A macro traffic flow generation method based on roadside equipment data statistics

Abstract

The present invention relates to the field of intelligent transportation technology, and specifically to a macro traffic flow generation method based on road-side equipment data statistics, comprising the following steps: S1, within a certain map area, using road-side equipment to obtain vehicle traffic information in a local area; S2, classifying the vehicle traffic information obtained in S1 according to vehicle types and time periods, and building a traffic flow database; S3, constructing a directed graph according to the road structure connection conditions of the selected area; the macro traffic flow generation method of the present invention generates complete traffic flow information covering the entire map range by calculation, and uses a data-driven approach to reconstruct the traffic environment of a certain area for an autonomous driving simulation vehicle, especially reproducing the characteristics of the macro traffic flow.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a macro traffic flow generation method based on road-end equipment data statistics. Background Art

[0002] Traffic flow simulation focuses on studying the operation of the traffic network from a macro perspective. By simulating the flow of vehicles in the entire traffic network, analyzing traffic congestion, signal timing or traffic distribution strategies, it can also reproduce the traffic environment in autonomous driving simulation technology and provide more realistic training and test data for autonomous driving algorithms. Under current technical conditions, the environmental vehicle data collected by road-side equipment or on-board sensors is limited to the driving trajectories of vehicles in the area within the sensor's field of view, and cannot effectively reflect the intersection traffic choices of all vehicle driving paths in the entire map. There is a problem of a narrow range of vehicle driving trajectories. Summary of the invention

[0003] The purpose of the present invention is to provide a macro traffic flow generation method based on road-side equipment data statistics, which uses calculation to generate complete traffic flow information covering the entire map range, thereby solving the problem of narrow vehicle driving trajectory range when using environmental vehicle data collected by road-side equipment or on-board sensors.

[0004] To achieve the above object, the present invention provides the following technical solution: a macro traffic flow generation method based on roadside equipment data statistics, specifically comprising the following steps:

[0005] S1, in a certain map area, use the roadside equipment to obtain the vehicle traffic information of the local area;

[0006] S2, classify the vehicle traffic information obtained in S1 according to vehicle type and time period, and build a traffic flow database;

[0007] S3, construct a directed graph according to the road structure connection of the selected area, regard the intersections as nodes, and the roads as edges connecting the nodes, abstract the road logic, and design a directed graph containing m nodes and n edges according to the selected map structure with m intersections and n roads;

[0008] The driving of vehicles on the road is abstracted as the orderly transfer of edges and nodes on a directed graph, including U-turn operations. A node connected to only one edge is defined as an edge node, and the complete driving trajectory of all vehicles needs to enter and exit from the edge node. If there is a parking lot with a large traffic volume, a node needs to be established for it and exist as an edge node.

[0009] S4, based on the collected vehicle traffic information, count the traffic choices of vehicles at the node and calculate the traffic matrix; specifically: at an intersection connected by n roads, an n*n matrix can be used to describe the direction selection ratio of all vehicles at this intersection, which is called the traffic matrix T. The element t of the traffic matrix T is ij Represents the frequency of driving from road i to road j at the current intersection; based on the collected vehicle position information and map information at each moment, it is mapped to the edges and nodes of the directed graph;

[0010] Count the traffic choices of vehicles at the intersection and calculate the traffic matrix. When there is a vehicle driving from road i connecting the intersection to road j, the element t of the matrix T ij Add 1;

[0011] According to the database classification, traffic matrices of intersections belonging to different time periods are established respectively;

[0012] S5, calculating the vehicle flow rate flowing into the road network at each edge node according to the collected vehicle traffic information;

[0013] S6, using a large batch simulation method, statistically obtains the transfer sequence and flow of all vehicles in the entire directed graph;

[0014] S7, in the traffic flow simulation software, reconstruct the map and traffic flow.

[0015] Preferably, the vehicle traffic information in S1 includes vehicle flow information, relative speed of vehicles, relative position of vehicles, data timestamp, vehicle size and vehicle model information.

[0016] Preferably, the vehicle types described in S2 include cars, trucks, bicycles, modified risk vehicles, motorcycles and buses, and the time periods include morning and evening rush hours on weekdays, idle time on weekdays and holidays. The data is classified and stored according to the time period information, a traffic flow database is constructed, and the data is processed separately, and finally a macro traffic flow that conforms to the characteristics of the corresponding time period is constructed.

[0017] Preferably, the method for mass simulation in S6 specifically includes:

[0018] Assume that a vehicle enters from an edge node and transfers from node to edge to node in the directed graph in sequence. Its transit selection at all nodes is performed according to the transit matrix until the vehicle exits from the edge node. The transfer sequence of the vehicle on the directed graph is obtained and counted.

[0019] Set a maximum number of transfers. Vehicles that exceed this number will no longer be transferred and their records will be deleted.

