Highway congestion prediction method, system and device integrating spatiotemporal correlation data

Abstract

The present invention discloses a highway congestion prediction method integrating spatiotemporal correlation data. The present invention uses highway gantry system data and toll flow data to calculate traffic flow characteristic parameters such as traffic flow, speed, q / C ratio, clusters the input data with the help of spectral clustering algorithm, uses machine learning algorithm to distinguish the traffic state of clustering results, and integrates each algorithm based on soft voting mechanism to construct a traffic state classification decision model based on crowd intelligence algorithm, introduces confusion matrix to obtain evaluation index to quantitatively evaluate the discrimination effect of the model, inputs predicted future traffic flow characteristic parameters to obtain high-precision traffic state prediction results of road sections in the future, and further predicts whether the road section is likely to be congested. The present invention greatly improves the efficiency of highway operation and management, and can be widely used to solve the problem of highway traffic congestion.

Description

Technical Field

[0001] The present invention relates to the fields of highway operation and management, artificial intelligence-intelligent transportation, and in particular to a highway congestion prediction method, system and device integrating spatiotemporal correlation data. Background Art

[0002] Expressways have played an extremely important role in promoting my country's economic development. In recent years, the demand for transportation has grown rapidly. Simply expanding the scale of road network construction cannot effectively solve the problems of imbalance between transportation supply and demand and increasingly serious traffic congestion. Therefore, how to achieve efficient operation and management of expressways under existing road conditions has become an urgent problem to be solved. Accurately grasping the traffic status of expressways is the key prerequisite for scientific management and decision-making, and is also the basic condition for road network guidance and traffic control.

[0003] Highway traffic status prediction is an important basis for highway operation and management, and is an important part of building an intelligent highway system. By providing accurate traffic status information and prediction information, it can provide scientific decision-making basis for management departments to formulate traffic control measures (such as pre-preventive measures, post-evacuation strategies, etc.), so that they can reasonably and effectively guide vehicles to pass, thereby achieving balanced traffic distribution on the highway network and maximizing the traffic capacity of existing roads. By publishing real-time and accurate traffic status and future trend information, traffic travelers can make more reasonable travel decision plans, dynamically adjust travel time, travel path and travel mode, and improve travel stability and travel quality.

[0004] Existing technologies mainly focus on the research of short-term prediction methods of traffic status indicator variables, but existing traffic flow prediction methods are often based on a single traffic flow parameter and only consider the parameter values ​​of the first few periods of the prediction period as features, while ignoring the inherent correlation between multiple traffic flow parameters and the influence of external characteristics such as weather and holidays. Therefore, when predicting traffic flow parameters, it is necessary to comprehensively consider the inherent correlation between different traffic flow parameters and introduce more characteristic factors to further improve the prediction accuracy, stability and applicability of the model.

[0005] Existing technologies focus more on algorithm innovation and improvement, but in fact, existing machine learning algorithms are relatively mature and can achieve good discrimination results even without innovation and improvement. The biggest problem is that the stability of the discrimination results of a single algorithm cannot be guaranteed, because traffic flow is uncertain. For traffic flow data under different time and space conditions, the discrimination accuracy of the classification algorithm also changes dynamically. Therefore, it is necessary to combine the advantages of different classification algorithms through ensemble learning to reduce the random error of a single algorithm, so as to further improve the stability of traffic state discrimination results (a highway traffic congestion prediction method based on deep learning (CN202211039892.7)). Summary of the invention

[0006] In order to solve at least one of the technical problems existing in the prior art to a certain extent, the purpose of the present invention is to provide a highway congestion prediction method integrating spatiotemporal correlation data, using highway gantry system data and toll flow data to calculate traffic flow characteristic parameters such as traffic flow, speed, q / C ratio, etc., clustering analysis is performed on the input data with the help of spectral clustering algorithm, and traffic state discrimination is performed on the clustering results using machine learning algorithm, and each algorithm is fused based on a soft voting mechanism to construct a traffic state classification decision model based on a crowd intelligence algorithm, and a confusion matrix is ​​introduced to obtain evaluation indicators to quantitatively evaluate the discrimination effect of the model, and the predicted future traffic flow characteristic parameters are input to obtain a high-precision prediction result of the future traffic state of the road section, and further predict whether the road section is likely to be congested.

