Urban-related road end data construction method, system, medium and equipment

Abstract

The invention provides a road end data construction method and system for a city, a medium and equipment, and the method comprises the steps: S1, collecting urban traffic data based on road end equipment, and automatically carrying out the data classification and fragmentation extraction through an event rule, and obtaining event data; s2, constructing an automatic driving simulation dynamic scene based on the preprocessed event data; and S3, constructing a scene feature library according to the key working condition parameters and the key traffic flow feature parameters of the dynamic scene, and constructing a simulation test scene library. According to the invention, the utilization rate of roadside end data in local automatic driving vehicle-road cooperation can be effectively improved, and the data value is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving testing, and in particular to a method, system, medium and device for constructing road-side data in a city. Background Art

[0002] At present, building an autonomous driving scenario library based on roadside data from the Internet of Vehicles has become a key data foundation for testing autonomous driving before its commercialization. The 24 / 7 processing of massive data by roadside equipment is the most critical issue currently faced, namely, how to obtain valuable data from massive data, how to transform data into effective test scenarios, and how to drive autonomous driving simulation testing and evaluation through data flywheels.

[0003] Among the published patents on roadside data construction methods, most only focus on how to generate scenarios from roadside data, ignoring how to propose rule definitions for effective scenarios from a testing perspective, how to extract scenario features based on already produced scenarios, and how to build a city-specific scenario library to realize a data flywheel.

[0004] Patent document CN117036733A discloses a method for extracting characteristic events in urban road scenes, which summarizes common and reliable event categories in urban road scenes, constructs a simulated data set containing category labels corresponding to characteristic events, and trains a classification model for obtaining characteristic events in road scenes. However, the invention does not propose a rule definition for effective scenes from a testing perspective. Summary of the invention

[0005] In view of the defects in the prior art, the purpose of the present invention is to provide a method, system, medium and equipment for constructing road end data of a city.

[0006] A method for constructing roadside data of a city provided by the present invention includes: Step S1: collecting urban traffic data based on road-side equipment, and automatically classifying and extracting data by using event rules to obtain event data; Step S2: constructing an autonomous driving simulation dynamic scene based on the preprocessed event data; Step S3: construct a scenario feature library based on key operating condition parameters and key traffic flow characteristic parameters of the dynamic scenario, and construct a simulation test scenario library.

[0007] Preferably, in step S1: Step S1.1: Based on the urban traffic data collected by the road-side equipment, including: Step S1.1.1: Classify and divide the points and coverage of the road-end equipment, determine the coverage of the local urban road conditions, and set the road condition type and road condition weight for the road-end equipment points; Step S1.1.2: define the content of the target-level urban traffic data, including target object identification time, target object unique identification code, target object type, target object size, target object speed, and target object location; Step S1.2: Event rules, including: Step S1.2.1: Define event types, including operating condition type events, traffic flow type events, and autonomous vehicle operation type events; Step S1.2.2: define rules corresponding to event types, including: time range rules, coverage rules, and valid event type rules; Step S1.2.3: Other event characteristics, including collection time, traffic characteristics, and traffic participation type; Step S1.3: Classification and fragmentation extraction of data, including: Step S1.3.1: Identify and extract data by event type and save as event data; Step S1.3.2: Different event types correspond to different data time ranges and data coverage; Step S1.4: Event data, including: acquisition time, target unique identification code, type, location, size, direction, and speed.

[0008] Preferably, in step S2: Step S2.1: Preprocessing of event data, including: data format normalization, data timing, trajectory smoothing, velocity smoothing, target jump tracking and fusion, invalid target data deletion, and target size matching; Step S2.2: Construction of dynamic scenarios. Using data tools, based on the simulation test platform and scenario production process specifications, the pre-processed event data is processed into simulation scenarios, and corresponding labeled scenarios are included, including: automatic scenario generation, scenario secondary processing, and scenario rationality verification.

[0009] Preferably, in step S3: Step S3.1: Scenario feature extraction: Use data tools to extract scenario features, and perform statistics and analysis on the scenario features. Combined with road condition features, parameter settings, parameter distribution, and parameter combination configurations are formed to provide a realistic design basis for the simulation test logic scenario, including: extraction of traffic participant behavior features, macro traffic flow features, working condition key features, and road condition key features. Step S3.2: Construction of urban simulation test scenario library: Based on the scenario feature library of real roadside data, the urban test scenario library is constructed, including: scenario library construction, scenario evaluation plan construction, test evaluation report construction, test plan and scenario configuration; Step S3.2.1: The construction of the scenario library includes: urban working condition scenario library and urban traffic flow scenario library.

