A smart traffic management method and system

By acquiring and analyzing traffic data, training a hierarchical model for intelligent signal light control, solving urban road congestion problems and improving traffic efficiency and quality of life.

CN115762149BActive Publication Date: 2024-11-22TIANDU SATELLITE APPL (CHANGCHUN) CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202211400513.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-11-09
Publication Date
2024-11-22
Estimated Expiration
2042-11-09

AI Technical Summary

Technical Problem

The existing urban road traffic light control method cannot be flexibly and intelligently regulated, which makes it difficult to effectively solve traffic congestion problems and affects traffic efficiency and quality of life.

Method used

By obtaining historical and real-time traffic data, training traffic intersection hierarchical models, using local and global neural network models for classification control, optimizing signal light cycles to reduce congestion and improve dredging speeds.

Benefits of technology

It realizes intelligent classification control based on real-time traffic information, reduces the frequency of traffic congestion and dredging speed, improves road traffic efficiency, and reduces energy waste and pollution.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115762149B_ABST
    Figure CN115762149B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of traffic management technology, and in particular to a smart traffic management method and system. First, a traffic intersection classification model is trained based on a historical data set to obtain the trained traffic intersection classification model; then, data of each traffic intersection is acquired according to the set acquisition cycle, and each traffic intersection is classified and categorized based on each traffic data and the trained traffic intersection classification model, and finally, the traffic lights of each traffic intersection are classified and controlled based on the classification results. The present invention classifies and controls the traffic lights of each traffic intersection according to real-time traffic information, reduces the occurrence of traffic congestion and improves the speed of clearing traffic congestion.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of traffic management, and in particular to an intelligent traffic management method and system. Background Art

[0002] With the rapid development of economy and the continuous acceleration of urbanization, all aspects are facing great pressure, especially urban transportation. With the increase of urban population and vehicles, traffic congestion has become a common phenomenon. The impacts of traffic congestion are as follows:

[0003] 1) Increased commuting time reduces the time available for work (production), causing economic losses to drivers and the region.

[0004] 2) It causes drivers to feel angry and irritated, increasing their stress and further causing adverse effects on their health.

[0005] 3) Waste of fuel and pollution: The engine continues to run in traffic jams, continuously consuming fuel. In addition, when there is a traffic jam, vehicles must constantly accelerate and brake, increasing fuel consumption. Therefore, traffic jams not only waste energy, but also cause air pollution.

[0006] 4) It causes the quality of life in urban areas to decline, resulting in a large number of residents moving to the suburbs (the so-called suburbanization).

[0007] 5) Difficulty in responding to emergencies: When there is an emergency, it may be difficult to reach the destination due to traffic congestion.

[0008] At present, there are two ways to control traffic lights at urban road intersections: timed control and irregular control. Among them, timed control refers to controlling traffic lights with the same signal cycle at any time, without considering the real-time changes in traffic flow at the intersection, which is prone to congestion. Irregular control refers to controlling traffic lights with different signal cycles set in different time periods, such as peak hours or non-peak hours. This method improves the traffic efficiency of the road to a certain extent, but the road conditions themselves are relatively complex and changeable. This control method is still not flexible enough and the effect is still not very obvious. Therefore, how to flexibly and intelligently control traffic lights to improve road traffic efficiency has become a technical problem that needs to be solved urgently. Summary of the invention

[0009] In view of this, the present invention provides an intelligent traffic management method and system, which performs classification control according to real-time traffic information, reduces the occurrence of traffic congestion and improves the speed of clearing traffic congestion.