[0007] The purpose of the present invention is achieved by at least one of the following technical solutions.

[0008] A highway congestion prediction method integrating spatiotemporal correlation data comprises the following steps:

[0009] S1. Based on the historical highway gantry traffic data collected by the highway ETC gantry, the basic characteristic parameter data of the road section traffic flow is obtained, including the traffic volume, the average vehicle travel speed, and the ratio of traffic volume to traffic capacity, i.e., the q / C ratio.

[0010] S2. To improve the reliability of highway traffic flow operation data status marking, a generative adversarial network is used to repair the data for the missing flow data, and a Lagrange linear interpolation method is used to repair the data for the missing vehicle average travel speed data;

[0011] S3. Use the spectral clustering algorithm to perform cluster analysis on the basic characteristic parameter data of historical highway traffic flow, determine the relative division criteria of five traffic states: smooth, basically smooth, light congestion, moderate congestion, and severe congestion, and finally obtain a traffic state data set consisting of five clusters;

[0012] S4. The traffic status dataset obtained by clustering is used as input, and four commonly used machine learning algorithms, namely random forest (RF), K nearest neighbor (KNN), support vector machine (SVM), and extreme gradient boosting (XGBoost), are used to classify and decide the traffic status. The soft voting mechanism is used to integrate the algorithms to build a highway traffic status classification decision model based on the crowd intelligence algorithm.

[0013] S5. Use the confusion matrix to quantitatively evaluate the classification effect of the classification decision model constructed in step S4. If the classification effect does not meet the user's requirements, readjust the model parameters and optimize the algorithm until the classification effect meets the discrimination requirements, thereby obtaining a trained classification decision model.

[0014] S6. Count the traffic data from the real-time traffic data of the highway gantry, including the traffic of the research section and the upstream section, and the time feature data (week, time period, holidays, etc.) as input, use the section traffic prediction model based on the multi-feature XGBoost algorithm to obtain the traffic prediction value of the section in the future time period, and calculate the q / C ratio of the section in the future time period;

[0015] S7, taking the speed data of the target section from the real-time traffic data of the expressway gantry and the predicted value of the section flow obtained in step S6 as input, and obtaining the predicted speed value by using the travel speed prediction model based on the TS-SVR algorithm;

[0016] S8. Take the predicted value of road traffic flow, predicted value of travel speed, and predicted value of q / C ratio as input, and rely on the trained classification decision model output in step S5 to obtain the category of traffic status in the future time period, thereby achieving effective prediction of highway traffic status to determine whether there is a possibility of congestion on the road section.

[0017] Furthermore, in step S1, a method for calculating the average travel speed of vehicles based on gantry traffic data includes:

[0018]

[0019] The vehicle passes through gantry A and gantry B successively. represents the average travel speed of vehicles between gantry A and gantry B, N represents the number of vehicles passing during the observation time, and t i，A represents the time when the vehicle passes through gantry A, t i，B It represents the moment when the vehicle passes through gantry B, and d represents the distance between gantry A and gantry B, which can be calculated based on the highway GIS section information and the latitude and longitude coordinates of gantry A and gantry B.

[0020] Furthermore, in step S2, the highway flow data is normalized using maximum and minimum normalization, as follows:

[0021]

[0022] Among them, x is the preprocessed flow data; x real Represents the original traffic data; x min is the minimum value in the original flow data; x max is the maximum value in the original flow data.

