Driving simulation scene generation method and device, equipment and storage medium

Abstract

The invention discloses a driving simulation scene generation method and device, equipment and a storage medium, and relates to the technical field of automatic driving, and the method comprises the steps: obtaining driving data, and determining a driving track of a current vehicle according to the driving data; dividing the driving track into a plurality of track fragments based on the road map information, and determining a driving behavior corresponding to each track fragment according to a preset driving behavior judgment condition; initializing a simulation scene and scene parameters according to the road map information and the driving behaviors corresponding to the track fragments; and updating the scene parameters according to a preset generalization mode, and generating a target driving simulation scene based on the simulation scene according to an updating result. According to the method, the driving track is divided into the track fragments, and the high-level driving behavior is captured based on the track fragments, so that the scene parameters of the simulation scene generated based on the driving behavior are updated according to the preset generalization mode, and the target driving simulation scene with higher adaptability and authenticity can be obtained.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a driving simulation scene generation method, device, equipment and storage medium. Background Art

[0002] With the rapid development of intelligent driving technology, simulation technology has become crucial in the development and testing of intelligent driving systems. Intelligent driving algorithms and systems can simulate real driving scenarios in a virtual environment through simulation technology to verify vehicle performance, providing a safe and efficient testing platform for intelligent driving systems.

[0003] Driving simulation scenarios can be obtained by reconstructing and generalizing real vehicle measurement data in real driving scenarios. However, since real vehicle measurement data is limited by a limited number of test vehicles, and the existing scenario reconstruction method only generalizes from the perspective of vehicle trajectory, it does not capture the high-level driving behavior of the vehicle, resulting in the lack of deep generalization of the driving behavior level in the obtained driving simulation scenarios, which in turn affects the adaptability and authenticity of the driving simulation scenarios. Summary of the invention

[0004] The main purpose of this application is to provide a driving simulation scene generation method, device, equipment and storage medium, aiming to solve the technical problem in the prior art that the driving simulation scene reconstructed according to the vehicle trajectory in the real vehicle measurement data lacks deep generalization at the driving behavior level, resulting in poor adaptability and authenticity of the driving simulation scene.

[0005] To achieve the above objectives, the present application proposes a driving simulation scene generation method, the driving simulation scene generation method comprising:

[0006] Acquiring driving data, and determining a current driving trajectory of the vehicle based on the driving data;

[0007] Dividing the driving trajectory into a plurality of trajectory segments based on the road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition;

[0008] Initialize simulation scenes and scene parameters according to the road map information and the driving behaviors corresponding to each of the trajectory segments;

[0009] The scenario parameters are updated in a preset generalized manner, and a target driving simulation scenario is generated based on the simulation scenario according to the update result.

[0010] In one embodiment, the current vehicle includes: a focused vehicle and a target vehicle, and the step of acquiring driving data and determining the driving trajectory of the current vehicle according to the driving data includes:

[0011] Acquire driving data of the vehicle of interest in the shadow mode through a vehicle sensor, and input the driving data into a preset target detection model to determine the target vehicle;

[0012] Based on the driving data, determining the trajectory data corresponding to the target vehicle according to a preset target tracking algorithm;

[0013] The driving trajectories of the vehicle of interest and the target vehicle are generated according to the trajectory data.

[0014] In one embodiment, the step of generating the driving trajectories of the vehicle of interest and the target vehicle according to the trajectory data includes:

[0015] Acquire the driving trajectory of the vehicle of interest according to the driving data, and determine the relative driving trajectory corresponding to the target vehicle according to the trajectory data;

[0016] The relative driving trajectory is matched with the driving trajectory of the vehicle of interest according to the absolute position and timestamp of the vehicle of interest, and the driving trajectory of the target vehicle is determined according to the matching result.

[0017] In one embodiment, the step of dividing the driving trajectory into a plurality of trajectory segments based on the road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition includes:

[0018] Determining road information and lane information related to the driving trajectory according to the road map information, and determining the current driving speed of the vehicle according to the road information and lane information;

[0019] Splitting the driving trajectory into a plurality of trajectory segments by combining the road information, the lane information and the driving speed;

[0020] The driving behavior corresponding to each of the trajectory segments is determined by presetting the driving behavior determination condition.

