Intelligent traffic flow prediction system based on big data analysis

Abstract

The invention relates to the technical field of traffic control, and particularly discloses a smart traffic flow prediction system based on big data analysis, which comprises a data acquisition system and a flow prediction system, and is characterized in that the data acquisition system acquires the number of vehicles and driving direction data through multiple channels, integrates weather information and event information, ensures that the data is comprehensive and updated in real time, and transmits the data to the traffic flow prediction system; the collected data is stored in a cloud database for subsequent analysis; according to the method, the importance of different regions is calculated through the graph attention convolutional network to capture the spatial-temporal characteristics, the problem of over-extrusion of a graph neural network is solved, a deeper network model is constructed, a semantic adjacency matrix is introduced to learn the spatial correlation of local regions and semantic information in a global space, and the spatial-temporal characteristics are captured. Traffic flow prediction from a local area to a global scale is realized, a causal relationship between long and short time in traffic flow data is captured by using a time causal convolutional network, and local and global spatial-temporal correlation in the traffic data is synchronously captured.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic control, and specifically relates to an intelligent traffic flow prediction system based on big data analysis. Background Art

[0002] Intelligent Transportation System (ITS) has become an important part of traffic network control. Detectors distributed on both sides of the road collect data generated in the traffic network. After data analysis and prediction, some useful information for traffic control is obtained. People apply this information to suitable places and provide corresponding countermeasures for different service groups. A core system in ITS is the traffic flow prediction system, which analyzes the traffic flow state based on historical traffic flow information and makes real-time predictions based on the analysis results. However, due to the complex spatial topological structure and dynamic spatiotemporal dependence of the traffic network, although researchers have done a lot of research on learning traffic spatiotemporal dynamics and long-range spatial correlation, there are still two key problems. On the one hand, traffic flow not only has short-term and long-term temporal dependence in time, but also has local and global spatial dependence in space. Most traffic prediction research focuses on capturing the spatial correlation between neighboring nodes in the local space, while ignoring the spatial correlation between long-range spatial nodes. On the other hand, the spatial characteristics of urban traffic flow will change dynamically over time. Due to the complex dynamic spatiotemporal characteristics of traffic flow, the predefined adjacency matrix will limit the model from learning the dynamic characteristics of traffic flow.

[0003] Therefore, it is necessary to propose an intelligent traffic flow prediction system based on big data analysis to accurately predict traffic flow, alleviate urban road traffic pressure, and reduce the incidence of traffic congestion accidents.

[0004] The above information disclosed in this background technology is only used to increase the understanding of the background technology of the present invention and therefore, it may include information that does not constitute the prior art known to ordinary technicians in this field. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent traffic flow prediction system based on big data analysis to accurately predict traffic flow, alleviate urban road traffic pressure, and reduce the incidence of traffic congestion accidents.

[0006] To achieve the above object, the present invention provides the following technical solutions: The intelligent traffic flow prediction system based on big data analysis includes: a data collection system and a traffic flow prediction system. The data collection system obtains vehicle quantity and driving direction data through multiple channels, integrates weather information and event information, ensures that the data is comprehensive and updated in real time, and the collected data will be stored in a cloud database for subsequent analysis; The data acquisition system includes a lane acquisition unit, a vehicle type acquisition unit, a detection range adjustment unit, and a road congestion detection warning unit. The lane acquisition unit is responsible for collecting lane-by-lane traffic flow data on the road where the surveillance video is located. The lane acquisition unit uses three-frame difference frame area flow acquisition to determine whether there is a vehicle passing through the target area by judging the overlap between the vehicle and the frame area, thereby realizing traffic flow statistics. The traffic prediction system uses deep learning and big data analysis technology to analyze data to identify traffic patterns and trends, predict traffic flow within a specific time period, and display the prediction results through visualization tools to help traffic management departments make timely decisions.

[0007] Preferably, the vehicle type acquisition unit is responsible for collecting traffic flow data by vehicle type on the road where the surveillance video is located. The vehicle type acquisition unit uses a vehicle type recognition method based on deep learning to identify the type of vehicle passing through the current road, and combines the three-frame difference frame area flow collection to realize traffic flow collection by vehicle type, and completes the vehicle flow statistics of the vehicle type.