[0023] Furthermore, in step S3, flow rate, average vehicle travel speed, and q / C ratio are selected as traffic state indicator variables to classify different traffic states, including:

[0024] Using relative measurement standards, rather than fixed index values ​​to classify traffic status levels, after fully considering various influencing factors, the measured traffic flow data of the road section is used to dynamically describe the traffic status of the road section under specific time and space. The basic idea is to directly classify according to the service level numbers stipulated internationally or domestically, or to combine the service levels of existing standards and use fewer levels to describe the traffic status of the road section; in order to be consistent with national technical standards and specifications, the traffic status of the basic sections of the expressway is divided into five levels: unblocked, basically unblocked, lightly congested, moderately congested, and severely congested, with reference to the service level numbers stipulated in the "Urban Traffic Status Evaluation Specifications", and then cluster analysis is performed on the historical traffic flow data under different time and space conditions using a cluster analysis algorithm, and the relative classification standards for different traffic states are determined based on the clustering results;

[0025] In traffic status judgment, traffic flow parameters are the most commonly used measurement indicators, but when applied, it is necessary to comprehensively consider the inherent correlation between the parameters, and a single parameter cannot be used as a judgment criterion;

[0026] When widely used application software such as Amap and Baidu Map divide road traffic status, they only consider speed indicators and ignore other indicators such as flow rate and occupancy rate. This is obviously incomplete because the same speed value may correspond to two different traffic status. Vehicles are also running at high speed, but driving in the "high speed + small flow" state and the "high speed + large flow" state will make travelers feel significantly different psychologically. The US "Road Capacity Manual" comprehensively considers multiple indicators such as flow rate, speed, saturation and load factor to divide traffic status. Among them, flow rate and speed indicators directly reflect the demand for traffic, the comfort and flexibility of travelers, and saturation and load factor effectively reflect the road service level. Compared with a single parameter, using multiple indicators to measure traffic status is obviously more comprehensive.

[0027] In summary, considering the requirement of data acquisition convenience from a practical point of view, three parameters, namely flow rate, average vehicle travel speed and q / C ratio, are selected to comprehensively characterize the traffic status of highways.

[0028] Furthermore, in step S4, a traffic state classification decision model of the new three traffic parameters is constructed based on the crowd intelligence algorithm, including:

[0029] With reference to previous studies and using existing data to conduct experimental tests, four classic machine learning classification algorithms, namely random forest (RF), support vector machine (SVR), K nearest neighbor (KNN), and extreme gradient boosting algorithm (XGBoost), are selected to identify the traffic state of traffic flow data, and the algorithms are integrated according to the soft voting mechanism to construct a traffic state classification decision model based on the crowd intelligence algorithm. By organically combining the results of clustering analysis and classification decision, a traffic state identification method based on spectral clustering and crowd intelligence algorithm is proposed. This method takes the flow of the target section, the average travel speed of the vehicle, and the q / C ratio as input, and the category of the traffic state as output.

[0030] Furthermore, in step S5, the traffic status data set consisting of five clusters includes indicator variables and status labels.

[0031] Furthermore, in step S5, the classification decision model performance evaluation based on the confusion matrix includes:

[0032] In order to quantitatively evaluate the effect of highway traffic status classification decision-making, the confusion matrix is ​​introduced to quantitatively analyze the accuracy of the classification decision-making model. The traffic status confusion matrix can intuitively reflect the distribution of the actual traffic status and the traffic status classification decision results, and reflects the traffic status classification accuracy of the classification decision-making model.

[0033] Each column of the confusion matrix represents the classification decision traffic state, and the total number of each column represents the number of data predicted as the traffic state; each row represents the actual traffic state, and the total number of data in each row represents the number of data instances of the traffic state. The traffic state is divided into five standards: unimpeded, basically unimpeded, light congestion, moderate congestion, and severe congestion;

[0034] Four evaluation indicators, namely accuracy, precision, recall and F-measure, are introduced into the confusion matrix to characterize the classification accuracy from different perspectives. jj represents the number of classification results when both the actual traffic state and the classified traffic state are j = {1, 2, 3, 4, 5};

[0035] The formula describing the overall discrimination accuracy Acc of the traffic state classification decision model is: Where P is the total number of samples in the test set;

[0036] The formula describing the percentage accuracy Prec of the correct classification in the classification results of the traffic state classification decision model is: Where P jis the number of test set samples with classified traffic state j;

[0037] The formula describing the accuracy recall rate Rec of each type of traffic state is: Where P j It refers to the number of test set samples with actual traffic state j;

[0038] The formula describing the stability F measure of the classification decision model is: It is assumed that precision and recall are equally important, and their weights are both 1, where F j It refers to the harmonic mean of the number of test set samples with actual traffic state j and the number of test set samples with classified traffic state j.

