Automatic driving simulation data acquisition method and device, equipment, medium and product

Abstract

The invention relates to an automatic driving simulation data acquisition method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring road side acquisition data and map data, and extracting a traffic scene based on the road side acquisition data and the map data to obtain scene data associated with at least one test vehicle on a road; the scene data comprises traffic element data and track data; screening the at least one test vehicle based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory satisfies a preset simulation test demand; and according to the road side acquisition data, the map data, and the traffic element data and the trajectory data associated with the target vehicle, performing road behavior and / or event analysis on a target road where the target vehicle is located, and obtaining automatic driving simulation data based on an analysis result. By adopting the method, the effectiveness of the automatic driving simulation test can be improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for acquiring autonomous driving simulation data. Background Art

[0002] With the transformation of the global automotive industry, the intelligent connected vehicle industry has become the main development direction. Generally speaking, a large number of autonomous driving simulation tests can be conducted to test, adjust and verify the operational feasibility of intelligent connected vehicles to ensure their smooth implementation.

[0003] To ensure the effectiveness of simulation testing, autonomous driving simulation testing usually requires the use of a large amount of complex and real traffic scene data. Usually, autonomous driving simulation data can be generated through map data, or through data collected by the test vehicle during driving. However, the autonomous driving simulation data obtained in the above way is difficult to cover a variety of traffic scenes, affecting the effectiveness of autonomous driving simulation testing. Summary of the invention

[0004] Based on this, in response to the above-mentioned technical problems, the present application at least provides a method, device, computer equipment, computer-readable storage medium and computer program product for acquiring autonomous driving simulation data that can improve the effectiveness of autonomous driving simulation testing.

[0005] In a first aspect, the present application provides a method for acquiring autonomous driving simulation data, comprising: acquiring road-side collected data and map data, extracting traffic scenes based on the road-side collected data and map data, and obtaining scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data; screening at least one test vehicle based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets preset simulation test requirements; analyzing road behavior and / or events on the target road where the target vehicle is located based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, and obtaining autonomous driving simulation data based on the analysis results.

[0006] The above-mentioned method for acquiring automatic driving simulation data can extract traffic scenes based on road-side collected data and map data to obtain scene data associated with at least one test vehicle on the road, wherein the scene data includes traffic element data and trajectory data. Further, the test vehicles will be accurately screened based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets the preset simulation test requirements, so as to achieve reasonable selection of the target vehicle, which is conducive to the subsequent effective analysis of automatic driving simulation data based on the road behavior and / or traffic behavior of the target vehicle on the target road. Further, the target road where the target vehicle is located can be analyzed for road behavior and / or events based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, so as to fully and accurately analyze the rich and complex traffic scenes on the target road and obtain automatic driving simulation data. Based on this, the automatic driving simulation data analyzed for at least one target vehicle can be obtained, so that the automatic driving simulation data covers rich and complex traffic scenes, which is conducive to improving the effectiveness of automatic driving simulation testing based on automatic driving simulation data.

[0007] In a specific example, traffic scene extraction is performed based on road-side collected data and map data to obtain scene data associated with at least one test vehicle on the road, including: extracting first vehicle status information of at least one test vehicle based on road-side collected data and map data; performing coordinate system conversion and interpolation processing on the first vehicle status information of at least one test vehicle to obtain second vehicle status information of at least one test vehicle suitable for simulation; constructing trajectory data and traffic element data associated with at least one test vehicle based on the second vehicle status information of at least one test vehicle, as well as the road-side collected data and map data.

[0008] In this embodiment, based on multi-source data (road-side collected data and map data), the first vehicle state information of each test vehicle can be fully extracted. After the first vehicle state information is converted into second vehicle state information suitable for simulation, rich scene data suitable for simulation (including trajectory data and traffic element data) can be extracted from the second vehicle state information, so as to facilitate the subsequent analysis of rich and complex autonomous driving simulation data based on the scene data, thereby improving the effectiveness of autonomous driving simulation testing based on the obtained autonomous driving simulation data.

[0009] In a specific example, traffic scene extraction is performed based on the road-side collected data and the map data to obtain scene data associated with at least one test vehicle on the road, and the method further includes: data segmentation of the road-side collected data and the map data according to time length to obtain at least one data block. Based on this, first vehicle status information of at least one test vehicle is extracted based on the road-side collected data and the map data, including: extracting the first vehicle status information of the test vehicle contained in each data block.

[0010] In this embodiment, before extracting traffic scenes from multi-source data, the multi-source data may be segmented according to time length to facilitate subsequent rapid reading, thereby improving the efficiency of extracting the first vehicle status information.

[0011] In some embodiments, the above-mentioned method for acquiring autonomous driving simulation data further includes: preprocessing traffic element data and / or preprocessing trajectory data. Preprocessing traffic element data includes: phase matching the road side collected data and the traffic element data in the map data. Preprocessing trajectory data includes: performing trajectory repair on the trajectory data.

[0012] In this embodiment, the traffic element data and / or trajectory data can be preprocessed to ensure the accuracy and completeness of the traffic element data and trajectory data, which is conducive to the subsequent accurate analysis of rich and complex traffic scenes on the target road based on the traffic element data and trajectory data, thereby obtaining autonomous driving simulation data, and further improving the effectiveness of autonomous driving simulation testing based on the autonomous driving simulation data.

[0013] In a possible implementation, performing trajectory repair on the trajectory data includes: performing trajectory quality detection on the trajectory data to obtain a trajectory quality detection result; performing interruption repair and defect filling on the trajectory data according to the trajectory quality detection result to obtain complete trajectory data; and performing smoothing on the complete trajectory data.

[0014] In this embodiment, the original trajectory data can be optimized into complete and smooth trajectory data by performing quality inspection, interruption repair, defect filling and smoothing on the trajectory data, which is conducive to the subsequent accurate and reasonable screening of each test vehicle based on the complete and smooth trajectory data, so as to screen out the target vehicle that meets the simulation test requirements, so that based on the traffic scene of the target vehicle on the target road, rich and complex autonomous driving simulation data can be analyzed to improve the effectiveness of the autonomous driving simulation test.

[0015] In one of the embodiments, at least one test vehicle is screened based on trajectory data associated with each test vehicle, including: based on the trajectory data of the at least one test vehicle, screening out at least one candidate vehicle from the at least one test vehicle whose trajectory starting point and trajectory ending point satisfy a first screening condition and whose trajectory maintains continuity between the trajectory starting point and the trajectory ending point; based on the trajectory data of the at least one candidate vehicle, screening out a test vehicle from the at least one candidate vehicle whose trajectory length and driving time satisfy a second screening condition as a target vehicle.

[0016] In this embodiment, the test vehicles can be accurately and reasonably screened based on the complete and smooth trajectory data associated with the test vehicles, and target vehicles that meet the requirements of the autonomous driving simulation test can be screened out.

[0017] In some embodiments, based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, the target road where the target vehicle is located is analyzed for road behavior and / or events to obtain autonomous driving simulation data, including: determining the target road where the target vehicle is located based on the map data and the trajectory data associated with the target vehicle, and determining the status and position of the traffic elements in the target road based on the traffic element data associated with the target vehicle; screening the traffic participant object data associated with the target road from the road-side collected data; analyzing the road behavior and / or events on the target road where the target vehicle is located based on the traffic participant object data, the trajectory data associated with the target vehicle, and the status and position of the traffic elements in the target road to obtain autonomous driving simulation data.

[0018] In this embodiment, by analyzing the road behavior and / or events on the target road where the target vehicle is located, the rich and complex traffic scenes on the target road can be fully and accurately analyzed, and autonomous driving simulation data covering rich and complex traffic scenes can be obtained, which is conducive to improving the effectiveness of subsequent autonomous driving simulation tests based on the autonomous driving simulation data.

