fusion model can more accurately extract features based on multi-source data to predict traffic flow.

[0062] Step S30: Based on the prediction results, dynamically adjust the traffic signal control parameters in the target area through a multi-agent learning algorithm.

[0063] Among them, after predicting the predicted structure of traffic flow in the target area through the learned spatiotemporal fusion model, the multi-agent learning algorithm will be used to dynamically adjust the traffic control signal parameters of multiple traffic control nodes in the target area, such as adjusting the red light duration / green light duration of traffic lights in single or multiple traffic control nodes, the turn signals of variable lanes, etc.

[0064] Specifically, when dynamically adjusting traffic control signal parameters through a multi-agent learning algorithm, the traffic control nodes are regarded as intelligent agents. First, the node characteristics of each traffic control node in the target area are output based on the traffic flow prediction results, and each traffic control node is controlled for centralized training and decentralized execution. The phase / green light duration of each traffic control node is output, and the global value score of each traffic control node is evaluated based on the phase / green light duration of each traffic control node. This usually depends on the location of each traffic control node and the supporting facilities in the surrounding area. Then, according to the global value score of each traffic control node, the global value scores corresponding to all traffic control nodes in the target area are sorted, and the phase / green light duration corresponding to each traffic control node is dynamically adjusted according to the time period characteristics.

[0065] More specifically, the cloud, or the backend server, re-predicts the traffic prediction results corresponding to the target area at the current time node based on the multi-source data previously obtained every first preset time, for example, every 5-10 minutes, and receives the traffic prediction results corresponding to the target area synchronously sent by the cloud, or the backend server, in real time, and then dynamically updates the global value score of each traffic control node according to the current traffic prediction results. After updating the global value score of each traffic control node, the control edge end regenerates the signal control plan for each traffic control node according to the updated global value score every second preset time, wherein the second preset time is the same as the first preset time, and the signal control plan includes the phase selection probability and the green light duration correction value, and then executes the signal control plan to achieve dynamic adjustment of the traffic signal.

[0066] The process of regenerating the signal control plan for each traffic control node at the control edge every second preset time period specifically includes the control edge running an actor network every second preset time period to regenerate the signal control plan for each traffic control node through the running of the actor network. The actor network traverses and queries the local state of a traffic control node in the target area and the neighbor states of other traffic control nodes in the target area, and outputs an action probability distribution. The action probability distribution includes a phase selection probability and a green light duration correction value. Specifically, the green light duration correction value for a traffic control node is calculated based on the phase selection probability of the traffic control node. For example, if the green light duration of the traffic control node's signal light is previously 50 seconds, then after querying the neighbor states of other traffic control nodes, it can be determined that the traffic flow in the target area has significantly increased or decreased. Based on the relationship between the traffic control node and the other traffic control nodes, the green light duration of the traffic control node's signal light is corrected, for example, to 80 seconds or 30 seconds, so that the traffic control node's corresponding signal light has a longer or shorter green light duration.

[0067] For example, when a traffic accident occurs at a certain intersection and causes congestion, the method shown in this embodiment quickly obtains the current multi-source data and constructs a space-time matrix, and then uses the space-time fusion model to predict the traffic flow in the target area. Finally, combined with the multi-agent learning algorithm, the phase selection probability of the upstream and downstream traffic control nodes and the green light duration of the traffic lights in the target area are dynamically adjusted. By quickly responding and formulating the optimal signal control plan, and combining it with the corresponding traffic condition feedback, it can effectively guide traffic and avoid worsening congestion.

[0068] Compared with the existing technology, the AI-based urban traffic flow prediction and dynamic signal optimization method shown in this embodiment has the following beneficial effects:

[0069] This embodiment combines a graph convolutional network to extract dynamic spatial dependencies in road networks and a Transformer to analyze multi-scale temporal features to achieve accurate traffic flow prediction. Furthermore, a multi-agent collaborative control architecture with centralized training and distributed execution is adopted, enabling each intersection to optimize signal parameters in real time based on local status and global reward signals, forming a closed-loop control system of "data perception-predictive modeling-autonomous decision-making." This overcomes the limitations of traditional methods, such as static spatial feature modeling and fragmented temporal correlation analysis, effectively resolving regional imbalances caused by single-point optimization. Furthermore, edge computing is used to reduce decision latency, achieving the triple technical benefits of improved prediction accuracy, enhanced real-time control, and optimized regional traffic efficiency, systematically alleviating urban traffic congestion.