[0008] Preferably, the detection range adjustment unit parses the received user request, determines the lane that needs to be detected and the coordinate ratio of the lane rectangular area based on the road network information, and the road congestion detection warning unit uses a congestion detection algorithm combined with traffic flow statistics to determine the traffic density, occupancy rate and road section driving speed of a certain lane, and promptly issues a warning to the management personnel when it exceeds the set standard.

[0009] Preferably, the traffic flow prediction system includes a traffic flow information management unit, a weather and meteorological information management unit, and a traffic flow prediction unit. The traffic flow information management unit is based on a timeline and integrates lane and vehicle type information to construct a traffic flow information statistics table, establishes a multi-condition combination of traffic flow information queries based on a search engine, and supports a multi-selection statistics function for vehicle types.

[0010] Preferably, the traffic flow information management unit is responsible for storing and managing real-time and historical traffic flow data, completing missing data and correcting erroneous data, and integrating multi-source data to ensure data security and accessibility. The traffic flow information management unit adopts a weighted method of historical and real-time data to complete missing data, and for erroneous data, first eliminates the erroneous data, and then completes the eliminated data according to the method of missing data completion.

[0011] Preferably, the weather and meteorological information management unit is responsible for storing and managing real-time and historical weather and meteorological data, establishing a correspondence with actual road weather conditions, and ensuring the authenticity and reliability of the data. The weather and meteorological information management unit obtains weather and meteorological forecast information for the current area from the Internet through a standard API interface and stores it in a MySQL database.

[0012] Preferably, the traffic flow prediction unit calculates the importance of different areas through a graph attention convolutional network to capture spatiotemporal features, thereby constructing a deeper network model, introducing a semantic adjacency matrix to learn the spatial correlation of local areas and semantic information in the global space, realizing traffic flow prediction from local areas to the global scale, and using a temporal causal convolutional network to capture the causal relationship between long-term and short-term time in traffic flow data.

[0013] The network model first explores hidden spatiotemporal features through a dual-path architecture. Each path shares the same structure to capture spatiotemporal dependencies. One path uses an adaptive adjacency matrix as input to model the spatial correlation between neighboring regions in a fine-grained manner. The other path uses regional functional similarity to construct a global semantic adjacency matrix to model the spatial dependencies of different regions from a global perspective. These two adjacency matrices are then input into the spatiotemporal attention block, which consists of a graph attention convolution layer sandwiched between two temporal causal convolution layers. The temporal causal convolution layer captures the causal relationship between local and global long-term and short-term time. Finally, it is input into the graph attention convolution layer to simultaneously capture long-range spatiotemporal dependencies. In order to ensure the effective training of deep network models, the model adopts a four-layer spatiotemporal attention block stacking structure, and uses residual connections and network layer normalization inside each spatiotemporal attention block. At the same time, in order to obtain comprehensive spatiotemporal features, the network layers in the spatiotemporal attention block use the same dimension.

[0014] Preferably, the communication module performs data request and transmission through the HTTP protocol, and adopts a layered strategy and a caching mechanism to reduce unnecessary information transmission and improve response speed. The service level evaluation module is used to measure and monitor the quality of system services, collect user feedback and evaluation results to optimize service processes and meet user needs.

[0015] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing any one of the above-mentioned intelligent traffic flow prediction systems based on big data analysis.

[0016] A terminal device includes a processor and a computer-readable storage medium, the processor is used to implement various instructions; the computer-readable storage medium is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executed by any one of the above-mentioned intelligent traffic flow prediction systems based on big data analysis.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention calculates the importance of different areas through a graph attention convolutional network to capture spatiotemporal features, solves the over-squeezing problem of graph neural networks, and thus constructs a deeper network model. It introduces a semantic adjacency matrix to learn the spatial correlation of local areas and the semantic information in the global space, and realizes traffic flow prediction from local areas to the global scale. It also uses a temporal causal convolutional network to capture the causal relationship between long-term and short-term time in traffic flow data, and simultaneously captures the local and global spatiotemporal correlation in traffic data. The present invention can accurately predict traffic conditions in real time, provide a scientific basis for traffic management departments to optimize traffic resource scheduling, and effectively alleviate urban traffic congestion.

