Autonomous driving traffic flow simulation method, system, device and storage medium

Abstract

The present application provides a method, system, electronic device and storage medium for simulating traffic flow in an autonomous driving, including: generating a road environment based on acquired map information and road network information; determining the generation range of traffic participants in the road environment and the configuration data of traffic participants corresponding to the generation range of traffic participants according to the type of traffic participants; generating a first traffic flow simulation scene based on the type of traffic participants, the configuration data of traffic participants and the road environment; wherein the configuration data of traffic participants is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the first traffic flow simulation scene; and is determined according to the generation range of traffic participants. According to the technical solution of the present application, different traffic flow simulation scenes can be automatically generated, thereby improving the scene generation efficiency.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to an autonomous driving traffic flow simulation method, system, device and storage medium. Background Art

[0002] Autonomous driving technology is a technology that installs sensors on vehicles and controls the vehicle through autonomous driving algorithms. Only autonomous driving algorithms that have passed simulation tests can be run on vehicles.

[0003] At present, test scenarios are derived from limited real road collection data and mainly rely on manually written script generation, which limits the types of generated simulation scenarios and reduces the efficiency of simulation scenario generation. Summary of the invention

[0004] In order to solve the above problems, the present application proposes an autonomous driving traffic flow simulation method, system, device and storage medium, which can automatically generate different traffic flow simulation scenarios and improve the scene generation efficiency.

[0005] According to a first aspect of an embodiment of the present application, a method for simulating traffic flow in an autonomous driving system is provided, comprising:

[0006] Generate a road environment based on the acquired map information and road network information;

[0007] According to the type of the traffic participant, determining a generation range of the traffic participant in the road environment, and configuration data of the traffic participant corresponding to the generation range of the traffic participant;

[0008] A first traffic flow simulation scenario is generated based on the type of traffic participants, the configuration data of traffic participants and the road environment; wherein the configuration data of traffic participants is at least one of vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the first traffic simulation scenario.

[0009] According to a second aspect of an embodiment of the present application, there is provided an autonomous driving traffic flow simulation system, comprising:

[0010] A traffic flow generation module, used to generate a road environment based on the acquired map information and road network information; for the type of the traffic participant, determine the generation range of the traffic participant in the road environment, and the configuration data of the traffic participant corresponding to the generation range of the traffic participant; based on the type of the traffic participant, the configuration data of the traffic participant and the road environment, generate a first traffic flow simulation scene; wherein the configuration data of the traffic participant is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required for generating the traffic simulation scene;

[0011] a traffic flow control module, configured to determine the target type of the traffic participant to be controlled based on the application scenario characteristics of the first autonomous driving algorithm, and to perform behavior control on the traffic participant of the target type in the first traffic flow simulation scenario to generate a second traffic flow simulation scenario;

[0012] It is also used to adjust the playback motion trajectory of other traffic participants when receiving the execution instruction of the closed-loop test if there is a motion trajectory intersection between the vehicle executing the first autonomous driving algorithm and other traffic participants;

[0013] A scene recording module, used for recording the first traffic flow simulation scene and the second traffic flow simulation scene to generate scene recording data;

[0014] A scene playback module, used to reconstruct the scene recording data in the scene recording module to obtain a traffic flow playback scene;

[0015] A collision detection module, used to detect the intersection of the motion trajectories of the vehicle executing the first autonomous driving algorithm and other traffic participants in the traffic flow playback scene;

[0016] An evaluation module is used to evaluate the second traffic flow simulation scenario based on the similarity between the second traffic flow simulation scenario and a preset real traffic flow scenario, the uniform speed driving time of the vehicle executing the first automatic driving algorithm, and the driving hazard state of the vehicle executing the first automatic driving algorithm to obtain an evaluation result.

[0017] A third aspect of the present application provides an electronic device, comprising: a control device;

[0018] The control device comprises: a memory and a processor;

[0019] The memory is connected to the processor and is used to store programs;

[0020] The processor implements the above-mentioned autonomous driving traffic flow simulation method by running the program in the memory.

[0021] A fourth aspect of the present application provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the above-mentioned autonomous driving traffic flow simulation method is implemented.

[0022] One embodiment of the above application has the following advantages or beneficial effects:

[0023] A road environment is generated based on the acquired map information and road network information, and a generation range of traffic participants is selected in the road environment according to the type of traffic participants, so that a first traffic flow simulation scene is generated according to the configuration data of the traffic participants corresponding to the generation range of the traffic participants, the type of traffic participants and the road environment. In this way, different traffic flow simulation scenes can be automatically generated based on the configuration data of the traffic participants, thereby improving the generation efficiency of the traffic flow simulation scenes, and because the configuration data of the traffic participants is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the traffic simulation scenes, the generated traffic flow simulation scenes have diversity, complexity and authenticity. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0025] Figure 1 A schematic diagram of an autonomous driving traffic flow simulation system provided in an embodiment of the present application;

[0026] Figure 2 A flow chart of a method for simulating traffic flow in an autonomous driving system provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of a specific process for testing according to a traffic flow simulation scenario provided in an embodiment of the present application;

[0028] Figure 4 A schematic diagram of the scene recording process provided in an embodiment of the present application;

[0029] Figure 5 A schematic diagram of a closed-loop test mode flow chart provided in an embodiment of the present application;

[0030] Figure 6 A schematic diagram of the process of the open-loop test mode provided in the embodiment of the present application;

[0031] Figure 7A schematic diagram of the structure of an automatic driving traffic flow simulation device provided in an embodiment of the present application;

[0032] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] The technical solution of the embodiment of the present application is applicable to various scene simulation scenarios, such as autonomous driving scenarios, intelligent traffic scenarios, etc. The technical solution of the embodiment of the present application can automatically generate different traffic flow simulation scenarios, thereby improving the scene generation efficiency.

[0034] The technical solution of the embodiment of the present application can be exemplarily applied to hardware devices such as processors, electronic devices, servers (including cloud servers), or packaged into software programs to be run. When the hardware device executes the processing of the technical solution of the embodiment of the present application, or the above-mentioned software program is run, the purpose of automatically generating different traffic flow simulation scenes is achieved. The embodiment of the present application only exemplarily introduces the specific processing of the technical solution of the present application, and does not limit the specific implementation form of the technical solution of the present application. Any technical implementation form that can execute the processing of the technical solution of the present application can be adopted by the embodiment of the present application.

