Traffic busy degree prediction method and system for traffic checkpoint
By constructing time series samples and using a long short-term memory network model to predict traffic congestion indicators, the problem of insufficient foresight in traffic checkpoint control systems is solved, and real-time optimization of traffic signals and navigation routes is achieved, thereby improving the efficiency and smoothness of the traffic system.
Patent Information
- Application Number
- CN202510812733.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-18
- Publication Date
- 2025-10-24
AI Technical Summary
The existing traffic control system lacks forward-looking prediction capabilities and is unable to dynamically adapt to the spatiotemporal changes in traffic flow, resulting in signal timing lags and route guidance failures, making it impossible to respond to sudden traffic jams in a timely manner.
By collecting raw data from traffic checkpoints, constructing time-series samples, and combining external auxiliary data for data cleaning, a long short-term memory network model is used to predict traffic congestion indicators, which are then sent to the traffic management platform for real-time control.
It achieves accurate prediction of traffic congestion at the minute level, supports real-time dynamic optimization of traffic signals and active congestion avoidance guidance of the navigation system, and improves road traffic efficiency and the balance of the regional transportation network.
Smart Images
Figure CN120833677A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of traffic control, in particular to a traffic congestion prediction method and system for traffic checkpoints. BACKGROUND
[0002] With the rapid development of urban traffic, traffic congestion has gradually become a major problem in urban management. As an important traffic management node, the traffic flow and passing efficiency of traffic checkpoints have an important influence on the smoothness of the entire city traffic. Traditional traffic congestion prediction methods mainly rely on static traffic flow data and manual experience, which often cannot adapt to complex and variable traffic conditions, resulting in low accuracy of prediction results and difficulty in realizing real-time dynamic regulation. With the continuous development of big data and artificial intelligence technology, data-driven traffic congestion prediction methods have gradually become an effective means to solve traffic management problems. At present, many traffic checkpoint management systems only rely on local traffic information, lacking comprehensive analysis of global traffic conditions, resulting in lag in traffic signal and path planning adjustment, and being unable to respond to sudden traffic congestion. At the same time, the influence of weather, holidays and road construction and other external factors on traffic flow is often not fully considered, resulting in traditional traffic management schemes failing to meet the growing traffic demand. Therefore, how to construct an efficient and accurate traffic congestion prediction method based on real-time traffic data and external auxiliary data, and then provide real-time and accurate regulation basis for traffic management platform, has become a technical problem to be solved in the current traffic management field. SUMMARY
[0003] The present application provides a traffic congestion prediction method and system for traffic checkpoints, aiming to solve the technical problem that existing traffic checkpoint regulation lacks forward-looking prediction ability, is difficult to dynamically adapt to the spatio-temporal variation characteristics of traffic flow, resulting in signal timing lag and path guidance failure.
[0004] The first aspect of the present application provides a traffic congestion prediction method for traffic checkpoints, the method comprising: collecting original traffic data of the traffic checkpoint to construct a time series sample; predicting a traffic congestion index in a target time period through the time series sample; sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths.
[0005] Another aspect of the present application provides a traffic congestion prediction system for traffic checkpoints, the system comprising: a data collection module for collecting original traffic data of the traffic checkpoint to construct a time series sample; a congestion prediction module for predicting a traffic congestion index in a target time period through the time series sample; a traffic regulation module for sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths.
[0006] The one or more technical solutions provided in the application have at least the following technical effects or advantages.
[0007] The traffic congestion prediction method for the traffic checkpoint first collects original traffic data of the traffic checkpoint and arranges the data into time sequence samples. Then, the traffic congestion index in a target time period is predicted by analyzing the time sequence samples. Finally, the predicted congestion index is sent to a traffic management platform for real-time regulation of traffic signals and navigation paths, so as to optimize traffic flow and reduce congestion.
[0008] The above description is only a summary of the technical solutions of the application. In order to more clearly understand the technical means of the application, the application can be implemented according to the content of the specification, and in order to make the above and other purposes, characteristics and advantages of the application more obvious and easy to understand, the following specific embodiments of the application are described. BRIEF DESCRIPTION OF DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.
[0010] Figure 1 A flowchart of the traffic congestion prediction method for the traffic checkpoint in an embodiment.
[0011] Figure 2 A system architecture diagram of the traffic congestion prediction system for the traffic checkpoint in an embodiment.