[0020] The proportion of vehicles entering each edge node is determined according to the proportion of traffic flow from the edge node to the road network;

[0021] When the simulation data is large enough, the transfer sequence and proportion of vehicles in the directed graph will stabilize, and the vehicle transfer sequence data corresponding to the directed graph and the time period classification will be obtained.

[0022] Preferably, S7 specifically includes:

[0023] By collecting high-precision map data or other methods, the map is rebuilt in the traffic flow simulation software, and the road and intersection names of the original map are matched to facilitate the mapping of vehicle driving trajectories;

[0024] According to the vehicle classification and time period classification of the collected data and the calculated vehicle transfer sequence data corresponding to different time periods, they are mapped into the vehicle's driving trajectory in the traffic flow simulation software map, and the vehicle composition, traffic flow and trajectory composition files are defined to reconstruct the traffic flow in the traffic flow simulation software.

[0025] Preferably, the road-side device for obtaining vehicle traffic information in S1 is connected to the Internet through a smart terminal. After the road-side device identifies and obtains the vehicle model information, the vehicle model information is sent to the smart terminal. The vehicle model is used as a clue to search for the vehicle model on the Internet through the smart terminal, and the vehicle size data of the vehicle model is downloaded from the Internet to the smart terminal.

[0026] At the same time, the road-side equipment sends the vehicle size data it recognizes to the smart terminal, which compares the vehicle size data downloaded from the Internet with the vehicle size data it recognizes, and obtains the error. .

[0027] Preferably, according to the sensor accuracy of the roadside equipment, a size tolerance range for identifying the vehicle is set. When it is less than the allowable error range, the data is confirmed.

[0028] Preferably, when When it is greater than the allowable error range, the road-side equipment is controlled to perform optical zoom, magnify the vehicle for secondary identification, obtain the vehicle model and vehicle size data again, and use the newly obtained vehicle model as a clue to search and download the vehicle size data of this model again on the Internet through the smart terminal.

[0029] Preferably, the vehicle size data obtained by the secondary identification is compared again with the vehicle size data downloaded from the Internet. When it is smaller than the allowable error range, the data is confirmed; when it is larger than the allowable error range, the vehicle is defined as a modified risk vehicle.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The macro traffic flow generation method of the present invention generates complete traffic flow information covering the entire map range by calculation, and uses a data-driven approach to reconstruct the traffic environment of a certain area for an autonomous driving simulation vehicle, especially reproducing the characteristics of macro traffic flow, and solving the problem of a narrow range of vehicle driving trajectories when using environmental vehicle data collected by road-side equipment or on-board sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The figure is a flow chart of a method for generating macro traffic flow based on road-end equipment data statistics according to the present invention.

[0033] Figure 2 It is a schematic diagram of a directed graph of the present invention. DETAILED DESCRIPTION

[0034] 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.

[0035] See also Figure 1 and Figure 2 The present invention provides a technical solution: a macro traffic flow generation method based on road-end equipment data statistics, which specifically includes the following steps:

[0036] S1, within a certain map area, use road-side equipment to obtain vehicle traffic information in the local area; vehicle traffic information includes traffic flow information, relative speed of vehicles, relative position of vehicles, data timestamp, vehicle size and vehicle model information. The road-side equipment is connected to the Internet through an intelligent terminal, which includes computers and other networked intelligent devices. After the road-side equipment identifies and obtains vehicle model information, it sends the vehicle model information to the intelligent terminal, and uses the vehicle model as a clue to search for the vehicle of this model on the Internet through the intelligent terminal, and downloads the vehicle size data of this model from the Internet to the intelligent terminal. The vehicle model includes the vehicle model and brand information, and the corresponding model under the brand is determined by comparing the vehicle appearance with the vehicle database;

[0037] At the same time, the road-side equipment sends the vehicle size data it recognizes to the smart terminal, which compares the vehicle size data downloaded from the Internet with the vehicle size data it recognizes, and obtains the error. .

[0038] According to the sensor accuracy of the roadside equipment, a permissible error range for identifying the vehicle size is set. When it is smaller than the allowable error range, the data is confirmed. The above data comparison process can be completed in only tens to hundreds of milliseconds in chip processing. The longest time comes from the Internet response. When the network is good, it can respond within one hundred milliseconds. Therefore, fast comparison can be achieved, and subsequent operations can be carried out before the vehicle leaves the detection range of the road-end equipment.

[0039] when When it is greater than the allowable error range, the road-side equipment is controlled to perform optical zoom, magnify the vehicle for secondary identification, obtain the vehicle model and vehicle size data again, and use the newly obtained vehicle model as a clue to search and download the vehicle size data of this model again on the Internet through the smart terminal.

