Expressway traffic flow prediction method and system

Abstract

The invention provides an expressway traffic flow prediction method and system, and the method comprises the steps: obtaining the position of a monitoring point in a target expressway region, and extracting the historical traffic flow data of the corresponding monitoring point in a database based on the position of the monitoring point, the monitoring point being a toll station or a portal arranged on an expressway; constructing a first traffic flow prediction vector based on historical traffic flow data of at least one monitoring point existing in a target road section in the target highway area, and inputting the first traffic flow prediction vector into a pre-trained prediction model to obtain a first predicted traffic flow of the target road section at a target time point; extracting external influence factors of the target road section, and calculating corresponding vectors of the external influence factors; and inputting the first predicted traffic flow at the target time point and the corresponding vector for calculating the external influence factors into a pre-trained correction model, and outputting a second predicted traffic flow after the first predicted traffic flow is corrected. According to the scheme, other external influence factors are further considered, and the prediction precision is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic management and planning, and in particular to a method and system for predicting highway traffic flow. Background Art

[0002] Predicting traffic flow on highway sections is a crucial part of modern traffic management and planning. Its importance is reflected in multiple dimensions. It not only concerns the improvement of traffic efficiency, but also directly affects road safety, environmental protection, economic benefits and many other aspects.

[0003] First, from the perspective of traffic management and diversion, accurate prediction of traffic flow is a prerequisite for formulating effective traffic management strategies. During peak hours or under special weather conditions, road capacity may drop significantly. If the trend of traffic flow cannot be predicted in advance, it will be difficult to implement targeted traffic diversion measures, such as adjusting the timing of signal lights, opening emergency lanes, or implementing temporary traffic control, which may lead to increased traffic congestion, longer delays, and even traffic accidents. By predicting traffic flow, traffic management departments can take precautions to ensure smooth roads and improve the public's travel experience. Secondly, traffic flow prediction is also of great significance to road safety. High traffic flow is often accompanied by a higher risk of traffic accidents, especially during peak travel periods such as holidays and weekends. Through accurate prediction of traffic flow, police resources can be reasonably deployed, patrol monitoring of key sections and time periods can be strengthened, traffic violations and potential safety hazards can be discovered and dealt with in a timely manner, the accident rate can be effectively reduced, and the safety of people's lives and property can be protected.

[0004] In summary, predicting the traffic flow on highway sections is not only a necessary means to improve the level of intelligent traffic management, but also the key to ensuring road safety, promoting environmental protection, optimizing resource allocation, and improving economic benefits. However, existing traffic flow predictions often use historical data of traffic flow for prediction, and the prediction accuracy is not accurate enough. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method for predicting highway traffic flow to eliminate or improve one or more defects existing in the prior art.

[0006] One aspect of the present invention provides a method for predicting highway traffic flow, the method comprising the steps of: Obtaining the location of a monitoring point in a target highway area, and extracting historical traffic flow data of the corresponding monitoring point from a database based on the location of the monitoring point, wherein the monitoring point is a toll station or a gantry set up on a highway; Based on historical traffic flow data of at least one monitoring point in a target section of a target highway area, a first traffic flow prediction vector is constructed and input into a pre-trained prediction model to obtain a first predicted traffic flow of the target section at a target time point; Extracting the external influencing factors of the target road section, calculating the corresponding vectors of the external influencing factors, the corresponding vectors of the external influencing factors include spatial dependence feature vectors and weather feature vectors, in the process of calculating the spatial dependence feature vectors, the spatial dependence feature vectors are calculated based on other target road sections that have a connectivity relationship with each target road section; in the process of calculating the weather feature vectors, the weather feature vectors are calculated based on historical weather data of the target road section; The first predicted traffic flow at the target time point and the corresponding vector of the calculated external influencing factors are input into the pre-trained correction model, and the second predicted traffic flow after the correction of the first predicted traffic flow is output.