[0070] Example 2

[0071] The second embodiment of the present invention also provides an AI-based urban traffic flow prediction and dynamic signal optimization method. The method shown in this embodiment is basically similar to the method shown in the first embodiment, except that:

[0072] In this embodiment, the steps of acquiring real-time traffic data, historical traffic data, and external environment data of a target area to obtain multi-source data and constructing a spatiotemporal matrix corresponding to the target area based on the multi-source data include:

[0073] Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data;

[0074] Extracting road network topology and time series data respectively according to the multi-source data;

[0075] Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data;

[0076] A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

[0077] Specifically, after obtaining the real-time traffic data, historical flow data and external environment data of the target area and fusing them to obtain multi-source data, the road network topology structure corresponding to the target area and the time series data of the corresponding traffic flow will be extracted based on the multi-source data. All intersections in the target area will be modeled into multiple graph nodes through the road network topology structure, and the spatial weights between adjacent intersections will be extracted. Then, the time series data will be periodically decomposed to extract the periodic features in the time series data. Finally, based on the spatial weight of each intersection and the periodic features of the corresponding traffic flow, a space-time matrix corresponding to the target area will be constructed. The space-time matrix will be used to express the dual correspondence between the traffic flow at each intersection in the target area and time and space.

[0078] More specifically, before each adjustment to the signal control scheme, it is necessary to obtain real-time multi-source data in the target area and then construct the corresponding space-time matrix to facilitate traffic flow prediction through the pre-built space-time fusion model, thereby dynamically adjusting the signal control scheme.

[0079] Example 3

[0080] See also Figure 2 The third embodiment of the present invention provides an AI-based urban traffic flow prediction and dynamic signal optimization system, which is applied to the method described in any of the above embodiments, and the system includes:

[0081] A data acquisition module 10 is used to acquire real-time traffic data, historical traffic data, and external environment data of a target area to obtain multi-source data, and to construct a spatiotemporal matrix corresponding to the target area based on the multi-source data;

[0082] A prediction output module 20 is configured to learn the spatiotemporal matrix using a preset spatiotemporal fusion model, and output a traffic flow prediction result for a preset future time period using the spatiotemporal fusion model obtained through learning; wherein the spatiotemporal fusion model is a combined architecture of a graph convolutional network and a Transformer network;

[0083] The dynamic optimization module 30 is used to dynamically adjust the traffic signal control parameters in the target area based on the prediction results through a multi-agent learning algorithm.

[0084] Wherein, the data acquisition module 10 is used to:

[0085] Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data;

[0086] Extracting road network topology and time series data respectively according to the multi-source data;

[0087] Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data;

[0088] A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

[0089] Compared with the existing technology, the AI-based urban traffic flow prediction and dynamic signal optimization system shown in this embodiment has the following beneficial effects:

[0090] This embodiment combines a graph convolutional network to extract dynamic spatial dependencies in road networks and a Transformer to analyze multi-scale temporal features to achieve accurate traffic flow prediction. Furthermore, a multi-agent collaborative control architecture with centralized training and distributed execution is adopted, enabling each intersection to optimize signal parameters in real time based on local status and global reward signals, forming a closed-loop control system of "data perception-predictive modeling-autonomous decision-making." This overcomes the limitations of traditional methods, such as static spatial feature modeling and fragmented temporal correlation analysis, effectively resolving regional imbalances caused by single-point optimization. Furthermore, edge computing is used to reduce decision latency, achieving the triple technical benefits of improved prediction accuracy, enhanced real-time control, and optimized regional traffic efficiency, systematically alleviating urban traffic congestion.

[0091] Example 4

[0092] A fourth embodiment of the present invention provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method described in any one of the above embodiments is implemented.

[0093] Example 5

[0094] A fifth embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any one of the above embodiments when executing the computer program.