[0019] In one embodiment, the road-side collected data includes traffic event data, and filtering the traffic participant data associated with the target road from the road-side collected data includes: filtering the behavior data of the traffic participant associated with the target road, and the traffic event data associated with the target vehicle and the traffic participant from the road-side collected data. Based on this, the target road where the target vehicle is located is analyzed for road behavior and / or events, including at least one of the following: based on the behavior data of the traffic participant, the trajectory data of the target vehicle, and the state and position of the traffic elements in the target road, analyzing the behavior of the traffic participant relative to the road and the behavior of the traffic participant relative to the target vehicle; based on the traffic event data associated with the target vehicle and the traffic participant, analyzing the traffic events on the target road.

[0020] In this embodiment, by analyzing the behavior of traffic participants relative to the road, the behavior of traffic participants relative to the target vehicle, and analyzing traffic incidents occurring on the target road, the rich and complex traffic scenes on the target road can be effectively mined, thereby obtaining autonomous driving simulation data covering rich and complex traffic scenes, thereby improving the effectiveness of autonomous driving simulation testing based on autonomous driving simulation data.

[0021] In the second aspect, the present application also provides an autonomous driving simulation data processing device, including: a scene data acquisition module, used to acquire road-side collected data and map data, extract traffic scenes based on the road-side collected data and map data, and obtain scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data; a target vehicle determination module, used to screen at least one test vehicle based on the trajectory data associated with each test vehicle, so as to determine at least one target vehicle whose trajectory meets the preset simulation test requirements; a simulation data acquisition module, used to analyze road behavior and / or events on the target road where the target vehicle is located according to the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, and obtain autonomous driving simulation data based on the analysis results.

[0022] On the third aspect, the present application also provides a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned autonomous driving simulation data acquisition method when executing the computer program.

[0023] In a fourth aspect, the present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method for acquiring autonomous driving simulation data is implemented.

[0024] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the above-mentioned autonomous driving simulation data acquisition method when executed by a processor.

[0025] The above-mentioned autonomous driving simulation data processing device, computer equipment, computer-readable storage medium and computer program product can obtain autonomous driving simulation data analyzed for at least one target vehicle, and make the autonomous driving simulation data cover rich and complex traffic scenarios, thereby improving the effectiveness of subsequent autonomous driving simulation tests based on the autonomous driving simulation data.

[0026] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and are used together with the specification to illustrate the technical solution of the present application.

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 A diagram showing an application environment of a method for acquiring autonomous driving simulation data in an embodiment;

[0030] Figure 2 A schematic diagram of a flow chart of a method for acquiring autonomous driving simulation data in one embodiment;

[0031] Figure 3 A schematic diagram of a process for extracting scene data from roadside collected data and map data in one embodiment;

[0032] Figure 4 A schematic diagram of a process for preprocessing trajectory data in one embodiment;

[0033] Figure 5 A schematic diagram of a process for reasonably selecting a target vehicle for an autonomous driving simulation test in one embodiment;

[0034] Figure 6 A schematic diagram of a process of performing a road behavior and / or event analysis on a target road where a target vehicle is located to obtain autonomous driving simulation data in one embodiment;

[0035] Figure 7 A schematic diagram of a flow chart of a method for acquiring autonomous driving simulation data in another embodiment;

[0036] Figure 8 A schematic diagram of a process for mining and managing autonomous driving simulation data in one embodiment;

[0037] Fig. 9 It is a structural block diagram of an automatic driving simulation data acquisition device in one embodiment;

[0038] Fig.10 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0040] The method for acquiring autonomous driving simulation data provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Specifically, the server 102 where the autonomous driving simulation platform is located can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The server 102 can communicate with the cloud control platform 104 through the network to obtain road-side collection data, and can also import high-precision map data from the map system 106 through the network. The cloud control platform 104 is interconnected with the collection equipment deployed on the road side, and can store a large amount of road-side collection data collected by the collection equipment. The collection equipment can specifically be: various devices that can collect data related to objects on the road, including but not limited to sensors, cameras, radars, etc. installed on the road side.

[0041] Among them, the road-side collected data and map data, as heterogeneous data sources, can provide environmental perception and road network information for subsequent processing and analysis, ensuring that complex traffic data can be accurately synchronized to the autonomous driving simulation platform. The road-side collected data can be specifically: MEC (Multi-access Edge Computing) data, such as data accessed from multiple collection devices. MEC is a distributed computing architecture that can move computing power from the central cloud to the edge of the network (such as roadside collection devices, base stations, etc.). Based on this, in vehicle-road collaboration, MEC can help vehicles obtain real-time computing services and data processing in a low-latency environment. High-precision map data can specifically include: HD-MAP (High Definition Map) data and V2X-MAP (map based on Vehicle To Everything technology) data. HD-MAP is a high-precision map provided for autonomous driving and advanced driver assistance systems. Unlike traditional navigation maps, it has centimeter-level accuracy and can contain rich road details, such as lane center lines, curbs, traffic signs, traffic lights, slopes and curves. Vehicle To Everything is a communication technology used in intelligent transportation and autonomous driving, which enables vehicles to exchange information with various entities in the surrounding environment in real time, enhancing driving safety, efficiency and traffic management. Based on this, V2X-MAP can provide road environment information such as lane information, traffic signs, and traffic signals to enhance the perception and decision-making capabilities of autonomous vehicles.

[0042] Exemplarily, the server 102 may import road-side collected data from the cloud control platform 104, and import high-precision map data from the map system 106. Furthermore, the server 102 may extract traffic scenes based on the road-side collected data and the map data, and obtain scene data associated with at least one test vehicle on the road, wherein the scene data includes traffic element data and trajectory data. The server 102 may screen at least one test vehicle based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets the preset simulation test requirements. Furthermore, the server may analyze the road behavior and / or events of the target road where the target vehicle is located based on the road-side collected data, the map data, and the traffic element data and trajectory data associated with the target vehicle, and obtain autonomous driving simulation data based on the analysis results for use in autonomous driving simulation testing on an autonomous driving simulation platform.

[0043] In an exemplary embodiment, Figure 2 As shown, a method for acquiring autonomous driving simulation data is provided, which can be applied to electronic devices with electrical data processing capabilities, such as Figure 1 The server 102 shown, or the terminal computing device on the road side, the vehicle-mounted electronic device, etc., are not particularly limited in this embodiment of the present application. Figure 1 The server 102 in the example is used as an example to illustrate, including the following steps 202 to 206. Among them:

[0044] Step 202, obtaining road-side collected data and map data, extracting traffic scenes based on the road-side collected data and map data, and obtaining scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data.

[0045] Among them, the data collected on the road side may include: the data of traffic participants on the road, the data of traffic events occurring on the road, and the traffic signal data on the road, which may be structured message data. Map data may include: road geometry data and data related to the road network. Traffic participants may include: motor vehicles, non-motor vehicles, pedestrians, etc. The test vehicle may specifically be a motor vehicle. The trajectory data of the test vehicle includes at least: the positioning of the test vehicle during driving and driving data such as acceleration, speed, and heading angle at each positioning point. Traffic element data includes at least: the position and status of traffic elements (including but not limited to traffic lights) on the road passed by the test vehicle. For each test vehicle, its associated traffic scene can be preliminarily characterized by the driving trajectory of the test vehicle and the traffic elements on the road side during driving.

[0046] Optionally, the roadside collection data collected for real traffic scenes can be imported from the cloud control platform in real time or periodically, and the latest high-precision map data can be imported from the map system. After the imported roadside collection data and map data are loaded before traffic scene extraction, traffic scene extraction can be performed based on the loaded roadside collection data and map data, that is, traffic scene mining can be performed to achieve data-to-scene conversion, and then the traffic element data and trajectory data associated with at least one test vehicle on the road can be obtained.

[0047] Specifically, the vehicle data in the road-side collected data can be extracted to determine the test vehicle, and the trajectory data related to the test vehicle in a certain road section and the traffic element data during the test vehicle's driving process can be extracted from the road-side collected data to achieve the extraction of scene data associated with the test vehicle in the traffic scene. Optionally, traffic event data on the road during the test vehicle's driving, such as data on traffic accidents, traffic jams, etc., can also be extracted as scene data associated with the test vehicle.

