A method, device, equipment and medium for managing traffic jams on expressways

Abstract

The present invention provides a method, device, equipment and medium for managing congestion points on expressways, including pre-congestion prediction and early warning and post-congestion relief plan intelligent matching. Pre-congestion prediction and early warning obtains the congestion situation prediction value of the easy-to-congestion point and the surrounding road network through two-level prediction, wherein the first-level single-point prediction can quickly screen out congestion points, thereby effectively reducing the probability of congestion events at certain important nodes of the expressway; the second-level road network prediction can determine the impact of congestion points on the surrounding road networks, thereby effectively predicting the congestion index of the road networks around the congestion points, and realizing timely warning of congestion in the road networks around the congestion points. Post-congestion relief plan intelligent matching selects the optimal control plan based on multidimensional vectors through a plan matching algorithm, based on the concept of "relieving congestion before congestion", forms a "one point one policy" congestion relief plan for controlling congestion and ensuring smooth flow, and realizes the optimal control measures for easy-to-congestion points.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, equipment and medium for managing traffic jams on expressways. Background Art

[0002] In the past, the impact of large traffic flow on traffic flow was often analyzed and predicted by traffic police based on their personal experience. It was difficult to comprehensively consider the congestion relief of the main line and the carrying capacity of the diversion section, and it was impossible to provide a scientific and reasonable basis for the formulation of control plans, which often added a lot of difficulties to the subsequent on-site management. In view of this, how the public security traffic management department can use traffic control technology to deal with large traffic flow on highways and ensure the safety and smoothness of roads has become an important task at present. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method, device, equipment and medium for managing congestion points on expressways. The congestion situation prediction values ​​of congestion-prone points and their surrounding road networks are obtained through two-level predictions, and then the optimal control plan is selected based on the multi-dimensional vector through the plan matching algorithm to form a "one point, one policy" congestion relief plan for controlling congestion and ensuring smooth traffic, thereby realizing the optimal control measures for congestion-prone points.

[0004] In a first aspect, the present invention provides a method for managing congestion points on a highway, including a two-level prediction process and a process for constructing a congestion point management plan;

[0005] The two-level prediction process includes:

[0006] First-level single-point prediction: Combined with the traffic data of the highways in the jurisdiction, the congestion situation of the points prone to congestion is predicted through the random forest algorithm in machine learning, and the single-point traffic prediction value of each point prone to congestion in the road network of the jurisdiction is obtained; according to each of the single-point traffic prediction values, it is judged whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction;

[0007] Secondary road network prediction: Select influential congestion points around the congestion point as associated points one by one, input the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index. The road network in the congestion point area is composed of the congestion point and the corresponding associated points;

[0008] The traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the congestion-prone points include main line points, ramp points, toll station points and hub points;

[0009] The process of constructing the blocking point management solution includes:

[0010] S31, describing the feature points of the congestion event on the expressway, wherein the feature points include the cause of congestion, the degree of congestion and the location of congestion; the cause of congestion includes road construction, traffic accidents and heavy traffic; the degree of congestion is the ratio of the number of affected lanes to the total number of lanes;

[0011] S32, constructing a multi-dimensional vector group of congestion events described by feature points;

[0012] S33, using a correlation matching algorithm to perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan to obtain a feature point matching result;

[0013] S34. Obtain a “one point, one policy” solution to the congestion point based on the feature point matching results.

[0014] In a second aspect, the present invention provides a device for managing congestion points on a highway, comprising:

[0015] The first-level single-point prediction module is used to combine the traffic data of the highways in the jurisdiction, and use the random forest algorithm in machine learning to predict the congestion situation of the points prone to congestion, and obtain the single-point traffic prediction value of each point prone to congestion in the road network of the jurisdiction; according to each of the single-point traffic prediction values, it is judged whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction;

[0016] A secondary road network prediction module is used to select influential and congested points around the congestion point as associated points one by one, input the distance between the associated point and the congestion point, and the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points;

[0017] The traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the congestion-prone points include main line points, ramp points, toll station points and hub points;

[0018] The plan matching module is used to perform the following processes:

[0019] S31, describing the feature points of the congestion event on the expressway, wherein the feature points include the cause of congestion, the degree of congestion and the location of congestion; the cause of congestion includes road construction, traffic accidents and heavy traffic; the degree of congestion is the ratio of the number of affected lanes to the total number of lanes;

[0020] S32, constructing a multi-dimensional vector group of congestion events described by feature points;

[0021] S33, using a correlation matching algorithm to perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan to obtain a feature point matching result;

[0022] S34. Obtain a “one point, one policy” solution to the congestion point based on the feature point matching results.