[0010] According to a roadside data construction system for a city provided by the present invention, executing the roadside data construction method for a city includes: Event rule module: quantify event rules into mathematical relations and formulas of target-level data; Data classification and fragmentation extraction module: Use event rule module and automatic data extraction algorithm to process road-end target-level data to obtain event data; Event data preprocessing module: uses automated data preprocessing algorithms to process event data and obtain standardized event data; Dynamic scene construction module: uses an automated scene construction algorithm to process the pre-processed normalized event data to obtain a dynamic scene; Dynamic scene key working condition parameter acquisition module: uses an automated working condition parameter statistical algorithm to process dynamic scenes and obtain key working condition parameters preset for dynamic scenes; Dynamic scene key traffic flow characteristic parameter acquisition module: uses automated traffic characteristic statistical algorithm to process dynamic scenes and obtains key traffic flow characteristic parameters preset for dynamic scenes; A typical working condition scenario library construction module for a city is used to construct a typical working condition scenario library preset for a city according to the key working condition parameters preset for the dynamic scenario; Urban traffic flow scenario library construction module: constructs an urban traffic flow scenario library according to the key traffic flow characteristic parameters preset in the dynamic scenario.

[0011] Preferably, in the urban typical operating scenario library construction module: The typical working condition scenario library preset in the city stores a variety of typical working condition scenarios preset in the city, including following scenarios, cut-in and cut-out scenarios, target vehicle lateral offset scenarios, vulnerable traffic participant crossing scenarios, obstacle scenarios, intersection scenarios, and special scenarios.

[0012] Preferably, the vulnerable traffic participant crossing scenarios include: zebra crossing scenarios, crossing scenarios without zebra crossings, and scenarios after being blocked; Obstacle scenarios include: illegal parking and road occupation scenarios, and construction road occupation scenarios; Intersection scenarios include: intersection merging scenarios, and traffic light scenarios with different states; Special scenarios include: mixed traffic scenarios of motor vehicles and vulnerable traffic participants, and slow-moving scenarios.

[0013] Preferably, in the urban traffic flow scene library construction module: The urban traffic flow scene library stores scenes of different time and different traffic characteristics, wherein the traffic characteristics cover traffic flow, traffic density and average vehicle speed, including low-density and low-flow and high-density and high-flow.

[0014] According to a computer-readable storage medium storing a computer program provided by the present invention, when the computer program is executed by a processor, the steps of the method for constructing road-end data about a city are implemented.

[0015] According to the present invention, a roadside data construction device for a city includes: a controller; The controller includes the computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for constructing road-end data about a city are implemented; or, the controller includes the system for constructing road-end data about a city.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses real urban traffic data collected by roadside equipment, and performs regular classification and fragmented extraction of urban traffic data. According to the key working condition parameters and key traffic flow characteristic parameters of dynamic scenes, a typical urban working condition scene library and an urban traffic flow scene library that are more in line with local urban characteristics are constructed as a test database for the landing application of local urban characteristics and a test database for the local urban autonomous driving road test management specifications, which improves the response capability of the autonomous driving system to local characteristic working conditions and promotes the commercial landing application of the autonomous driving system in the local area. At the same time, it can also effectively improve the utilization rate of the roadside data in the local autonomous driving vehicle-road collaboration and enhance the data value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A schematic diagram of the construction method of the present invention; Figure 2 It is a schematic diagram of the structure of the device constructed in the present invention (the dotted frame part). DETAILED DESCRIPTION

[0018] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0019] Embodiment 1: In view of the development trend of existing scenario libraries and the shortcomings of existing patents, the present invention proposes a method, device, equipment and storage medium for constructing road-side data in cities, starting from the definition of effective scenario rules, to data preprocessing and scenario construction, and finally proposing a city-specific test scenario library and scenario test evaluation scheme based on scenario characteristics, forming a complete data flywheel and test closed loop based on real road-side data. The present invention improves the efficiency of using massive road-side data and the degree of automation of simulation scenario production based on road-side data, thereby improving the realism and complexity of simulation scenarios, and using simulation scenarios constructed based on real data to conduct simulation tests of intelligent connected vehicles, effectively replacing some road tests, and improving the safety of urban road driving faced during commercial application.