[0048] In this step, the road-side collected data and map data can be combined to solve the problem of difficulty in obtaining scene data, and the collected data can be updated late, avoiding the lack of richness and coverage of scene data, or the inability of scene data to reflect the latest road traffic conditions, which affects the effectiveness and accuracy of the autonomous driving simulation test. In addition, the road-side collected data in this step integrates the data collected by a large number of collection devices deployed on the road side, which can avoid data missing, jitter or inconsistency, and can ensure the quality of the road-side collected data, thereby ensuring the diversity of the extracted scene data.

[0049] Step 204 , screening at least one test vehicle based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets a preset simulation test requirement.

[0050] The target vehicle may specifically be: the vehicle that is mainly analyzed during the autonomous driving simulation test. The simulation test requirements can be flexibly preset according to the actual application scenario.

[0051] Alternatively, at least one test vehicle can be screened based on the trajectory data associated with each test vehicle, so as to reasonably determine at least one target vehicle whose trajectory meets the preset simulation test requirements, thereby solving the problem of selecting the main vehicle (target vehicle) during the autonomous driving simulation test.

[0052] Step 206, based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, analyze the road behavior and / or events of the target road where the target vehicle is located to obtain automatic driving simulation data.

[0053] Optionally, the target road where the target vehicle is located can be analyzed for road behavior and / or road events based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, so as to fully explore the traffic behavior associated with the target road and the traffic events occurring on the target road. Based on this, at least one reference scenario data covering rich and complex traffic scenarios can be extracted based on the collected analysis results of the road behavior and / or road events, and the key traffic scenarios in the reference scenario data can be automatically identified and annotated, so that when used, the features that have a key impact on the vehicle's driving decision and driving behavior can be quickly extracted based on the annotated information.

[0054] Furthermore, for each reference scenario data (with key traffic scenarios marked), the reference scenario data can be converted into a format acceptable to the autonomous driving simulation platform to obtain autonomous driving simulation data. For example, simulation use cases (CASE) and human-machine interaction (HMI, Human-Machine Interface) preview files are generated based on the reference scenario data for subsequent testing and analysis on the autonomous driving simulation platform. Furthermore, the obtained autonomous driving simulation data can be collected and stored in the background database as a long-term data asset for subsequent simulation testing, scenario reproduction or scenario mining improvement, forming a series of scenario libraries that can be called by corporate customers.

[0055] The above-mentioned method for acquiring automatic driving simulation data can extract traffic scenes based on road-side collected data and map data to obtain scene data associated with at least one test vehicle on the road, wherein the scene data includes traffic element data and trajectory data. Further, the test vehicles will be accurately screened based on the trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets the preset simulation test requirements, so as to achieve reasonable selection of the target vehicle, which is conducive to the subsequent effective analysis of automatic driving simulation data based on the road behavior and / or traffic behavior of the target vehicle on the target road. Further, the target road where the target vehicle is located can be analyzed for road behavior and / or events based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, so as to fully and accurately analyze the rich and complex traffic scenes on the target road and obtain automatic driving simulation data. Based on this, the automatic driving simulation data analyzed for at least one target vehicle can be obtained, so that the automatic driving simulation data covers rich and complex traffic scenes, which is conducive to improving the effectiveness of automatic driving simulation testing based on automatic driving simulation data.

[0056] In a specific example, traffic scene extraction is performed based on road-side collected data and map data to obtain scene data associated with at least one test vehicle on the road, including: extracting first vehicle status information of at least one test vehicle based on road-side collected data and map data; performing coordinate system conversion and interpolation processing on the first vehicle status information of at least one test vehicle to obtain second vehicle status information of at least one test vehicle suitable for simulation; constructing trajectory data and traffic element data associated with at least one test vehicle based on the second vehicle status information of at least one test vehicle, as well as the road-side collected data and map data.

[0057] The first vehicle status information may include: the vehicle's position, speed, acceleration, etc.

[0058] Alternatively, the road-side acquisition data and map data can be read in parallel, and the first vehicle state information of at least one test vehicle can be extracted therefrom to provide a basis for the subsequent generation of trajectory data. Then, the first vehicle state information of at least one test vehicle is respectively subjected to coordinate system conversion and interpolation processing. Specifically, the coordinate system conversion includes: converting the position in the first vehicle state information from the map coordinate system to the simulation coordinate system to ensure that it can be used in a simulation environment. The interpolation processing includes: interpolating the speed and acceleration in the first vehicle state information respectively to ensure the continuity and smoothness of the generated trajectory data, eliminate possible noise and avoid incomplete data, so as to obtain the second vehicle state information suitable for simulation for at least one test vehicle. Among them, the interpolation processing is used to estimate the value of unknown data between known data, and can be implemented based on at least the following methods: linear interpolation method, piecewise interpolation method, Lagrange interpolation method, and Newton interpolation method.

[0059] Furthermore, the trajectory data associated with at least one test vehicle can be generated based on the position, speed, acceleration and other information in the second vehicle status information of at least one test vehicle, and the actual traffic elements on the road side of the test vehicle's driving trajectory, i.e., traffic element data, can be obtained from the road side collection data and map data in combination with the driving trajectory represented by the trajectory data. Based on this, the trajectory data and traffic element data associated with at least one test vehicle can be collected to obtain the preliminary extracted scene data.

[0060] In this embodiment, based on multi-source data (road-side collected data and map data), the first vehicle state information of each test vehicle can be fully extracted. After the first vehicle state information is converted into second vehicle state information suitable for simulation, rich scene data suitable for simulation can be extracted from the second vehicle state information, so as to facilitate the subsequent analysis of rich and complex autonomous driving simulation data based on the scene data, thereby improving the effectiveness of autonomous driving simulation testing based on the obtained autonomous driving simulation data.

[0061] In a specific example, traffic scene extraction is performed based on the road-side collected data and the map data to obtain scene data associated with at least one test vehicle on the road, and the method further includes: data segmentation of the road-side collected data and the map data according to time length to obtain at least one data block. Based on this, first vehicle status information of at least one test vehicle is extracted based on the road-side collected data and the map data, including: extracting the first vehicle status information of the test vehicle contained in each data block.

[0062] Alternatively, after importing the roadside collected data and map data, the duration of data segmentation can be determined for the roadside data and map data according to the time scale required for simulation, and then the roadside data and map data can be segmented according to the duration to obtain at least one data block, so that each data block contains the roadside collected data and map data within a period of time. After obtaining at least one data block, the data blocks generated by the segmentation can be read in parallel, and then the first vehicle status information of the test vehicle contained in each data block can be extracted respectively to obtain the first vehicle status information of at least one test vehicle. Among them, the duration of data segmentation can be flexibly configured according to the actual simulation requirements, for example, including but not limited to 7 minutes and 30 seconds, 10 minutes, 11 minutes, etc.

[0063] For example, Figure 3As shown, taking the road-side collected data including MEC data and the map data including V2X-MAP data as an example, a flow chart of extracting scene data from the road-side collected data and the map data is provided. Specifically, after importing the road-side collected data and the map data, the imported road-side collected data and the map data can be segmented according to the duration of the data segmentation to generate N data blocks (data block 1 to data block N). Further, the N data blocks can be read in parallel, and the first vehicle state information of the test vehicle can be extracted from the read data blocks, and then the coordinates in the first vehicle state information are converted to a coordinate system, and the speed and acceleration in the first vehicle state information are interpolated to obtain the second vehicle state information. Furthermore, according to the second vehicle state information, the trajectory data of at least one test vehicle is determined, and then the traffic element data associated with at least one test vehicle is extracted from the road-side collected data according to the road corresponding to the trajectory data, so as to obtain the scene data.

[0064] In this embodiment, before extracting the traffic scene from the multi-source data, the multi-source data may be segmented to facilitate subsequent rapid reading, thereby improving the efficiency of extracting the first vehicle status information.