[0039] Furthermore, in step S5, adjusting the model parameters and optimizing the algorithm include:

[0040] A random sampling method is used to extract 70% of the data from the traffic status dataset as a training set, and 30% as a test set. The training set is input into the constructed traffic status classification decision model, that is, four base classifiers constructed based on RF, KNN, SVM, and XGBoost algorithms, respectively. During the process of model training and parameter tuning, the KNN model uses cross-validation method to determine the K value, and the RF, SVM, and XGBoost models all use grid search method to determine model parameters.

[0041] A highway congestion prediction system integrating spatiotemporal correlation data, comprising:

[0042] The data processing module is used to clean the highway gantry traffic data, calculate the basic characteristic parameter data of traffic flow, including flow, speed, q / C ratio, and pre-process the data;

[0043] A traffic state classification module is used to determine the relative classification standards of five traffic states: unblocked, basically unblocked, lightly congested, moderately congested, and severely congested, and to perform cluster analysis on the input data using a spectral clustering algorithm to make classification decisions on the traffic state, thereby obtaining a traffic state data set consisting of five clusters and including indicator variables and state labels;

[0044] The traffic prediction module is used to input the indicator variable values ​​such as section flow and speed extracted from the real-time collected highway traffic flow data into the trained section flow prediction model and the travel speed short-term prediction model to obtain the future flow, speed and q / C ratio prediction data of each section.

[0045] The traffic status identification module is a highway traffic status classification decision model based on the swarm intelligence algorithm. It takes the future flow, speed, and q / C ratio prediction data of the road section obtained by the traffic prediction module as input, and outputs the future traffic status prediction of the road section to determine whether the road section is likely to be congested.

[0046] A highway congestion prediction device integrating spatiotemporal correlation data, comprising:

[0047] at least one processor;

[0048] at least one memory for storing at least one program;

[0049] When the at least one program is executed by the at least one processor, the at least one processor implements a highway congestion prediction method that integrates spatiotemporal correlation data.

[0050] The beneficial effects of the present invention are:

[0051] The present invention uses spectral clustering to determine the relative division standard for traffic status, and then uses the classification decision model to realize real-time judgment of highway traffic status. The traffic status judgment method based on spectral clustering and swarm intelligence algorithm effectively reduces the judgment error of a single classification algorithm, further improves the robustness and generalization ability of the classification decision model, and can well judge the traffic status of basic sections of highways.

[0052] The present invention takes a large amount of highway toll data as the research object, and the accuracy and reliability of the data lay a solid foundation for the implementation of the present invention;

[0053] Based on the data of the highway gantry system, the factors affecting the traffic flow of the road section were analyzed, and four new features were found: upstream road section traffic flow, day of the week, time period, and holidays. New features were introduced on the basis of the XGBoost algorithm, and then a short-term prediction model for road section traffic flow based on the multi-feature XGBoost algorithm was established.

[0054] In order to predict the travel speed, considering the strong correlation between speed and flow, the flow prediction is combined with the speed prediction, and a short-term travel speed prediction model based on the TS-SVR algorithm is constructed.

[0055] Based on the above traffic and speed prediction model, a new highway congestion prediction method integrating spatiotemporal correlation data is constructed with three new traffic parameters, namely traffic, speed and q / C ratio, as indicator variables.