[0021] In one embodiment, the preset driving behavior determination conditions include: a lateral behavior determination condition, a longitudinal behavior determination condition, and a trajectory behavior determination condition; the step of determining the driving behavior corresponding to each trajectory segment by the preset driving behavior determination conditions includes:

[0022] Acquire trajectory data corresponding to each of the trajectory segments;

[0023] Determining whether the trajectory data satisfies the lateral behavior determination condition and / or the longitudinal behavior determination condition;

[0024] If yes, determining the driving behavior corresponding to each of the trajectory segments according to the judgment result;

[0025] If not, the trajectory data is converted into a driving behavior of traveling along a preset fixed point according to the trajectory behavior determination condition, and is determined as the driving behavior corresponding to the trajectory segment.

[0026] In one embodiment, the step of initializing the simulation scene and scene parameters according to the road map information and the driving behavior corresponding to each of the trajectory segments includes:

[0027] Initialize a static road structure of a simulation scene according to the road map information and each of the trajectory segments;

[0028] Initialize a scene entity of a simulation scene based on the static road structure and in combination with the current vehicle;

[0029] The driving behaviors corresponding to the trajectory segments are associated with the scene entities, and an initial simulation scene file and scene parameters are generated according to the association results.

[0030] In one embodiment, the step of updating the scenario parameters in a preset generalization manner and generating a target driving simulation scenario based on the simulation scenario according to the update result includes:

[0031] Determining a preset generalization method according to the current generalization requirement, wherein the preset generalization method includes: at least one of a road parameter adjustment method, an environment parameter adjustment method, and a vehicle state parameter adjustment method;

[0032] Determining a scene parameter to be adjusted from each of the scene parameters according to the preset generalization method, and updating the scene parameter to be adjusted, wherein the scene parameter includes: a road parameter, an environmental parameter, and a vehicle state parameter;

[0033] The initial simulation scene file is updated according to the scene parameter update result to obtain the target driving simulation scene.

[0034] In addition, to achieve the above-mentioned purpose, the present application also proposes a driving simulation scene generation device, the device comprising:

[0035] A driving trajectory generation module, used to obtain driving data and determine the current vehicle driving trajectory according to the driving data;

[0036] A driving behavior extraction module, used to divide the driving trajectory into a plurality of trajectory segments based on road map information, and determine the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition;

[0037] A simulation scene reconstruction module, used to initialize the simulation scene and scene parameters according to the road map information and the driving behavior corresponding to each of the trajectory segments;

[0038] The simulation scenario generalization module is used to update the scenario parameters according to a preset generalization method, and generate a target driving simulation scenario based on the simulation scenario according to the update result.

[0039] In addition, to achieve the above-mentioned purpose, the present invention also proposes a driving simulation scene generation device, which includes: a memory, a processor, and a driving simulation scene generation program stored in the memory and executable on the processor, and the driving simulation scene generation program is configured to implement the steps of the driving simulation scene generation method described above.

[0040] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which a driving simulation scenario generation program is stored, and when the driving simulation scenario generation program is executed by a processor, the steps of the driving simulation scenario generation method described above are implemented.

[0041] The present application proposes a method for generating a driving simulation scene, including: obtaining driving data, and determining the driving trajectory of the current vehicle based on the driving data; dividing the driving trajectory into a number of trajectory segments based on road map information, and determining the driving behavior corresponding to each trajectory segment based on preset driving behavior judgment conditions; initializing the simulation scene and scene parameters based on the road map information and the driving behavior corresponding to each trajectory segment; updating the scene parameters in a preset generalization manner, and generating a target driving simulation scene based on the simulation scene according to the update result. Since the present application can divide the driving trajectory of the current vehicle into trajectory segments based on the road map information, it can capture the high-level driving behavior of the vehicle based on the trajectory segments. Compared with the existing limitation of only generalizing the vehicle trajectory, the scene parameters of the simulation scene generated based on the driving behavior can be updated in a preset generalization manner, which is conducive to obtaining a target driving simulation scene with higher adaptability and authenticity. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0043] 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0044] Figure 1 This is a flow chart of the first embodiment of the driving simulation scene generation method of the present application;

[0045] Figure 2This is a flow chart of the second embodiment of the driving simulation scene generation method of the present application;