[0020] According to a method for constructing road end data of a city provided by the present invention, Figure 1-Figure 2 As shown, including: Step S1: collecting urban traffic data based on road-side equipment, and automatically classifying and extracting data by using event rules to obtain event data; Specifically, in step S1: Step S1.1: Based on the urban traffic data collected by the road-side equipment, including: Step S1.1.1: Classify and divide the points and coverage of the road-end equipment, determine the coverage of the local urban road conditions, and set the road condition type and road condition weight for the road-end equipment points; According to the road elements, the road condition type is set for the roadside equipment points, and the road condition type code is formed to assign the roadside equipment point attributes. Road elements include: the number of lanes, the central separation method, and the number of roads entering the intersection (for example, the intersection has 4 roads entering). The road condition type code is encoded according to ["intersection code", "number of lanes in the southeast, northwest and east directions", "central separation method code in the southeast, northwest and east directions"]. For example, the road condition type code of an intersection with two lanes in both directions is 422221111.

[0021] Intersection code table

[0022] Central partition code table

[0023] When the road condition type code assigned by the intersection point equipment is the same code value, or the code consistency can be achieved through rotation (for example, the intersection 44242111 with four lanes in both directions and two lanes in both directions, and the intersection 424241111 with four lanes in both directions and two lanes in both directions, can be made consistent by rotating 90 degrees, that is, shifting the digits backward. These two intersections can be regarded as the same code value and have the same road condition characteristics), it means that the road condition characteristics of the intersection are highly consistent and can be regarded as intersections of the same intersection type.

[0024] By counting the intersection type code values ​​of point equipment in the entire area and based on the distribution of the number of intersection characteristics in the area, a regional road condition weight table can be formed according to the regional characteristics. The road condition weight can be taken into consideration when analyzing the regional traffic flow characteristics.

[0025] The formula for calculating the road condition weight is: number of points with the same code value / total number of points. For example, if there are 20 intersections in the project area, 3 of them are intersections 44242111 with four lanes in both directions and two lanes in both directions, and they have the same code value. Therefore, the weight of the road condition obtained by the roadside equipment at the intersection with the code value of 44242111 is 4 / 20=20%.

[0026] Step S1.1.2: define the content of the target-level urban traffic data, including target object identification time, target object unique identification code, target object type, target object size, target object speed, and target object location; Step S1.2: Event rules, including: Step S1.2.1: Define event types, including operating condition type events, traffic flow type events, and autonomous vehicle operation type events; Step S1.2.2: define rules corresponding to event types, including: time range rules, coverage rules, and valid event type rules; Step S1.2.3: Other event characteristics, including collection time, traffic characteristics, and traffic participation type; Step S1.3: Classification and fragmentation extraction of data, including: Step S1.3.1: Identify and extract data by event type and save as event data; Step S1.3.2: Different event types correspond to different data time ranges and data coverage; Step S1.4: Event data, including: acquisition time, target unique identification code, type, location, size, direction, and speed.

[0027] Step S2: constructing an autonomous driving simulation dynamic scene based on the preprocessed event data; Specifically, in step S2: Step S2.1: Preprocessing of event data, including: data format normalization, data timing, trajectory smoothing, velocity smoothing, target jump tracking and fusion, invalid target data deletion, and target size matching; Step S2.2: Construct dynamic scenarios using data tools (since roadside data is regularized data, such as json format, and dynamic scenarios are also regularized data, such as openscenario format, it is possible to write roadside data into dynamic scenarios by writing python scripts or C++ scripts based on personal abilities and habits). Based on the simulation test platform and scenario production process specifications, the pre-processed event data is processed into simulation scenarios, and corresponding labeled scenarios are generated, including: automatic scenario generation, scenario secondary processing, and scenario rationality verification.

[0028] Among them, the data tool writes roadside data into dynamic scenes by writing Python scripts or C++ scripts. The specific process includes: data collection: obtaining data from sensors, databases or other data sources in real time; data processing: cleaning, converting and formatting the collected data for subsequent use; scene construction: building dynamic scenes based on the processed data, including drawing graphics, updating UI elements or simulating physical behaviors; real-time update: updating the scene periodically or based on event triggers to keep it synchronized with real-time data; user interaction: processing user input, including mouse clicks and keyboard keys, to change the scene or obtain more information.

[0029] Step S3: construct a scenario feature library based on key operating condition parameters and key traffic flow characteristic parameters of the dynamic scenario, and construct a simulation test scenario library.