[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0036] Exemplary Systems

[0037] Figure 1 : is a flow chart of an autonomous driving traffic flow simulation system according to an embodiment of the present application. In an exemplary embodiment, an autonomous driving traffic flow simulation system is provided, comprising:

[0038] A traffic flow generation module, used to generate a road environment based on the acquired map information and road network information; for the type of the traffic participant, determine the generation range of the traffic participant in the road environment, and the configuration data of the traffic participant corresponding to the generation range of the traffic participant; based on the type of the traffic participant, the configuration data of the traffic participant and the road environment, generate a first traffic flow simulation scene; wherein the configuration data of the traffic participant is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required for generating the first traffic simulation scene;

[0039] a traffic flow control module, configured to determine the target type of the traffic participant to be controlled based on the application scenario characteristics of the first autonomous driving algorithm, and to perform behavior control on the traffic participant of the target type in the first traffic flow simulation scenario to generate a second traffic flow simulation scenario;

[0040] It is also used to adjust the playback motion trajectory of other traffic participants when receiving the execution instruction of the closed-loop test if there is a motion trajectory intersection between the vehicle executing the first autonomous driving algorithm and other traffic participants;

[0041] A scene recording module, used for recording the first traffic flow simulation scene and the second traffic flow simulation scene to generate scene recording data;

[0042] A scene playback module, used to reconstruct the scene recording data in the scene recording module to obtain a traffic flow playback scene;

[0043] A collision detection module, used to detect the intersection of the motion trajectories of the vehicle executing the first autonomous driving algorithm and other traffic participants in the traffic flow playback scene;

[0044] An evaluation module is used to evaluate the second traffic flow simulation scenario based on the similarity between the second traffic flow simulation scenario and a preset real traffic flow scenario, the uniform speed driving time of the vehicle executing the first automatic driving algorithm, and the driving hazard state of the vehicle executing the first automatic driving algorithm to obtain an evaluation result.

[0045] Specifically, the traffic flow generation module and the traffic flow control module initialize the vehicle parameter information of the traffic flow scene, select different vehicle control models and pedestrian control models to complete the planning and control of the vehicle, thereby generating a complete traffic flow scene, and the simulation scene recording module will record the simulation test process, and can also use open-loop testing or closed-loop testing in the traffic flow playback scene to test the autonomous driving algorithm. Finally, the traffic flow evaluation module evaluates the traffic flow simulation scene. In this way, the autonomous driving traffic flow simulation system can generate a variety of different traffic flow simulation scenes, and can reconstruct the same simulation scene through the scene playback module to repeatedly test the autonomous driving algorithm, so as to timely discover the problems of the autonomous driving algorithm and the improvement effect. At the same time, the traffic flow simulation scene can be evaluated in a timely manner.

[0046] Exemplary Methods

[0047] Figure 2 : is a flow chart of an autonomous driving traffic flow simulation method according to an embodiment of the present application. In an exemplary embodiment, an autonomous driving traffic flow simulation method is provided, comprising:

[0048] S110, generating a road environment based on the acquired map information and road network information;

[0049] S120, selecting a generation range of traffic participants in the road environment and configuration data of traffic participants corresponding to the generation range of traffic participants according to the type of traffic participants;

[0050] S130. Generate a first traffic flow simulation scenario based on the type of traffic participant, the configuration data of the traffic participant and the road environment; wherein the configuration data of the traffic participant is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the first traffic simulation scenario.

[0051] It should be noted that steps S110-S130 can be run in a simulation system, the operating system of the simulation system can be Ubuntu 22.04, the memory can be set to 16257868kB, and the graphics card can be Nvidia 3060. Other hardware environments can also be used, which are not limited here.

[0052] In step S110, illustratively, the map information represents the surface conditions of the region (such as mountains, rivers, villages, cities, etc.). The road network information represents a traffic network composed of many roads such as main roads, secondary roads, branch roads, forked roads, etc., and traffic signs (such as lane lines, zebra crossings, traffic lights, and pedestrian lines, etc.). Optionally, the map information and the road network information can be data extracted from real road collection data, wherein the real road collection data can be a road surface video or road surface image taken by a camera. Specifically, the simulation system can be started, and the map information and the road network information can be used as parameters of the calling interface (Application Programming Interface, API) to generate a road environment.

[0053] In step S120, illustratively, traffic participants refer to objects that appear in the traffic flow scene and participate in the actual traffic process, such as motor vehicles, non-motor vehicles, pedestrians, road obstacles, animals, etc. on the road. The types of traffic participants may include life forms and vehicles, wherein life forms include pedestrians and animals, etc. Taking the types of traffic participants as pedestrians and vehicles as an example, the traffic participant generation range includes at least a pedestrian generation range and a vehicle generation range. Optionally, in the case where the traffic participant generation range is a pedestrian generation range, it may be a specified path or range in the road, etc. The configuration parameter corresponding to the pedestrian generation range is the number of pedestrians. For example, all generated pedestrians are in the specified path.

[0054] Preferably, when the type of the traffic participant is a vehicle, the step of selecting a generation range of the traffic participant in the road environment for the type of the traffic participant includes:

[0055] At least one path is selected in the road environment to be determined as the generation range of the traffic participant; or, a target range is generated in the road environment with a preset vehicle spacing as a radius, and the target range is determined as the generation range of the traffic participant; or, the road environment is determined as the generation range of the traffic participant. Among them, the preset vehicle spacing can be set according to actual needs and is not limited here. The target range can be a circular range generated with the vehicle spacing as a radius. In this way, traffic participants can be generated in different areas, thereby forming different traffic flow simulation scenes.

[0056] Furthermore, the configuration data of the traffic participants represent the situation of the traffic participants in the traffic flow scene. Specifically, the configuration data of the traffic participants include: at least one of vehicle configuration parameters, map configuration parameters, behavior control configuration parameters, and traffic rules configuration parameters. For example, the first traffic flow simulation scene can be generated according to any one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters, and traffic rules configuration parameters combined with the type of traffic participants and the road environment. It is also possible to select multiple configuration parameters from the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters, and traffic rules configuration parameters, and combine the multiple configuration parameters with the type of traffic participants and the road environment to generate the first traffic flow simulation scene.

[0057] Specifically, the map configuration information represents the map information obtained in step S110, that is, the map used to generate the road environment. The vehicle configuration parameters may include: vehicle-specific configuration parameters and vehicle-general configuration parameters. Specifically, each designated vehicle range generation method may correspond to different vehicle-specific configuration parameters, which may be set according to actual conditions and are not limited here. Specifically, if the vehicle generation range is determined in the road environment by a specified radius, the corresponding vehicle-specific configuration parameter is the distance between vehicles. If the vehicle generation range is determined in the road environment by a specified path, the corresponding vehicle-specific configuration parameter is the total number of vehicles or vehicle flow or vehicle density, that is, after determining the designated path, vehicles are generated on the designated path according to the total number of vehicles or vehicle flow or vehicle density. If the global is used as the vehicle generation range, there is no need to set vehicle-specific configuration parameters, and vehicles are randomly generated globally.