[0012] Explanation of reference signs: data acquisition module 11, congestion prediction module 12, traffic regulation module 13. DETAILED DESCRIPTION
[0013] The embodiments of the application provide a traffic congestion prediction method and system for a traffic checkpoint, which solve the technical problem that the existing traffic checkpoint regulation lacks forward-looking prediction ability and is difficult to dynamically adapt to the space-time variation characteristics of traffic flow, resulting in signal timing lag and path guidance failure.
[0014] The technical solutions in the embodiments of the application will be described clearly and completely in combination with the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.
[0015] It is to be understood that the terms "including", "comprising", "having" and
[0016] As shown in Figure 1, the present application provides a traffic congestion prediction method for traffic checkpoints, which comprises: Figure 1
[0017] Collecting original traffic data of the traffic checkpoint, and constructing time series samples.
[0018] In the embodiments of the present application, real-time traffic flow, vehicle speed, lane occupancy rate and other data are obtained from various sensors or monitoring devices of the traffic checkpoint. By recording these data in real time, an original data set is formed. Subsequently, the original traffic data collected in real time is organized in time sequence to construct time series samples for subsequent analysis and prediction. These samples will reflect the changing trend of traffic conditions and provide a data basis for subsequent traffic congestion prediction.
[0019] Further, the present application provides collecting original traffic data of the traffic checkpoint, and constructing time series samples, which comprises:
[0020] Obtaining external auxiliary data, the external auxiliary data including weather data, holiday information, road construction information and traffic event data affecting traffic; performing data cleaning processing on the original traffic data and the external auxiliary data to construct the time series samples, wherein the original traffic data includes traffic flow, average vehicle speed, lane occupancy rate and other basic information.
[0021] Preferably, first, external auxiliary data is obtained, including weather conditions (such as temperature, precipitation, wind speed, etc.), holiday information (such as public holidays and special festivals), road construction conditions (such as road closures, construction progress, etc.), and event data that may affect traffic flow (such as traffic accidents, major activities, etc.), which may have a direct impact on traffic conditions and therefore need to be included in data analysis. Subsequently, combined with the original traffic data and external auxiliary data, the two are subjected to data cleaning, which includes removing erroneous or incomplete records, filling in missing data, standardizing the formats of different data sources, and other steps. The cleaned data is more accurate and consistent, facilitating further analysis. Finally, based on the cleaned data, time series samples are constructed, which will combine the cleaned original traffic data (such as traffic volume, average speed, lane occupancy rate, etc.) with external auxiliary data in chronological order to form continuous, time-labeled data sequences, providing reliable data support for subsequent traffic congestion prediction.
[0022] The traffic congestion index in the target time period is predicted by the time series sample.
[0023] In one embodiment, after obtaining the time series sample, feature variables such as historical period statistical features and time periodicity features are extracted from the constructed time series sample, and the extracted features are input into a traffic prediction model. The model analyzes historical data patterns, trends, and external factors (such as weather, holidays, etc.) to predict future traffic key indicators such as traffic volume. Finally, combining road traffic levels and prediction results, the traffic congestion index is calculated to reflect the expected changes in traffic conditions during the time period, such as whether traffic congestion will occur or traffic will flow smoothly, thereby providing effective decision-making basis for traffic management departments.
[0024] Further, the present application provides a method for predicting a traffic congestion index in a target time period by a time series sample, comprising:
[0025] The continuous traffic data in the time series sample is subjected to sliding window processing to form a sample sequence, wherein the sample sequence has a time label; feature variables of the sample sequence are extracted, wherein the feature variables include historical period statistical features, time periodicity features, upstream and downstream tollgate flow difference features, and weather event embedding vectors; the feature variables are input into a traffic prediction model, and a traffic volume prediction value of a target time period is output, wherein the traffic prediction model is a long short-term memory network; and the traffic congestion index is calculated based on the traffic volume prediction value.
[0026] Preferably, firstly, the continuous traffic data in the time series sample is processed by a sliding window, and the time series data is divided into multiple small time periods, each containing a certain number of continuous data points. Each time the window slides forward, the data contained therein is updated, thereby forming multiple sample sequences, each containing a time label identifying the traffic condition of the time period. Subsequently, feature variables are extracted from these sample sequences, which are key indicators for describing traffic conditions, including historical period statistical features (such as average traffic flow and speed in the past few hours or days), time periodicity features (such as different time periods in a day or periodic changes in a week), upstream and downstream traffic flow difference features (i.e. traffic flow changes between adjacent checkpoints), and weather event embedding vectors (weather data is integrated into the sample as a feature to reflect the impact of weather on traffic). After that, these feature variables are input into the traffic prediction model for analysis. This traffic prediction model is obtained by using historical traffic features and historical traffic flow data (such as historical traffic flow, historical confidence score, historical abnormal fluctuation label) to perform forward propagation, loss calculation (mean square error), back propagation, parameter optimization (Adam) and other steps on the long short-term memory network (LSTM) for iterative training. The traffic prediction model outputs the predicted traffic flow value of the target time period based on the input feature variables and the learned mapping relationship. Finally, according to the predicted traffic flow value, the traffic congestion index is calculated, which reflects the degree of traffic congestion in the target time period, usually calculated in combination with road traffic levels and other factors to provide a reference for traffic management decisions.