[0030] Specifically, in step S3: Step S3.1: Scenario feature extraction: Use data tools (write the required Python script or C++ script according to the above rules and requirements based on personal ability) to extract scenario features, and perform statistics and analysis on the scenario features. Combined with road condition features, parameter settings, parameter distribution and parameter combination configuration are formed to provide a realistic design basis for the simulation test logic scenario, including: extraction of traffic participant behavior features, macro traffic flow features, working condition key features, and road condition key features. Among them, the data tool writes roadside data into dynamic scenes by writing Python scripts or C++ scripts; specifically, it includes: data preprocessing: using Python libraries including Pandas, NumPy or C++ standard libraries to clean, format and convert the collected data; data visualization: using libraries including Matplotlib, Folium, Pyecharts to visualize the processed data into dynamic scenes; data writing into dynamic scenes: using Web frameworks including Flask, Django or real-time data push technologies including WebSocket to push processed data to the front end; the front end uses technologies including JavaScript and HTML5 Canvas to dynamically render data into scenes; Among them, feature selection requires selecting the most appropriate feature extraction algorithm according to the specific application scenario, including SIFT and SURF; image preprocessing requires image preprocessing before feature extraction, including denoising, contrast enhancement, and resizing to improve the effect of feature extraction; visualization functions are used to display key points and descriptors to deeply understand the process and results of feature extraction; Step S3.2: Construction of urban simulation test scenario library: Based on the scenario feature library of real roadside data, the urban test scenario library is constructed, including: scenario library construction, scenario evaluation plan construction, test evaluation report construction, test plan and scenario configuration; Step S3.2.1: The construction of the scenario library includes: urban working condition scenario library and urban traffic flow scenario library.

[0031] Embodiment 2: Embodiment 2 is a preferred example of Embodiment 1, and is used to illustrate the present invention in more detail.

[0032] The present invention also provides a road-end data construction system for a city. The road-end data construction system for a city can be implemented by executing the process steps of the road-end data construction method for a city, that is, those skilled in the art can understand the road-end data construction method for a city as a preferred implementation of the road-end data construction system for a city.

[0033] According to a roadside data construction system for a city provided by the present invention, executing the roadside data construction method for a city includes: Event rule module: quantify event rules into mathematical relations and formulas of target-level data; The event rule module includes: conditional selection of the affected vehicle (i.e. the main vehicle) in the event, conditional selection of the key influencing vehicle (i.e. the key target vehicle) in the event, and event rule data screening. The conditional selection of the main vehicle and the key target vehicle is not necessary and can be used in conjunction with the event rules.

[0034] Condition selection of the affected vehicle (i.e., the main vehicle) in the event: The conditions include the vehicle type of the main vehicle, the minimum duration requirement for the main vehicle to exist in the data segment, the shortest mileage requirement for the main vehicle to travel in the data segment, and the minimum proportion requirement for the main vehicle's non-stationary time in the data segment. These four conditions can be customized according to test requirements. The default values ​​are that the main vehicle type is a sedan, the shortest duration requirement for the main vehicle to exist is 15 seconds, the shortest mileage requirement for the main vehicle to exist is 100 meters, and the minimum proportion of the main vehicle's non-stationary time is 50%. All targets in the data segment are conditionally screened, and each unique ID is regarded as a target. Vehicles that meet the conditions can be selected as the main vehicle. When the main vehicle condition selection is checked and set, each condition can be set separately or together. Only the main vehicle that meets the conditions can be transferred and included in the next event rule data screening.

[0035] Conditional selection of key influencing vehicles (i.e. key target vehicles) in the event: The conditions include the vehicle type of the target vehicle, the minimum duration requirement for the target vehicle to exist in the data segment, and the minimum mileage requirement for the target vehicle to travel in the data segment. These three conditions can be customized according to the test requirements. The default values ​​are that the target vehicle type is a sedan, the minimum duration requirement for the target vehicle to exist is 5 seconds, and the minimum mileage requirement for the target vehicle to exist is 20 meters. All target objects in the data segment are conditionally screened, and each unique ID is regarded as a target object. Vehicles that meet the conditions can be selected as key target vehicles. When the key target vehicle condition selection is checked and set, each condition can be set separately or together. Only key target vehicles that meet the conditions can be transferred and included in the next event rule data screening.

[0036] Event rule data screening includes: congestion in the driving area of ​​the main vehicle, the presence of a specific target vehicle close to the main vehicle, and changes in the external input conditions of the main vehicle.

[0037] The basic principle of event rule data screening is: select the vehicles that meet the main vehicle conditions one by one as the main vehicle for data processing. Data processing requires converting the coordinates of other vehicles into the main vehicle coordinate system coordinates, and using the xy values ​​in the main vehicle coordinate system as the basis for calculating the lateral and longitudinal motion relationship relative to the main vehicle (x corresponds to the lateral distance, y corresponds to the longitudinal distance). At the same time, the speed also needs to be converted to the speed component in the main vehicle coordinate system (vx represents the lateral speed, vy represents the longitudinal speed). All subsequent data calculations of event rules are based on the main vehicle coordinate system.