[0058] The vehicle general configuration parameters for each specified vehicle range generation method include: vehicle type ratio (vehicles-ratio) and vehicle speed factor (vehicles-speed-factor).

[0059] Optionally, the vehicle type ratio indicates the ratio of generated vehicles. The parameter type is int / list and the default value is 1. For example, there are five types of vehicles (cars, motorcycles, trucks, vans, bicycles). The input number can be greater than 1, and the program will automatically normalize the input number. If only one number is entered, only cars will be generated.

[0060] The vehicle speed factor represents the speed factor of the generated vehicle type. The parameter type is float / list and the default value is 1.0. For example, there are five types of vehicles (cars, motorcycles, trucks, vans, bicycles). This parameter type is list, and the order corresponds to the vehicles-speed-factor of each type of vehicle. If only one integer is given, then the vehicles-speed-factor for all vehicle types is the integer, otherwise it is set to 1.0 by default, and on this road, 95% of the vehicle speeds will be between 80% and 120% of the vehicles-speed-factor*the speed limit of this section.

[0061] Traffic rules configuration parameters include: break-law-possibility, which represents the coefficient of changing the possibility of breaking the law (running a red light), that is, the proportion of vehicles that will definitely run a red light in the total number of generated vehicles. The parameter type is float_range[0,1), and the default value is 0.0.

[0062] Behavior control configuration parameters include: lateral resolution (lateral-resolution) and simulation step length (step-length).

[0063] Optionally, the lateral resolution indicates that the lane is implicitly divided into one or more sub-lanes, that is, each vehicle can occupy multiple sub-lanes, describing the lane information in a finer granularity. This parameter can simulate the behavior of multiple two-wheeled vehicles driving in parallel on a single lane, or vehicles overtaking on a single lane. The parameter type is float_range[0,3), and the default value is 0.1.

[0064] The simulation step size indicates the step size of the control model in the traffic flow control module to control the NPC, in seconds. The parameter type is float, and the default value is 0.1.

[0065] In step S130, illustratively, the first traffic flow simulation scene represents a simulation scene generated between at least two traffic participants. Optionally, a traffic flow simulation scene can be formed between every two traffic participants. For example, when two motor vehicles are driving forward and backward on the same lane, a simple traffic flow simulation scene of two vehicles driving forward and backward can be formed. Furthermore, when the positions, driving speeds, and driving directions of the two vehicles are different, other traffic flow simulation scenes can be derived, such as the front vehicle cutting in, the front vehicle braking, the rear vehicle overtaking, etc. It can be understood that the types of traffic participants in the first traffic flow simulation scene can be different, such as motor vehicles and bicycles, motor vehicles and pedestrians, motor vehicles and obstacles, etc. Various motor vehicles and other traffic participants can constitute a traffic flow simulation scene.

[0066] Specifically, the acquired types of traffic participants, configuration data and road environment may be input into a pre-trained scene generation model to obtain a first traffic flow simulation scene; wherein the pre-trained scene generation model may be obtained by training a neural network model according to traffic participant type samples, configuration data samples and road environment samples. Alternatively, a generation model of pedestrians may be obtained by training a neural network according to pedestrian configuration data samples, and a generation model of vehicles may be obtained by training a neural network according to vehicle configuration data samples. A generation model may be selected according to the type of traffic participants, and the configuration data of traffic participants may be input into a generation model of pedestrians or a generation model of vehicles, and the corresponding traffic participants may be output, and the traffic participants may be integrated with the road environment to obtain the first traffic flow simulation scene.

[0067] In the technical solution of the present application, a road environment is generated based on the acquired map information and road network information, and the generation range of traffic participants is selected in the road environment according to the type of traffic participants. In this way, the configuration data of the traffic participants is determined according to the generation range of the traffic participants, and then the first traffic flow simulation scene is generated based on the type of traffic participants, the configuration data of the traffic participants and the road environment. In this way, different traffic flow simulation scenes can be automatically generated based on the configuration data of the traffic participants, and the generation efficiency of the traffic flow simulation scenes is improved. Moreover, since the configuration data of the traffic participants is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the traffic simulation scene, the generated traffic flow simulation scene has diversity, complexity and authenticity.

[0068] In one embodiment, the method further comprises:

[0069] Based on the application scenario characteristics of the first autonomous driving algorithm, the target type of the traffic participant to be controlled is determined, and the behavior of the traffic participants of the target type in the first traffic flow simulation scenario is controlled to generate a second traffic flow simulation scenario.

[0070] Exemplarily, the first autonomous driving algorithm represents the algorithm applied when the vehicle is autonomously driving. Optionally, the first autonomous driving algorithm can be any algorithm and is not limited here. The application scenario characteristics include the driving characteristics of other vehicles in the scenario and the road environment characteristics of the vehicle executing the first autonomous driving algorithm, wherein other vehicles represent vehicles in the scenario other than the vehicle executing the first autonomous driving algorithm.

[0071] For example, the driving characteristics of other vehicles include: aggressive driving and normal driving. Among them, aggressive driving means that there are vehicles in the scene that are often accompanied by violent vehicle movements such as sudden acceleration, sudden deceleration and sharp turns. Normal driving means that the vehicles in the execution scene are usually accompanied by uniform vehicle movements. The road environment characteristics of the vehicle executing the first autonomous driving algorithm represent the traffic sign characteristics in the scene. For example, if the traffic light at the crosswalk is red, then the target type of the traffic participant to be controlled is determined to be a pedestrian, and the pedestrian is controlled to stop crossing the crosswalk. For another example, if vehicle A is likely to collide with the vehicle executing the first autonomous driving algorithm, then the target type of the traffic participant to be controlled is determined to be vehicle A, and vehicle A is controlled to change lanes. Behavior control means changing the actions, speeds, and paths of traffic participants. For example, microscopic actions of pedestrians, speeds of vehicles, etc. Specifically, behavior control includes: behavior control of pedestrians and behavior control of vehicles.

[0072] Optionally, the behavior control of pedestrians can adopt a pedestrian control model, such as a pedestrian following model (LogisticRegression), which extracts feature vectors of pedestrians and surrounding traffic participants (including vehicles and pedestrians), outputs the acceleration or deceleration features of pedestrians, controls the acceleration or deceleration features of pedestrians, and thus controls the microscopic movements of pedestrians in the simulation, thereby generating high-likelihood pedestrian trajectories.