[0023] In a third aspect, 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 the first aspect when executing the program.

[0024] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0025] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages: 1. Pre-judgment and early warning of congestion

[0026] Traffic congestion situation prediction is one of the important steps to solve the traffic congestion problem. Through the prediction of high-volume traffic congestion situation of the highway network, we can effectively predict the travel traffic situation in the future and alleviate the traffic congestion pressure on the highway. The first-level single-point prediction can effectively reduce the probability of congestion events at certain important nodes of the highway; the second-level road network prediction can effectively predict the congestion index of the road network around the congestion point and realize timely warning of road network congestion. Through the two-level prediction of the traffic congestion situation of the highway network, it is possible to effectively predict the occurrence of congestion and take preventive measures for possible congestion in time, effectively improve the traffic efficiency of the highway, save travel time, and reduce environmental pollution and energy waste.

[0027] 2. Intelligent matching of post-event congestion relief plans

[0028] Relying on the correlation matching algorithm, the congestion relief and smooth flow plan is launched early for important highway traffic nodes / areas. Based on the concept of "decongestion before congestion", through the intelligent matching of the post-congestion relief plan, a refined "one point, one policy" congestion point management plan is formed to provide decision-making assistance for police officers, and further distribute the control tasks and plans to the corresponding police terminals to guide police officers to carry out traffic organization work efficiently, so as to improve the standardization level of traffic organization work, strengthen the control level of highway network congestion relief and smooth flow, and improve the modern highway traffic management capabilities.

[0029] In summary, the highway congestion management method based on machine learning proposed in the present invention can realize early warning and prediction of congestion at congestion-prone points on highways and intelligent matching of congestion relief plans through machine learning, providing a scientific solution for the management of congested sections of highways.

[0030] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The present invention will be further described below in conjunction with embodiments with reference to the accompanying drawings.

[0032] Figure 1 is a flow chart of the method in Embodiment 1 of the present invention;

[0033] Figure 2 This is a flow chart of a first-level single-point prediction process according to an embodiment of the present invention;

[0034] Figure 3 A flowchart of a secondary road network prediction process according to an embodiment of the present invention;

[0035] Figure 4 It is a flowchart of the process of constructing the blocking point management solution according to an embodiment of the present invention;

[0036] Figure 5 A flowchart of a Hamming distance algorithm according to an embodiment of the present invention;

[0037] Figure 6 A flowchart of a process for constructing a multi-dimensional vector group of a control plan according to an embodiment of the present invention;

[0038] Figure 7 It is a structural schematic diagram of the device in the second embodiment of the present invention;

[0039] Figure 8 This is a schematic diagram of the structure of an electronic device in Embodiment 3 of the present invention;

[0040] Fig. 9 Schematic diagram of the structure of the medium in the fourth embodiment of the present invention. DETAILED DESCRIPTION

[0041] The embodiment of the present application provides a method, device, equipment and medium for managing congestion points on expressways, obtains the predicted values ​​of the congestion situation of congestion-prone points and their surrounding road networks through two-level prediction, and then selects the optimal control plan based on the multi-dimensional vector through the plan matching algorithm, so as to form a "one point, one policy" congestion relief plan for controlling congestion and ensuring smooth traffic, thereby realizing the optimal control measures for congestion-prone points.

[0042] The technical solution in the embodiment of the present application has the following general idea: it includes advance prediction and warning of congestion beforehand and intelligent matching of congestion relief plans afterwards. Advance prediction and warning of congestion beforehand obtains the predicted value of congestion situation of congestion-prone points and their surrounding road networks through two-level prediction, among which, the first-level single-point prediction can quickly screen out congestion points, thereby effectively reducing the probability of congestion events at certain important nodes of the highway; the second-level road network prediction can determine the impact of congestion points on surrounding congestion-prone points, thereby effectively predicting the congestion index of the road network around the congestion points, and realizing timely warning of congestion in the road network around the congestion points. Intelligent matching of congestion relief plans afterwards uses a plan matching algorithm to select the optimal control plan based on multi-dimensional vectors, based on the concept of "relieving congestion before it is blocked", to form a "one point, one policy" congestion relief plan to control congestion and ensure smooth flow, and realize the optimal control measures for congestion-prone points. Embodiment 1

[0043] like Figure 1 As shown, this embodiment provides a method for managing congestion points on a highway, including a two-level prediction process and a process for constructing a congestion point management plan.