[0038] (1) Congestion in the driving area of ​​the main vehicle: With the main vehicle coordinates as the bottom center of the rectangle, the length TJ_Length and width TJ_Width of the rectangle determined by the driving area are given. The rectangle is the future forward driving area of ​​the main vehicle. Given the time TJ_Duration for maintaining the calculation state, determine whether the xy value of the target vehicle in the main vehicle coordinate system falls within the rectangular driving area and satisfies the falling time ≥ 0.5*TJ_Duration. If so, the target vehicle is in the driving area in front of the main vehicle. Calculate the number of target vehicles that meet the conditions and determine whether the number meets the vehicle threshold of the congestion condition. The recommended calculation method for the vehicle congestion threshold is: according to the standard of 5 meters long and 2 meters wide for a car, the number of vehicles that can exist in the longitudinal direction is TJ_Length / 5, and the number of vehicles that can exist in the lateral direction is TJ_Width / 2. Considering the parking space of vehicles, etc., the threshold is recommended to be set to [0.5*(TJ_Length / 5)]*[0.5* TJ_Width / 2]. The entire data segment is calculated by sliding window one by one according to the given TJ_Duration based on the main vehicle. When the number of target vehicles in a certain TJ_Duration segment of a main vehicle is ≥ the vehicle congestion threshold, the time segment of the main vehicle is the data segment extracted by the event rule.

[0039] (2) There is a specific target vehicle that is relatively close to the main vehicle: With the main vehicle coordinates as the center, the Euclidean distance from the target object to the main vehicle is calculated, and the nearest Euclidean distance setting and target object type setting are given. When a frame of the target vehicle meets the minimum Euclidean distance requirement, the 2 seconds before and after this time point of the target vehicle is the data segment that the event rule needs to extract.

[0040] (3) Changes in external input conditions of the host vehicle: Based on the horizontal / longitudinal speed of the target object in the host vehicle coordinate system mentioned above, subtract the speed of the host vehicle to convert it into relative horizontal / longitudinal speed. The horizontal / longitudinal distance of the target vehicle in the host vehicle coordinate system is the relative horizontal / longitudinal distance. Based on the given horizontal and longitudinal distance change rate and horizontal and longitudinal speed change rate (the default recommended values ​​are the relative horizontal distance change rate of 1m / s, the relative longitudinal distance change rate of 2m / s, and the relative horizontal / longitudinal speed change rate of 1.5m / s 2 ), calculate the state change of the target vehicle in each frame. When the target vehicle meets the condition of the state change rate under a given number of consecutive frames, the time period of the target vehicle is the data segment that needs to be extracted by the event rule.

[0041] Data classification and fragmentation extraction module: Use event rule module and automatic data extraction algorithm to process road-end target-level data to obtain event data; Event data preprocessing module: uses automated data preprocessing algorithms to process event data and obtain standardized event data; Dynamic scene construction module: uses an automated scene construction algorithm to process the pre-processed normalized event data to obtain a dynamic scene; Dynamic scene key working condition parameter acquisition module: uses an automated working condition parameter statistical algorithm to process dynamic scenes and obtain key working condition parameters preset for dynamic scenes; Dynamic scene key traffic flow characteristic parameter acquisition module: uses automated traffic characteristic statistical algorithm to process dynamic scenes and obtains key traffic flow characteristic parameters preset for dynamic scenes; A typical working condition scenario library construction module for a city is used to construct a typical working condition scenario library preset for a city according to the key working condition parameters preset for the dynamic scenario; Urban traffic flow scenario library construction module: constructs an urban traffic flow scenario library according to the key traffic flow characteristic parameters preset in the dynamic scenario.

[0042] Specifically, in the urban typical operating scenario library construction module: The typical working condition scenario library preset in the city stores a variety of typical working condition scenarios preset in the city, including following scenarios, cut-in and cut-out scenarios, target vehicle lateral offset scenarios, vulnerable traffic participant crossing scenarios, obstacle scenarios, intersection scenarios, and special scenarios.

[0043] Specifically, the crossing scenarios of vulnerable traffic participants include: crossing at zebra crossings, crossing without zebra crossings, and appearing after being blocked; Obstacle scenarios include: illegal parking and road occupation scenarios, and construction road occupation scenarios; Intersection scenarios include: intersection merging scenarios, and traffic light scenarios with different states; Special scenarios include: mixed traffic scenarios of motor vehicles and vulnerable traffic participants, and slow-moving scenarios.

[0044] Specifically, in the urban traffic flow scenario library construction module: The urban traffic flow scene library stores scenes of different time and different traffic characteristics, wherein the traffic characteristics cover traffic flow, traffic density and average vehicle speed, including low-density and low-flow and high-density and high-flow.