[0027] Further, the application provides a method for forming a sample sequence, comprising:
[0028] performing sliding window processing based on a preset overlap rate to obtain an enhanced number of samples, wherein the enhanced number of samples has a time semantic feature; and obtaining a sample sequence by multi-thread parallel execution of the enhanced number of samples.
[0029] Optionally, in the sliding window processing process, the time series data is divided into multiple small time windows based on a preset overlap rate, which determines the overlapping part of data between each window. For example, if the overlap rate is 50%, each window will share half of the data with the previous window. In this way, the number of generated sample sequences is increased, which can cover more time periods and different traffic states, ensuring that the model can learn more historical rules. Moreover, these enhanced number of samples have time semantic features, meaning that each sample not only contains the original traffic data but also includes context information for the corresponding time period, such as periodicity, trends, and changes in time, thereby enhancing the understanding of the time dimension. To improve processing efficiency, these enhanced samples are processed in a multi-thread parallel execution manner, with each thread responsible for processing a portion of the samples, thereby speeding up the sample generation and data preprocessing process. Through parallel processing, complete sample sequences can be quickly obtained, further improving the speed and effectiveness of subsequent model analysis.
[0030] Further, the application provides that the traffic flow prediction value includes a confidence score and an abnormal fluctuation marker, the confidence score is generated according to the statistical distribution of the traffic prediction model, and the abnormal fluctuation marker is used to identify short-time mutations. The traffic flow prediction value has a time validity label.
[0031] Optionally, the traffic flow prediction value not only includes the predicted traffic flow but also contains two additional information, namely, the confidence score and the abnormal fluctuation marker. The confidence score is generated according to the statistical distribution of the traffic prediction model, indicating the reliability of the prediction result. The model calculates a score by analyzing the distribution rule of historical data and the matching degree of the current prediction result with the historical data, indicating the credibility of the prediction result. If the confidence score is high, it indicates that the prediction result is more accurate, otherwise it indicates that the prediction uncertainty is larger. The abnormal fluctuation marker is used to identify short-time mutations of traffic flow, for example, when the traffic flow in a certain time period has a sudden and large fluctuation (such as due to accidents, traffic control, etc.), this marker can immediately identify these abnormal fluctuations, so as to respond in time. In addition, the traffic flow prediction value also has a time validity label, indicating the time range or validity period to which the prediction value is applicable. This is to ensure that in the time series, the prediction result can be updated synchronously with the actual traffic situation, ensuring the accuracy and effectiveness of the prediction value in the future period of time.
[0032] Further, the application provides that the congestion index is calculated by combining the road traffic level and the traffic flow prediction value, and the congestion index is converted into the traffic congestion index. The congestion index is obtained by comparing the ratio of the road traffic level and the traffic flow prediction value with the preset congestion index level. The congestion index includes smooth, light congestion, moderate congestion, and severe congestion.
[0033] Optionally, when calculating the traffic congestion index, first combine the road traffic level and the traffic flow prediction value to calculate a congestion index. The road traffic level generally represents the carrying capacity of the road and is divided into different levels, such as high-level expressway or highway, urban trunk road, secondary trunk road, etc. Different levels of roads have different carrying capacities and traffic flow management. The traffic flow prediction value reflects the traffic flow situation in the target time period. The calculation method of the congestion index is to calculate the ratio of the traffic flow prediction value to the road traffic level, and then compare it with the preset congestion index level. The larger the traffic flow prediction value and the lower the road traffic level, the higher the congestion index, indicating that the road congestion is more serious. Conversely, if the traffic flow is less or the road traffic level is higher, the congestion index is lower. Finally, the calculated congestion index is used as the traffic congestion index to quantify the traffic congestion level. Different congestion indexes correspond to different traffic states, such as smooth, light congestion, moderate congestion, and severe congestion. Each traffic state has a congestion index interval. In summary, through the above method, the quantification of the traffic congestion index can more effectively evaluate the current and future traffic conditions, timely discover potential traffic congestion problems, and provide accurate data support for traffic signal regulation, route planning, and traffic flow optimization, which helps to improve road traffic efficiency and reduce traffic congestion.