[0044] The two-level prediction process is to predict and warn of congestion in advance. Through the two-level prediction of the traffic congestion situation of the highway network, timely prediction and warning of road network congestion can be achieved, and the highway management department can be actively reminded to command and control the congestion points. First, combined with the historical / real-time data of the highways in the jurisdiction, the change trend of the congestion-prone points is predicted through machine learning technology. Secondly, the overall congestion change trend of the highway network around the congestion point is predicted based on the operating characteristics of the road network area.

[0045] The process of constructing the congestion point management plan is to combine the road network situation prediction value scientifically judged in advance, and deploy control measures in advance for the points / areas of the highway that are about to be congested, so as to form a new mode of large-volume control of road sections that transforms from passive control to active intervention, and achieve the positive effect of "unblocking before congestion" in highway traffic. Adopting the intelligent matching of post-congestion relief plans, combined with data such as congestion causes, congestion levels, and congestion points, the optimal control plan is selected based on the correlation matching algorithm of Hamming distance in multi-dimensional vector space, forming a "one point, one policy" congestion relief plan to control congestion and ensure smooth traffic, and realizing the optimal control measures for easy congestion points. At the same time, the optimal control measures are issued by the police in the command center to the front-line police, which effectively serves the front-line police in practical applications, improves the control level of highway network congestion relief and smooth traffic, and realizes scientific diversion control.

[0046] The two-level prediction process includes:

[0047] First-level single-point prediction: Combined with the traffic data of the expressways in the jurisdiction, the random forest algorithm in machine learning is used to predict the congestion situation of the points prone to congestion, and the single-point traffic prediction value of each point prone to congestion in the road network in the jurisdiction is obtained; according to each of the single-point traffic prediction values, it is determined whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction; points prone to congestion refer to points that are prone to congestion, usually including main lines, ramps, toll stations, hubs, interchanges and other points.

[0048] like Figure 2 As shown, the first-level single-point prediction specifically includes the following steps:

[0049] S11, selection of characteristic data and setting of traffic congestion index;

[0050] The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway time, weather conditions, and holiday factors;

[0051] The traffic congestion index is an indicator that comprehensively reflects the traffic operation status of the road network. Different intervals are divided according to the size of the traffic congestion index, and each interval corresponds to a congestion status label;

[0052] S12. Construct a random forest regression model, randomly extract data based on the Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities;

[0053] S13, inputting the traffic data of the current congestion-prone points into the random forest regression model to predict the congestion situation of the current congestion-prone points, and obtaining the single-point traffic prediction value of each congestion-prone point in the road network of the jurisdiction according to the set traffic congestion index;

[0054] S14. Determine whether the corresponding traffic point prone to congestion is a congestion point based on the traffic prediction value of each single point. If so, start the secondary road network prediction. If not, wait for a cycle and return to step 13 to enter the next round of primary single point prediction. The so-called congestion point is a congestion point prone to congestion whose prediction value reaches a higher congestion level.

[0055] Secondary road network prediction: Select influential congestion-prone points around the congestion point one by one as associated points, input the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index. The road network in the congestion point area is composed of the congestion point and the corresponding associated points.

[0056] Among them, under the premise that computing power allows, it is possible to further determine whether new congestion points appear at the associated points around the congestion point. If so, a secondary road network prediction centered on the new congestion point is initiated. If not, the process ends.

[0057] The secondary road network prediction specifically includes the following steps:

[0058] S21, selection of feature data;

[0059] The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway, weather conditions, holiday factors, distances between points in the road network within the jurisdiction, and traffic forecast values ​​for each point in the road network within the jurisdiction;

[0060] S22, build a random forest regression model, randomly extract data based on Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities;

[0061] S23, selecting influential congestion-prone points around the congestion point one by one as associated points, inputting the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtaining the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points.

[0062] Among them, the traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the points prone to congestion include main line points, ramp points, toll station points and hub points; the main factor for judging the associated points is the distance between the points prone to congestion and the congestion points. Generally, the shorter the distance, the greater the impact of the congestion point on the points prone to congestion.