[0045] According to a computer-readable storage medium storing a computer program provided by the present invention, when the computer program is executed by a processor, the steps of the method for constructing road-end data about a city are implemented.

[0046] According to the present invention, a roadside data construction device for a city includes: a controller; The controller includes the computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for constructing road-end data about a city are implemented; or, the controller includes the system for constructing road-end data about a city.

[0047] Embodiment 3: Embodiment 3 is a preferred example of Embodiment 1, and is used to illustrate the present invention in more detail.

[0048] The present invention adopts the following technical solutions: In a first aspect, the present invention provides a method for constructing roadside data of a city. The method for constructing a dynamic scene of the roadside data of a city comprises: 1. Based on the urban traffic data collected by road-side equipment, event rules are used to automatically classify and extract fragments of data to obtain event data.

[0049] 1.1. Based on the urban traffic data collected by roadside equipment, including: 1.1.1. Classify and divide the locations and coverage of road-end equipment, determine the coverage of typical road conditions in the local city, and set the road condition type and weight for the road-end equipment locations.

[0050] 1.1.2. Define the content of target-level urban traffic data, including at least the target identification time (such as UNIX time), target unique identification code, target type, target size, target speed, and target location.

[0051] 1.2. Event rules, including: 1.2.1. Define event types, including operating condition type events, traffic flow type events, and autonomous vehicle operation type events.

[0052] 1.2.2. Define the rules corresponding to the event type. The rules include: time range rules, coverage rules, and valid event type rules.

[0053] 1.2.3. Other event characteristics. Including collection time, traffic characteristics (traffic flow, traffic density), and traffic participation type.

[0054] 1.3. Data classification and fragmentation extraction, including: 1.3.1. Identify and extract data by event type and save it as event data.

[0055] 1.3.2. Different event types correspond to different data time ranges and data coverage.

[0056] 1.4. Event data, including: collection time, target unique identification code, type, location, size, direction, and speed.

[0057] 2. Build dynamic scenarios for autonomous driving simulation based on preprocessed event data.

[0058] 2.1. Preprocessing of event data, including: data format normalization, data timing, trajectory smoothing, velocity smoothing, target jump tracking and fusion, invalid target data deletion, and target size matching.

[0059] 2.2. Construction of dynamic scenarios. Using specific data tools, based on the simulation test platform and scenario production process specifications, the pre-processed event data is processed into scenarios that can be used for simulation, and corresponding labeled scenarios are included. This includes: automatic scenario generation, scenario secondary processing, and scenario rationality verification.

[0060] 3. Build a scenario feature library based on the key operating parameters and key traffic flow characteristic parameters of dynamic scenarios, and build a simulation test scenario library with urban characteristics.

[0061] 3.1. Scenario feature extraction. Use specific data tools to extract scenario features, and perform statistics and analysis on the scenario features. Combined with road condition features, form realistic and reasonable parameter settings, parameter distributions, and parameter combination configurations, providing a realistic design basis for simulation test logic scenarios. This includes: extraction of traffic participant behavior features, macro traffic flow feature extraction, working condition key feature extraction, road condition key feature extraction, etc.

[0062] 3.2. Construction of a simulation test scenario library with city characteristics. Based on the scenario feature library of real roadside data, a test scenario library with city characteristics is constructed to help the commercialization of intelligent connected vehicles in the local area. Including: scenario library construction, scenario evaluation plan construction, test evaluation report construction, test plan and scenario configuration, etc.

[0063] 3.2.1. The construction of scenario library includes: urban typical working condition scenario library and urban traffic flow scenario library.

[0064] In a second aspect, the present invention provides a device for constructing roadside data of a city. The device for constructing dynamic scenes of the roadside data of a city comprises: 1. An event rule module is configured to quantify event rules into mathematical relations and formulas of target-level data.

[0065] 2. The data classification and fragmentation extraction module is configured to process the road-end target-level data using the event rule module and the automated data extraction algorithm to obtain event data.

[0066] 3. An event data preprocessing module is configured to process event data using an automated data preprocessing algorithm to obtain standardized event data.

[0067] 4. A dynamic scene construction module is configured to process the pre-processed normalized event data using an automated scene construction algorithm to obtain a dynamic scene.

[0068] 5. A dynamic scene key operating condition parameter acquisition module is configured to process the dynamic scene using an automated operating condition parameter statistical algorithm to obtain the dynamic scene key operating condition parameters.