[0007] The above scheme is adopted. In the process of predicting the traffic flow, the detection data of the monitoring points of the toll stations and the gantries installed on the highway are first integrated to make a preliminary prediction of the traffic flow of the target section. In the preliminary prediction process, all the monitoring points in the target highway area can be integrated, and the data is not limited to a single section. This scheme further considers other external influencing factors, including spatial dependence and weather characteristics, and corrects the initially predicted traffic flow of the target section through the spatial dependence feature vector and the weather feature vector to ensure the prediction accuracy.

[0008] In some embodiments of the present invention, the method further comprises: comparing the second predicted traffic volume with a preset congestion threshold to determine whether to issue a congestion warning to the corresponding target road section.

[0009] In some embodiments of the present invention, in the step of constructing a first traffic flow prediction vector based on historical traffic flow data of at least one monitoring point existing in a target section of a target highway area, the number of vehicles passing through the monitoring point existing in the target section of the target highway area during a preset first historical time period is used as the value of a dimension of the first traffic flow prediction vector to construct the first traffic flow prediction vector.

[0010] In some embodiments of the present invention, based on historical traffic flow data of at least one monitoring point in a target section of a target highway area, a first traffic flow prediction vector is constructed and input into a pre-trained prediction model, and the step of obtaining a first predicted traffic flow of the target section at a target time point includes: Sequentially number each monitoring point based on the connectivity relationship between the monitoring points; Arrange the first vehicle flow prediction vectors of the monitoring points in a longitudinal order based on the serial number sequence to obtain a vehicle flow prediction matrix; The vehicle flow prediction matrix is ​​input into a pre-trained prediction model, and the prediction model outputs a prediction result matrix, wherein each row in the prediction result matrix corresponds to a first predicted vehicle flow of a target road section.

[0011] In some embodiments of the present invention, in the step of sequentially numbering the monitoring points based on the connectivity relationship between the monitoring points, the monitoring points are sequentially numbered from upstream to downstream based on the upstream and downstream relationship between the monitoring points.

[0012] In some embodiments of the present invention, in the step of calculating the spatial dependency characteristic vector based on other target road segments that are connected to each target road segment, the spatial dependency characteristic vector is calculated based on the spatial dependency weights set for the other target road segments that are connected to each target road segment, and the first predicted traffic flow of the other target road segments that are connected to each target road segment.

[0013] In some embodiments of the present invention, in the step of calculating the spatial dependency feature vector based on the spatial dependency weights set for other target road segments that are connected to each target road segment and the first predicted traffic flow of other target road segments that are connected to each target road segment, the following formula is used to calculate the spatial dependency feature vector: in, represents the spatially dependent eigenvector, , and Indicates the spatial dependency weights of other target road segments that are connected to the target road segment. , and Indicates the first predicted traffic flow corresponding to other target road sections that are connected to the target road section.

[0014] In some embodiments of the present invention, the step of calculating the weather feature vector based on the historical weather data of the target road section includes: Obtaining weather forecast data for a target road section, and encoding various items of the weather forecast data, wherein the items of the weather forecast data include wind speed, temperature, humidity, and precipitation probability; The coded data of each item are combined to obtain the weather feature vector.

[0015] In some embodiments of the present invention, the corresponding vector of the external influencing factors includes an event influence vector. In the calculation step of the event influence vector, the event influence vector includes a first dimension and a second dimension. The first dimension corresponds to holidays, and is 1 if the target time point is a holiday, otherwise it is 0; the second dimension corresponds to large-scale activities, and is 1 if there are large-scale activities in the target highway area in a preset time period before and after the target time point, otherwise it is 0.

[0016] The second aspect of the present invention also provides a highway traffic flow prediction system, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the system implements the steps implemented by the method described above.

[0017] The third aspect of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps implemented by the aforementioned highway traffic flow prediction method.

[0018] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention can be specifically pointed out and obtained in the specification and the accompanying drawings.

[0019] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention.