[0010] To achieve the above object, the present invention provides the following solutions:

[0011] A smart traffic management method, comprising:

[0012] Acquire a historical data set, and segment the historical data set according to a set acquisition cycle to obtain a plurality of segmented data sets;

[0013] Training a traffic intersection classification model based on each of the segmented data sets to obtain a trained traffic intersection classification model;

[0014] Acquire real-time data of each traffic intersection according to the set collection period, and obtain traffic data based on the real-time data;

[0015] Classifying each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection;

[0016] Classifying each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set;

[0017] Based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model, a local optimal control set corresponding to each traffic intersection in the local optimal set is obtained; based on each of the local optimal control sets, a traffic light at each traffic intersection in the local optimal set is controlled;

[0018] Based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model, a global optimal control set corresponding to each traffic intersection in the global optimal set is obtained; and based on each of the global optimal control sets, the traffic lights at each traffic intersection in the global optimal set are controlled.

[0019] Preferably, the historical data set is obtained, and the historical data set is segmented according to a set acquisition period to obtain a plurality of segmented data sets, including:

[0020] Acquire an initial historical data set; the initial historical data set includes historical video data, historical important building data, and historical traffic intersection grade data;

[0021] Based on the historical video data, historical traffic data is obtained; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection level in the historical traffic intersection level data includes expressway intersection, main road intersection, secondary road intersection and branch road intersection;

[0022] The historical data set is segmented according to a set collection cycle to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

[0023] Preferably, the training of the traffic intersection classification model based on each of the segmented data sets to obtain the trained traffic intersection classification model includes:

[0024] Inputting any of the segmented data sets into the traffic intersection classification model to obtain a training classification result for each traffic intersection;

[0025] Classifying each traffic intersection based on the training classification results to obtain a training local optimal set and a training global optimal set;

[0026] Based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, a training local optimal control set corresponding to each traffic intersection in the training local optimal set is obtained;

[0027] Based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, a training global optimal control set corresponding to the traffic intersection in the training global optimal set is obtained;

[0028] Perform a traffic virtual simulation experiment based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result;

[0029] Evaluate the degree of solving traffic congestion based on the virtual simulation experiment results to obtain evaluation results;

[0030] Adjusting the parameters of the traffic intersection classification model based on the evaluation results;

[0031] The above process is repeatedly performed to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

[0032] Preferably, the traffic data includes vehicle flow and traffic retention.

[0033] The present invention also provides a smart traffic management system, comprising: a data acquisition module, a model training module, a plurality of edge terminal modules and a central processing unit;

[0034] The data acquisition module is used to acquire a historical data set, and segment the historical data set according to a set acquisition cycle to obtain a plurality of segmented data sets;

[0035] The model training module is used to train the traffic intersection classification model based on each segmented data set to obtain the trained traffic intersection classification model;

[0036] One edge terminal module is provided at each traffic intersection; each edge terminal module is used to obtain real-time data of each traffic intersection according to the set collection period, and obtain traffic data based on the real-time data;

[0037] Each of the edge terminal modules is also used to classify each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection;

[0038] The central processing unit is used to classify each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set;

[0039] The central processor is also used to obtain a global optimal control set corresponding to each traffic intersection in the global optimal set based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model;

[0040] The central processor is also used to control the signal lights at each traffic intersection in the global optimal set based on each of the global optimal control sets;

[0041] Each of the edge terminal modules is further used to obtain a local optimal control set corresponding to each traffic intersection in the local optimal set based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model;

[0042] Each of the edge terminal modules is also used to control the traffic lights at each traffic intersection in the local optimal set based on each of the local optimal control sets.

[0043] Preferably, the data acquisition module is specifically:

[0044] Acquire an initial historical data set; the initial historical data set includes historical video data, historical important building data, and historical traffic intersection grade data;

[0045] Based on the historical video data, historical traffic data is obtained; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection level in the historical traffic intersection level data includes expressway intersection, main road intersection, secondary road intersection and branch road intersection;

[0046] The historical data set is segmented according to a set collection cycle to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

[0047] Preferably, the model training module is specifically:

[0048] Inputting any of the segmented data sets into the traffic intersection classification model to obtain a training classification result for each traffic intersection;

[0049] Classifying each traffic intersection based on the training classification results to obtain a training local optimal set and a training global optimal set;