[0063] The process of constructing the blocking point management solution includes:

[0064] S31. Describe the feature points of the congestion events on the expressway, where the feature points include the cause of congestion, the degree of congestion and the location of congestion. The cause of congestion includes road construction, traffic accidents and heavy traffic. The degree of congestion is the ratio of the number of affected lanes to the total number of lanes, which is the traffic congestion index, and can be classified according to the size range and labeled as shown in the following table:

[0065] Table 1 Traffic congestion index

[0066]

[0067] S32, constructing a multi-dimensional vector group of congestion events described by feature points;

[0068] S33, using a correlation matching algorithm to perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan to obtain a feature point matching result;

[0069] like Figure 5 As shown, the correlation matching algorithm is a Hamming distance algorithm, which is calculated by calculating the number of differences between different characters at the same position in the descriptor of each feature point pair, and then calculating the Hamming distance of each feature point; the feature point pair refers to two feature points of the same dimension in the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan; the descriptor refers to a way of expressing the feature point;

[0070] The Hamming distance calculation formula is as follows: ;

[0071] represents the Hamming distance, x and y represent the n-dimensional descriptors that describe the feature points, and ⊕ represents the XOR operation, where: The smaller it is, the smaller the difference between the two feature points is. The larger the value, the greater the difference between the two feature points.

[0072] For example, if the feature point is congestion, if the cause of the congestion is road construction, the descriptor is "road construction"; if the cause of the congestion is a traffic accident, the descriptor is "traffic accident"; if the cause of the congestion is heavy traffic, the descriptor is "heavy traffic".

[0073] S34. Obtain a “one point, one policy” solution to the congestion point based on the feature point matching results.

[0074] like Figure 6 As shown, the construction of the multi-dimensional vector group of the control plan specifically includes the following processes:

[0075] S41, constructing a multidimensional vector of three characteristic points: congestion cause, congestion degree, and congestion point;

[0076] Congestion causes A={A1,A2,A3}; A1 represents road construction, A2 represents traffic accidents, and A3 represents heavy traffic;

[0077] Congestion level B={B1,B2,B3,B4,B5}; among which B1=0.25, B2=0.33, B3=0.5, B4=0.75, B5=1;

[0078] The congestion point C={C1,C2,C3,C4}; C1 represents the main line, C2 represents the ramp, C3 represents the toll station, and C4 represents the hub.

[0079] S42. Construct a three-dimensional vector group Wi of management and control plans. Each three-dimensional vector group Wi of management and control plans corresponds to a management and control plan. The multi-dimensional vector group Wi of management and control plans includes: W1={A1, B1, C1}; W2={A1, B1, C2}; W3={A1, B1, C3}; W4={A1, B1, C4}; W5={A1, B2, C1}; W6={A1, B2, C2}; W7={A1, B2, C3}; W8={A1, B2, C4}; W9={A1, B3, C1}, ..., and so on.

[0080] For example, the characteristic points of a congestion event on a highway are described as follows:

[0081] The cause of congestion is road construction, and the cause of congestion A = {A1};

[0082] The congestion level is mild congestion, and the congestion level B={B3};

[0083] The congestion point is the main line, and the congestion point C = {C1};

[0084] Then the multi-dimensional vector group of congestion events constructed is: {A1, B3, C1};

[0085] Substitute the Hamming distance calculation formula for calculation, perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan, and obtain the feature point matching result: W9={A1,B3,C1}

[0086] According to W9, you can go to the plan library to obtain the corresponding control plan.

[0087] If there are two vectors of a certain dimension for a congestion event on a highway, for example, the causes of congestion include road construction and traffic accidents, two multi-dimensional vector groups of congestion events can be constructed to match two corresponding control plans. The two control plans can be fused, for example, by taking the union of the two control plans, to obtain a new plan.