[0046] Figure 3 A schematic diagram of the complete trajectory generation process for the target vehicle;

[0047] Figure 4 This is a flow chart of the third embodiment of the driving simulation scene generation method of the present application;

[0048] Figure 5 Reconstruct the comparison map of the required source scene for dynamic scenes;

[0049] Figure 6 This is a schematic diagram of the driving behavior extraction process in the third embodiment of the driving simulation scene generation method of the present application;

[0050] Figure 7 This is a flow chart of a fourth embodiment of the driving simulation scene generation method of the present application;

[0051] Figure 8 This is a schematic diagram of simulation scene reconstruction in the fourth embodiment of the driving simulation scene generation method of the present application;

[0052] Fig. 9 This is an example diagram of a simulation scene in the fourth embodiment of the driving simulation scene generation method of the present application;

[0053] Fig.10 A schematic diagram of the entire process of the driving simulation scenario generation method of this application;

[0054] Fig.11 This is a schematic diagram of the module structure of the first embodiment of the driving simulation scene generation device of the present application;

[0055] Fig.12 This is a schematic diagram of the structure of the driving simulation scene generation device for this application.

[0056] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0057] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0058] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0059] The present application embodiment provides a driving simulation scene generation method, referring to Figure 1 , Figure 1This is a flow chart of the first embodiment of the driving simulation scene generation method of the present application. In this embodiment, the method includes: Steps S10 to S40:

[0060] Step S10: Acquire driving data, and determine the current driving trajectory of the vehicle according to the driving data.

[0061] It should be noted that the execution subject of the method of this embodiment can be a computing electronic device with data processing, program running, and network communication functions, such as a mobile phone, a tablet computer, a vehicle terminal, a simulation environment management server, etc., and can also be other electronic devices that can access the driving simulation scene platform. The following takes a driving simulation scene generation device (hereinafter referred to as a "generation device") as an example to explain this embodiment and the following embodiments.

[0062] It can be understood that the current vehicle may include a focus vehicle and a target vehicle. The focus vehicle may be regarded as the source vehicle of the driving data, i.e., the ego vehicle, and the target vehicle may be several non-ego vehicles sensed by the vehicle sensor during the driving of the focus vehicle.

[0063] During the driving process of the vehicle, the driving data can be collected by the vehicle sensors of the concerned vehicle in the shadow mode, and the concerned vehicle can automatically upload the driving data to the generating device when the shadow mode is triggered.

[0064] It should be understood that shadow mode refers to the technology in which the intelligent driving system still operates but does not actually control the vehicle when someone is driving. Automated data collection based on shadow mode can reduce the time and manpower costs required for real vehicle measurement in traditional methods, and is conducive to collecting a large amount of driving data from real roads, covering a wider range of driving environments and scenarios, including complex and rare traffic conditions.

[0065] Exemplarily, the vehicle sensor may include, for example, a camera, a millimeter wave radar, a GPS, an IMU, etc. Therefore, the driving data may be multi-sensor fusion driving data formed by the above-mentioned different vehicle sensors.

[0066] Specifically, after receiving the driving data uploaded by the vehicle of interest in the shadow mode, the generating device can first perform vehicle detection based on the driving data from the camera and millimeter-wave radar to determine several target vehicles related to the vehicle of interest; and then determine the driving trajectory of the vehicle of interest and the target vehicle based on the driving data from the GPS and IMU.

[0067] Step S20: dividing the driving trajectory into a plurality of trajectory segments based on the road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition.

[0068] It can be understood that the road map information can be a high-precision map provided by a map provider, or constructed in real time by vehicle sensors. The road map information contains detailed road information (road starting point, road end point, road reference line, road ID) and lane information (lane starting point, lane end point, lane center line, lane ID) of the vehicle's driving route.

[0069] Specifically, the current vehicle's driving trajectory can be matched with the road map information to determine all the roads and lanes that the current vehicle has traveled. Then, each driving trajectory is segmented at the road and / or lane change locations to obtain a number of trajectory segments. Finally, based on the trajectory segments, driving behavior extraction is performed according to preset driving behavior judgment conditions to determine the driving behavior corresponding to each trajectory segment, as well as the triggering condition and termination condition of each driving behavior.

[0070] Step S30: Initializing simulation scenes and scene parameters according to the road map information and the driving behaviors corresponding to the trajectory segments.