[0018] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a framework diagram of the intelligent traffic flow prediction system based on big data analysis of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be further understood that the term “and / or” used in the specification of this application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0022] Embodiment 1:

[0023] See also Figure 1As shown, the intelligent traffic flow prediction system based on big data analysis includes: the data acquisition system obtains vehicle quantity and driving direction data through multiple channels, integrates weather information and event information, ensures that the data is comprehensive and updated in real time, and the collected data will be stored in the cloud database for subsequent analysis.

[0024] The data collection system is responsible for collecting data from channels including mobile microwave traffic detection, intersection induction coil traffic detection and video traffic detection, and then converts the traffic information into a unified data format. The information obtained is the number of various types of vehicles (small trucks, medium trucks, large trucks, container trucks, small and medium-sized buses and large buses) passing through the lanes involved in units of hours, as well as the weather and meteorological information covering the lanes involved during this period.

[0025] Video traffic flow detection example: Video-based traffic flow feature data collection can be realized for 8 real-time images; "virtual" vehicle monitoring areas are set on the display in combination with road monitoring point images. Each virtual vehicle monitoring area can simulate an induction coil. When a vehicle passes through any virtual vehicle monitoring area, a detection signal is generated, and the processor analyzes the input video image to generate the required traffic data.

[0026] The data collection system includes a lane collection unit, a vehicle type collection unit, a detection range adjustment unit, and a road congestion detection warning unit. The lane collection unit uses three-frame difference frame area traffic collection to determine the overlap between the vehicle and the frame area to determine whether there is a vehicle passing through the target area, thereby realizing traffic flow statistics.

[0027] The vehicle type collection unit uses a vehicle type recognition method based on deep learning to identify the type of vehicle passing through the current road, and combines the three-frame difference frame area flow collection to realize traffic flow collection by vehicle type and complete the vehicle flow statistics for the vehicle type.

[0028] The detection range adjustment unit parses the received user requests and determines the lanes that need to be detected and the coordinate ratios of the lane rectangular area based on the road network information. The road congestion detection warning unit uses a congestion detection algorithm combined with traffic flow statistics to determine the traffic density, occupancy rate and road speed of a certain lane. When it exceeds the set standards, it will promptly issue a warning to the management personnel.

[0029] Embodiment 2:

[0030] The intelligent traffic flow prediction system based on big data analysis includes: the traffic prediction system uses deep learning and big data analysis technology to analyze data to identify traffic patterns and trends, predict traffic flow within a specific time period, and display the prediction results through visualization tools to help traffic management departments make timely decisions.

[0031] The traffic flow prediction system includes a traffic flow information management unit, a weather and meteorological information management unit, and a traffic flow prediction unit. The traffic flow information management unit is based on the timeline and integrates lane and vehicle type information to construct a traffic flow information statistics table. It establishes a multi-condition combination of traffic flow information query based on the search engine and supports multiple selection statistics of vehicle types.

[0032] The traffic flow information management unit adopts the weighted method of historical and real-time data to fill in the missing data. This method is based on the trend method. It uses the trend method to fill in the data and the data of the previous moment to weightedly calculate the current missing data. For the wrong data, the wrong data is first eliminated, and then the eliminated data is supplemented according to the method of missing data filling.

[0033] The weather and meteorological information management unit obtains the weather and meteorological forecast information of the current area from the Internet through the standard API interface and stores it in the MySQL database. The information obtained includes temperature, humidity, PM2.5AQI value, weather conditions (such as sunny, rainy, foggy, snowy, etc.), wind direction and wind force, etc. The historical weather and meteorological information of the current area is obtained in the same way and stored in the same file location as the weather and meteorological forecast information. The stored weather and meteorological information is used to perform statistics according to multiple conditions such as start time, end time, statistical type, display method and area name. The statistical types include weather type, wind force, AQI level, and support statistics in the form of bar charts and pie charts. The weather types include sunny, cloudy, overcast, showers, thunderstorms, thunderstorms, hail, sleet, light rain, moderate rain, heavy rain, heavy rain, heavy rain, light snow, moderate snow, heavy snow, blizzard fog, etc.