[0021] Figure 1 A schematic diagram of an implementation of a method for predicting highway traffic flow according to the present invention; Figure 2 Schematic diagram of the processing architecture of the highway traffic flow prediction method of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0023] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0024] In modern traffic management and planning, accurate prediction of highway traffic volume is crucial for traffic safety, congestion relief and resource allocation. Traditional highway traffic volume prediction methods mainly rely on historical data and empirical models, such as time series analysis and regression analysis. These methods can usually provide a certain degree of prediction effect, but have the following limitations: Data limitations: Traditional methods mainly rely on historical data for prediction, and the use of real-time data is relatively limited. With the development of information technology, real-time traffic data on highways has become more abundant and available, and traditional methods are difficult to fully utilize these real-time data for prediction.

[0025] Model complexity: The prediction models in traditional methods are usually simple and often fail to fully consider the complexity and nonlinear characteristics of the traffic system. For example, traffic flow is affected by many factors, including time, weather, road conditions, accidents, etc. Traditional models often cannot capture the complex relationship between these factors well.

[0026] Prediction accuracy: Due to the above limitations, the prediction accuracy of traditional methods is often not high enough, especially in the face of emergencies or special circumstances, its prediction ability is greatly challenged. This makes it difficult for traffic managers to respond to traffic congestion and emergencies in a timely and effective manner.

[0027] Based on the above problems and challenges, a new highway traffic flow forecasting method needs to be proposed to make full use of real-time data and complex models to improve forecasting accuracy and practicality. Such a new method should be able to better meet the needs of modern traffic management and provide more reliable decision support for traffic managers.

[0028] like Figure 1 and 2 As shown, the present invention proposes a method for predicting highway traffic flow, the method comprising the steps of: Step S100, obtaining the location of a monitoring point in a target highway area, and extracting historical vehicle flow data of the corresponding monitoring point from a database based on the location of the monitoring point, wherein the monitoring point is a toll station or a gantry set up on a highway; In the specific implementation process, the target high-speed area is an area within a range set by the user, and the historical traffic flow data is the traffic flow data of the monitoring point in historical events. Specifically, the database adopts a 5-layer hierarchical structure, consisting of the original data layer (ods), standard detailed data layer (dwd), light summary data layer (dws), theme data layer (dwt), and application service data layer (ads). In the original data layer, there are original data such as the issuance business library, the clearing business library, and the road business library, which can be used to obtain the road section detector data, video surveillance data and other information required for traffic flow prediction, providing a solid foundation and strong support for more accurate prediction of road network traffic flow. In the subsequent layers, each business data is processed according to the standard, and dimensional modeling is performed according to the needs of the business segment. Table structures such as fact tables and dimension tables are established, and ETL is used to convert standard data into dimensional modeling data and store it in the data detail table.

[0029] Step S200, constructing a first vehicle flow prediction vector based on historical vehicle flow data of at least one monitoring point existing in a target section of a target highway area, and inputting the vector into a pre-trained prediction model to obtain a first predicted vehicle flow of the target section at a target time point; In the specific implementation process, the prediction model is a convolutional neural network model (CNN), a support vector machine (SVM), a K-nearest neighbor algorithm (KNN), a random forest (RF), GBDT or an XGBoost model; the support vector machine algorithm can obtain extremely high accuracy, but the effectiveness of the algorithm will be affected by the previous process data, and continuously adding continuously acquired information to the model will also affect the complexity of the data. The K-nearest neighbor algorithm model combined with spatiotemporal correlation can achieve a lower root mean square error (RMSE). The random forest algorithm (RF) achieves the purpose of predicting traffic flow by constructing multiple decision trees for ensemble learning. The model training and application based on this algorithm have a great impact in the field of traffic prediction. GBDT and XGBoost are gradient boosting tree algorithms and extended versions that can better handle large-scale data and improve prediction accuracy.

[0030] From the comparison results, it can be seen that the accuracy of the GBDT algorithm model is relatively low, while the accuracy of the XGBoost algorithm and the RF algorithm is relatively high. During the training process, compared with the other two algorithms, the RF algorithm model runs slowly and has a large model, which is not the best prediction algorithm model. The present invention uses the XGBoost algorithm to tune the model parameters and generate the final model.