[0065] In some embodiments, the above-mentioned method for acquiring autonomous driving simulation data further includes: preprocessing traffic element data and / or preprocessing trajectory data. Preprocessing traffic element data includes: phase matching the road side collected data and the traffic element data in the map data. Preprocessing trajectory data includes: performing trajectory repair on the trajectory data.

[0066] Alternatively, after obtaining the traffic element data and trajectory data associated with at least one test vehicle on the road, considering that the initially extracted traffic element data and / or trajectory data may have missing or erroneous data, it is necessary to preprocess the traffic element data and / or preprocess the trajectory data to make the traffic element data and / or trajectory data more complete and accurate, and then put the preprocessed traffic element data and / or trajectory data into use.

[0067] Taking the preprocessing of traffic element data as an example, phase matching includes: traffic element state matching and traffic element position matching. Specifically, the actual state and actual position of the traffic elements on both sides of the target road where the target vehicle is located can be determined based on the road-side collected data and map data associated with the traffic elements during the driving process of the target vehicle, so as to adjust the state of the traffic elements in the traffic element data associated with the target vehicle to be consistent with the actual state (realizing traffic element state matching), and adjust the position of the traffic elements in the traffic element data associated with the target vehicle to be consistent with the actual position (realizing traffic element position matching), so as to ensure the correct interpretation of the traffic elements and make the traffic element data more complete and accurate.

[0068] Taking the preprocessing of trajectory data as an example, trajectory repair of trajectory data includes: detecting defects in the trajectory data, and then repairing the trajectory data according to the defects, so that the trajectory data is more complete, accurate, and in line with the laws of vehicle dynamics, so that at least one test vehicle can be accurately screened according to the trajectory data of each test vehicle, and then the target vehicle whose trajectory meets the preset simulation test requirements can be determined.

[0069] In this embodiment, the traffic element data and / or trajectory data can be preprocessed to ensure the accuracy and completeness of the traffic element data and trajectory data, which is conducive to the subsequent accurate analysis of rich and complex traffic scenes on the target road based on the traffic element data and trajectory data, thereby obtaining autonomous driving simulation data, and further improving the effectiveness of autonomous driving simulation testing based on the autonomous driving simulation data.

[0070] In a possible implementation, performing trajectory repair on the trajectory data includes: performing trajectory quality detection on the trajectory data to obtain a trajectory quality detection result; performing interruption repair and defect filling on the trajectory data according to the trajectory quality detection result to obtain complete trajectory data; and performing smoothing on the complete trajectory data.

[0071] Optionally, the trajectory data associated with each test vehicle is repaired, which can be specifically achieved in the following manner: first, the trajectory data is subjected to a trajectory quality test to obtain a trajectory quality test result, so as to evaluate the quality of the trajectory data, check its integrity and accuracy, and ensure that the trajectory data finally obtained can reflect the actual motion trajectory of the test vehicle. Furthermore, the trajectory interruption part and the data missing part in the trajectory data can be determined according to the trajectory quality test result, so as to perform interruption repair on the trajectory interruption part and fill in the missing part of the data, thereby obtaining complete trajectory data, solving the trajectory discontinuity problem caused by data loss or sampling anomaly, and ensuring that the trajectory data for subsequent analysis is continuous. Furthermore, the velocity and acceleration in the complete trajectory data can be preliminarily smoothed, and the heading angle data can be extracted from the complete trajectory data after the preliminarily smoothed processing, and then the heading angle extracted from the heading angle data can be smoothed to obtain the preprocessed trajectory data.

[0072] In a specific example, the following rules can be used to repair the track interruption: (1) the previous track does not intersect the next track; (2) the lateral distance between the end point of the previous track and the starting point of the next track does not exceed 1 m; (3) the longitudinal distance between the end point of the previous track and the starting point of the next track does not exceed 1 m; (4) the relative speed between the end point of the previous track and the starting point of the next track does not exceed 0.3 m / s; (5) the heading angle error between the end point of the previous track and the starting point of the next track does not exceed 5°.

[0073] Alternatively, for each test vehicle, the complete trajectory data obtained after repair can be smoothed through the following process to obtain trajectory data: (1) The trajectory, speed, and acceleration in the complete trajectory data are preliminarily smoothed to eliminate noise and unreasonable fluctuations so as to be more in line with the motion law of the real vehicle. (2) Based on the complete trajectory data after preliminary smoothing, the heading angle data is generated, and the heading angle is extracted from the heading angle data, and then the heading angle is smoothed to avoid sudden changes in the heading angle and ensure that the angle changes in the trajectory are smooth and natural. (3) Combine the processed trajectory, speed, acceleration, heading angle and other information to generate preprocessed trajectory data. Among them, the trajectory, speed, and acceleration in the complete trajectory data can be preliminarily smoothed by constructing a kinematic optimization objective function. The kinematic optimization objective function can be specifically shown as function (1):

[0074] (1)

[0075] In function (1), the coefficient =50, coefficient =0.01, is the lateral trajectory coordinate before optimization, is the optimized lateral trajectory coordinate, is the longitudinal trajectory coordinate before optimization, is the optimized longitudinal trajectory coordinate, is the heading angle, is the steering angle, is the point sequence of the optimized coordinate points. , , Used to configure kinematic constraints. Kinematic constraints can include: , to constrain lateral motion, is the time interval; , to constrain longitudinal motion; , to constrain the plane rotation, ; , to ensure rapid tracking of the target vertically; , to limit the front wheel turning angle; to avoid oversteering.

[0076] For example, Figure 4 As shown, a schematic diagram of the process of preprocessing trajectory data is provided, which mainly includes the following processes: (1) For each test vehicle, the trajectory data associated with the test vehicle can be first subjected to a trajectory quality test. (2) According to the trajectory quality test result, the trajectory interruption part and the data missing part in the trajectory data are determined, so as to repair the interruption part of the trajectory and fill the missing part of the data to obtain the complete trajectory data associated with the test vehicle. (3) Furthermore, the trajectory, speed, and acceleration in the complete trajectory data can be preliminarily smoothed. (4) Based on the complete trajectory data after the preliminary smoothing process, the heading angle is generated. (5) The heading angle is smoothed. (6) The preprocessed trajectory data is generated.

[0077] In this embodiment, the original trajectory data can be optimized into complete and smooth trajectory data by performing quality inspection, data completion and smoothing on the trajectory data, which is beneficial for subsequent accurate and reasonable screening of each test vehicle based on the complete and smooth trajectory data, thereby screening out target vehicles that meet the simulation test requirements, so that rich and complex autonomous driving simulation data can be analyzed based on the traffic scene on the target road where the target vehicle is located, thereby improving the effectiveness of the autonomous driving simulation test.

[0078] In one of the embodiments, at least one test vehicle is screened based on trajectory data associated with each test vehicle, including: based on the trajectory data of the at least one test vehicle, screening out at least one candidate vehicle from the at least one test vehicle whose trajectory starting point and trajectory ending point satisfy a first screening condition and whose trajectory maintains continuity between the trajectory starting point and the trajectory ending point; based on the trajectory data of the at least one candidate vehicle, screening out a test vehicle from the at least one candidate vehicle whose trajectory length and driving time satisfy a second screening condition as a target vehicle.

[0079] Among them, the first screening condition may specifically refer to: the starting point of the trajectory is located in the import road area of ​​the road network to ensure that the starting position of the vehicle is at the starting point of the road or the entrance of the intersection, and the end point of the trajectory is located in the exit road area of ​​the road network to ensure that the end position of the vehicle is at the end of the road or the exit of the intersection. The first screening condition can make the screened vehicles have a complete driving path, enter the road from the entrance, and exit from the exit road after passing smoothly. Maintaining trajectory continuity may specifically refer to: during the driving process, the vehicle's driving trajectory does not suddenly interrupt, jump or leave the road.