[0071] It should be noted that the reconstruction of the simulation scene can be divided into static scene reconstruction and dynamic scene reconstruction. The static scene reconstruction of the simulation scene can be performed through the road map information, and the dynamic behavior in the simulation scene can be defined according to the driving behavior corresponding to each track segment, thereby completing the dynamic scene reconstruction.

[0072] It should also be noted that the generating device can realize static scene reconstruction and dynamic scene reconstruction based on the international authoritative autonomous driving simulation scene standards OpenDrive and OpenScenario. OpenDrive is an open standard for describing road networks. OpenDrive files can use XML format to describe static content in simulation scenes; OpenScenario defines a data model and a file format based on it, namely OpenScenario files, which can be used to describe dynamic content in simulation scenes.

[0073] In the specific implementation, the generating device will focus on the starting point of the vehicle's travel as the map origin of the static scene, and record the road information that the current vehicle has traveled on in XML format to the OpenDrive file; then generate the OpenScenario file based on the driving behavior corresponding to each trajectory segment. Each driving behavior contains the scenario parameters related to it. Obtaining the OpenDrive file and the OpenScenario file completes the initialization of the simulation scene and scenario parameters.

[0074] Step S40: updating the scenario parameters in a preset generalization manner, and generating a target driving simulation scenario based on the simulation scenario according to the update result.

[0075] It should be noted that the scene parameters are variable parameters in a modifiable scene, and the preset generalization method may be a combination and variation of variable parameters determined by the user based on different generalization requirements.

[0076] Exemplarily, scene parameters may be divided into: road parameters, environment parameters, and vehicle status parameters, and may be further subdivided based on these.

[0077] In the specific implementation, users selectively combine and change the parameters of each scenario in the simulation scenario based on different generalization requirements, thereby generating a variety of target driving simulation scenarios to ensure that the subsequent intelligent driving system can be comprehensively tested and verified in various driving scenarios.

[0078] This embodiment can divide the current vehicle's driving trajectory into trajectory segments according to the road map information, so that the high-level driving behavior of the vehicle can be captured based on the trajectory segments. Compared with the existing limitation of only generalizing the vehicle trajectory, the scene parameters of the simulation scene generated based on the driving behavior can be updated according to the preset generalization method, which is conducive to obtaining a target driving simulation scene with higher adaptability and authenticity.

[0079] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2 , Figure 2 This is a flow chart of the second embodiment of the driving simulation scene generation method of the present application.

[0080] In this embodiment, in order to specifically explain how to obtain an accurate driving trajectory according to driving data, step S10 specifically includes: steps S101 to S103:

[0081] Step S101: Acquire driving data of the vehicle of interest in the shadow mode through a vehicle sensor, and input the driving data into a preset target detection model to determine the target vehicle.

[0082] It should be understood that when the expected action command of the intelligent driving system of the concerned vehicle differs significantly from the actual control command of the vehicle driver, the shadow mode trigger system of the concerned vehicle records the difference generation data and calculates the deviation rate between the expected action command and the actual control command (by comparing the numerical differences of the vehicle steering angle, vehicle speed, etc. for weighted average calculation to obtain the overall deviation rate). When the deviation rate is greater than the preset deviation threshold, that is, in the case of a high deviation rate, the concerned vehicle sends the difference generation data as driving data to the generating device.

[0083] It will be appreciated that the driving data may include camera images and millimeter-wave radar point clouds collected by external sensors (such as cameras and millimeter-wave radars), as well as GPS / IMU positioning data collected by internal body sensors (such as GPS and IMU).

[0084] It should be understood that the camera image and millimeter-wave radar point cloud in the driving data can be input into a preset target detection model, which obtains input fusion features by respectively extracting features of the camera image and millimeter-wave radar point cloud (image features and point cloud features), and then detects the target vehicle perceived by the external sensor during the driving process of the vehicle according to the input fusion features.

[0085] It should also be noted that the preset target detection model can be a target detection model based on deep learning, such as a multimodal fusion model such as Camera Radar Fusion Network (CRFNet), Center Fusion Neural Network (Center Fusion), and Multi-sensor Fusion Detection Model (CRAFT). The CRFNet can identify the appearance information of the target based on image features and provide the spatial position information of the target through point cloud features, thereby generating a three-dimensional detection frame for the perceived target (including target vehicles, pedestrians, cyclists, etc.) by fusing the two, and then determining the target vehicle.