[0034] The traffic flow prediction unit captures spatiotemporal features by calculating the importance of different areas through a graph attention convolutional network, thereby building a deeper network model, introducing a semantic adjacency matrix to learn the spatial correlation of local areas and the semantic information in the global space, and realizing traffic flow prediction from local areas to the global scale. It also uses a temporal causal convolutional network to capture the causal relationship between long-term and short-term time in traffic flow data.

[0035] The graph attention convolutional layer is a graph attention convolutional neural network with a self-attention layer added. The parameters of the graph attention network are shared in the spatiotemporal attention block to alleviate the over-squeezing of the graph neural network, enabling the model to make accurate long-range spatial predictions. The graph attention network adds a hidden self-attention network layer on the basis of the graph convolutional network. In the convolution process, the spatial neighborhood information of different sizes is processed by superimposing the self-attention layer, and the importance weights are assigned to different nodes in the neighborhood.

[0036] The input of the graph attention network is a set of node features. ,in For different node sequences, features Expressed as , is the number of nodes, At each stage of time The number of features of each node in . First, Transpose and reshape into , and then generate a new set of node features , and finally as output.

[0037] Here, a learnable linear transformation function is used to transform the input features into higher-dimensional feature representations. , a shared linear transformation parameterized by the weight matrix Applied to each node; then a shared attention mechanism is used , in any pair and The following self-attention coefficient calculation is performed on the node:

[0038] In the formula, and are the weight matrix components, is the attention mechanism operation function.

[0039] In order to make the attention coefficients between different nodes easy to distinguish, the softmx function is used to normalize the attention coefficient score between node neighborhoods, and the following output is obtained:

[0040] In the formula, Nodes in the graph Based on this attention weight, the neighbor set is integrated with all neighbor node information to obtain the updated new node features:

[0041] In the formula, is the activation function, It is The training parameters of each node.

[0042] The spatiotemporal causal convolution layer consists of a causal convolution layer and a dilation factor that can multiply the receptive field. The dilated causal convolution formula of the spatiotemporal causal convolution layer is as follows: In the formula, is the expansion factor, is the filter size, is a time series, is the convolution operator, is the expansion factor Dilated causal operators that control jump distances, is the convolution kernel.

[0043] Model Input It is a three-dimensional traffic map data signal, and the extended causal convolution is used to generate the output features of the node. The formula is as follows:

[0044] In the formula, , is the number of output features, Is a node In time No. Output characteristics, Is a node In time No. Input features, kernel It is trainable. is the number of output features.

[0045] From the above, it can be seen that the present invention captures spatiotemporal features by calculating the importance of different areas through a graph attention convolutional network, solves the over-squeezing problem of graph neural networks, and thus constructs a deeper network model. It introduces a semantic adjacency matrix to learn the spatial correlation of local areas and the semantic information in the global space, and realizes traffic flow prediction from local areas to the global scale. It also uses a temporal causal convolutional network to capture the causal relationship between long-term and short-term time in traffic flow data, and simultaneously captures the local and global spatiotemporal correlation in traffic data. The present invention can accurately predict traffic conditions in real time, provide a scientific basis for traffic management departments to optimize traffic resource scheduling, and effectively alleviate urban traffic congestion.

[0046] Embodiment 3: The intelligent traffic flow prediction system based on big data analysis also includes: a communication module and a service level evaluation module. The communication module requests and transmits data through the HTTP protocol, and adopts a layered strategy and a cache mechanism to reduce unnecessary information transmission and improve response speed. The service level evaluation module is used to measure and monitor the quality of system services, collect user feedback and evaluation results to optimize service processes and meet user needs.

[0047] An embodiment of the present invention also provides a terminal device, including a processor and a computer-readable storage medium, the processor being used to implement various instructions; the computer-readable storage medium being used to store multiple instructions, the instructions being suitable for being loaded by the processor and executing any one of the above-mentioned intelligent traffic flow prediction systems based on big data analysis.

[0048] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each process of the above-mentioned intelligent traffic flow prediction system embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. Among them, the computer-readable storage medium, such as read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random ACGess Memory, referred to as RAM), a disk or an optical disk, etc.