[0040] Each set of road-side equipment uses at least two sets of cameras with visual recognition functions, and at least one of the cameras has servo steering and optical zoom functions. It can track and identify the vehicle through servo steering during secondary identification, and at the same time, magnify the vehicle image through optical zoom to improve the recognition clarity. The reason for setting two sets of cameras for each set of road-side equipment is to enable one set of cameras to track the identified vehicle for the second time while the other set of cameras can maintain the original traffic flow information, relative speed of the vehicle, relative position of the vehicle, data timestamp, vehicle size and vehicle model information recognition work. When the traffic volume is large, the number of cameras with servo steering and optical zoom functions can be increased.

[0041] The vehicle size data obtained by the secondary identification is compared again with the vehicle size data downloaded from the Internet. If it is less than the allowable error range, the data is confirmed; if it is greater than the allowable error range, the vehicle is defined as a modified risk vehicle. Through the above automatic judgment, not only can the accuracy of obtaining vehicle model and vehicle size data be improved, but also modified risk vehicles can be classified, thereby adding a column for modified risk vehicles in the vehicle model information.

[0042] S2, classifies the vehicle traffic information obtained in S1 according to vehicle type and time period, and builds a traffic flow database; vehicle types include cars, trucks, bicycles, modified risk vehicles, motorcycles and buses, and time periods include morning and evening rush hours on weekdays, idle time on weekdays and weekends. The data is classified and stored according to the time period information to build a traffic flow database, and is processed separately, and finally a macro traffic flow that conforms to the characteristics of the corresponding time period is constructed.

[0043] S3, constructs a directed graph based on the road structure connectivity of the selected area, such as Figure 2 As shown in , the intersections are regarded as nodes, and the roads are regarded as the edges connecting the nodes. The road logic is abstracted, and according to the selected map structure with m intersections and n roads, a directed graph with m nodes and n edges is designed.

[0044] The driving of vehicles on the road is abstracted as the orderly transfer of edges and nodes on a directed graph, including U-turn operations. A node connected to only one edge is defined as an edge node, and the complete driving trajectory of all vehicles needs to enter and exit from the edge node. If there is a parking lot with a large traffic volume, a node needs to be established for it and exist as an edge node.

[0045] S4, based on the collected vehicle traffic information, count the traffic choices of vehicles at the node and calculate the traffic matrix; specifically: at an intersection connected by n roads, an n*n matrix can be used to describe the direction selection ratio of all vehicles at this intersection, which is called the traffic matrix T. The element t of the traffic matrix T is ij Represents the frequency of driving from road i to road j at the current intersection; based on the collected vehicle position information and map information at each moment, it is mapped to the edges and nodes of the directed graph;

[0046] Count the traffic choices of vehicles at the intersection and calculate the traffic matrix. When there is a vehicle driving from road i connecting the intersection to road j, the element t of the matrix T ij Add 1;

[0047] According to the database classification, the traffic matrix of the intersection belonging to different time periods is established;

[0048] S5, calculating the vehicle flow rate flowing into the road network at each edge node according to the collected vehicle traffic information;

[0049] S6, using a large batch simulation method, statistically obtain the transfer sequence and flow of all vehicles in the entire directed graph; the large batch simulation method specifically includes:

[0050] Assume that a vehicle enters from an edge node and transfers from node to edge to node in the directed graph. Its transit selection at all nodes is performed according to the transit matrix until the vehicle exits from the edge node. The transfer sequence of the vehicle on the directed graph is obtained and counted.

[0051] Set a maximum number of transfers. Vehicles that exceed this number will no longer be transferred and their records will be deleted.

[0052] The proportion of vehicles entering each edge node is determined according to the proportion of traffic flow from the edge node to the road network;

[0053] When the simulation data is large enough, the transfer sequence and proportion of vehicles in the directed graph will stabilize, and the vehicle transfer sequence data corresponding to the directed graph and the time period classification will be obtained.

[0054] S7, in the traffic flow simulation software, reconstructing the map and the traffic flow, the traffic flow simulation software may adopt SUMO traffic flow simulation software; by collecting high-precision map data or other methods, reconstructing the map in the traffic flow simulation software, and matching the road and intersection names of the original map, so as to facilitate the mapping of vehicle driving trajectories;

[0055] According to the vehicle classification and time period classification of the collected data and the calculated vehicle transfer sequence data corresponding to different time periods, they are mapped into the vehicle's driving trajectory in the traffic flow simulation software map, and the vehicle composition, traffic flow and trajectory composition files are defined to reconstruct the traffic flow in the traffic flow simulation software.