[0050] Based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, a training local optimal control set corresponding to each traffic intersection in the training local optimal set is obtained;

[0051] Based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, a training global optimal control set corresponding to the traffic intersection in the training global optimal set is obtained;

[0052] Perform a traffic virtual simulation experiment based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result;

[0053] Evaluate the degree of solving traffic congestion based on the virtual simulation experiment results to obtain evaluation results;

[0054] Adjusting the parameters of the traffic intersection classification model based on the evaluation results;

[0055] The above process is repeatedly performed to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

[0056] Preferably, the traffic data includes vehicle flow and traffic retention.

[0057] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0058] The present invention relates to a smart traffic management method and system, the method comprising: obtaining a historical data set, and segmenting the historical data set according to a set collection cycle to obtain a plurality of segmented data sets; training a traffic intersection classification model based on each of the segmented data sets to obtain the trained traffic intersection classification model; obtaining real-time data of each traffic intersection according to the set collection cycle, and obtaining traffic data based on the real-time data; grading each traffic intersection based on each of the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection; classifying each traffic intersection based on each of the real-time classification results to obtain a local optimal set and a global optimal set; obtaining a local optimal control set corresponding to each traffic intersection in the local optimal set based on the traffic data and the local neural network model corresponding to each traffic intersection in the local optimal set; controlling the traffic lights of each traffic intersection in the local optimal set based on each of the local optimal control sets; obtaining a global optimal control set corresponding to each traffic intersection in the global optimal set based on the traffic data and the global neural network model corresponding to each traffic intersection in the global optimal set; controlling the traffic lights of each traffic intersection in the global optimal set based on each of the global optimal control sets. The present invention performs classification control according to real-time traffic information, reduces the occurrence of traffic congestion and improves the speed of clearing traffic congestion. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0060] Figure 1 This is a flow chart of the intelligent traffic management method of the present invention;

[0061] Figure 2 This is a structural diagram of the intelligent traffic management system of the present invention.

[0062] Explanation of symbols: 1. Data acquisition module; 2. Model training module; 3. Edge terminal module; 4. Central processing unit. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] The purpose of the present invention is to provide an intelligent traffic management method and system, which performs classified control according to real-time traffic information, reduces the occurrence of traffic congestion and improves the speed of clearing traffic congestion.

[0065] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] Figure 1 This is a flow chart of the intelligent traffic management method of the present invention. Figure 1 As shown, the present invention provides a smart traffic management method, comprising:

[0067] Step S1, obtaining a historical data set, and segmenting the historical data set according to a set collection cycle to obtain a plurality of segmented data sets.

[0068] Specifically, the step S1 includes:

[0069] Step S11, obtaining an initial historical data set; the initial historical data set includes historical video data, historical important building data and historical traffic intersection level data.

[0070] Step S12, obtaining historical traffic data based on the historical video data; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection levels in the historical traffic intersection level data include express intersections, main road intersections, secondary road intersections and branch intersections. The important buildings in the historical important building data include:

[0071] 1) Office buildings at the municipal level and above.

[0072] 2) Indoor places with large crowds, such as gymnasiums, halls, theaters, entertainment venues, stations, stock exchanges, etc., with a design capacity or number of seats exceeding 1,500 people.

[0073] 3) Libraries with a collection of more than 500,000 books; cultural relics, historical sites, museums, exhibition halls, archives and other buildings at the municipal level and above.

[0074] 4) Office buildings of banks and other financial institutions at the provincial level and above, and radio and television buildings at the provincial level and above.

[0075] 5) Open-air stadiums, open-air swimming pools and other open-air public gathering and entertainment venues designed to accommodate more than 5,000 people.

[0076] 6) Primary and secondary schools and other schools for minors with more than 500 students; kindergartens, nurseries and rehabilitation facilities for the disabled with more than 200 students; nursing homes and outpatient and inpatient buildings of hospitals with 50 beds or more.