[0056] The present invention combines the advantages of different classification algorithms through ensemble learning, reduces the random error of a single algorithm, and further improves the stability of traffic state discrimination results. When predicting traffic flow parameters, the intrinsic correlation between different traffic flow parameters should be comprehensively considered, and more characteristic factors should be introduced. Compared with the existing method model prediction results, the prediction accuracy, stability and applicability are higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the embodiments of the present invention or the drawings of related technical solutions in the prior art are introduced below. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 It is an implementation flow chart of a highway congestion prediction method integrating spatiotemporal correlation data in an embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of calculating the average travel speed in an embodiment of the present invention;

[0060] Figure 3 Schematic diagram of traffic state prediction effect evaluation index in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, and embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limitations of the present invention. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0062] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0063] In the description of the present invention, "several" means one or more, "more" means more than two, "greater than", "less than", "exceed" etc. are understood as not including the number itself, and "above", "below", "within" etc. are understood as including the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0064] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.

[0065] Example:

[0066] like Figure 1 As shown, this embodiment provides a highway congestion prediction method integrating spatiotemporal correlation data, including the following steps:

[0067] S1. The section between Dayouyuan Interchange and the intersection of Dongguan-Shenzhen Expressway and Songshan Lake was selected as the research section from May 30 to June 5, 2021. The pre-processed historical traffic flow data of this section was used to construct a sample set consisting of three parameters: section flow, average vehicle travel speed, and q / C ratio.

[0068] In one embodiment, a method for calculating the average travel speed of vehicles based on gantry traffic data is provided, such as Figure 2 As shown, including:

[0069]

[0070] The vehicle passes through gantry A and gantry B successively. represents the average travel speed of vehicles between gantry A and gantry B, N represents the number of vehicles passing during the observation time, and t i，A represents the time when the vehicle passes through gantry A, t i，B It represents the moment when the vehicle passes through gantry B, and d represents the distance between gantry A and gantry B, which can be calculated based on the highway GIS section information and the latitude and longitude coordinates of gantry A and gantry B.

[0071] S2. Perform cluster analysis on the sample set data by using a spectral clustering algorithm, determine the relative metrics of the five traffic states of the highway, and then use the traffic state data set obtained by clustering to train the classification decision model and save it. In one embodiment, the sample size of the traffic state data obtained by clustering is shown in Table 1;

[0072] Table 1 Sample size of each traffic state (Dayouyuan Interchange and Dongguan-Shenzhen Expressway Interchange Connection-Songshan Lake)

[0073]

[0074] S3, taking the traffic data (study section traffic and upstream section traffic) of the first 10 weeks of the prediction period after cleaning in step S1 and the time feature data (week, time period, holidays, etc.) as input, predicting the section traffic based on the multi-feature XGBoost model and calculating the q / C ratio;

[0075] S4, using the speed data of the first 10 weeks of the prediction period after cleaning in step S1 and the predicted value of the road section flow obtained in step S3 as input, and using the TS-SVR model to predict the speed;

[0076] S5. Take the predicted value of road traffic flow, the predicted value of average vehicle travel speed, and the predicted value of q / C ratio as input, and rely on the output of the swarm intelligence model trained in step S2 to obtain the category of traffic status in the future time period, thereby achieving effective prediction of traffic status.

[0077] In one embodiment, analyzing Table 2 and Figure 3 It can be seen that the overall prediction accuracy of the highway congestion prediction method integrating spatiotemporal correlation data proposed by the present invention is as high as 92.86%. The prediction accuracy of this method under different traffic conditions is above 80%, and it even reaches 100% for severe congestion conditions, indicating that it is feasible to use this method to solve the traffic state prediction problem, and it has a good and stable prediction effect.

[0078] Table 2 Performance evaluation index of multi-classification decision model Traffic status data set obtained by clustering

[0079]

[0080] A highway congestion prediction system integrating spatiotemporal correlation data of this embodiment can execute a highway congestion prediction method integrating spatiotemporal correlation data provided by a method embodiment of the present invention, can execute any combination of implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0081] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.