[0086] In addition, the detection results of continuous frames can be obtained by inputting the camera images of continuous frames and the corresponding millimeter-wave radar point clouds into a preset target detection model, that is, the driving speed of the target vehicle can be calculated based on the displacement of the target vehicle between continuous frames; or the driving speed of the target vehicle can be obtained directly based on the millimeter-wave radar.

[0087] Step S102: Based on the driving data, determine the trajectory data corresponding to the target vehicle according to a preset target tracking algorithm.

[0088] It should be understood that after the target detection is completed, the detection results of the above continuous frames can be input into a target tracking model constructed based on a preset target tracking algorithm. The preset target tracking algorithm can include: a Simple Online and Realtime Tracking (SORT) algorithm and a Deep SORT (DeepSORT) algorithm.

[0089] For example, the SORT algorithm can assign a unique ID to each target vehicle, and perform target association based on the appearance features (e.g., visual features) or motion features (e.g., speed features, acceleration features) of each target vehicle, thereby ensuring that the same target vehicle can be tracked in consecutive frames. Furthermore, it can also be combined with trajectory prediction models such as Kalman filters to handle short-term occlusion or target loss, maintain continuous tracking of the target vehicle, and then generate a complete trajectory for each target vehicle, that is, determine the trajectory data corresponding to the target vehicle.

[0090] Here you can combine Figure 3 A process for tracking a target vehicle in consecutive frames to generate a complete trajectory of the target vehicle is described. Figure 3 Schematic diagram of the complete trajectory generation process for the target vehicle.

[0091] exist Figure 3 In the process, the detection results of consecutive frames are used as the input of the target tracking model. When the target tracking starts, a unique ID can be assigned to each target vehicle in the detection results of consecutive frames, and the trajectory can be recorded (trajectory initialization). In each consecutive frame, the SORT algorithm is used to associate the trajectory points of the detected target vehicles. If the association fails, the trajectory record of the target vehicle is maintained. If the association succeeds, the trajectory record of the target vehicle is updated. At the same time, the trajectory prediction model can also be used to predict the next trajectory point of the target vehicle to deal with possible short-term occlusion or loss of the target vehicle. Finally, the complete trajectory of each target vehicle is generated and output.

[0092] Step S103: generating the driving trajectories of the vehicle of interest and the target vehicle according to the trajectory data.

[0093] It should be noted that the trajectory data of the target vehicle is the target vehicle's driving trajectory relative to the vehicle of interest (the complete trajectory of the target vehicle obtained above), so the relative driving trajectory of the target vehicle can be converted into an absolute trajectory in combination with the driving trajectory of the vehicle of interest, and the driving trajectory of the vehicle of interest can be directly obtained based on the driving data from the internal body sensor. Therefore, step S103 specifically includes: steps S1031 to S1032:

[0094] Step S1031: Acquire the driving trajectory of the vehicle of interest according to the driving data, and determine the relative driving trajectory corresponding to the target vehicle according to the trajectory data.

[0095] It should be understood that the driving trajectory of the vehicle of interest in the global coordinate system can be calculated using a multi-sensor fusion algorithm (such as a Kalman filter algorithm or an extended Kalman filter algorithm) based on the GPS / IMU positioning data in the driving data. Specifically, the position of the vehicle of interest at the start time in the GPS / IMU positioning data can be used as the driving origin, and the absolute position of the vehicle of interest at each time stamp relative to the driving origin can be directly calculated to form the driving trajectory of the vehicle of interest, which includes the timestamps corresponding to the vehicle of interest at each absolute position.

[0096] It should also be noted that the above trajectory data is the driving trajectory of the target vehicle relative to the vehicle of interest, that is, the relative driving trajectory corresponding to the target vehicle, and the relative driving trajectory includes the timestamps corresponding to the target vehicle at each relative position.

[0097] Step S1032: Match the relative driving trajectory with the driving trajectory of the target vehicle according to the absolute position and timestamp of the target vehicle, and determine the driving trajectory of the target vehicle according to the matching result.