[0077] 7) Store (shopping mall) buildings with a total construction area exceeding 20,000 square meters, and comprehensive buildings with a construction area of ​​commercial business premises exceeding 15,000 square meters.

[0078] 8) Subway entrances and exits, tunnel entrances and exits.

[0079] Step S13, segmenting the historical data set according to a set collection period to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

[0080] Step S2: training the traffic intersection classification model based on each of the segmented data sets to obtain the trained traffic intersection classification model.

[0081] Furthermore, the step S2 comprises:

[0082] Step S21, input any of the segmented data sets into the traffic intersection classification model to obtain a training classification result for each traffic intersection. In this embodiment, the training classification results include level one, level two and level three.

[0083] Step S22, classify each traffic intersection based on the training classification result to obtain a training local optimal set and a training global optimal set. In this embodiment, the traffic intersection corresponding to the first level of the training classification result is used as the training local optimal set, and the traffic intersection corresponding to the second and third levels of the training classification result is used as the training global optimal set.

[0084] Step S23, based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, obtain the training local optimal control set corresponding to each traffic intersection in the training local optimal set.

[0085] Step S24, based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, obtain the training global optimal control set corresponding to the traffic intersection in the training global optimal set.

[0086] Step S25, performing a traffic virtual simulation experiment based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result.

[0087] Step S26, evaluating the degree of solving traffic congestion based on the virtual simulation experiment results to obtain an evaluation result.

[0088] Step S27: adjusting the parameters of the traffic intersection classification model based on the evaluation result.

[0089] Step S28, repeatedly executing "step S21-step S27" to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

[0090] Step S3, acquiring real-time data of each traffic intersection according to the set collection cycle, and obtaining traffic data based on the real-time data. In this embodiment, the traffic data includes vehicle flow and retention.

[0091] Step S4, classifying each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection.

[0092] Step S5, classifying each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set.

[0093] Step S6, based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model, obtain the local optimal control set corresponding to each traffic intersection in the local optimal set; and control the traffic lights at each traffic intersection in the local optimal set based on each local optimal control set. Taking traffic intersection A in the local optimal set as an example, the local neural network model obtains the local optimal control set A corresponding to traffic intersection A based only on the traffic data corresponding to traffic intersection A, and controls the traffic lights corresponding to traffic intersection A based on the local optimal control set A.

[0094] Step S7, based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model, obtain the global optimal control set corresponding to each traffic intersection in the global optimal set; and control the traffic lights at each traffic intersection in the global optimal set based on each of the global optimal control sets. Assuming that the global optimal set contains M traffic intersections, the global neural network model performs a global optimal solution based on the M traffic data corresponding to the M traffic intersections to obtain a global optimal solution, namely the global optimal control set.

[0095] Figure 2 This is the structure diagram of the intelligent traffic management system of the present invention. Figure 2 As shown, the present invention provides an intelligent traffic management system, including: a data acquisition module 1, a model training module 2, a plurality of edge terminal modules 3 and a central processor 4. Figure 2 Where N is the number of edge terminal modules.

[0096] The data acquisition module 11 is used to acquire a historical data set, and segment the historical data set according to a set acquisition cycle to obtain a plurality of segmented data sets.

[0097] The model training module 22 is used to train the traffic intersection classification model based on each of the segmented data sets to obtain the trained traffic intersection classification model.

[0098] One edge terminal module 3 is provided at each traffic intersection; each edge terminal module 3 is used to obtain real-time data of each traffic intersection according to the set collection period, and obtain traffic data based on the real-time data.

[0099] Each of the edge terminal modules 3 is also used to classify each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection.

[0100] The central processor 4 is used to classify each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set.

[0101] The central processor 4 is also used to obtain a global optimal control set corresponding to each traffic intersection in the global optimal set based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model.

[0102] The central processor 4 is also used to control the traffic lights at each traffic intersection in the global optimal set based on each of the global optimal control sets.