[0088] Based on the same inventive concept, the present application also provides a device corresponding to the method in Example 1, see Example 2 for details. Embodiment 2

[0089] like Figure 7 As shown, in this embodiment, a device for managing congestion points on a highway is provided, comprising:

[0090] The first-level single-point prediction module is used to combine the traffic data of the highways in the jurisdiction, and use the random forest algorithm in machine learning to predict the congestion situation of the points prone to congestion, and obtain the single-point traffic prediction value of each point prone to congestion in the road network of the jurisdiction; according to each of the single-point traffic prediction values, it is judged whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction;

[0091] like Figure 2 As shown, the first-level single-point prediction module is specifically used to perform the following process:

[0092] S11, selection of characteristic data and setting of traffic congestion index;

[0093] The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway time, weather conditions, and holiday factors;

[0094] The traffic congestion index is an indicator that comprehensively reflects the traffic operation status of the road network. Different intervals are divided according to the size of the traffic congestion index, and each interval corresponds to a congestion status label;

[0095] S12. Construct a random forest regression model, randomly extract data based on the Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities;

[0096] S13, inputting the traffic data of the current prone-to-congestion point into the random forest regression model to predict the congestion situation of the current prone-to-congestion point, and obtaining the single-point traffic prediction value of each prone-to-congestion point in the road network of the jurisdiction according to the set traffic congestion index,

[0097] S14, judging whether the corresponding traffic point prone to congestion is a congestion point according to the traffic prediction value of each single point, if so, starting the secondary road network prediction, if not, waiting for a cycle, returning to step 13, and entering the next round of primary single point prediction;

[0098] The secondary road network prediction module is used to select influential congestion-prone points around the congestion point one by one as associated points, input the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index. The road network in the congestion point area is composed of the congestion point and the corresponding associated points.

[0099] Similarly, if the computing power of the system allows, it can be further determined whether there are new congestion points around the congestion point. If so, a secondary road network prediction centered on the new congestion point is initiated. If not, the prediction ends.

[0100] like Figure 3 As shown, the secondary road network prediction module is specifically used to perform the following process:

[0101] S21, selection of feature data;

[0102] The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway, weather conditions, holiday factors, distances between points in the road network within the jurisdiction, and traffic forecast values ​​for each point in the road network within the jurisdiction;

[0103] S22, build a random forest regression model, randomly extract data based on Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities;

[0104] S23, selecting influential congestion-prone points around the congestion point one by one as associated points, inputting the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtaining the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points.

[0105] The traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the congestion-prone points include main line points, ramp points, toll station points and hub points;

[0106] Plan matching module, such as Figure 4 As shown, it is used to perform the following process:

[0107] S31. Describe the feature points of the congestion events on the expressway, where the feature points include the cause of congestion, the degree of congestion and the location of congestion; the cause of congestion includes road construction, traffic accidents and heavy traffic; the degree of congestion is the ratio of the number of affected lanes to the total number of lanes; the ratio is the traffic congestion index, which can be classified according to the size range and labeled as shown in the following table:

[0108] Table 1 Traffic congestion index

[0109]

[0110] S32, using a correlation matching algorithm to perform feature point matching on the feature point description to obtain a feature point matching result; Figure 5As shown, the correlation matching algorithm is a Hamming distance algorithm, which is calculated by calculating the number of differences between different characters at the same position in the descriptor of each feature point pair, and then calculating the Hamming distance of each feature point; the feature point pair refers to two feature points of the same dimension in the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan; the descriptor refers to a way of expressing the feature point;

[0111] The Hamming distance calculation formula is as follows: ;

[0112] represents the Hamming distance, x and y represent the n-dimensional descriptors that describe the feature points, and ⊕ represents the XOR operation, where: The smaller it is, the smaller the difference between the two feature points is. The larger the value, the greater the difference between the two feature points.

[0113] S33. Obtain a “one point, one policy” solution to the congestion point based on the feature point matching results.

[0114] like Figure 6 As shown, the construction of the multi-dimensional vector group of the control plan specifically includes the following processes:

[0115] S41, constructing a multidimensional vector of three characteristic points: congestion cause, congestion degree, and congestion point;

[0116] Congestion causes A={A1,A2,A3}; A1 represents road construction, A2 represents traffic accidents, and A3 represents heavy traffic;

[0117] Congestion level B={B1,B2,B3,B4,B5}; among which B1=0.25, B2=0.33, B3=0.5, B4=0.75, B5=1;

[0118] The congestion point C={C1,C2,C3,C4}; C1 represents the main line, C2 represents the ramp, C3 represents the toll station, and C4 represents the hub.

[0119] S42. Construct a three-dimensional vector group Wi of management and control plans. Each three-dimensional vector group Wi of management and control plans corresponds to a management and control plan. The multi-dimensional vector group Wi of management and control plans includes: W1={A1, B1, C1}; W2={A1, B1, C2}; W3={A1, B1, C3}; W4={A1, B1, C4}; W5={A1, B2, C1}; W6={A1, B2, C2}; W7={A1, B2, C3}; W8={A1, B2, C4}; W9={A1, B3, C1}, ..., and so on.