[0098] It should be noted that the driving trajectory of the concerned vehicle and the relative driving trajectory of the target vehicle can be matched to the nearest neighbor based on the timestamp (i.e., the timestamp of the concerned vehicle closest to the timestamp of the given target vehicle is found to complete the matching), and the driving trajectory (absolute driving trajectory) of each target vehicle is obtained. This embodiment ensures that the driving trajectory of the concerned vehicle and the relative driving trajectory of the target vehicle correspond precisely in time by utilizing the synchronous time information (based on timestamp alignment), thereby improving the accuracy of the generated driving trajectory of the target vehicle.

[0099] Furthermore, considering that the number of target vehicles is usually not unique, the target vehicle's driving trajectory can be post-processed to filter out unreasonable or scene-irrelevant target vehicles. The deletion criteria can be pre-set as follows: 1. Target vehicles that are too far away from the target vehicle and have no interaction; 2. Target vehicles that are only detected for a short period of time; 3. Target vehicles that show unreasonable speed or acceleration changes.

[0100] Through the above-mentioned screening of target vehicles, it can be ensured that the driving trajectories of the focus vehicle and the screened target vehicles meet the vehicle dynamics constraints (such as maximum acceleration, steering angle, etc.), which is conducive to improving the authenticity of the subsequently generated target driving simulation scene.

[0101] This embodiment inputs the camera image and millimeter wave radar point cloud in the driving data into the preset target detection model, detects and outputs the three-dimensional detection frame of each target vehicle, and tracks and updates each target vehicle in real time in continuous frames. In this way, the detection result of the target vehicle is combined with the precise positioning data of the vehicle of interest to obtain the driving trajectory of the vehicle of interest and multiple target vehicles. The accuracy and continuity of the generated driving trajectory are ensured.

[0102] Based on the first and second embodiments of the present application, in the third embodiment of the present application, the same or similar contents as those of the first and second embodiments can be referred to the above description, and will not be described in detail later. Figure 4 , Figure 4 This is a flow chart of the third embodiment of the driving simulation scene generation method of the present application.

[0103] In this embodiment, considering that the existing method usually directly uses the original vehicle trajectory data to reconstruct the driving simulation scene, it fails to effectively abstract the core elements of the driving behavior in the complex scene. Therefore, in order to specifically explain how to extract the driving behavior, step S20 specifically includes: steps S201 to S203:

[0104] Step S201: determining the road information and lane information related to the driving trajectory according to the road map information, and determining the driving speed of the current vehicle according to the road information and lane information.

[0105] It should be understood that the driving trajectory of the current vehicle can be matched with the high-precision map to obtain road map information, which may include all road IDs (and corresponding road starting points, road end points, road reference lines) and lane IDs (and corresponding lane starting points, lane end points, lane center lines) that the current vehicle (focus vehicle and target vehicle) has traveled.

[0106] Specifically, the spatial coordinates of the current vehicle in the driving trajectory can be converted into the road ID and lane ID of the trajectory point, as well as the driving distance (Local_Y) relative to the starting point of the road and the lateral offset value (Local_X) relative to the center line of the lane; then, the speed and acceleration of the current vehicle are calculated based on the distance and timestamp interval between adjacent trajectory points in the same driving trajectory; finally, the speed and acceleration of the current vehicle can be decomposed into lateral speed (V_X), longitudinal speed (V_Y), lateral acceleration (A_X) and longitudinal acceleration (A_Y) in combination with the direction of the road reference line.

[0107] Step S202: Split the driving trajectory into a plurality of trajectory segments based on the road information, the lane information and the driving speed.

[0108] It should be understood that the driving trajectory can be sliced ​​according to the vehicle state according to the preset switching logic, which may include: 1. Road ID or lane ID switching: When the current vehicle switches roads or lanes, the driving trajectory before and after the switch is immediately divided into two trajectory segments. 2. Significant changes in lateral and longitudinal speed and acceleration: When the lateral and longitudinal speed and acceleration of the current vehicle exceed the set threshold, the trajectory segment is segmented.

[0109] Through the preset slicing logic, the complete driving trajectory is divided into multiple smaller trajectory segments, so as to facilitate the subsequent driving behavior extraction based on the trajectory segments.

[0110] Step S203: Determine the driving behavior corresponding to each of the trajectory segments by using preset driving behavior determination conditions.