[0103] Each of the edge terminal modules 3 is also used to obtain a local optimal control set corresponding to each traffic intersection in the local optimal set based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model.

[0104] Each of the edge terminal modules 3 is also used to control the traffic lights at each traffic intersection in the local optimal set based on each of the local optimal control sets.

[0105] Optionally, the data acquisition module 1 is specifically:

[0106] Acquire an initial historical data set; the initial historical data set includes historical video data, historical important building data and historical traffic intersection level data.

[0107] Historical traffic data is obtained based on the historical video data; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection levels in the historical traffic intersection level data include express intersections, main road intersections, secondary road intersections and branch intersections.

[0108] The historical data set is segmented according to a set collection cycle to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

[0109] Optionally, the model training module 2 is specifically:

[0110] Any of the segmented data sets is input into the traffic intersection classification model to obtain a training classification result for each traffic intersection.

[0111] Each traffic intersection is classified based on the training classification result to obtain a training local optimal set and a training global optimal set.

[0112] Based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, a training local optimal control set corresponding to each traffic intersection in the training local optimal set is obtained.

[0113] Based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, a training global optimal control set corresponding to the traffic intersection in the training global optimal set is obtained.

[0114] A traffic virtual simulation experiment is performed based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result.

[0115] Based on the virtual simulation experiment results, the degree of solving traffic congestion is evaluated to obtain an evaluation result.

[0116] The parameters of the traffic intersection classification model are adjusted based on the evaluation results.

[0117] The above process is repeatedly performed to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

[0118] Optionally, the traffic data includes vehicle flow and traffic congestion.

[0119] The edge terminal module of the present invention can reduce the load requirement of the central processing unit and further reduce the performance requirement of the central processing unit. In addition, the edge terminal module can also perform individual control on traffic intersections with severe traffic congestion or located next to important buildings, thereby clearing such traffic intersections as quickly as possible.

[0120] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0121] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A smart traffic management method, characterized in that: include: Acquire a historical data set, and segment the historical data set according to a set acquisition period to obtain a plurality of segmented data sets; Training a traffic intersection classification model based on each of the segmented data sets to obtain a trained traffic intersection classification model; Acquire real-time data of each traffic intersection according to the set collection period, and obtain traffic data based on the real-time data; Classifying each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection; Classifying each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set; Based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model, a local optimal control set corresponding to each traffic intersection in the local optimal set is obtained; based on each of the local optimal control sets, a traffic light at each traffic intersection in the local optimal set is controlled; Based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model, a global optimal control set corresponding to each traffic intersection in the global optimal set is obtained; and based on each of the global optimal control sets, the traffic lights at each traffic intersection in the global optimal set are controlled.

2. The intelligent traffic management method according to claim 1, characterized in that: The historical data set is obtained, and the historical data set is segmented according to a set acquisition period to obtain a plurality of segmented data sets, including: Acquire an initial historical data set; the initial historical data set includes historical video data, historical important building data, and historical traffic intersection grade data; Based on the historical video data, historical traffic data is obtained; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection level in the historical traffic intersection level data includes expressway intersection, main road intersection, secondary road intersection and branch road intersection; The historical data set is segmented according to a set collection cycle to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

3. The intelligent traffic management method according to claim 2, characterized in that: The step of training the traffic intersection classification model based on each of the segmented data sets to obtain the trained traffic intersection classification model comprises: Inputting any of the segmented data sets into the traffic intersection classification model to obtain a training classification result for each traffic intersection; Classifying each traffic intersection based on the training classification results to obtain a training local optimal set and a training global optimal set; Based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, a training local optimal control set corresponding to each traffic intersection in the training local optimal set is obtained; Based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, a training global optimal control set corresponding to the traffic intersection in the training global optimal set is obtained; Perform a traffic virtual simulation experiment based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result; Evaluate the degree of solving traffic congestion based on the virtual simulation experiment results to obtain evaluation results; Adjusting the parameters of the traffic intersection classification model based on the evaluation results; The above process is repeatedly performed to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

4. The intelligent traffic management method according to claim 1, characterized in that: The traffic data includes vehicle flow and traffic congestion.