[0034] The traffic congestion index is sent to a traffic management platform to regulate traffic signals and navigation paths.
[0035] In one embodiment, after the calculated traffic congestion index is sent to the traffic management platform, the platform can regulate traffic signals and navigation paths according to these indexes. Specifically, the traffic management platform will adjust the change strategy of traffic signal lights according to the real-time traffic congestion, such as timely extending or shortening the green light time, optimizing the traffic priority of the lane, to alleviate traffic congestion or improve the smoothness. At the same time, the platform can also adjust the navigation path according to the traffic congestion index, such as providing alternative paths or detour suggestions for drivers when the traffic congestion in a certain area is high, avoiding traffic-intensive areas, thereby dispersing traffic and reducing congestion. This regulation can respond to traffic changes in real time and help improve the efficiency and capacity of the entire traffic system.
[0036] Further, the present application provides sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths, including:
[0037] The traffic control center is linked with the intelligent signal light control system to provide lane traffic priority adjustment suggestions according to the traffic congestion index; navigation suggestions are planned based on the lane traffic priority adjustment suggestions; and navigation paths are regulated according to the navigation suggestions.
[0038] Preferably, after the traffic control center is linked with the intelligent signal light control system, the traffic control center provides real-time lane priority adjustment suggestions according to the traffic congestion index received by the traffic control center, for example, during a traffic rush period, some lanes are suggested to be given priority for passing, or the cycle of the signal light is adjusted according to the change of traffic flow to ensure that the traffic flow flows more smoothly. According to these lane priority adjustment suggestions, the traffic control center will further plan navigation suggestions. Specifically, new driving routes will be provided for drivers to avoid high-flow or congested road sections, thereby reducing traffic pressure and improving traffic efficiency. These suggestions are usually pushed to drivers in real time through a navigation system. Finally, based on these navigation suggestions, the traffic control center will regulate the navigation path to ensure that vehicles travel according to the optimized path and avoid congested areas, thereby achieving more efficient traffic management and traffic scheduling. This intelligent linkage mechanism can dynamically respond to the actual traffic situation, effectively alleviate traffic congestion and improve road use efficiency.
[0039] In summary, the embodiments of the present application have at least the following technical effects:
[0040] The embodiments of the present application first collect the original traffic data of the traffic checkpoint to construct a time series sample, then predict the traffic congestion index in the target time period through the time series sample, and finally send the traffic congestion index to the traffic management platform to regulate the traffic signal and the navigation path. These technical effects collectively solve the technical problems that the existing traffic checkpoint regulation lacks forward-looking prediction ability, is difficult to dynamically adapt to the spatiotemporal variation characteristics of traffic flow, and leads to signal timing lag and path guidance failure, achieving the technical effects of realizing minute-level traffic congestion precision prediction through time series deep learning, supporting real-time dynamic optimization of traffic signals and active congestion avoidance guidance of navigation systems, and improving road traffic efficiency and regional traffic network balance.
[0041] Embodiment two, based on the same inventive concept as the traffic congestion prediction method for traffic checkpoints in the foregoing embodiments, as shown in Figure 2 The present application provides a traffic congestion prediction system for traffic checkpoints, which comprises: a data acquisition module 11 for acquiring original traffic data of the traffic checkpoint to construct a time series sample; a congestion prediction module 12 for predicting a traffic congestion index in a target time period through the time series sample; and a traffic regulation module 13 for sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths.
[0042] Further, the data acquisition module 11 is further used to perform the following method:
[0043] Obtaining external auxiliary data, the external auxiliary data including weather data, holiday information, road construction information and traffic event data; performing data cleaning processing on the original traffic data and the external auxiliary data to construct the time series sample, wherein the original traffic data includes traffic flow, average speed, lane occupancy rate and other basic information.
[0044] Further, the congestion prediction module 12 is further configured to perform the following method:
[0045] Performing sliding window processing on the continuous traffic data in the time series sample to form a sample sequence, wherein the sample sequence has a time label; extracting feature variables of the sample sequence, wherein the feature variables include historical period statistical features, time periodicity features, upstream and downstream toll gate flow difference features and weather event embedding vectors; inputting the feature variables into a traffic prediction model to output a traffic flow prediction value of a target time period, wherein the traffic prediction model is a long short-term memory network; and calculating the traffic congestion index based on the traffic flow prediction value.