[0031] During the specific implementation process, each target road section has at least one monitoring point.

[0032] In view of the actual needs of the current high-speed road network traffic flow planning, the traffic flow situation after 7 days and the traffic flow situation after 24 hours are of great significance for route planning and traffic control. Therefore, the present invention divides the traffic flow prediction model into the traffic flow prediction model for the next 7 days and the traffic flow prediction model for the next 24 hours. At the same time, in order to better fit the traffic flow situation of the road section, the present invention adopts the method of combining the gantry traffic flow prediction with the toll station traffic flow prediction to realize the road section traffic flow prediction.

[0033] The present invention introduces R2 accuracy as an evaluation index of the vehicle flow prediction model, which is a commonly used evaluation index of the vehicle flow prediction model and is convincing. At the same time, in order to obtain a model that meets the requirements of production applications, the present invention introduces two indicators, model size and model running time (the sum of model generation time and model prediction time), to evaluate the availability of the model.

[0034] This model uses the training set to iteratively train the model, adjust the hyperparameters and characteristic variables of the three algorithms, and obtain the toll station traffic flow situation in the next 7 days, the toll station traffic flow situation in the next 24 hours, the gantry traffic flow situation in the next 7 days, and the gantry traffic flow situation in the next 24 hours including the three algorithm models.

[0035] Step S300, extracting the external influencing factors of the target road section, calculating the corresponding vectors of the external influencing factors, the corresponding vectors of the external influencing factors include spatial dependence feature vectors and weather feature vectors, in the process of calculating the spatial dependence feature vectors, the spatial dependence feature vectors are calculated based on other target road sections that have a connectivity relationship with each target road section; in the process of calculating the weather feature vectors, the weather feature vectors are calculated based on historical weather data of the target road section; In the specific implementation process, the external influencing factors often have an objective impact on the traffic flow. This solution comprehensively considers various influencing factors to further ensure the prediction accuracy.

[0036] Step S400: input the first predicted traffic flow at the target time point and the corresponding vector of the calculated external influencing factors into the pre-trained correction model, and output the second predicted traffic flow after the correction of the first predicted traffic flow.

[0037] In the specific implementation process, the correction model is a convolutional neural network model.

[0038] The above scheme is adopted. In the process of predicting the traffic flow, the detection data of the monitoring points of the toll stations and the gantries installed on the highway are first integrated to make a preliminary prediction of the traffic flow of the target section. In the preliminary prediction process, all the monitoring points in the target highway area can be integrated, and the data is not limited to a single section. This scheme further considers other external influencing factors, including spatial dependence and weather characteristics, and corrects the initially predicted traffic flow of the target section through the spatial dependence feature vector and the weather feature vector to ensure the prediction accuracy.

[0039] In some embodiments of the present invention, the method further comprises: comparing the second predicted traffic volume with a preset congestion threshold to determine whether to issue a congestion warning to the corresponding target road section.

[0040] In a specific implementation process, if the second predicted traffic volume is greater than a preset congestion threshold, a congestion warning is issued to the corresponding target road section.

[0041] In some embodiments of the present invention, in the step of constructing a first traffic flow prediction vector based on historical traffic flow data of at least one monitoring point existing in a target section of a target highway area, the number of vehicles passing through the monitoring point existing in the target section of the target highway area during a preset first historical time period is used as the value of a dimension of the first traffic flow prediction vector to construct the first traffic flow prediction vector.

[0042] In some embodiments of the present invention, based on historical traffic flow data of at least one monitoring point in a target section of a target highway area, a first traffic flow prediction vector is constructed and input into a pre-trained prediction model, and the step of obtaining a first predicted traffic flow of the target section at a target time point includes: Sequentially number each monitoring point based on the connectivity relationship between the monitoring points; Arrange the first vehicle flow prediction vectors of the monitoring points in a longitudinal order based on the serial number sequence to obtain a vehicle flow prediction matrix; The vehicle flow prediction matrix is ​​input into a pre-trained prediction model, and the prediction model outputs a prediction result matrix, wherein each row in the prediction result matrix corresponds to a first predicted vehicle flow of a target road section.