[0120] For specific examples, please refer to the relevant description in Embodiment 1.

[0121] Since the device introduced in the second embodiment of the present invention is a device used to implement the method of the first embodiment of the present invention, those skilled in the art can understand the specific structure and deformation of the device based on the method introduced in the first embodiment of the present invention, so it is not described here in detail. All devices used in the method of the first embodiment of the present invention belong to the scope of protection of the present invention.

[0122] Based on the same inventive concept, the present application provides an electronic device embodiment corresponding to the first embodiment, see the third embodiment for details. Embodiment 3

[0123] This embodiment provides an electronic device, such as Figure 8 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any implementation method in the first embodiment can be implemented.

[0124] Since the electronic device introduced in this embodiment is a device used to implement the method in the first embodiment of the present application, based on the method introduced in the first embodiment of the present application, a person skilled in the art can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as a person skilled in the art implements the device used by the method in the embodiment of the present application, it belongs to the scope of protection of the present application.

[0125] Based on the same inventive concept, the present application provides a storage medium corresponding to the first embodiment, see the fourth embodiment for details. Embodiment 4

[0126] This embodiment provides a computer-readable storage medium, such as Fig. 9 As shown, a computer program is stored thereon, and when the computer program is executed by a processor, any implementation method in Example 1 can be implemented.

[0127] The methods, devices, systems, equipment, and media provided in the embodiments of the present application have at least the following technical effects or advantages:

[0128] 1. Predict and warn of congestion in advance

[0129] Traffic congestion situation prediction is one of the important steps to solve the traffic congestion problem. Through the prediction of high-volume traffic congestion situation of the highway network, we can effectively predict the travel traffic situation in the future and alleviate the traffic congestion pressure on the highway. The first-level single-point prediction can effectively reduce the probability of congestion events at certain important nodes of the highway; the second-level road network prediction can effectively predict the congestion index of the road network around the congestion point and realize timely warning of road network congestion. Through the two-level prediction of the traffic congestion situation of the highway network, it is possible to effectively predict the occurrence of congestion and take preventive measures for possible congestion in time, effectively improve the traffic efficiency of the highway, save travel time, and reduce environmental pollution and energy waste.

[0130] 2. Intelligent matching of post-event congestion relief plans

[0131] Relying on the correlation matching algorithm, the congestion relief and smooth flow plan is launched early for important highway traffic nodes / areas. Based on the concept of "decongestion before congestion", through the intelligent matching of the post-congestion relief plan, a refined "one point, one policy" congestion point management plan is formed to provide decision-making assistance for police officers, and further distribute the control tasks and plans to the corresponding police terminals to guide police officers to carry out traffic organization work efficiently, so as to improve the standardization level of traffic organization work, strengthen the control level of highway network congestion relief and smooth flow, and improve the modern highway traffic management capabilities.

[0132] In summary, the highway congestion management method based on machine learning proposed in the present invention can realize early warning and prediction of congestion at congestion-prone points on highways and intelligent matching of congestion relief plans through machine learning, providing a scientific solution for the management of congested sections of highways.

[0133] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices or systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0135] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0136] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0137] Although the specific implementation modes of the present invention are described above, those skilled in the art should understand that the specific implementation modes described are only illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A method for managing congestion points on a highway, characterized in that: It includes the two-level prediction process and the process of constructing a solution for bottleneck management; The two-level prediction process includes: First-level single-point prediction: Combined with the traffic data of the highways in the jurisdiction, the congestion situation of the points prone to congestion is predicted through the random forest algorithm in machine learning, and the single-point traffic prediction value of each point prone to congestion in the road network of the jurisdiction is obtained; according to each of the single-point traffic prediction values, it is judged whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction; Secondary road network prediction: Select influential congestion points around the congestion point as associated points one by one, input the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index. The road network in the congestion point area is composed of the congestion point and the corresponding associated points; The traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the congestion-prone points include main line points, ramp points, toll station points and hub points; The process of constructing the blocking point governance solution includes: S31, describing the feature points of the congestion event on the expressway, wherein the feature points include the cause of congestion, the degree of congestion and the location of congestion; the cause of congestion includes road construction, traffic accidents and heavy traffic; the degree of congestion is the ratio of the number of affected lanes to the total number of lanes; S32, constructing a multi-dimensional vector group of congestion events described by feature points; S33, using a correlation matching algorithm to perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan to obtain a feature point matching result; S34. Obtain a "one point, one policy" solution to the congestion point based on the feature point matching results.