[0073] The behavior control of the vehicle can adopt the vehicle control model, wherein the vehicle control model adopts the GAIL algorithm. It can be understood that the vehicle control model can control all vehicles in the traffic flow simulation scene except the vehicle executing the automatic driving algorithm. The vehicle control model extracts the feature vectors of the self-vehicle (i.e., the vehicle to be controlled) and the surrounding vehicles, and outputs the vehicle acceleration or deceleration and the vehicle steering angle. By controlling the vehicle acceleration or deceleration and the vehicle steering angle, the microscopic action of the vehicle in the simulation can be controlled.

[0074] Furthermore, different vehicle control models can be used based on different application scenario characteristics, and the vehicle control models include: an aggressive driving control model and a normal driving control model. Among them, the aggressive driving control model is a control model with adjustable aggressiveness, which can adjust the vehicle characteristics within a range according to given parameters (i.e., the configuration data of the traffic participant) to make it meet the application scenario characteristics.

[0075] Specifically, when the application scenario feature is the aggressive driving of any vehicle in the scenario, the target type is determined to be a vehicle, and the driving aggressiveness of the vehicle is controlled according to the aggressive driving control model in the first simulation scenario to generate a second traffic flow simulation scenario. It is understandable that if the aggressiveness is higher, the vehicle controlled by the aggressive driving model is more likely to collide with the vehicle executing the first autonomous driving algorithm in situations such as following and changing lanes, thereby causing the vehicle executing the first autonomous driving algorithm to form a variety of driving behaviors to generate more simulation scenarios. When the application scenario feature is the normal driving of other vehicles in the scenario, the target type is determined to be a vehicle, and in the first simulation scenario, the normal driving control model that meets the high likelihood of driver behavior is directly used to control other vehicles in the scene to generate a second traffic flow simulation scenario.

[0076] In one implementation, the generating of the first traffic flow simulation scenario based on the type of the traffic participant, the configuration data of the traffic participant and the road environment includes:

[0077] determining a motion state of the traffic participant based on the configuration data of the traffic participant;

[0078] generating a traffic participant in a generation range of the traffic participant by utilizing the movement state of the traffic participant according to the type of the traffic participant;

[0079] The first traffic flow simulation scenario is generated using the road environment and the traffic participants.

[0080] Exemplarily, the movement state of a traffic participant includes: location information of the traffic participant (such as longitude and latitude in a map), speed information (such as starting speed, etc.), and movement direction (such as heading angle, etc.).

[0081] Specifically, the simulation system is started, and the vehicle driving environment data is used as a parameter of the calling interface (Application Programming Interface, API) to generate a road environment. The configuration data of the traffic participants is input into a preset AI generation model (such as variational autoencoders, VAE), and the motion state of the traffic participants (such as vehicle position, starting speed, heading angle, number of pedestrians, and other information) is output. The motion state of the traffic participants is used as a calling interface parameter to generate the traffic participants, thereby combining the road environment with the traffic participants to obtain the first traffic flow simulation scene. In this way, there is no need to manually write scripts to generate test scenario use cases, which improves the efficiency of writing test scenario use cases and makes the test scenario use cases diverse.

[0082] In this embodiment, the configuration data of traffic participants can be divided into the configuration data of pedestrians and the configuration data of vehicles. The corresponding variational autoencoders are trained for the configuration data of pedestrians and the configuration data of vehicles, respectively, so that the motion state of pedestrians and the motion state of vehicles can be output respectively. Furthermore, different models are trained for the configuration data of vehicles corresponding to different vehicle generation ranges. Optionally, if the vehicle generation range is determined in the road environment by specifying a radius, the vehicle generation model can be trained according to the distance between vehicles and general configuration data; if the vehicle generation range is determined in the road environment by specifying a path, the vehicle generation model can be trained according to the total number of vehicles or vehicle flow or vehicle density and general configuration data; if the global is used as the vehicle generation range, the vehicle generation model can be trained according to the general configuration data. For example, the model can use a variational autoencoder to determine the vehicle generation range for a specified radius, train the variational autoencoder according to the vehicle spacing samples and general configuration data samples, and obtain a first vehicle generation model.

[0083] In this way, after determining the determination method of the vehicle generation range, the corresponding configuration data is obtained according to the determination method. The configuration data is input into the corresponding vehicle generation model, and the motion state of the vehicle (such as vehicle position, starting speed, heading angle, etc.) is output. The motion state of the vehicle and the motion state of the pedestrian are used as calling interface parameters to generate vehicles and pedestrians (i.e., traffic participants), so as to combine the road environment with the traffic participants to obtain the first traffic flow simulation scene.

[0084] In one embodiment, the method further comprises:

[0085] According to the received execution instruction of the test mode, a first vehicle is selected in the second traffic flow simulation scenario to execute the first autonomous driving algorithm, and the first vehicle is tested in the second traffic flow simulation scenario to obtain a first autonomous driving test result.

[0086] Exemplarily, the test mode includes an open-loop test mode and a closed-loop test mode. Optionally, after the traffic flow simulation scene is generated, a test mode option is sent to the user for the user to select. According to the test mode selected by the user, an execution instruction of the corresponding test mode is generated to test the vehicle executing the autonomous driving algorithm in the generated traffic flow simulation scene.

[0087] Exemplarily, the first vehicle may be a vehicle specified in the first traffic flow simulation scenario, or may be a vehicle specified in the second traffic flow simulation scenario. Optionally, if the first vehicle is a vehicle specified in the first traffic flow simulation scenario, then the behavior of vehicles other than the first vehicle in the first traffic flow simulation scenario and the second traffic flow simulation scenario may be controlled. If the first vehicle is a vehicle specified in the second traffic flow simulation scenario, then the behavior of all vehicles in the first traffic flow simulation scenario may be controlled. The autonomous driving test result indicates the situation of the first autonomous driving algorithm in the second traffic flow simulation scenario, such as driving time, whether a collision occurred, the cause of the collision, driving speed, driving route, etc.

[0088] Specifically, after receiving the execution instruction of the test mode, the first vehicle deletes the motion track in the second traffic flow simulation scene and executes the first automatic driving algorithm. Then, the first vehicle performs automatic driving in the second traffic flow simulation scene until the simulation end condition is completed, and the test ends and the first automatic driving test result is output. Among them, the simulation end condition of the first vehicle may include that the driving time reaches a preset driving time threshold or the driving path meets the requirements of reaching the second location from the first location. Among them, the driving time threshold can be set according to actual needs and is not limited here.