[0080] The second screening condition may specifically include: the length of the trajectory of the vehicle reaches a preset trajectory length, and the continuous driving time of the vehicle reaches a preset driving time, so as to provide sufficient data for testing. The preset trajectory length and the preset driving time can be flexibly configured according to the needs of the actual application scenario, including but not limited to: the preset trajectory length is 80 meters, 90 meters, etc., and the preset driving time is 2 seconds, 5 seconds, etc.

[0081] Optionally, based on the preprocessed trajectory data associated with at least one test vehicle, at least one candidate vehicle whose trajectory starting point and trajectory ending point meet the first screening condition and whose trajectory is continuous between the trajectory starting point and trajectory ending point can be screened out from at least one test vehicle, so that the driving trajectory of the candidate vehicle is complete, continuous and reasonable. Furthermore, based on the trajectory data of at least one candidate vehicle, a test vehicle whose trajectory length and driving time meet the second screening condition can be screened out from at least one candidate vehicle as a target vehicle, so that the target vehicle can more fully participate in rich and complex traffic scenes, and it is convenient to analyze the road behavior and / or events on the target road where the target vehicle is located, and obtain automatic driving simulation data covering rich and complex traffic scenes.

[0082] Exemplarily, after screening out test vehicles whose trajectory starting points and trajectory ending points meet the first screening condition, it is possible to verify whether the trajectory continuity is maintained between the trajectory starting point and the trajectory ending point of the test vehicle in the following manner: based on the preprocessed trajectory data associated with the test vehicle, the position of each data point between the trajectory starting point and the trajectory ending point is checked to determine whether there is any positional inconsistency between the data points.

[0083] Alternatively, after at least one candidate vehicle is screened, the length of its travel trajectory can be calculated by a positioning system deployed on the candidate vehicle, and the travel time of the candidate vehicle can be calculated by using timestamp information carried by preprocessed trajectory data associated with the candidate vehicle.

[0084] In some embodiments, Figure 5 As shown in the figure, taking the preset trajectory length of 80 meters and the preset driving time of 2 seconds as an example, a schematic diagram of the process of reasonably selecting target vehicles for autonomous driving simulation testing is provided, which mainly includes the following screening process: (1) determining that the starting point of the trajectory is located in the import road area of ​​the road network; (2) determining that the end point of the trajectory is located in the export road area of ​​the road network; (3) determining that the trajectory continuity is maintained between the starting point and the end point of the trajectory; (4) determining that the trajectory length is ≥80 meters; (5) determining that the driving time is ≥2 seconds; (6) screening out at least one target vehicle.

[0085] In this embodiment, the test vehicles can be accurately and reasonably screened based on the complete and smooth trajectory data corresponding to the test vehicles, and target vehicles with reasonable, continuous, complete trajectories that meet the requirements for autonomous driving simulation testing can be screened out, so as to facilitate subsequent analysis of road behaviors and / or events on the target roads where the target vehicles are located, and obtain autonomous driving simulation data covering rich and complex traffic scenarios.

[0086] In some embodiments, based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, the target road where the target vehicle is located is analyzed for road behavior and / or events to obtain autonomous driving simulation data, including: determining the target road where the target vehicle is located based on the map data and the trajectory data associated with the target vehicle, and determining the status and position of the traffic elements in the target road based on the traffic element data associated with the target vehicle; screening the traffic participant object data associated with the target road from the road-side collected data; analyzing the road behavior and / or events on the target road where the target vehicle is located based on the traffic participant object data, the trajectory data associated with the target vehicle, and the status and position of the traffic elements in the target road to obtain autonomous driving simulation data.

[0087] Among them, the traffic participants associated with the target road may include: vehicles (including motor vehicles and non-motor vehicles) and / or pedestrians on the target road, vehicles (including motor vehicles and non-motor vehicles) and / or pedestrians on other roads connected to the target road. Traffic participant object data may refer to: location data and movement data of traffic participant objects on the road. Map data includes high-precision data such as the location, geometry, and topological connection relationship of the road, and the trajectory data associated with the target vehicle includes the positioning data of the target vehicle.

[0088] Exemplarily, the target road where the target vehicle is located and the road type of the target road can be determined based on the map data and the trajectory data associated with the target vehicle, thereby determining the state and position of the traffic elements in the target road where the target vehicle is located during the driving process of the target vehicle based on the traffic element data associated with the target vehicle (including the position and state of the traffic elements on both sides of the target road where the target vehicle is located). Among them, determining the target road at least includes: determining the position, geometric shape, topological connection relationship, etc. of the target road, thereby determining the road type of the target road based on the position, geometric shape, and topological connection relationship of the target road. Road types include at least: straight roads, curves, diverging sections, merging sections, merging sections, merging sections, cross intersections, and T-shaped intersections.

[0089] Furthermore, the collection equipment on the road side can collect data generated by various objects on different roads. After determining the target road where the target vehicle is located, other roads connected to the target road can be determined based on the map data, and then the traffic participant object data of each traffic participant object associated with the target road can be obtained from the road side collection data, such as the traffic participant object data of the traffic participant object on the target road and the traffic participant object data of the traffic participant object on other roads connected to the target road. Among them, the number of traffic participant objects associated with the target road can be used as a measure of the complexity of the traffic scene. The more the number of traffic participant objects, the more complex the traffic scene involved in the target road.

[0090] Furthermore, considering that traffic participants that are farther away from the target vehicle have less impact on the driving behavior of the target vehicle, key traffic participants whose distance to the target vehicle is less than a preset distance can be screened out from the traffic participants associated with the target road based on the traffic participant data associated with the target road.

[0091] Based on this, the road behavior and / or events of the target vehicle on the target road can be analyzed according to the traffic participant data of key traffic participants, the trajectory data associated with the target vehicle, and the status and position of traffic elements in the target road, so as to fully explore the traffic scenarios involved in the target road and obtain autonomous driving simulation data.

[0092] In this embodiment, by analyzing the road behavior and / or events on the target road where the target vehicle is located, the rich and complex traffic scenes on the target road can be fully and accurately analyzed, and autonomous driving simulation data covering rich and complex traffic scenes can be obtained, which is conducive to improving the effectiveness of subsequent autonomous driving simulation tests based on the autonomous driving simulation data.

[0093] In one of the embodiments, the road-side collected data includes traffic event data. Filtering the traffic participant object data associated with the target road from the road-side collected data may also include: filtering the behavior data of the traffic participant objects associated with the target road, and the traffic event data associated with the target vehicle and the traffic participant objects from the road-side collected data. Based on this, the target road where the target vehicle is located is analyzed for road behavior and / or events, including at least one of the following: based on the behavior data of the traffic participant objects, the trajectory data of the target vehicle, and the state and position of the traffic elements in the target road, analyzing the behavior of the traffic participant objects relative to the road and the behavior of the traffic participant objects relative to the target vehicle; based on the traffic event data associated with the target vehicle and the traffic participant objects, analyzing the traffic events of the target road.

[0094] Alternatively, analyzing the behavior of traffic participants relative to the road may include: analyzing the behavior of a target vehicle on the target road relative to the target road, analyzing the behavior of other traffic participants on the target road (excluding the target vehicle) relative to the target road, and analyzing the behavior of traffic participants on other roads connected to the target road relative to the road on which they are located.

[0095] Optionally, analyzing the behavior of traffic participants relative to the target vehicle may include: analyzing the behavior of other traffic participants on the target road relative to the target vehicle, and analyzing the behavior of traffic participants on other roads connected to the target road relative to the target vehicle.

[0096] Based on this, the road behavior and / or events of the target road where the target vehicle is located can be analyzed to determine the behavior of traffic participants relative to the road, the behavior of traffic participants relative to the target vehicle, and traffic events occurring on the target road. Combined with the above analysis results, the rich and complex traffic scenes occurring on the target road can be fully explored, thereby obtaining autonomous driving simulation data covering the explored traffic scenes.