2. The method according to claim 1, characterized in that: The first-level single-point prediction specifically includes the following steps: S11, selection of characteristic data and setting of traffic congestion index; The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway time, weather conditions, and holiday factors; The traffic congestion index is an indicator that comprehensively reflects the traffic operation status of the road network. Different intervals are divided according to the size of the traffic congestion index, and each interval corresponds to a congestion status label; S12. Construct a random forest regression model, randomly extract data based on the Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities; S13, inputting the traffic data of the current congestion-prone points into the random forest regression model to predict the congestion situation of the current congestion-prone points, and obtaining the single-point traffic prediction value of each congestion-prone point in the road network of the jurisdiction according to the set traffic congestion index; S14, judging whether the corresponding traffic point prone to congestion is a congestion point according to the traffic prediction value of each single point, if so, starting the secondary road network prediction, if not, waiting for a cycle, returning to step 13, and entering the next round of primary single point prediction; The secondary road network prediction specifically includes the following steps: S21, selection of feature data; The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway, weather conditions, holiday factors, distances between points in the road network within the jurisdiction, and traffic forecast values ​​for each point in the road network within the jurisdiction; S22, build a random forest regression model, randomly extract data based on Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities; S23, selecting influential congestion-prone points around the congestion point one by one as associated points, inputting the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtaining the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points.

3. The method according to claim 1, characterized in that: The construction of the multi-dimensional vector group of the control plan specifically includes the following process: S41, constructing a multidimensional vector of three characteristic points: congestion cause, congestion degree, and congestion point; Congestion causes A={A1,A2,A3}; A1 represents road construction, A2 represents traffic accidents, and A3 represents heavy traffic; Congestion level B={B1,B2,B3,B4,B5}; among which B1=0.25, B2=0.33, B3=0.5, B4=0.75, B5=1; Congestion point C = {C1, C2, C3, C4}; C1 represents the main line, C2 represents the ramp, C3 represents the toll station, and C4 represents the hub; S42. Construct a three-dimensional vector group Wi of management and control plans. Each three-dimensional vector group Wi of management and control plans corresponds to a management and control plan. The multi-dimensional vector group Wi of management and control plans includes: W1={A1, B1, C1}; W2={A1, B1, C2}; W3={A1, B1, C3}; W4={A1, B1, C4}; W5={A1, B2, C1}; W6={A1, B2, C2}; W7={A1, B2, C3}; W8={A1, B2, C4}; W9={A1, B3, C1}, ..., and so on.

4. The method according to claim 1, characterized in that: The correlation matching algorithm is a Hamming distance algorithm, which is calculated by calculating the number of differences between different characters in the same position in the descriptor of each feature point pair, and then calculating the Hamming distance of each feature point; the feature point pair refers to two feature points of the same dimension in the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan; the descriptor refers to a way of expressing the feature point; the Hamming distance calculation formula is as follows: ; represents the Hamming distance, x and y represent the n-dimensional descriptors that describe the feature points, and ⊕ represents the XOR operation, where: The smaller it is, the smaller the difference between the two feature points is. The larger the value, the greater the difference between the two feature points.

5. A device for managing congestion points on a highway, characterized in that: include: The first-level single-point prediction module is used to combine the traffic data of the highways in the jurisdiction, and use the random forest algorithm in machine learning to predict the congestion situation of the points prone to congestion, and obtain the single-point traffic prediction value of each point prone to congestion in the road network of the jurisdiction; according to each of the single-point traffic prediction values, it is judged whether the corresponding point prone to congestion is a congestion point. If so, the second-level road network prediction is started. If not, wait for a cycle and enter the next round of first-level single-point prediction; A secondary road network prediction module is used to select influential and easily congested points around the congestion point one by one as associated points, input the distance between the associated point and the congestion point, and the single-point traffic prediction values ​​and traffic data of the associated point and the congestion point into a random forest regression model, and obtain the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points; The traffic data includes historical congestion data, historical traffic flow data, historical weather data, real-time congestion data, real-time traffic flow data and real-time weather data; the congestion-prone points include main line points, ramp points, toll station points and hub points; The building blocks of the bottleneck management solution are used to perform the following processes: S31, describing the feature points of the congestion event on the expressway, wherein the feature points include the cause of congestion, the degree of congestion and the location of congestion; the cause of congestion includes road construction, traffic accidents and heavy traffic; the degree of congestion is the ratio of the number of affected lanes to the total number of lanes; S32, constructing a multi-dimensional vector group of congestion events described by feature points; S33, using a correlation matching algorithm to perform feature point matching on the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan to obtain a feature point matching result; S34. Obtain a "one point, one policy" solution to the congestion point based on the feature point matching results.