[0089] For example, Figure 3 As shown, the user starts the traffic flow function, inputs the configuration data, map information and road network information of the traffic participants, and generates a first traffic flow simulation scene based on the above data. The first vehicle can perform an autonomous driving test in the first traffic flow simulation scene, and outputs the initialization autonomous driving test result when the simulation end condition is met. If the simulation end condition is not met, the next frame of the traffic flow simulation scene continues to be output.

[0090] When the user needs a different test case scenario, the behavior of the traffic participants in the first traffic flow simulation scenario can be controlled to generate a second traffic flow simulation scenario. The first vehicle can perform an autonomous driving test in the second traffic flow simulation scenario, and output the first autonomous driving test result if the simulation end condition is met. If the simulation end condition is not met, the next frame of the traffic flow simulation scenario continues to be output.

[0091] In one embodiment, the method further comprises:

[0092] Selecting a corresponding traffic flow playback scene using a second autonomous driving algorithm; wherein the second autonomous driving algorithm is obtained by optimizing the first autonomous driving algorithm; and the traffic flow playback scene is obtained by reconstructing a second traffic flow simulation scene corresponding to the first autonomous driving algorithm;

[0093] According to the received execution instruction of the test mode, a second vehicle is selected in the traffic flow playback scenario to execute the second autonomous driving algorithm, and the second vehicle is tested in the traffic flow playback scenario to obtain a second autonomous driving test result.

[0094] Exemplarily, the traffic flow playback scene is obtained by reproducing the poses of all traffic participants in the second traffic flow simulation scene corresponding to the first autonomous driving algorithm frame by frame. That is, the traffic flow playback scene is the same as the second traffic flow simulation scene corresponding to the first autonomous driving algorithm. The second vehicle is any vehicle, or a specified vehicle. For example, the second vehicle is the same as the vehicle executing the first autonomous driving algorithm (i.e., the first vehicle).

[0095] Specifically, obtain the second autonomous driving algorithm and its corresponding first autonomous driving algorithm, determine the corresponding second traffic flow simulation scene according to the first autonomous driving algorithm, and replay the scene (LogSim) for all traffic participants in the second traffic flow simulation scene to obtain the traffic flow playback scene. Then select the second vehicle in the traffic flow playback scene and delete the motion trajectory of the second vehicle. Make the second vehicle execute the second autonomous driving algorithm, and test the second vehicle in the traffic flow playback scene until the simulation end condition is met, and output the second autonomous driving test result. In this way, the same simulation scene is used to repeatedly test the autonomous driving algorithms before and after optimization, so that problems of the autonomous driving algorithm can be discovered, and the improvement effect of the optimized autonomous driving algorithm can be detected.

[0096] Preferably, the traffic flow playback scene is obtained by reconstructing the second traffic flow simulation scene corresponding to the first autonomous driving algorithm, including:

[0097] Using the second autonomous driving algorithm, selecting corresponding scene recording data in a preset database; wherein the scene recording data represents data obtained by recording a second traffic flow simulation scene corresponding to the first autonomous driving algorithm;

[0098] The scene recording data is reconstructed to generate the traffic flow playback scene.

[0099] Exemplarily, the preset database is used to save all scene recording data. For example, the preset database can be expressed as a playback log. The scene recording data includes individual data of traffic participants, behavioral data of traffic participants, and road environment data. Optionally, individual data of traffic participants include: length, width, type, color and other information of all traffic participants. The behavioral data of participants include: three-dimensional position coordinates, three-dimensional angle coordinates, three-dimensional speed, three-dimensional acceleration and other information of all traffic participants in each frame. Road environment data includes map information, road network information and other information. Optionally, the scene recording data can be obtained by recording the first traffic flow simulation scene, or by recording the second traffic flow simulation scene. In addition, if there are modifications to the playback scene, the scene recording data can be recorded and output for the modified playback scene, and saved to the database. For example, if Figure 4 As shown, the scene recording function is turned on. Before the simulation end conditions are met, the traffic flow simulation scene is recorded and saved frame by frame until the simulation ends.

[0100] Specifically, a second autonomous driving algorithm and its corresponding first autonomous driving algorithm are obtained, and a corresponding second traffic flow simulation scene is determined according to the first autonomous driving algorithm. Scene recording data corresponding to the second traffic flow simulation scene is searched in a database. The pose of each frame of the scene recording data is reproduced to generate a traffic flow playback scene. In this way, the same scene can be used to test the autonomous driving algorithms before and after optimization, so that the improvement of the optimized autonomous driving algorithm can be detected.

[0101] In one implementation, the step of selecting a second vehicle in the traffic flow playback scenario to execute the second autonomous driving algorithm according to the received execution instruction of the test mode, and testing the second vehicle in the traffic flow playback scenario to obtain a second autonomous driving test result includes:

[0102] deleting the movement track of the second vehicle in the traffic flow playback scene;

[0103] When the second vehicle executes the second automatic driving algorithm, predicting a motion trajectory of the second vehicle to obtain a predicted motion trajectory of the second vehicle;

[0104] In response to the execution instructions of the test mode, a second autonomous driving test result is determined using the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene.

[0105] Exemplarily, the playback motion trajectory of the second vehicle represents the original motion trajectory of the second vehicle in the traffic flow playback scene. The test mode includes an open-loop test mode and a closed-loop test mode.

[0106] Specifically, the replayed motion trajectory of the second vehicle is deleted in the traffic flow playback scene. The optimized autonomous driving algorithm is executed in the second vehicle. The trajectory of the second vehicle is predicted using the obstacle information and map information (such as vehicles and pedestrians) around the second vehicle to obtain a predicted motion trajectory. For example, a pre-trained trajectory prediction model is used to predict the obstacles and map information around the second vehicle, and the predicted motion trajectory is output; wherein the pre-trained trajectory prediction model may be obtained by using a neural network and training the neural network based on obstacle sample information and map sample information. The second driving test result corresponding to the optimized autonomous driving algorithm is then determined based on the predicted motion trajectory and the replayed motion trajectories of other traffic participants in the traffic flow playback scene.

[0107] Preferably, the execution instruction for the test mode performs a simulation test on the second vehicle based on the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene to obtain a second autonomous driving test result, including:

[0108] In the case where the execution instruction of the test mode is an execution instruction of a closed-loop test mode, if there is a motion trajectory intersection between the second vehicle and the other traffic participants, adjusting the playback motion trajectory of the other traffic participants to obtain a third traffic flow simulation scene;

[0109] The second vehicle is controlled to perform a simulation test in the third traffic flow simulation scenario until a preset simulation end condition is met, and a second autonomous driving test result is output.