[0097] For example, analyzing the behavior of traffic participants relative to the road and the behavior of traffic participants relative to the target vehicle may specifically include: (1) judging the traffic behavior of the target vehicle on the target road based on the behavior data of the traffic participants, the trajectory data of the target vehicle, and the status and position of the traffic elements on the target road, including but not limited to judging whether the target vehicle is waiting for a red light, passing when the light is green, running a red light, slowing down, accelerating, stopping, changing lanes, etc., and further analyzing the rationality of the traffic behavior of the target vehicle on the target road in combination with the type of the target road. Based on this, the behavior of the target vehicle relative to the target road can be determined. (2) performing collision analysis based on the behavior data of the traffic participants and the trajectory data of the target vehicle to determine whether the target vehicle collides with the key traffic participants or whether there is a potential collision risk between the two. (3) determining the road type of the participating road where the key traffic participants are located based on the behavior data of the key traffic participants and the map data, wherein the participating road can be the target road or other roads connected to the target road. (4) Based on the behavior data of the key traffic participants, the trajectory data of the target vehicle and the map data, calculate the relative distance between the target vehicle and the key traffic participants during the driving process. (5) Combined with the road type of the participant road and the behavior data of the key traffic participants, analyze the traffic behavior of the key traffic participants on the participant road, that is, analyze the behavior of the key traffic participants relative to the road, and analyze the rationality of their traffic behavior, including but not limited to: whether the key traffic participants run a red light, whether they wait for a red light, whether they stop moving, whether they change speed, whether they turn, whether they accelerate, etc. Among them, if the key traffic participant is on the target road, the analysis is the behavior of the key traffic participant relative to the target road. (6) Based on the relative distance between the target vehicle and the key traffic participant during the driving process, analyze the interactive behavior between the target vehicle and the key traffic participant to determine the behavior of the key traffic participant relative to the target vehicle.

[0098] In an exemplary embodiment, Figure 6As shown, a flow chart of analyzing the road behavior and / or events of the target road where the target vehicle is located to obtain the autonomous driving simulation data is provided, which mainly includes the following processes: (1) For each target vehicle, the road type of the target road where the target vehicle is located can be determined based on the map data and the positioning data of the target vehicle in the trajectory data associated with the target vehicle. (2) Further, based on the traffic element data associated with the target vehicle, the state and position of the traffic elements in the target road where the target vehicle is located during the driving process of the target vehicle are determined. (3) In combination with the state and position of the traffic elements in the target road where the target vehicle is located and the trajectory data of the target vehicle, the behavior of the target vehicle relative to the target road is analyzed. (4) Further, based on the data collected on the road side, the traffic participants associated with the target road are screened. (5) The key traffic participants are screened from the traffic participants. (6) Based on the traffic participant data of the key traffic participant and the trajectory data of the target vehicle, it is analyzed whether the target vehicle has collided with the key traffic participant or whether there is a collision risk. (7) Based on the traffic participant data of the key traffic participant, the road type of the participant road where the key traffic participant is located is determined. (8) Further calculate the relative distance between the key traffic participants and the target vehicle during the driving process of the target vehicle. (9) Analyze the behavior of the key traffic participants relative to the participating road based on the traffic participant data of the key traffic participants. (10) Combine the traffic participant data of the key traffic participants and the trajectory data of the target vehicle to analyze the behavior of the traffic participants relative to the target vehicle. (11) Gather the above analysis results, and fully explore the rich and complex traffic scenes to obtain reference scene data, and annotate the key traffic scenes carried in the reference scene data. Then, convert the annotated reference scene data into a format acceptable to the autonomous driving simulation platform to obtain the corresponding autonomous driving simulation data, and store the autonomous driving simulation data in the background data.

[0099] In a specific example, the behaviors of key traffic participants relative to the target vehicle include but are not limited to: the vehicle in front of the target vehicle slows down / brakes suddenly / stops, the target vehicle has a road conflict with the oncoming vehicle, the target vehicle cuts in front of other vehicles, other vehicles cut in front of the target vehicle, pedestrians cross the zebra crossing in front of the target vehicle, the target vehicle has a road conflict with the target non-motor vehicle, etc. Specifically, the behaviors of key traffic participants relative to the target vehicle can be analyzed by the following steps (using the international standard unit system):

[0100] (1) The target vehicle's front low speed can be determined by the following constraints: 0.3m / s≤front vehicle speed≤5m / s; front vehicle low speed duration T≥5s; front vehicle lateral distance dx≤1m; front vehicle longitudinal distance dy≤15m; 4s≤ target vehicle and front vehicle expected collision time TTC≤8s. TTC (Time to Collision) is a key indicator commonly used in autonomous driving and advanced driver assistance systems. It is used to measure the expected time between two vehicles or an object and a vehicle to collide. It is calculated by dividing the relative distance by the relative speed.

[0101] (2) The following constraints can be used to determine the sudden braking of the vehicle in front of the target vehicle: the duration of the deceleration of the vehicle in front T ≥ 1s; the deceleration of the vehicle in front a ≥ 1m / s 2 ; Final speed of the front vehicle v ≥ 0.3m / s; Lateral distance of the front vehicle dx <= 1m; Longitudinal distance of the front vehicle dy ≤ 50m; Estimated collision time between the vehicle and the front vehicle TTC ≤ 4s.

[0102] (3) The following constraints can be used to determine that the target vehicle’s front vehicle has stopped: the duration of the front vehicle’s low speed T ≥ 3 s; the front vehicle’s speed v ≤ 0.3 m / s; the front vehicle’s lateral distance dx ≤ 1 m; the front vehicle’s longitudinal distance dy ≤ 50 m; and the estimated collision time TTC between the target vehicle and the front vehicle ≤ 4 s.

[0103] (4) The following constraints can be used to determine the road conflict between the target vehicle and the oncoming vehicle: the trajectories of the target vehicle and the oncoming vehicle intersect; the time interval between the intersection of the target vehicle and the oncoming vehicle is T≤2s; the target vehicle and the oncoming vehicle are in the intersection; the minimum relative distance between the target vehicle and the oncoming vehicle is dr≤20m; the heading angle error between the target vehicle and the oncoming vehicle is h≥45°; and the estimated collision time TTC between the target vehicle and the oncoming vehicle is TTC≤4s.

[0104] (5) The following constraints can be used to determine that the target vehicle cuts in front of other vehicles: the trajectories of the target vehicle and other vehicles have an intersection; the target vehicle and other vehicles do not intersect before the intersection; the target vehicle arrives at the intersection before other vehicles; the time interval between the target vehicle and other vehicles reaching the intersection of the trajectories is T≤2s; the minimum relative distance between the target vehicle and other vehicles is dr≤20m; the heading angle error between the target vehicle and other vehicles is h<=27°; and the estimated collision time between the target vehicle and other vehicles is TTC≤4s.

[0105] (6) The following constraints can be used to determine whether other vehicles cut in front of the target vehicle: the trajectories of the target vehicle and other vehicles have an intersection; the target vehicle and other vehicles do not intersect before the intersection; the target vehicle arrives at the intersection later than other vehicles; the time interval between the target vehicle and other vehicles when they arrive at the intersection is T≤2s; the minimum relative distance between the target vehicle and other vehicles is dr≤20m; the heading angle error between the target vehicle and other vehicles is h≤27°; and the estimated collision time TTC between the target vehicle and other vehicles is TTC≤4s.

[0106] (7) The following constraints can be used to determine that a pedestrian is crossing the zebra crossing in front of the target vehicle: the target vehicle and the pedestrian have intersections; the number of intersections between the target vehicle and the pedestrian is P = 1; the target vehicle arrives at the intersection later than the pedestrian; the time interval between the target vehicle and the pedestrian arriving at the intersection is T ≤ 5 s; the minimum relative distance between the target vehicle and other vehicles is dr ≤ 20 m; and the pedestrian is within the zebra crossing area.