5. A smart traffic management system, characterized in that: include: Data acquisition module, model training module, several edge terminal modules and central processing unit; The data acquisition module is used to acquire a historical data set, and segment the historical data set according to a set acquisition cycle to obtain a plurality of segmented data sets; The model training module is used to train the traffic intersection classification model based on each segmented data set to obtain the trained traffic intersection classification model; One edge terminal module is provided at each traffic intersection; each edge terminal module is used to obtain real-time data of each traffic intersection according to the set collection period, and obtain traffic data based on the real-time data; Each of the edge terminal modules is also used to classify each traffic intersection based on the traffic data and the trained traffic intersection classification model to obtain a real-time classification result for each traffic intersection; The central processing unit is used to classify each traffic intersection based on the real-time classification results to obtain a local optimal set and a global optimal set; The central processor is also used to obtain a global optimal control set corresponding to each traffic intersection in the global optimal set based on the traffic data corresponding to each traffic intersection in the global optimal set and the global neural network model; The central processor is also used to control the signal lights at each traffic intersection in the global optimal set based on each of the global optimal control sets; Each of the edge terminal modules is further used to obtain a local optimal control set corresponding to each traffic intersection in the local optimal set based on the traffic data corresponding to each traffic intersection in the local optimal set and the local neural network model; Each of the edge terminal modules is also used to control the traffic lights at each traffic intersection in the local optimal set based on each of the local optimal control sets.

6. The intelligent traffic management system according to claim 5, characterized in that: The data acquisition module is specifically: Acquire an initial historical data set; the initial historical data set includes historical video data, historical important building data, and historical traffic intersection grade data; Based on the historical video data, historical traffic data is obtained; the historical data set includes the historical traffic data, the historical important building data and the historical traffic intersection level data; the traffic intersection level in the historical traffic intersection level data includes expressway intersection, main road intersection, secondary road intersection and branch road intersection; The historical data set is segmented according to a set collection cycle to obtain a plurality of segmented data sets; the segmented data sets include segmented data corresponding to each traffic intersection.

7. The intelligent traffic management system according to claim 6, characterized in that: The model training module is specifically: Inputting any of the segmented data sets into the traffic intersection classification model to obtain a training classification result for each traffic intersection; Classifying each traffic intersection based on the training classification results to obtain a training local optimal set and a training global optimal set; Based on the segmentation data and the local neural network model corresponding to each traffic intersection in the training local optimal set, a training local optimal control set corresponding to each traffic intersection in the training local optimal set is obtained; Based on the segmentation data corresponding to each traffic intersection in the training global optimal set and the global neural network model, a training global optimal control set corresponding to the traffic intersection in the training global optimal set is obtained; Perform a traffic virtual simulation experiment based on each of the training local optimal control sets and the training global optimal control set to obtain a virtual simulation experiment result; Evaluate the degree of solving traffic congestion based on the virtual simulation experiment results to obtain evaluation results; Adjusting the parameters of the traffic intersection classification model based on the evaluation results; The above process is repeatedly performed to iteratively optimize the traffic intersection classification model until a set number of iterations is reached or the evaluation result is greater than a set threshold, thereby obtaining a trained traffic intersection classification model.

8. The intelligent traffic management system according to claim 5, characterized in that: The traffic data includes vehicle flow and traffic congestion.

Citation Information

Patent Citations

  • Hierarchical decision-making method for realizing real-time intelligent traffic management under vehicle networking

    CN110021168A

  • Traffic signal control method and system based on artificial intelligence two-way progression

    CN110223513A

  • Traffic signal lamp control method based on cooperative multi-agent reinforcement learning

    CN115083174A

  • Method and system for intersection group-based traffic control

    WO2017166474A1