[0110] For example, the intersection of the motion trajectories of the second vehicle and other traffic participants indicates that there is a risk of collision between the second vehicle and other traffic participants. Figure 5 As shown, the user starts the scene playback (LogSim). In the closed-loop test mode, if the second vehicle has a collision risk with other traffic participants, the ID of the traffic participant with a collision risk is read and taken over. At the same time, the playback motion trajectory of the traffic participant with a collision risk is controlled according to the vehicle control model to prevent it from colliding with the second vehicle and other vehicles. It should be noted that there may be multiple vehicles that have a collision risk with the second vehicle. At this time, the trajectory of multiple vehicles that have a collision risk with the second vehicle can be adjusted at the same time. It is also possible to adjust the trajectory of multiple vehicles with the greatest collision risk with the second vehicle, which is not limited here. In this way, the occurrence of vehicle collisions in non-realistic situations during the test can be reduced, and the accuracy of the test can be improved.

[0111] Then, continue to read the next frame of scene recording data and replay the scene to perform simulation testing on the second vehicle in the third traffic flow simulation scene. At the same time, determine whether there is a risk of collision between the second vehicle and other traffic participants and whether vehicle takeover is required for each frame of scene recording data, until the preset simulation end conditions are met and the second autonomous driving test result is output.

[0112] Preferably, for a preset test mode, determining a second autonomous driving test result using the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene includes:

[0113] like Figure 6 As shown, when the preset test mode is the open-loop test mode, if there is an intersection of the motion trajectories of the second vehicle and other traffic participants in the traffic flow playback scene, the collision situation will be recorded as collision information that does not require reference. Then continue to read the scene recording data to test the second vehicle until the simulation end condition is met, and use the information recorded during the simulation as the second autonomous driving test result. It can be understood that in the open-loop test mode, other traffic participants in the traffic flow playback scene do not have the perception ability, so they cannot perceive the movement of the second vehicle (such as changing lanes, overtaking, etc.), and other traffic participants in the traffic flow playback scene still move according to the trajectory in the traffic flow playback scene. At this time, the collision between other traffic participants in the traffic flow playback scene and the second vehicle is a passive collision (i.e., a collision that is not in a real situation), not a collision actively caused by the second vehicle, so there is no need to consider it.

[0114] In one embodiment, the method further comprises:

[0115] The second traffic flow simulation scenario is evaluated based on the similarity between the second traffic flow simulation scenario and a preset real traffic flow scenario, the uniform speed driving time of the first vehicle, and the driving danger state of the first vehicle to obtain an evaluation result.

[0116] For example, the preset real traffic flow scene refers to the traffic flow scene collected from the road. The driving danger state refers to the situation where the vehicle collides.

[0117] Specifically, the similarity between the second traffic flow simulation scene and the preset real traffic flow scene includes: behavior decision likelihood, motion performance naturalness, average displacement error, and end point offset error. Among them, the behavior decision likelihood is to calculate the proportion of the number of time slices in which the output behavior of the scene generation model and the actual behavior are consistent with the total number of time slices for a certain traffic participant on a road sampling trajectory of a duration of T. The behavior can be left lane change, right lane change, acceleration, deceleration, constant speed, etc. The naturalness of motion performance is to quantitatively compare the difference between the motion characteristics generated by the scene generation model and the distribution of the motion behavior characteristics of the real driver through Kullback-Leibler divergence. The average displacement error is the average offset of the vehicle position controlled by the scene generation model and the road sampling data position at each moment. The end point offset error is the offset of the vehicle position controlled by the scene generation model and the road sampling data position at the end point. Driving danger states include: collision rate, number of lane change deceleration times, and number of acceleration times in high-density traffic flow; among them, the collision rate is the ratio of the number of vehicle pairs that collide with each other to the total number of vehicles among traffic participants, which measures the simulation reactivity and the interaction ability between vehicles. The number of lane change decelerations is the number of times car A brakes before changing lanes to car B. The number of accelerations in high-density traffic flow is that in high-density traffic flow, most vehicles follow the vehicle in front. If they are considered to maintain the original speed within a certain range, they should maintain the original speed most of the time. For vehicles with many accelerations, their acceleration times are recorded. In addition, vehicles with many accelerations are more dangerous and are more likely to collide with the vehicle in front. The comfortable speed ratio is the ratio of the vehicle's uniform driving time to the total scene time. It is understandable that uniform driving makes users feel comfortable, so the comfortable speed time is the uniform driving time. The above indicators are recorded during the simulation test, and the second traffic flow simulation scene can be evaluated according to the above indicators in an offline state to obtain an evaluation result. For example, a corresponding score is set for each indicator, and the total score of the second traffic flow simulation scene is obtained as the evaluation result according to the score of each indicator. It should be noted that each traffic flow simulation scene can be evaluated once, and each time the traffic flow simulation scene is updated, the updated traffic flow simulation scene is evaluated once.

[0118] Exemplary Devices

[0119] Accordingly, Figure 7 1 is a schematic diagram of the structure of an automatic driving traffic flow simulation device according to an embodiment of the present application. In an exemplary embodiment, an automatic driving traffic flow simulation device is provided, comprising:

[0120] A generation module 610, used to generate a road environment based on the acquired map information and road network information;

[0121] A selection module 620, for determining a generation range of traffic participants in the road environment and configuration data of traffic participants corresponding to the generation range of traffic participants according to the type of traffic participants;

[0122] The simulation module 630 is used to generate a first traffic flow simulation scenario based on the type of the traffic participant, the configuration data of the traffic participant and the road environment; wherein the configuration data of the traffic participant is at least one of the vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the traffic simulation scenario.

[0123] In one embodiment, the device also includes: a control module, which is used to determine the target type of the traffic participant to be controlled based on the application scenario characteristics of the first autonomous driving algorithm, and to perform behavior control on the traffic participants of the target type in the first traffic flow simulation scene to generate a second traffic flow simulation scene.

[0124] In one embodiment, the simulation module 630 is further configured to:

[0125] determining a motion state of the traffic participant based on the configuration data of the traffic participant;

[0126] generating a traffic participant in a generation range of the traffic participant by utilizing the movement state of the traffic participant according to the type of the traffic participant;

[0127] The first traffic flow simulation scenario is generated using the road environment and the traffic participants.

[0128] In one implementation, when the type of the traffic participant is a vehicle, the selection module 620 is further configured to:

[0129] Selecting at least one path in the road environment to be determined as a generation range of the traffic participant;

[0130] or, generating a target range in the road environment with a preset vehicle spacing as a radius, and determining the target range as a generated range of the traffic participant;

[0131] Or, the road environment is determined as a generation range of the traffic participant.