[0107] (8) The following constraints can be used to determine the road conflict between the target vehicle and the non-motor vehicle: the trajectories of the target vehicle and the non-motor vehicle have an intersection; the target vehicle and the non-motor vehicle do not intersect before the intersection; the time interval between the target vehicle and the non-motor vehicle arriving at the intersection is T≤5s; the minimum relative distance between the target vehicle and the non-motor vehicle is dr≤20m; the estimated collision time between the target vehicle and the non-motor vehicle is TTC≤4s; the non-motor vehicle is not within the zebra crossing area.

[0108] In this embodiment, by analyzing the behavior of traffic participants relative to the road, the behavior of traffic participants relative to the target vehicle, and analyzing traffic incidents occurring on the target road, the rich and complex traffic scenes on the target road can be effectively mined, thereby obtaining autonomous driving simulation data covering rich and complex traffic scenes, thereby improving the effectiveness of autonomous driving simulation testing based on autonomous driving simulation data.

[0109] In a specific example, based on the above embodiments, Figure 7 As shown, a flowchart of another method for acquiring autonomous driving simulation data is provided, which mainly includes the following steps:

[0110] (1) Obtain multi-source heterogeneous data, including roadside collection data (MEC data), map data (V2X-MAP data and HD-MAP data), and other configuration parameters imported according to simulation test requirements.

[0111] (2) Analyze and load the traffic scene of multi-source heterogeneous data to obtain scene data. The scene data includes: traffic element data and trajectory data associated with at least one test vehicle.

[0112] (3) Preprocessing the scene data. Preprocessing includes: phase matching the roadside data and the traffic element data in the map data, and trajectory repair of the trajectory data.

[0113] (4) Based on the trajectory data, select N target vehicles (target vehicle 1 to target vehicle N) from at least one test vehicle.

[0114] (5) For each target vehicle, based on the multi-source heterogeneous data, as well as the traffic element data and trajectory data associated with the target vehicle, the road behavior and / or events of the target vehicle are analyzed on the target road, at least one reference scene data is obtained, and the key traffic scenes in each reference scene data are labeled.

[0115] (6) Converting the reference scenario data into a format acceptable to the autonomous driving simulation platform to obtain autonomous driving simulation data, wherein the autonomous driving simulation data includes: generating a simulation use case (CASE) and a human-machine interaction (HMI) preview file based on the reference scenario data.

[0116] (7) The simulation use case (CASE) and human-machine interaction (HMI) preview files are stored in the background database as long-term data assets for subsequent simulation testing, scenario reproduction or scenario mining and improvement, forming a series of scenario libraries that can be called by corporate customers.

[0117] In a specific example, Figure 8 As shown, a process diagram of mining and managing autonomous driving simulation data is provided by combining the process of mining autonomous driving simulation data on the autonomous driving simulation test platform side and the process of managing autonomous driving simulation data on the user side.

[0118] For the mining process on the autonomous driving simulation platform side, first, the autonomous driving simulation platform can import map data in real time / periodically, and import road-side collected data in real time / periodically as basic data. Then, data fusion and reasoning are performed based on the imported data to make the data continuous, reasonable, and complete, and then the autonomous driving simulation scene data is automatically mined based on the fused and reasoned data.

[0119] For the user side, such as the operation and maintenance personnel of the autonomous driving simulation platform, the operation and maintenance personnel can monitor and manage the mining process of autonomous driving simulation data through the autonomous driving simulation platform so that they can manually deal with it in time when an abnormality occurs. The operation and maintenance personnel can also perform sampling quality inspection on the autonomous driving simulation data and perform targeted operations based on the sampling quality inspection results. The operation and maintenance personnel can also manage the scenarios in detail through functions such as scenario library management and scenario management to form a scenario library for corporate customers.

[0120] In this embodiment, the above-mentioned autonomous driving simulation data acquisition method is integrated into the autonomous driving simulation platform to help verify and test the performance of the autonomous driving system. In the simulation environment provided by the autonomous driving simulation platform, the autonomous driving system can be operated and tested in various complex traffic scenarios, for example, testing under different road types, road scenarios, and traffic densities to ensure the safety and reliability of the autonomous driving process. In addition, users can use the scenario library built based on the above-mentioned autonomous driving simulation acquisition method to conveniently test the autonomous driving system, for example, to conduct tests in some rare or dangerous traffic scenarios, to help the development team verify and optimize the performance of the autonomous driving system in different situations, so as to improve the safety and reliability of the autonomous driving process.

[0121] In a specific application scenario, a scenario algorithm tool chain for the autonomous driving scenario library can be constructed based on the above-mentioned autonomous driving simulation acquisition method. That is, based on the real traffic flow data obtained on the roadside of the demonstration area, services such as scenario data preprocessing, scenario mining, scenario generation, scenario quality inspection, and scenario conversion are constructed. By calling the scene mining algorithm, a tool chain providing basic scenario data and scenario analysis for the construction of the demonstration area scenario library system is provided. Through the scenario algorithm tool chain, data in the demonstration area can be accumulated continuously and automatically and efficiently, providing data support and algorithm support for the construction of the scenario library, and continuously constructing a demonstration area characteristic scenario library system including a real traffic event scenario library, a characteristic road condition scenario library, and a standard and regulatory scenario library.

[0122] Taking the test of data collected on a single day at a single intersection in a certain intelligent connected demonstration area as an example, there are a total of 500 test samples, and the scene data distribution is approximately 8:4:2:3:3:12:12:22. The test results are shown in Table 1. The test conclusions are as follows: (1) The number of simulation scene types (mining operators) that support automatic mining is ≥8; (2) The efficiency of random sampling of automatic mining simulation scenes is ≥90%; (3) The number of automatically generated simulation scenes is ≥200 / day*intersection.

[0123] Table 1 Scene mining correctness test table

[0124]

[0125] It can be seen that the scenario algorithm tool chain constructed using the above-mentioned autonomous driving simulation acquisition method can effectively perform scenario mining, thereby improving the effectiveness of autonomous driving simulation testing.

[0126] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0127] Based on the same inventive concept, the embodiment of the present application also provides an autonomous driving simulation data processing device for implementing the autonomous driving simulation data acquisition method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more autonomous driving simulation data acquisition device embodiments provided below can refer to the limitations of the autonomous driving simulation data acquisition method above, and will not be repeated here.

[0128] In an exemplary embodiment, Fig. 9 As shown, an automatic driving simulation data processing device is provided, comprising: a scene data acquisition module 902, a target vehicle determination module 904 and a simulation data acquisition module 906. Among them:

[0129] A scene data acquisition module is used to acquire road-side collected data and map data, extract traffic scenes based on the road-side collected data and map data, and obtain scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data;

[0130] A target vehicle determination module, used to screen at least one test vehicle based on trajectory data associated with each test vehicle to determine at least one target vehicle whose trajectory meets a preset simulation test requirement;

[0131] The simulation data acquisition module is used to analyze the road behavior and / or events of the target vehicle on the target road based on the road-side collected data, map data, and traffic element data and trajectory data associated with the target vehicle, and obtain automatic driving simulation data based on the analysis results.

[0132] The above-mentioned automatic driving simulation data acquisition device can extract traffic scenes based on the road side collection data and map data to obtain scene data associated with at least one test vehicle on the road, and the scene data includes traffic element data and trajectory data. Further, based on the trajectory data associated with each test vehicle, the test vehicle will be accurately screened to determine at least one target vehicle whose trajectory meets the preset simulation test requirements, so as to achieve reasonable selection of the target vehicle, which is conducive to the subsequent road behavior and / or traffic behavior of the target vehicle on the target road, and effectively extract the automatic driving simulation data. Further, according to the road side collection data, map data, and the traffic element data and trajectory data associated with the target vehicle, the target road where the target vehicle is located can be analyzed for road behavior and / or events, so as to fully and accurately analyze the rich and complex traffic scenes on the target road, and obtain the automatic driving simulation data. Based on this, the automatic driving simulation data analyzed for at least one target vehicle can be obtained, so that the automatic driving simulation data covers rich and complex traffic scenes, which is conducive to improving the effectiveness of the automatic driving simulation test based on the automatic driving simulation data.