[0043] Adopting the above scheme, in the process of calculating the first predicted traffic flow, this scheme first constructs a first traffic flow prediction vector for the traffic flow of the monitoring points of each target road section. However, in the process of calculating the first predicted traffic flow, a prediction result matrix is ​​further constructed. The first traffic flow prediction vector is arranged in the prediction result matrix according to the connectivity relationship between the monitoring points. Therefore, in the process of calculating the first predicted traffic flow, not only the monitoring points in the target road section are considered, but also the monitoring points in other locations are taken into consideration, thereby preliminarily ensuring the calculation accuracy of the first predicted traffic flow.

[0044] In some embodiments of the present invention, in the step of sequentially numbering the monitoring points based on the connectivity relationship between the monitoring points, the monitoring points are sequentially numbered from upstream to downstream based on the upstream and downstream relationship between the monitoring points.

[0045] In some embodiments of the present invention, in the step of calculating the spatial dependency characteristic vector based on other target road segments that are connected to each target road segment, the spatial dependency characteristic vector is calculated based on the spatial dependency weights set for the other target road segments that are connected to each target road segment, and the first predicted traffic flow of the other target road segments that are connected to each target road segment.

[0046] In a specific implementation process, the spatial dependency weight is a weight value preset for each target road section.

[0047] In some embodiments of the present invention, in the step of calculating the spatial dependency feature vector based on the spatial dependency weights set for other target road segments that are connected to each target road segment and the first predicted traffic flow of other target road segments that are connected to each target road segment, the following formula is used to calculate the spatial dependency feature vector: in, represents the spatially dependent eigenvector, , and Indicates the spatial dependency weights of other target road segments that are connected to the target road segment. , and Indicates the first predicted traffic flow corresponding to other target road sections that are connected to the target road section.

[0048] Adopting the above scheme, this scheme further considers the influence degree of other target sections that are connected with the target section on the target section through the spatial dependency weight, and combines the first predicted traffic flow of other target sections that are connected with the target section to obtain the spatial dependency feature vector of the corresponding target section.

[0049] In some embodiments of the present invention, the step of calculating the weather feature vector based on the historical weather data of the target road section includes: Obtaining weather forecast data for a target road section, and encoding various items of the weather forecast data, wherein the items of the weather forecast data include wind speed, temperature, humidity, and precipitation probability; The coded data of each item are combined to obtain the weather feature vector.

[0050] In some embodiments of the present invention, the corresponding vector of the external influencing factors includes an event influence vector. In the calculation step of the event influence vector, the event influence vector includes a first dimension and a second dimension. The first dimension corresponds to holidays, and is 1 if the target time point is a holiday, otherwise it is 0; the second dimension corresponds to large-scale activities, and is 1 if there are large-scale activities in the target highway area in a preset time period before and after the target time point, otherwise it is 0.

[0051] Adopting the above scheme, this scheme further considers the impact of weather and events on traffic flow among external influencing factors, combines the factors of force majeure and the impact of large-scale activities (such as concerts and sports events) to ensure the accuracy of correction.

[0052] In summary, in order to obtain the best traffic flow prediction effect, the present invention conducts research according to the following parts: In the training data preparation part, the present invention applies a data collection strategy suitable for a big data platform, collects real and reliable data of the road network, and pre-processes the data. This part of the data preparation provides a solid foundation for the authenticity and usability of subsequent model training.

[0053] In terms of model design, the present invention first classifies the data collection according to the characteristic factors, and then divides it into different characteristic variables according to different characteristic factors, and finally obtains the corresponding data from the database and external websites. Traffic transaction data, toll station data, gantry data, etc. can all be directly obtained from the big data platform business database, and the data is accurate and easy to obtain. The direct acquisition of high-weighted characteristic variables such as whether it is morning peak, whether it is evening peak, whether it is a free holiday, whether it is a non-free holiday, etc., lays the foundation for accurate model training. Other dimensional data mainly include major event dates, license plate restriction dates, road construction information, weather conditions, etc. These data can be obtained from the corresponding official website, which is authoritative and also ensures the accuracy of the training data.