[0132] In one embodiment, the device further comprises a testing module, further configured to:

[0133] According to the received execution instruction of the test mode, a first vehicle is selected in the second traffic flow simulation scenario to execute the first autonomous driving algorithm, and the first vehicle is tested in the second traffic flow simulation scenario to obtain a first autonomous driving test result.

[0134] In one embodiment, the device further includes a scene playback module, further configured to:

[0135] Selecting a corresponding traffic flow playback scene using a second autonomous driving algorithm; wherein the second autonomous driving algorithm is obtained by optimizing the first autonomous driving algorithm; and the traffic flow playback scene is obtained by reconstructing a second traffic flow simulation scene corresponding to the first autonomous driving algorithm;

[0136] According to the received execution instruction of the test mode, a second vehicle is selected in the traffic flow playback scenario to execute the second autonomous driving algorithm, and the second vehicle is tested in the traffic flow playback scenario to obtain a second autonomous driving test result.

[0137] In one implementation, the traffic flow playback scene is obtained by reconstructing a second traffic flow simulation scene corresponding to the first autonomous driving algorithm, including:

[0138] Using the second autonomous driving algorithm, selecting corresponding scene recording data in a preset database; wherein the scene recording data represents data obtained by recording a second traffic flow simulation scene corresponding to the first autonomous driving algorithm;

[0139] The scene recording data is reconstructed to generate the traffic flow playback scene.

[0140] In one implementation, the step of selecting a second vehicle in the traffic flow playback scenario to execute the second autonomous driving algorithm according to the received execution instruction of the test mode, and testing the second vehicle in the traffic flow playback scenario to obtain a second autonomous driving test result includes:

[0141] deleting the movement track of the second vehicle in the traffic flow playback scene;

[0142] When the second vehicle executes the second automatic driving algorithm, predicting a motion trajectory of the second vehicle to obtain a predicted motion trajectory of the second vehicle;

[0143] In response to the execution instructions of the test mode, a second autonomous driving test result is determined using the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene.

[0144] In one embodiment, the execution instruction for the test mode performs a simulation test on the second vehicle based on the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene to obtain a second autonomous driving test result, including:

[0145] In the case where the execution instruction of the test mode is an execution instruction of a closed-loop test mode, if there is a motion trajectory intersection between the second vehicle and the other traffic participants, adjusting the playback motion trajectory of the other traffic participants to obtain a third traffic flow simulation scene;

[0146] The second vehicle is controlled to perform a simulation test in the third traffic flow simulation scenario until a preset simulation end condition is met, and a second autonomous driving test result is output.

[0147] In one embodiment, the device also includes: an evaluation module, which is also used to evaluate the second traffic flow simulation scene based on the similarity between the second traffic flow simulation scene and the preset real traffic flow scene, the constant speed driving time of the first vehicle, and the driving hazard state of the first vehicle to obtain an evaluation result.

[0148] The autonomous driving traffic flow simulation device provided in this embodiment belongs to the same application concept as the autonomous driving traffic flow simulation method provided in the above embodiments of this application, and can execute the autonomous driving traffic flow simulation method provided in any of the above embodiments of this application, and has the corresponding functional modules and beneficial effects of executing the autonomous driving traffic flow simulation method. For technical details not fully described in this embodiment, please refer to the specific processing content of the autonomous driving traffic flow simulation method provided in the above embodiments of this application, which will not be repeated here.

[0149] Exemplary Systems

[0150] Another embodiment of the present application also provides an electronic device, see Figure 8 As shown, the device includes:

[0151] Memory 700 and processor 710;

[0152] The memory 700 is connected to the processor 710 and is used to store programs;

[0153] The processor 710 is used to implement the autonomous driving traffic flow simulation method disclosed in any of the above embodiments by running the program stored in the memory 700.

[0154] Specifically, the electronic device may further include: a bus, a communication interface 720 , an input device 730 and an output device 740 .

[0155] The processor 710, the memory 700, the communication interface 720, the input device 730 and the output device 740 are connected to each other via a bus.

[0156] A bus may include a pathway that transfers information between components of a computer system.

[0157] Processor 710 may be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the program of the scheme of the present invention. It may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0158] The processor 710 may include a main processor, and may also include a baseband chip, a modem, and the like.

[0159] The memory 700 stores a program for executing the technical solution of the present invention, and may also store an operating system and other key services. Specifically, the program may include a program code, and the program code includes a computer operation instruction. More specifically, the memory 700 may include a read-only memory (ROM), other types of static storage devices that can store static information and instructions, a random access memory (RAM), other types of dynamic storage devices that can store information and instructions, a disk storage, a flash, and the like.

[0160] The input device 730 may include a device for receiving data and information input by a user, such as a keyboard, a mouse, a camera, a scanner, a light pen, a voice input device, a touch screen, a pedometer, or a gravity sensor.

[0161] Output device 740 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0162] The communication interface 720 may include any transceiver or the like to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0163] The processor 710 executes the program stored in the memory 700 and calls other devices, which can be used to implement the various steps of any autonomous driving traffic flow simulation method provided in the above embodiments of the present application.

[0164] Exemplary computer program products and storage media

[0165] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the autonomous driving traffic flow simulation method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.

[0166] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0167] In addition, an embodiment of the present application may also be a storage medium on which a computer program is stored, and the computer program is executed by a processor to execute the steps of the autonomous driving traffic flow simulation method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification. The specific working content of the above-mentioned electronic device, as well as the specific working content of the above-mentioned computer program product and the computer program on the storage medium when being executed by the processor, can all be referred to the contents of the above-mentioned method embodiment, and will not be repeated here.

[0168] For the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0169] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0170] The steps in the methods of each embodiment of the present application can be adjusted in order, combined and deleted according to actual needs, and the technical features recorded in each embodiment can be replaced or combined.

[0171] The modules and sub-modules in the devices and terminals of the various embodiments of the present application can be combined, divided and deleted according to actual needs.

[0172] In the several embodiments provided in the present application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the terminal embodiments described above are only schematic, for example, the division of modules or submodules is only a logical function division, and there may be other division methods in actual implementation, for example, multiple submodules or modules can be combined or integrated into another module, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0173] The modules or submodules described as separate components may or may not be physically separated, and the components of the modules or submodules may or may not be physical modules or submodules, that is, they may be located in one place, or they may be distributed on multiple network modules or submodules. Some or all of the modules or submodules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0174] In addition, each functional module or submodule in each embodiment of the present application may be integrated into one processing module, or each module or submodule may exist physically separately, or two or more modules or submodules may be integrated into one module. The above-mentioned integrated modules or submodules may be implemented in the form of hardware or in the form of software functional modules or submodules.