[0069] 6. A dynamic scene key traffic flow characteristic parameter acquisition module is configured to process the dynamic scene using an automated traffic characteristic statistical algorithm to obtain the dynamic scene key traffic flow characteristic parameters.

[0070] 7. A module for constructing a library of typical urban working conditions is configured to construct a library of typical urban working conditions based on the key working condition parameters of the dynamic scenes; the library of typical urban working conditions stores a variety of typical urban working conditions scenarios, including following scenarios, cut-in and cut-out scenarios, target vehicle lateral offset scenarios, vulnerable traffic participant crossing scenarios (zebra crossing scenarios, crossing scenarios without zebra crossings, and scenarios appearing after being blocked), obstacle scenarios (illegal parking and road occupation scenarios, construction road occupation scenarios, etc.), intersection scenarios (intersection merging scenarios, traffic scenes with different states of traffic lights), special scenarios (mixed traffic scenarios of motor vehicles and vulnerable traffic participants, slow-moving scenarios), etc.

[0071] 8. An urban traffic flow scenario library construction module is configured to construct an urban traffic flow scenario library based on the key traffic flow characteristic parameters of the dynamic scenario; the urban traffic flow scenario library stores scenes of different times and different traffic characteristics, wherein the traffic characteristics cover traffic flow, traffic density and average vehicle speed, such as low-density and low-flow (mostly occurring at night) and high-density and high-flow (mostly occurring during peak hours in the morning and evening on weekdays).

[0072] In addition, to achieve the above-mentioned purpose, the present invention also proposes a scene construction device, which includes: a memory, a processor, and a scene construction program stored in the memory and executable on the processor, and the scene construction program implements the scene construction method described above when executed by the processor.

[0073] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which a scene construction program is stored, and when the scene construction program is executed by a processor, the scene construction method as described above is implemented.

[0074] The method and device proposed in this method are all implemented in Python. During the implementation process, we used Python 3.9 as the interpreter and PyCharm as the main operating software. In order to implement this method, we need the following Python packages: pandas, numpy, xml.etree.ElementTree, scipy, math, and pyproj. These packages provide us with functions such as data processing, mathematical calculations, and geographic information processing.

[0075] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0076] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for constructing roadside data in a city, characterized in that: include: Step S1: collecting urban traffic data based on road-side equipment, and automatically classifying and extracting data by using event rules to obtain event data; Step S2: constructing an autonomous driving simulation dynamic scene based on the preprocessed event data; Step S3: constructing a scenario feature library according to key operating condition parameters and key traffic flow characteristic parameters of the dynamic scenario, and constructing a simulation test scenario library; In step S1: Step S1.1: Based on the urban traffic data collected by the road-side equipment, including: Step S1.1.1: Classify and divide the points and coverage of the road-end equipment, determine the coverage of the local urban road conditions, and set the road condition type and road condition weight for the road-end equipment points; Step S1.1.2: define the content of the target-level urban traffic data, including target object identification time, target object unique identification code, target object type, target object size, target object speed, and target object location; Step S1.2: Event rules, including: Step S1.2.1: Define event types, including operating condition type events, traffic flow type events, and autonomous vehicle operation type events; Step S1.2.2: define rules corresponding to event types, including: time range rules, coverage rules, and valid event type rules; Step S1.2.3: Other event characteristics, including collection time, traffic characteristics, and traffic participation type; Step S1.3: Classification and fragmentation extraction of data, including: Step S1.3.1: Identify and extract data by event type and save as event data; Step S1.3.2: Different event types correspond to different data time ranges and data coverage; Step S1.4: Event data, including: acquisition time, target unique identification code, type, location, size, direction, and speed.

2. The method for constructing road end data about a city according to claim 1, characterized in that: In step S2: Step S2.1: Preprocessing of event data, including: data format normalization, data timing, trajectory smoothing, velocity smoothing, target jump tracking and fusion, invalid target data deletion, and target size matching; Step S2.2: Construction of dynamic scenarios, using data tools, based on the simulation test platform and scenario production process specifications, processing the pre-processed event data into simulation scenarios, and corresponding labeled scenarios, including: automatic scenario generation, scenario secondary processing, and scenario rationality verification; Among them, the data tool writes roadside data into dynamic scenes by writing Python scripts or C++ scripts. The specific process includes: data collection: obtaining data from sensors, databases or other data sources in real time; data processing: cleaning, converting and formatting the collected data for subsequent use; scene construction: building dynamic scenes based on the processed data, including drawing graphics, updating UI elements or simulating physical behaviors; real-time update: updating the scene periodically or based on event triggers to keep it synchronized with real-time data; user interaction: processing user input, including mouse clicks and keyboard keys, to change the scene or obtain more information.