[0054] In terms of prediction accuracy, based on a complete and reliable data foundation and a complex model design with multi-dimensional influencing factors and characteristic variables, the present invention trains the gantry vehicle flow prediction model and the toll station vehicle flow prediction model based on the RF algorithm, GBDT algorithm, and XGBoost algorithm, respectively, which greatly improves the prediction accuracy.

[0055] Traffic flow prediction technology provides a scientific basis for decision makers and ensures the accuracy of the decision-making process. Through this technology, traffic resources are optimally allocated, road traffic efficiency is significantly improved, and traffic congestion is effectively prevented, thereby greatly improving the overall efficiency of traffic management. The following are the main manifestations of this technology in practical applications: 1. Advance planning and adjustment: Short-term traffic forecasting models allow traffic management departments to predict traffic flows in the next 24 hours and 7 days, enabling them to plan traffic flows in advance, adjust traffic signals and road use strategies, and reduce congestion and delays.

[0056] 2. Optimal resource allocation: Based on the prediction results, traffic management departments can allocate traffic resources more reasonably, such as increasing police deployment during peak hours or adding temporary traffic signs at key nodes.

[0057] 3. Improve traffic efficiency: The prediction model helps toll stations know the peak traffic hours in advance, arrange staff and equipment reasonably, quickly process vehicle traffic, and reduce waiting time.

[0058] 4. Vehicle type-specific management measures: The vehicle type prediction model enables traffic management departments to understand the flow distribution of different vehicle types and formulate targeted management measures, such as setting up special lanes for large trucks or providing priority access to public transport vehicles.

[0059] 5. Improve service quality: By predicting the traffic needs of different types of vehicles, toll stations can prepare corresponding toll collection equipment and tickets in advance, improve toll collection efficiency, ensure service quality, and enhance user satisfaction.

[0060] 6. Data-driven decision support: The data support provided by the predictive model helps traffic management departments make evidence-based decisions, improve decision-making transparency and public trust.

[0061] 7. Transportation policy evaluation and formulation: Use long-term traffic forecast results to evaluate the effectiveness of existing transportation policies and provide a basis for the formulation of new policies.

[0062] 8. Public Information Services: Provide real-time and predicted traffic information to the public through various media to help drivers make better travel decisions.

[0063] 9. Emergency response and accident prevention: When a possible traffic accident or road congestion is predicted, the emergency plan is quickly activated to reduce the impact of the accident on traffic flow.

[0064] The beneficial effects of the present invention include: 1. The present invention uses three different algorithms to predict the traffic flow of the expressway network, realizing the prediction of the traffic flow of the gantry and the toll station. It also generates a 24-hour prediction model and a 7-day prediction model as needed, providing an effective reference for expressway travel planning.

[0065] 2. The feature vector used in the present invention combines the hot feature variables in the current research field, such as weather factor feature variables and holiday factor feature variables, and also integrates the industry data in the highway field as feature variables, such as toll station characteristics and traffic volume characteristics, providing research ideas for subsequent highway network traffic prediction.

[0066] An embodiment of the present invention also provides a highway traffic flow prediction system, which includes a computer device, wherein the computer device includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method described above.

[0067] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps implemented by the above-mentioned highway traffic flow prediction method are implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0068] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0069] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0070] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.

[0071] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting highway traffic flow, characterized in that: The steps of the method include: Obtaining the location of a monitoring point in a target highway area, and extracting historical traffic flow data of the corresponding monitoring point from a database based on the location of the monitoring point, wherein the monitoring point is a toll station or a gantry set up on a highway; Based on historical traffic flow data of at least one monitoring point in a target section of a target highway area, a first traffic flow prediction vector is constructed and input into a pre-trained prediction model to obtain a first predicted traffic flow of the target section at a target time point; Extracting the external influencing factors of the target road section, calculating the corresponding vectors of the external influencing factors, the corresponding vectors of the external influencing factors include spatial dependence feature vectors and weather feature vectors, in the process of calculating the spatial dependence feature vectors, the spatial dependence feature vectors are calculated based on other target road sections that have a connectivity relationship with each target road section; in the process of calculating the weather feature vectors, the weather feature vectors are calculated based on historical weather data of the target road section; The first predicted traffic flow at the target time point and the corresponding vector of the calculated external influencing factors are input into the pre-trained correction model, and the second predicted traffic flow after the correction of the first predicted traffic flow is output.