[0095] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0096] The above-described embodiments merely illustrate several implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. An AI-based urban traffic flow prediction and dynamic signal optimization method, characterized in that: The method comprises: Acquire real-time traffic data, historical flow data, and external environment data of a target area to obtain multi-source data, and construct a spatiotemporal matrix corresponding to the target area based on the multi-source data; The spatiotemporal matrix is ​​learned using a preset spatiotemporal fusion model, and the spatiotemporal fusion model obtained through learning outputs a traffic flow prediction result within a preset time period in the future; wherein the spatiotemporal fusion model is a joint architecture of a graph convolutional network and a Transformer network; Based on the prediction results, traffic signal control parameters in the target area are dynamically adjusted through a multi-agent learning algorithm.

2. The AI-based urban traffic flow prediction and dynamic signal optimization method according to claim 1 is characterized in that: The steps of acquiring real-time traffic data, historical traffic data, and external environment data of a target area to obtain multi-source data and constructing a spatiotemporal matrix corresponding to the target area based on the multi-source data include: Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data; Extracting road network topology and time series data respectively according to the multi-source data; Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data; A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

3. The AI-based urban traffic flow prediction and dynamic signal optimization method according to claim 1 is characterized in that: The steps of learning the space-time matrix using a preset space-time fusion model and outputting a prediction result of traffic flow in a future preset time period using the learned space-time fusion model include: Using a preset spatiotemporal fusion model to learn the spatial weights and periodic features in the spatiotemporal matrix to obtain a learned spatiotemporal fusion model; The spatiotemporal fusion model is used to output the traffic flow prediction results within a preset time period in the future.

4. The AI-based urban traffic flow prediction and dynamic signal optimization method according to claim 1 is characterized in that: The step of dynamically adjusting the traffic signal control parameters within the target area by a multi-agent learning algorithm based on the prediction result includes: Based on the prediction results, output node features of each traffic control node in the target area through a multi-agent learning algorithm; Control each traffic control node for centralized training and decentralized execution, output the phase / green light duration of each traffic control node, and evaluate the global value score of each traffic control node; According to the global value score of the traffic control node, the phase / green light duration corresponding to the traffic control node is adjusted according to the time period characteristics.

5. The AI-based urban traffic flow prediction and dynamic signal optimization method according to claim 4 is characterized in that: The step of adjusting the phase / green light duration corresponding to the traffic control node according to the global value score of the traffic control node and the time period characteristics includes: receiving a traffic prediction result corresponding to a target area from a cloud at a first preset time interval, so as to update a global value score of the traffic control node according to the traffic prediction result; The control edge end regenerates a signal control scheme for each traffic control node every second preset time, wherein the signal control scheme includes a phase selection probability and a green light duration correction value.

6. The AI-based urban traffic flow prediction and dynamic signal optimization method according to claim 5 is characterized in that: The control edge end regenerates a signal control scheme for each traffic control node every second preset time, wherein the signal control scheme includes a phase selection probability and a green light duration correction value, including: The control edge runs the Actor network every second preset time and regenerates the signal control plan for each traffic control node through the Actor network; Among them, the Actor network outputs the action probability distribution by inputting the local state and neighbor state in the target area. The action probability distribution includes the phase selection probability and the green light duration correction value.

7. An AI-based urban traffic flow prediction and dynamic signal optimization system, characterized by: The method according to any one of claims 1 to 6, wherein the system comprises: A data acquisition module is used to acquire real-time traffic data, historical flow data and external environment data of the target area to obtain multi-source data, and to construct a spatiotemporal matrix corresponding to the target area based on the multi-source data; A prediction output module is used to learn the spatiotemporal matrix using a preset spatiotemporal fusion model, and output a prediction result of traffic flow within a preset time period in the future using the spatiotemporal fusion model obtained by learning; wherein the spatiotemporal fusion model is a joint architecture of a graph convolutional network and a Transformer network; A dynamic optimization module is used to dynamically adjust the traffic signal control parameters in the target area based on the prediction results through a multi-agent learning algorithm.

8. The AI-based urban traffic flow prediction and dynamic signal optimization system according to claim 7 is characterized in that: The data acquisition module is used for: Acquire real-time traffic data, historical flow data, and external environment data of a target area, and fuse the real-time traffic data, historical flow data, and external environment data to obtain multi-source data; Extracting road network topology and time series data respectively according to the multi-source data; Modeling intersections within a target area as graph nodes using the road network topology, extracting spatial weights between adjacent intersections, and performing periodic decomposition on the time series data to extract periodic features from the time series data; A space-time matrix corresponding to the target area is constructed according to the spatial weight and the periodic characteristics.

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

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images