3. The method for constructing roadside data of a city according to claim 1, characterized in that: In step S3: Step S3.1: Scenario feature extraction: Use data tools to extract scenario features, and perform statistics and analysis on the scenario features. Combined with road condition features, parameter settings, parameter distribution, and parameter combination configurations are formed to provide a realistic design basis for the simulation test logic scenario, including: extraction of traffic participant behavior features, macro traffic flow features, working condition key features, and road condition key features. Among them, the data tool writes roadside data into dynamic scenes by writing Python scripts or C++ scripts; specifically, it includes: data preprocessing: using Python libraries including Pandas, NumPy or C++ standard libraries to clean, format and convert the collected data; data visualization: using libraries including Matplotlib, Folium, Pyecharts to visualize the processed data into dynamic scenes; data writing into dynamic scenes: using Web frameworks including Flask, Django or real-time data push technologies including WebSocket to push processed data to the front end; the front end uses technologies including JavaScript and HTML5Canvas to dynamically render data into scenes; Among them, feature selection requires selecting the most appropriate feature extraction algorithm according to the specific application scenario, including SIFT and SURF; image preprocessing requires image preprocessing before feature extraction, including denoising, contrast enhancement, and resizing to improve the effect of feature extraction; visualization functions are used to display key points and descriptors to deeply understand the process and results of feature extraction; Step S3.2: Construction of urban simulation test scenario library: Based on the scenario feature library of real roadside data, the urban test scenario library is constructed, including: scenario library construction, scenario evaluation plan construction, test evaluation report construction, test plan and scenario configuration; Step S3.2.1: The construction of the scenario library includes: urban working condition scenario library and urban traffic flow scenario library.

4. A roadside data construction system for a city, characterized in that: The method for constructing road-end data of a city according to any one of claims 1 to 3 comprises: Event rule module: quantify event rules into mathematical relations and formulas of target-level data; Data classification and fragmentation extraction module: Use event rule module and automatic data extraction algorithm to process road-end target-level data to obtain event data; Event data preprocessing module: uses automated data preprocessing algorithms to process event data and obtain standardized event data; Dynamic scene construction module: uses an automated scene construction algorithm to process the pre-processed normalized event data to obtain a dynamic scene; Dynamic scene key working condition parameter acquisition module: uses an automated working condition parameter statistical algorithm to process dynamic scenes and obtain key working condition parameters preset for dynamic scenes; Dynamic scene key traffic flow characteristic parameter acquisition module: uses automated traffic characteristic statistical algorithm to process dynamic scenes and obtains key traffic flow characteristic parameters preset for dynamic scenes; A typical working condition scenario library construction module for a city is used to construct a typical working condition scenario library preset for a city according to the key working condition parameters preset for the dynamic scenario; Urban traffic flow scenario library construction module: constructs an urban traffic flow scenario library according to the key traffic flow characteristic parameters preset in the dynamic scenario.

5. The roadside data construction system for a city according to claim 4, characterized in that: In the typical working condition scenario library construction module of the city: The typical working condition scenario library preset in the city stores a variety of typical working condition scenarios preset in the city, including following scenarios, cut-in and cut-out scenarios, target vehicle lateral offset scenarios, vulnerable traffic participant crossing scenarios, obstacle scenarios, intersection scenarios, and special scenarios.

6. The road end data construction system for a city according to claim 5, characterized in that: The crossing scenarios for vulnerable traffic participants include: crossing at a zebra crossing, crossing without a zebra crossing, and appearing after being blocked; Obstacle scenarios include: illegal parking and road occupation scenarios, and construction road occupation scenarios; Intersection scenarios include: intersection merging scenarios, and traffic light scenarios with different states; Special scenarios include: mixed traffic scenarios of motor vehicles and vulnerable traffic participants, and slow-moving scenarios.

7. The roadside data construction system for a city according to claim 4, characterized in that: In the urban traffic flow scenario library construction module: The urban traffic flow scene library stores scenes of different time and different traffic characteristics, wherein the traffic characteristics cover traffic flow, traffic density and average vehicle speed, including low-density and low-flow and high-density and high-flow.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for constructing road end data for a city described in any one of claims 1 to 3 are implemented.

9. A device for constructing roadside data of a city, characterized in that: include: Controller; The controller includes the computer-readable storage medium storing the computer program as described in claim 8, and when the computer program is executed by the processor, the steps of the method for constructing road-end data for a city as described in any one of claims 1 to 3 are implemented; or, the controller includes the system for constructing road-end data for a city as described in any one of claims 4 to 7.

Images