[0049] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples" or "some 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 representations of the above terms do 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. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0050] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a program, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A way to specify functions in one or more boxes.

[0051] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0052] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0053] In the drawings of the embodiments disclosed in the present invention, only the structures involved in the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other.

[0054] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0055] 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 appended claims and their equivalents.

Claims

1. Intelligent traffic flow prediction system based on big data analysis, including: Data collection system and traffic prediction system, characterized in that the data collection system obtains vehicle quantity and driving direction data through multiple channels, integrates weather information and event information, ensures that the data is comprehensive and updated in real time, and the collected data will be stored in a cloud database for subsequent analysis; The data collection system includes a lane collection unit, a vehicle type collection unit, a detection range adjustment unit, and a road congestion detection warning unit. The lane collection unit uses three-frame difference frame area flow collection to determine whether there is a vehicle passing through the target area by judging the overlap between the vehicle and the frame area, thereby realizing vehicle flow statistics. The traffic prediction system uses deep learning and big data analysis technology to analyze data to identify traffic patterns and trends, predict traffic flow within a specific time period, and display the prediction results through visualization tools to help traffic management departments make timely decisions.

2. The intelligent traffic flow prediction system based on big data analysis according to claim 1 is characterized by: The vehicle type collection unit uses a vehicle type recognition method based on deep learning to identify the type of vehicle passing through the current road, and combines the three-frame difference frame area flow collection to realize vehicle type traffic flow collection and complete the vehicle flow statistics of the vehicle type.

3. The intelligent traffic flow prediction system based on big data analysis according to claim 2 is characterized by: The detection range adjustment unit analyzes the received user request and determines the lane that needs to be detected and the coordinate ratio of the lane rectangular area based on the road network information. The road congestion detection warning unit uses a congestion detection algorithm combined with traffic flow statistics to determine the traffic density, occupancy rate and road section driving speed of a certain lane, and promptly issues a warning to the management personnel when it exceeds the set standard.

4. The intelligent traffic flow prediction system based on big data analysis according to claim 3 is characterized by: The traffic flow prediction system includes a traffic flow information management unit, a weather and meteorological information management unit, and a traffic flow prediction unit. The traffic flow information management unit is based on a timeline and integrates lane and vehicle type information to construct a traffic flow information statistics table, establishes a traffic flow information query with multiple condition combinations based on a search engine, and supports a multi-selection statistics function for vehicle types.

5. The intelligent traffic flow prediction system based on big data analysis according to claim 4 is characterized by: The traffic flow information management unit adopts the weighted method of historical and real-time data to fill in the missing data. For the erroneous data, the erroneous data is first eliminated, and then the eliminated data is filled in according to the method of missing data filling. The weather and meteorological information management unit obtains the weather and meteorological forecast information of the current area from the Internet through the standard API interface and stores it in the MySQL database.

6. The intelligent traffic flow prediction system based on big data analysis according to claim 5 is characterized by: The traffic flow prediction unit captures spatiotemporal features by calculating the importance of different areas through a graph attention convolutional network, thereby building a deeper network model, introducing a semantic adjacency matrix to learn the spatial correlation of local areas and the semantic information in the global space, and realizing traffic flow prediction from local areas to the global scale, and using a temporal causal convolutional network to capture the causal relationship between long-term and short-term time in traffic flow data.

7. The intelligent traffic flow prediction system based on big data analysis according to claim 6 further comprises: Communication module and service level evaluation module, characterized in that the communication module performs data request and transmission through HTTP protocol, and adopts layered strategy and cache mechanism to reduce unnecessary information transmission and improve response speed; the service level evaluation module is used to measure and monitor the quality of system services, collect user feedback and evaluation results to optimize service processes and meet user needs.

8. A computer-readable storage medium, characterized in that: There are multiple instructions, which are suitable for being loaded by a processor of a terminal device and executed by the intelligent traffic flow prediction system based on big data analysis as described in any one of claims 1-7.

9. A terminal device, characterized in that: It includes a processor and a computer-readable storage medium, the processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to implement the intelligent traffic flow prediction system based on big data analysis described in any one of claims 1-7.

Images