6. The device according to claim 5, characterized in that: The first-level single-point prediction module is specifically used to perform the following process: S11, selection of characteristic data and setting of traffic congestion index; The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway time, weather conditions, and holiday factors; The traffic congestion index is an indicator that comprehensively reflects the traffic operation status of the road network. Different intervals are divided according to the size of the traffic congestion index, and each interval corresponds to a congestion status label; S12. Construct a random forest regression model, randomly extract data based on the Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities; S13, inputting the traffic data of the current prone-to-congestion point into the random forest regression model to predict the congestion situation of the current prone-to-congestion point, and obtaining the single-point traffic prediction value of each prone-to-congestion point in the road network of the jurisdiction according to the set traffic congestion index, S14, judging whether the corresponding traffic point prone to congestion is a congestion point according to the traffic prediction value of each single point, if so, starting the secondary road network prediction, if not, waiting for a cycle, returning to step 13, and entering the next round of primary single point prediction; The secondary road network prediction module is specifically used to perform the following process: S21, selection of feature data; The characteristic data include traffic flow, vehicle speed, number of lanes, headway, headway, weather conditions, holiday factors, distances between points in the road network within the jurisdiction, and traffic forecast values ​​for each point in the road network within the jurisdiction; S22, build a random forest regression model, randomly extract data based on Bagging and Random Subspace methods to generate decision trees, so that each decision tree has different regression capabilities; S23, selecting influential congestion-prone points around the congestion point one by one as associated points, inputting the distance between the associated point and the congestion point, as well as the single-point traffic prediction value and traffic data of the associated point and the congestion point into the random forest regression model, and obtaining the traffic prediction value of the overall congestion situation of the road network in the congestion point area according to the traffic congestion index, wherein the road network in the congestion point area is composed of the congestion point and the corresponding associated points.

7. The device according to claim 5, characterized in that: The construction of the multi-dimensional vector group of the control plan specifically includes the following process: S41, constructing a multidimensional vector of three characteristic points: congestion cause, congestion degree, and congestion point; Congestion causes A={A1,A2,A3}; A1 represents road construction, A2 represents traffic accidents, and A3 represents heavy traffic; Congestion level B={B1,B2,B3,B4,B5}; among which B1=0.25, B2=0.33, B3=0.5, B4=0.75, B5=1; Congestion point C = {C1, C2, C3, C4}; C1 represents the main line, C2 represents the ramp, C3 represents the toll station, and C4 represents the hub; S42. Construct a three-dimensional vector group Wi of management and control plans. Each three-dimensional vector group Wi of management and control plans corresponds to a management and control plan. The multi-dimensional vector group Wi of management and control plans includes: W1={A1, B1, C1}; W2={A1, B1, C2}; W3={A1, B1, C3}; W4={A1, B1, C4}; W5={A1, B2, C1}; W6={A1, B2, C2}; W7={A1, B2, C3}; W8={A1, B2, C4}; W9={A1, B3, C1}, ..., and so on.

8. The device according to claim 5, characterized in that: The correlation matching algorithm is a Hamming distance algorithm, which is to calculate the number of differences between different characters in the same position in the descriptor of each feature point pair, and then calculate the Hamming distance of each feature point; the feature point pair refers to two feature points of the same dimension in the multidimensional vector group of the congestion event and the multidimensional vector group of the control plan; the descriptor refers to a way of expressing the feature point; The Hamming distance calculation formula is as follows: ; represents the Hamming distance, x and y represent the n-dimensional descriptors that describe the feature points, and ⊕ represents the XOR operation, where: The smaller it is, the smaller the difference between the two feature points is. The larger the value, the greater the difference between the two feature points.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images