[0133] In a specific example, the scene data acquisition module may include: a first vehicle status information extraction unit, used to extract the first vehicle status information of at least one test vehicle based on road-side collected data and map data; a second vehicle status information acquisition unit, used to perform coordinate system conversion and interpolation processing on the first vehicle status information of at least one test vehicle, so as to obtain the second vehicle status information of at least one test vehicle suitable for simulation; a scene data extraction unit, used to construct trajectory data and traffic element data associated with at least one test vehicle based on the second vehicle status information of at least one test vehicle, as well as the road-side collected data and map data.

[0134] In a specific example, the scene data acquisition module may further include a data segmentation unit, which is used to segment the roadside collected data and the map data according to the time length to obtain at least one data block. Based on this, the first vehicle state information extraction unit may also be used to extract the first vehicle state information of the test vehicle contained in each data block.

[0135] In a possible implementation, the trajectory data preprocessing unit is specifically used to: perform trajectory quality detection on the trajectory data to obtain a trajectory quality detection result; perform interruption repair and defect filling on the trajectory data according to the trajectory quality detection result to obtain complete trajectory data; and perform smoothing on the complete trajectory data.

[0136] In one of the embodiments, the target vehicle determination module is specifically used to: based on the trajectory data of at least one test vehicle, select from at least one test vehicle at least one candidate vehicle whose trajectory starting point and trajectory ending point meet the first screening condition and whose trajectory maintains continuity between the trajectory starting point and the trajectory ending point; based on the trajectory data of at least one candidate vehicle, select from at least one candidate vehicle a test vehicle whose trajectory length and driving time meet the second screening condition as the target vehicle.

[0137] In some embodiments, the simulation data acquisition module includes: a first processing unit, used to determine the target road where the target vehicle is located based on the map data and the trajectory data associated with the target vehicle, and to determine the status and position of the traffic elements in the target road based on the traffic element data associated with the target vehicle; a second processing unit, used to filter the traffic participant object data associated with the target road from the road side collected data; a third processing unit, used to analyze the road behavior and / or events of the target road where the target vehicle is located based on the traffic participant object data, the trajectory data associated with the target vehicle, and the status and position of the traffic elements in the target road to obtain autonomous driving simulation data.

[0138] In one embodiment, the road-side collected data includes traffic event data, and the second processing unit is specifically used to: filter the behavior data of traffic participants associated with the target road, and the traffic event data associated with the target vehicle and the traffic participants from the road-side collected data. Based on this, the third processing unit is specifically used to: analyze the behavior of the traffic participants relative to the road and the behavior of the traffic participants relative to the target vehicle based on the behavior data of the traffic participants, the trajectory data of the target vehicle, and the status and position of the traffic elements in the target road; analyze the traffic events of the target road based on the traffic event data associated with the target vehicle and the traffic participants.

[0139] Each module in the above-mentioned automatic driving simulation data processing device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0140] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.10As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store autonomous driving simulation data processing data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for acquiring autonomous driving simulation data is implemented.

[0141] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0142] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0143] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0144] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0145] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0146] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0147] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0148] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for acquiring autonomous driving simulation data, characterized in that: The method comprises: Acquire road-side collected data and map data, extract traffic scenes based on the road-side collected data and the map data, and obtain scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data; Screening the at least one test vehicle based on the trajectory data associated with each of the test vehicles to determine at least one target vehicle whose trajectory meets a preset simulation test requirement; According to the road-side collected data, the map data, and the traffic element data and trajectory data associated with the target vehicle, the road behavior and / or events of the target road on which the target vehicle is located are analyzed, and the autonomous driving simulation data is obtained based on the analysis results.

2. The method according to claim 1, characterized in that The extracting of traffic scenes based on the road-side collected data and the map data to obtain scene data associated with at least one test vehicle on the road includes: extracting first vehicle status information of each of the at least one test vehicle based on the roadside collected data and the map data; Performing coordinate system conversion and interpolation processing on the first vehicle state information of each of the at least one test vehicle to obtain second vehicle state information of each of the at least one test vehicle suitable for simulation; According to the second vehicle status information of each of the at least one test vehicle, the road-side collected data and the map data, trajectory data and traffic element data associated with each of the at least one test vehicle are constructed.

3. The method according to claim 2, characterized in that The extracting of traffic scenes based on the road side collected data and the map data to obtain scene data associated with at least one test vehicle on the road also includes: Segmenting the road-side collected data and the map data according to time length to obtain at least one data block; The extracting first vehicle status information of each of the at least one test vehicle based on the road side collected data and the map data includes: The first vehicle status information of the test vehicle contained in each of the data blocks is extracted respectively.

4. The method according to claim 1, characterized in that: The method further comprises: preprocessing the traffic element data and / or preprocessing the trajectory data; The preprocessing of the traffic element data includes: Phase matching of the road side collected data and the traffic element data in the map data; The preprocessing of the trajectory data comprises: Performing trajectory repair on the trajectory data.

5. The method according to claim 4, characterized in that The performing trajectory repair on the trajectory data comprises: Performing a trajectory quality test on the trajectory data to obtain a trajectory quality test result; According to the trajectory quality detection result, the trajectory data is repaired and the defects are supplemented to obtain complete trajectory data; The complete trajectory data is smoothed.

6. The method according to claim 1, characterized in that The screening of the at least one test vehicle based on the trajectory data associated with each of the test vehicles comprises: According to the trajectory data of the at least one test vehicle, at least one candidate vehicle is selected from the at least one test vehicle, the trajectory starting point and the trajectory ending point of which satisfy a first screening condition and the trajectory of the candidate vehicle is continuous between the trajectory starting point and the trajectory ending point; According to the trajectory data of the at least one candidate vehicle, a test vehicle whose trajectory length and driving time meet the second screening condition is screened out from the at least one candidate vehicle as a target vehicle.

7. The method according to claim 1, characterized in that The analyzing the road behavior and / or events of the target road where the target vehicle is located according to the road side collected data, the map data, and the traffic element data and trajectory data associated with the target vehicle to obtain the automatic driving simulation data includes: Determine a target road on which the target vehicle is located based on the map data and the trajectory data associated with the target vehicle, and determine the state and position of traffic elements in the target road based on the traffic element data associated with the target vehicle; Filtering the traffic participant object data associated with the target road from the road side collected data; According to the traffic participant data, the trajectory data associated with the target vehicle, and the status and position of the traffic elements in the target road, the road behavior and / or events of the target road where the target vehicle is located are analyzed to obtain automatic driving simulation data.

8. The method according to claim 7, characterized in that The road-side collected data includes traffic event data, and the filtering of the traffic participant object data associated with the target road from the road-side collected data includes: filtering the behavior data of the traffic participant object associated with the target road and the traffic event data associated with the target vehicle and the traffic participant object from the road-side collected data; The analyzing of the road behavior and / or event on the target road where the target vehicle is located includes at least one of the following: Analyze the behavior of the traffic participant relative to the road and the behavior of the traffic participant relative to the target vehicle based on the behavior data of the traffic participant, the trajectory data of the target vehicle, and the status and position of the traffic elements in the target road; Based on the traffic event data associated with the target vehicle and the traffic participant objects, a traffic event on the target road is analyzed.

9. An automatic driving simulation data acquisition device, characterized in that: The device comprises: A scene data acquisition module, used to acquire road-side collected data and map data, extract traffic scenes based on the road-side collected data and the map data, and obtain scene data associated with at least one test vehicle on the road; the scene data includes traffic element data and trajectory data; A target vehicle determination module, configured to screen the at least one test vehicle based on the trajectory data associated with each of the test vehicles to determine at least one target vehicle whose trajectory meets a preset simulation test requirement; A simulation data acquisition module is used to analyze the road behavior and / or events of the target road where the target vehicle is located based on the road-side collected data, the map data, and the traffic element data and trajectory data associated with the target vehicle, and obtain automatic driving simulation data based on the analysis results.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images