[0046] Further, the congestion prediction module 12 is further configured to perform the following method:
[0047] Performing sliding window processing based on a preset overlap rate to obtain an enhanced number of samples, wherein the enhanced number of samples has time semantic features; and obtaining a sample sequence through multi-thread parallel execution of the enhanced number of samples.
[0048] Further, the congestion prediction module 12 is further configured to perform the following method:
[0049] The traffic flow prediction value includes a confidence score and an abnormal fluctuation mark, the confidence score is generated according to the statistical distribution of the traffic prediction model, the abnormal fluctuation mark is used to identify short-time mutations, and the traffic flow prediction value has a time validity label.
[0050] Further, the congestion prediction module 12 is further configured to perform the following method:
[0051] Combining the road traffic level and the traffic flow prediction value to calculate a congestion index, and converting the congestion index into the traffic congestion index, wherein the congestion index is obtained by comparing the ratio of the road traffic level and the traffic flow prediction value with a preset congestion index level, and the congestion index includes smooth, light congestion, moderate congestion and severe congestion.
[0052] Further, the traffic control module 13 is further configured to perform the following method:
[0053] The traffic control center is linked with the intelligent signal light control system, and provides lane traffic priority adjustment suggestion according to the traffic congestion index; plans a navigation suggestion based on the lane traffic priority adjustment suggestion; and regulates a navigation path according to the navigation suggestion.
[0054] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.
[0055] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
[0056] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that fall within the scope of the present application are considered to be covered by the present application. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is intended to include these modifications and variations.
Claims
1. A traffic congestion prediction method for a traffic checkpoint, characterized by, The system comprises: collecting original traffic data of a traffic checkpoint to construct a time series sample; predicting a traffic congestion index in a target time period through the time series sample; sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths.
2. The traffic congestion prediction method for traffic kiosks as claimed in claim 1 wherein, The method comprises: collecting original traffic data of a traffic checkpoint to construct a time series sample, comprising: obtaining external auxiliary data, the external auxiliary data including weather data, holiday information, road construction information, and traffic event data; 3. The traffic congestion prediction method for traffic kiosks as claimed in claim 1 wherein, performing data cleaning processing on the original traffic data and the external auxiliary data to construct the time series sample, wherein the original traffic data includes traffic volume, average speed, lane occupancy rate, and other basic information. The method comprises: performing sliding window processing on continuous traffic data in the time series sample to form a sample sequence, wherein the sample sequence has a time label; extracting feature variables of the sample sequence, wherein the feature variables include historical period statistical features, time periodicity features, upstream and downstream checkpoint traffic difference features, and weather event embedding vectors; inputting the feature variables into a traffic prediction model to output a traffic volume prediction value of the target time period, wherein the traffic prediction model is a long short-term memory network; 4. The traffic congestion prediction method for traffic kiosks as claimed in claim 3, wherein, calculating the traffic congestion index based on the traffic volume prediction value. The method comprises: performing sliding window processing based on a preset overlap rate to obtain an enhanced number of samples, wherein the enhanced number of samples has a time semantic feature; 5. The traffic congestion prediction method for traffic kiosks as claimed in claim 3, wherein, the enhanced number of samples is obtained by multi-thread parallel execution to obtain a sample sequence.
6. The traffic congestion prediction method for traffic kiosks as claimed in claim 3, wherein, The traffic volume prediction value includes a confidence score and an abnormal fluctuation mark, the confidence score is generated according to the statistical distribution of the traffic prediction model, the abnormal fluctuation mark is used to identify short-time mutations, and the traffic volume prediction value has a time validity label.
7. The traffic congestion prediction method for traffic kiosks as claimed in claim 1 wherein, The method comprises: combining road traffic levels and traffic volume prediction values to calculate a congestion index, and converting the congestion index into the traffic congestion index, wherein the congestion index is calculated by comparing the ratio of road traffic levels and traffic volume prediction values with a preset congestion index level, and the congestion index includes smooth, light congestion, moderate congestion, and severe congestion. The method comprises: The traffic command center is linked with an intelligent signal lamp control system to provide lane traffic priority adjustment suggestions according to the traffic congestion index; 8. A traffic congestion prediction system for traffic kiosks, characterized by, planning navigation suggestions based on the lane traffic priority adjustment suggestions; regulating navigation paths according to the navigation suggestions. The system is used to execute the traffic congestion prediction method for a traffic checkpoint according to any one of claims 1-7, comprising: a data collection module for collecting original traffic data of a traffic checkpoint to construct a time series sample; a congestion prediction module for predicting a traffic congestion index in a target time period through the time series sample; a traffic regulation module for sending the traffic congestion index to a traffic management platform to regulate traffic signals and navigation paths.