2. The method for predicting highway traffic flow according to claim 1, characterized in that: The method further includes comparing the second predicted traffic volume with a preset congestion threshold to determine whether to issue a congestion warning to the corresponding target road section.

3. The method for predicting highway traffic flow according to claim 1, characterized in that: In the step of constructing a first traffic flow prediction vector based on historical traffic flow data of at least one monitoring point existing in a target section in a target highway area, the number of vehicles passing through the monitoring point existing in the target section in a preset first historical time period is used as the value of a dimension of the first traffic flow prediction vector to construct the first traffic flow prediction vector.

4. The method for predicting highway traffic flow according to claim 1, characterized in that: Based on historical traffic flow data of at least one monitoring point in a target section of a target highway area, a first traffic flow prediction vector is constructed and input into a pre-trained prediction model to obtain a first predicted traffic flow of the target section at a target time point, the steps include: Sequentially number each monitoring point based on the connectivity relationship between the monitoring points; Arrange the first vehicle flow prediction vectors of the monitoring points in a longitudinal order based on the serial number sequence to obtain a vehicle flow prediction matrix; The vehicle flow prediction matrix is ​​input into a pre-trained prediction model, and the prediction model outputs a prediction result matrix, wherein each row in the prediction result matrix corresponds to a first predicted vehicle flow of a target road section.

5. The method for predicting highway traffic flow according to claim 4, characterized in that: In the step of sequentially numbering the monitoring points based on the connectivity relationship between the monitoring points, the monitoring points are sequentially numbered from upstream to downstream based on the upstream and downstream relationship between the monitoring points.

6. The method for predicting highway traffic flow according to claim 1, characterized in that: In the step of calculating the spatial dependency characteristic vector based on other target road segments that are connected to each target road segment, the spatial dependency characteristic vector is calculated based on the spatial dependency weights set for the other target road segments that are connected to each target road segment, and the first predicted traffic flow of the other target road segments that are connected to each target road segment.

7. The method for predicting highway traffic flow according to claim 6, characterized in that: In the step of calculating the spatial dependency feature vector based on the spatial dependency weights set for other target road segments that are connected to each target road segment and the first predicted traffic flow of other target road segments that are connected to each target road segment, the spatial dependency feature vector is calculated using the following formula: in, represents the spatially dependent eigenvector, , and Indicates the spatial dependency weights of other target road segments that are connected to the target road segment. , and Indicates the first predicted traffic flow corresponding to other target road sections that are connected to the target road section.

8. The method for predicting highway traffic flow according to claim 1, characterized in that: The step of calculating the weather feature vector based on the historical weather data of the target road section comprises: Obtaining weather forecast data for the target road section, and encoding various items of the weather forecast data, wherein the items of the weather forecast data include wind speed, temperature, humidity, and precipitation probability; The coded data of each item are combined to obtain the weather feature vector.

9. The method for predicting highway traffic flow according to claim 1, characterized in that: The corresponding vector of the external influencing factors includes an event influence vector. In the calculation step of the event influence vector, the event influence vector includes a first dimension and a second dimension. The first dimension corresponds to holidays. If the target time point is a holiday, it is 1, otherwise it is 0; the second dimension corresponds to large-scale activities. If there is a large-scale activity in the target highway area in a preset time period before and after the target time point, it is 1, otherwise it is 0.

10. A highway traffic flow prediction system, characterized in that: The system includes a computer device, which includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method as described in any one of claims 1 to 9.

Images