[0175] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0176] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly by hardware, software units executed by a processor, or a combination of the two. The software units may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0177] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0178] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for simulating traffic flow for autonomous driving, characterized in that: include: Generate a road environment based on the acquired map information and road network information; Determining, for the type of traffic participant, a generation range of the traffic participant in the road environment, and configuration data of the traffic participant corresponding to the generation range of the traffic participant; Generate a first traffic flow simulation scenario based on the type of the traffic participant, the configuration data of the traffic participant and the road environment, wherein the configuration data of the traffic participant is at least one of a vehicle configuration parameter, a map configuration parameter, a behavior control configuration parameter and a traffic rule configuration parameter required to generate the first traffic flow simulation scenario; Based on the application scenario characteristics of the first autonomous driving algorithm, determine the target type of the traffic participant to be controlled, and perform behavior control on the traffic participant of the target type in the first traffic flow simulation scenario to generate a second traffic flow simulation scenario; According to the received execution instruction of the test mode, a first vehicle is selected in the second traffic flow simulation scenario to execute the first autonomous driving algorithm, and the first vehicle is tested in the second traffic flow simulation scenario to obtain a first autonomous driving test result.

2. The method according to claim 1, characterized in that The generating a first traffic flow simulation scenario based on the type of the traffic participant, the configuration data of the traffic participant and the road environment includes: determining a motion state of the traffic participant based on the configuration data of the traffic participant; generating a traffic participant in a generation range of the traffic participant by utilizing the movement state of the traffic participant according to the type of the traffic participant; The first traffic flow simulation scenario is generated using the road environment and the traffic participants.

3. The method according to claim 1, characterized in that In the case where the type of the traffic participant is a vehicle, determining a generation range of the traffic participant in the road environment for the type of the traffic participant includes: Selecting at least one path in the road environment to be determined as a generation range of the traffic participant; or, generating a target range in the road environment with a preset vehicle spacing as a radius, and determining the target range as a generated range of the traffic participant; Or, the road environment is determined as a generation range of the traffic participant.

4. The method according to claim 1, characterized in that: The method further comprises: Selecting a corresponding traffic flow playback scene using a second autonomous driving algorithm; wherein the second autonomous driving algorithm is obtained by optimizing the first autonomous driving algorithm; and the traffic flow playback scene is obtained by reconstructing a second traffic flow simulation scene corresponding to the first autonomous driving algorithm; According to the received execution instruction of the test mode, a second vehicle is selected in the traffic flow playback scenario to execute the second autonomous driving algorithm, and the second vehicle is tested in the traffic flow playback scenario to obtain a second autonomous driving test result.

5. The method according to claim 4, characterized in that The traffic flow playback scene is obtained by reconstructing the second traffic flow simulation scene corresponding to the first autonomous driving algorithm, including: Using the second autonomous driving algorithm, selecting corresponding scene recording data in a preset database; wherein the scene recording data represents data obtained by recording a second traffic flow simulation scene corresponding to the first autonomous driving algorithm; The scene recording data is reconstructed to generate the traffic flow playback scene.

6. The method according to claim 4, characterized in that The step of selecting a second vehicle in the traffic flow playback scene to execute the second autonomous driving algorithm according to the received execution instruction of the test mode, and testing the second vehicle in the traffic flow playback scene to obtain a second autonomous driving test result includes: deleting the movement track of the second vehicle in the traffic flow playback scene; When the second vehicle executes the second automatic driving algorithm, predicting a motion trajectory of the second vehicle to obtain a predicted motion trajectory of the second vehicle; In response to the execution instructions of the test mode, a second autonomous driving test result is determined using the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene.

7. The method according to claim 6, characterized in that The execution instruction for the test mode performs a simulation test on the second vehicle based on the predicted motion trajectory of the second vehicle and the replayed motion trajectories of other traffic participants in the traffic flow replay scene to obtain a second automatic driving test result, including: In the case where the execution instruction of the test mode is an execution instruction of a closed-loop test mode, if there is a motion trajectory intersection between the second vehicle and the other traffic participants, adjusting the playback motion trajectory of the other traffic participants to obtain a third traffic flow simulation scene; The second vehicle is controlled to perform a simulation test in the third traffic flow simulation scenario until a preset simulation end condition is met, and a second autonomous driving test result is output.

8. The method according to claim 1, characterized in that The method further comprises: The second traffic flow simulation scenario is evaluated based on the similarity between the second traffic flow simulation scenario and a preset real traffic flow scenario, the uniform speed driving time of the first vehicle, and the driving danger state of the first vehicle to obtain an evaluation result.

9. A traffic flow simulation system based on an autonomous driving scenario, characterized in that: include: A traffic flow generation module, used to generate a road environment based on the acquired map information and road network information; for the type of traffic participants, determine the generation range of traffic participants in the road environment, and the configuration data of traffic participants corresponding to the generation range of traffic participants; based on the type of traffic participants, the configuration data of traffic participants and the road environment, generate a first traffic flow simulation scene; wherein the configuration data of traffic participants is at least one of vehicle configuration parameters, map configuration parameters, behavior control configuration parameters and traffic regulations configuration parameters required to generate the first traffic flow simulation scene; a traffic flow control module, configured to determine the target type of the traffic participant to be controlled based on the application scenario characteristics of the first autonomous driving algorithm, and to perform behavior control on the traffic participant of the target type in the first traffic flow simulation scenario to generate a second traffic flow simulation scenario; It is also used to adjust the playback motion trajectory of other traffic participants when receiving the execution instruction of the closed-loop test if there is a motion trajectory intersection between the vehicle executing the first autonomous driving algorithm and other traffic participants; A scene recording module, used for recording the first traffic flow simulation scene and the second traffic flow simulation scene to generate scene recording data; A scene playback module, used to reconstruct the scene recording data in the scene recording module to obtain a traffic flow playback scene; A collision detection module, used to detect the intersection of the motion trajectories of the vehicle executing the first autonomous driving algorithm and other traffic participants in the traffic flow playback scene; An evaluation module is used to evaluate the second traffic flow simulation scenario based on the similarity between the second traffic flow simulation scenario and a preset real traffic flow scenario, the uniform speed driving time of the vehicle executing the first automatic driving algorithm, and the driving hazard state of the vehicle executing the first automatic driving algorithm to obtain an evaluation result.

10. An electronic device, characterized in that: include: Control devices; The control device comprises: a memory and a processor; The memory is connected to the processor and is used to store programs; The processor implements the autonomous driving traffic flow simulation method as claimed in any one of claims 1 to 8 by running the program in the memory.

11. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the method for simulating traffic flow of autonomous driving as claimed in any one of claims 1 to 8 is implemented.

Images