[0056] 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. A macro traffic flow generation method based on roadside equipment data statistics, characterized by: The specific steps include: S1, in a certain map area, use the roadside equipment to obtain the vehicle traffic information of the local area; S2, classify the vehicle traffic information obtained in S1 according to vehicle type and time period, and build a traffic flow database; S3, construct a directed graph according to the road structure connection of the selected area, regard the intersections as nodes, and the roads as edges connecting the nodes, abstract the road logic, and design a directed graph containing m nodes and n edges according to the selected map structure with m intersections and n roads; The driving of vehicles on the road is abstracted as the orderly transfer of edges and nodes on a directed graph, including U-turn operations. A node connected to only one edge is defined as an edge node, and the complete driving trajectory of all vehicles needs to enter and exit from the edge node. If there is a parking lot, a node needs to be established for it and exist as an edge node. S4, based on the collected vehicle traffic information, count the traffic choices of vehicles at the node and calculate the traffic matrix; specifically: at an intersection connected by n roads, an n*n matrix can be used to describe the direction selection ratio of all vehicles at this intersection, which is called the traffic matrix T. The element t of the traffic matrix T is ij Represents the frequency of driving from road i to road j at the current intersection; based on the collected vehicle position information and map information at each moment, it is mapped to the edges and nodes of the directed graph; Count the traffic choices of vehicles at the intersection and calculate the traffic matrix. When there is a vehicle driving from road i connecting the intersection to road j, the element t of the matrix T ij Add 1; According to the database classification, traffic matrices of intersections belonging to different time periods are established respectively; S5, calculating the vehicle flow rate flowing into the road network at each edge node according to the collected vehicle traffic information; S6, using a large batch simulation method, statistically obtains the transfer sequence and flow of all vehicles in the entire directed graph; S7, in the traffic flow simulation software, reconstruct the map and traffic flow.

2. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 1, characterized in that: The vehicle traffic information in S1 includes vehicle flow information, relative speed of vehicles, relative position of vehicles, data timestamp, vehicle size and vehicle model information.

3. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 1, characterized in that: The vehicle types described in S2 include cars, trucks, bicycles, modified risk vehicles, motorcycles and buses. The time periods include morning and evening rush hours on weekdays, idle time on weekdays and weekends. The data is classified and stored according to the information of the time period, and a traffic flow database is constructed. The data is processed separately, and finally a macro traffic flow that conforms to the characteristics of the corresponding time period is constructed.

4. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 1, characterized in that: The methods of large-scale simulation in S6 specifically include: Assume that a vehicle enters from an edge node and transfers from node to edge to node in the directed graph in sequence. Its transit selection at all nodes is performed according to the transit matrix until the vehicle exits from the edge node. The transfer sequence of the vehicle on the directed graph is obtained and counted. Set a maximum number of transfers. Vehicles that exceed this number will no longer be transferred and their records will be deleted. The proportion of vehicles entering each edge node is determined according to the proportion of traffic flow from the edge node to the road network; When the simulation data is large enough, the transfer sequence and proportion of vehicles in the directed graph will stabilize, and the vehicle transfer sequence data corresponding to the directed graph and the time period classification will be obtained.

5. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 4, characterized in that: S7 specifically includes: By collecting high-precision map data or other methods, the map is rebuilt in the traffic flow simulation software, and the road and intersection names of the original map are matched to facilitate the mapping of vehicle driving trajectories; According to the vehicle classification and time period classification of the collected data and the calculated vehicle transfer sequence data corresponding to different time periods, they are mapped into the vehicle's driving trajectory in the traffic flow simulation software map, and the vehicle composition, traffic flow and trajectory composition files are defined to reconstruct the traffic flow in the traffic flow simulation software.

6. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 2, characterized in that: The road-side device for obtaining vehicle traffic information in S1 is connected to the Internet through the smart terminal. After the road-side device identifies and obtains the vehicle model information, the vehicle model information is sent to the smart terminal. The vehicle model is used as a clue to search for the vehicle model on the Internet through the smart terminal, and the vehicle size data of the vehicle model is downloaded from the Internet to the smart terminal; At the same time, the road-side equipment sends the vehicle size data it recognizes to the smart terminal, which compares the vehicle size data downloaded from the Internet with the vehicle size data it recognizes, and obtains the error. .

7. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 6, characterized in that: According to the sensor accuracy of the roadside equipment, a permissible error range for identifying the vehicle size is set. When it is less than the allowable error range, the data is confirmed.

8. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 7, characterized in that: when When it is greater than the allowable error range, the road-side equipment is controlled to perform optical zoom, magnify the vehicle for secondary identification, obtain the vehicle model and vehicle size data again, and use the newly obtained vehicle model as a clue to search and download the vehicle size data of this model again on the Internet through the smart terminal.

9. The method for generating macro traffic flow based on roadside equipment data statistics according to claim 8, characterized in that: The vehicle size data obtained by the secondary identification is compared again with the vehicle size data downloaded from the Internet. When it is smaller than the allowable error range, the data is confirmed; when it is larger than the allowable error range, the vehicle is defined as a modified risk vehicle.

Images