[0082] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0083] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0084] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0085] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0086] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0087] In the above description of this specification, the description with reference to the terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0088] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0089] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A highway congestion prediction method integrating spatiotemporal correlation data, characterized in that: The following steps are involved: S1. Based on the historical highway gantry traffic data collected by the highway ETC gantry, the basic characteristic parameter data of the road section traffic flow is obtained, including the traffic volume, the average vehicle travel speed, and the ratio of traffic volume to traffic capacity, i.e., the q / C ratio. S2. To improve the reliability of highway traffic flow operation data status marking, a generative adversarial network is used to repair the data for the missing flow data, and a Lagrange linear interpolation method is used to repair the data for the missing vehicle average travel speed data; S3. Use the spectral clustering algorithm to perform cluster analysis on the basic characteristic parameter data of historical highway traffic flow, determine the relative division criteria of five traffic states: smooth, basically smooth, light congestion, moderate congestion, and severe congestion, and finally obtain a traffic state data set consisting of five clusters; S4. The traffic status dataset obtained by clustering is used as input, and four machine learning algorithms, namely, random forest RF, K nearest neighbor KNN, support vector machine SVM, and extreme gradient boosting XGBoost, are used to classify and decide the traffic status. The soft voting mechanism is used to integrate the algorithms to build a highway traffic status classification decision model based on the crowd intelligence algorithm. S5. Use the confusion matrix to quantitatively evaluate the classification effect of the classification decision model constructed in step S4. If the classification effect does not meet the user's requirements, readjust the model parameters and optimize the algorithm until the classification effect meets the discrimination requirements, thereby obtaining a trained classification decision model. S6. Count the flow data from the real-time traffic data of the highway gantry, including the flow of the research section and the flow of the upstream section, and the time feature data as input, use the section flow prediction model based on the multi-feature XGBoost algorithm to obtain the flow prediction value of the section in the future period, and calculate the q / C ratio of the section in the future period; S7, taking the speed data of the target section from the real-time traffic data of the expressway gantry and the predicted value of the section flow obtained in step S6 as input, and obtaining the predicted speed value by using the travel speed prediction model based on the TS-SVR algorithm; S8. Take the predicted value of road traffic flow, predicted value of travel speed, and predicted value of q / C ratio as input, and rely on the trained classification decision model output in step S5 to obtain the category of traffic status in the future time period, thereby achieving effective prediction of highway traffic status to determine whether there is a possibility of congestion on the road section.

2. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S1, a method for calculating the average travel speed of vehicles based on gantry traffic data includes: Among them, the vehicle passes through gantry A and gantry B successively, v AB represents the average travel speed of vehicles between gantry A and gantry B, N represents the number of vehicles passing during the observation time, and t i,A represents the time when the vehicle passes through gantry A, t i,B It represents the moment when the vehicle passes through gantry B, and d represents the distance between gantry A and gantry B, which can be calculated based on the highway GIS section information and the latitude and longitude coordinates of gantry A and gantry B.

3. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S2, the highway flow data is normalized using maximum and minimum standardization, as follows: Among them, x is the preprocessed flow data; x real Represents the original traffic data; x min is the minimum value in the original flow data; x max is the maximum value in the original flow data.

4. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S3, flow rate, average vehicle travel speed, and q / C ratio are selected as traffic state indicator variables to classify different traffic states, including: Relative measurement standards are used instead of fixed index values ​​to divide traffic status levels. Instead, after fully considering various influencing factors, the measured traffic flow data of the road section is used to dynamically describe the traffic status of the road section under specific time and space. The classification is directly based on the number of service levels stipulated internationally or domestically, or the service levels of existing standards are combined to describe the traffic status of the road section with fewer levels. The traffic status of the basic sections of the expressway is divided into five levels: unblocked, basically unblocked, lightly congested, moderately congested, and severely congested. The cluster analysis algorithm is then used to cluster the historical traffic flow data under different time and space conditions, and the relative division standards for different traffic status are determined based on the clustering results. Three parameters, flow rate, average vehicle travel speed, and q / C ratio, are selected to comprehensively characterize the traffic status of the expressway.

5. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S4, a traffic state classification decision model of the new three traffic parameters is constructed based on the crowd intelligence algorithm, including: Experimental tests were carried out using existing data, and four classic machine learning classification algorithms, including random forest RF, support vector machine SVR, K nearest neighbor KNN, and extreme gradient boosting algorithm XGBoost, were selected to identify the traffic state of traffic flow data. The algorithms were integrated according to the soft voting mechanism to construct a traffic state classification decision model based on the swarm intelligence algorithm. By organically combining the results of clustering analysis and classification decision-making, a traffic state discrimination method based on spectral clustering and swarm intelligence algorithm was proposed. This method takes the flow rate of the target section, the average travel speed of the vehicle, and the q / C ratio as input, and the category to which the traffic state belongs as output.

6. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S5, the traffic status data set consisting of five clusters includes indicator variables and status labels.

7. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S5, the performance evaluation of the classification decision model based on the confusion matrix includes: In order to quantitatively evaluate the effect of highway traffic status classification decision-making, the confusion matrix is ​​introduced to quantitatively analyze the accuracy of the classification decision-making model. The traffic status confusion matrix can intuitively reflect the distribution of the actual traffic status and the traffic status classification decision results, and reflects the traffic status classification accuracy of the classification decision-making model. Each column of the confusion matrix represents the classification decision traffic state, and the total number of each column represents the number of data predicted as the traffic state; each row represents the actual traffic state, and the total number of data in each row represents the number of data instances of the traffic state; the traffic state is divided into five standards: unblocked, basically unblocked, light congestion, moderate congestion, and severe congestion; Four evaluation indicators, namely accuracy, precision, recall and F-measure, are introduced into the confusion matrix to characterize the classification accuracy from different perspectives. jk Indicates the number of classification results when the actual traffic state is j∈{1,2,3,4,5} and the classified traffic state is k∈{1,2,3,4,5}; The formula describing the overall discrimination accuracy Acc of the traffic state classification decision model is: Where P is the total number of samples in the test set; The formula describing the percentage accuracy Prec of the correct classification in the classification results of the traffic state classification decision model is: Where j = k, P k is the number of test set samples with classified traffic state k; The formula describing the accuracy recall rate Rec of each type of traffic state is: Where k = j, P j It refers to the number of test set samples with actual traffic state j; The formula describing the stability F measure of the classification decision model is: It is assumed that precision and recall are equally important, and their weights are both 1, where F jk It refers to the harmonic mean of the number of test set samples with actual traffic state j and the number of test set samples with classified traffic state k.

8. The highway congestion prediction method integrating spatiotemporal correlation data according to claim 1 is characterized in that: In step S5, the adjusting of model parameters and optimization algorithm include: A random sampling method is used to extract 70% of the data from the traffic status dataset as a training set, and 30% as a test set. The training set is input into the constructed traffic status classification decision model, that is, four base classifiers constructed based on RF, KNN, SVM, and XGBoost algorithms, respectively. During the process of model training and parameter tuning, the KNN model uses cross-validation method to determine the K value, and the RF, SVM, and XGBoost models all use grid search method to determine model parameters.

9. A highway congestion prediction system integrating spatiotemporal correlation data, characterized in that: include: The data processing module is used to clean the highway gantry traffic data, calculate the basic characteristic parameter data of traffic flow, including flow, speed, q / C ratio, and pre-process the data; The traffic status classification module is used to determine the relative classification standards of the five traffic statuses: smooth, basically smooth, lightly congested, moderately congested, and severely congested. The spectral clustering algorithm is used to perform cluster analysis on the input data, and the traffic status is classified and decided to obtain a traffic status data set consisting of five clusters containing indicator variables and status labels. Traffic prediction module, used to input the road section flow and speed variable values ​​extracted from the real-time collected highway traffic flow data into the trained road section flow prediction model and the travel speed short-term prediction model to obtain the future flow, speed and q / C ratio prediction data of each road section; The traffic status identification module is a highway traffic status classification decision model based on the swarm intelligence algorithm. It takes the future flow, speed, and q / C ratio prediction data of the road section obtained by the traffic prediction module as input, and outputs the future traffic status prediction of the road section to determine whether the road section is likely to be congested.

10. A highway congestion prediction device integrating spatiotemporal correlation data, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the highway congestion prediction method integrating spatiotemporal correlation data as described in any one of claims 1 to 8.

Images