[0111] It should be understood that, based on the trajectory segments, the driving behaviors corresponding to the trajectory segments can be extracted from the perspectives of lateral behavior, longitudinal behavior and trajectory behavior. Therefore, the preset driving behavior determination conditions include: lateral behavior determination conditions, longitudinal behavior determination conditions and trajectory behavior determination conditions. Accordingly, step S203 specifically includes: steps S2031 to S2034:

[0112] Step S2031: Acquire the trajectory data corresponding to each of the trajectory segments.

[0113] It should be noted that the trajectory data may include the trajectory points, road information, lane information, and driving speed of the current vehicle in the trajectory segment.

[0114] Step S2032: determining whether the trajectory data satisfies the lateral behavior determination condition and / or the longitudinal behavior determination condition.

[0115] Step S2033: If yes, determine the driving behavior corresponding to each of the trajectory segments according to the judgment result.

[0116] It should be noted that lateral behavior may include: lane change, lateral deviation and lateral obstacle avoidance. The lateral behavior determination condition may be expressed as follows:

[0117] a. Lane change: When the lane ID of the current vehicle changes and the offset (Local_X) relative to the center of the lane changes significantly, it can be determined that the current vehicle has changed lanes.

[0118] b. Lateral offset: Whether the lateral offset value (Local_X) of the current vehicle relative to the lane centerline exceeds a preset threshold without changing the lane ID. If so, it is determined that the current vehicle has undergone lateral offset or lane deviation.

[0119] c. Lateral obstacle avoidance: If the lateral speed (V_X) of the current vehicle increases or decreases rapidly without changing the lane ID, and the corresponding lateral acceleration (A_X) is greater than the set threshold, obstacle avoidance may have occurred.

[0120] It should also be noted that the longitudinal behavior may include: acceleration / deceleration, maintaining vehicle distance, stopping or starting. The longitudinal behavior determination condition may be expressed as follows:

[0121] d. Acceleration / deceleration: If the longitudinal speed (V_Y) of the current vehicle changes beyond the set threshold, it can be determined that the current vehicle is accelerating or decelerating. For example, a certain speed increment threshold is set. When the longitudinal speed increment is greater than the speed increment threshold, it is determined that the current vehicle has an acceleration behavior in the corresponding trajectory segment; conversely, when the longitudinal speed increment is less than a certain speed decrement threshold, it is determined that the current vehicle has a deceleration behavior in the corresponding trajectory segment.

[0122] e. Keeping the distance between vehicles: By monitoring the relative distance and longitudinal speed difference between the current vehicle and the vehicle in front of it on its driving track, it is detected whether the current vehicle is keeping a safe distance. If the ratio of the longitudinal speed to the change in the distance to the vehicle in front is within a certain preset reasonable range, it can be determined that the current vehicle is keeping the distance between vehicles.

[0123] f. Stop or start: When the longitudinal speed (V_Y) of the current vehicle is zero and the duration reaches a preset threshold, it can be determined that the current vehicle is stationary; if the longitudinal speed (V_Y) increases from zero, it can be determined that the current vehicle is started.

[0124] Step S2034: If not, converting the trajectory data into a driving behavior of traveling along a preset fixed point according to the trajectory behavior determination condition, and determining it as the driving behavior corresponding to the trajectory segment.

[0125] It should be noted that in the absence of a high-precision map or the inability to extract the vehicle's lateral and longitudinal driving behavior from the trajectory data, the trajectory data can be converted into a driving behavior along a preset fixed point based on the trajectory behavior determination conditions.

[0126] The trajectory behavior determination condition can be further divided into determination conditions along a fixed heading path point or along a fixed time path point based on the difference of the preset fixed points, and can be expressed as follows:

[0127] g. Heading tracking: In the absence of a high-precision map or when the vehicle's lateral and longitudinal driving behavior cannot be extracted from the trajectory data, and when the current vehicle's driving speed and acceleration change little, the trajectory data is converted into a driving behavior of driving along fixed heading waypoints at a specified speed.

[0128] h. Path tracking: In the absence of a high-precision map or when the lateral and longitudinal driving behaviors of the vehicle cannot be extracted from the trajectory data, and when the changes in the current vehicle's speed and acceleration exceed a certain preset threshold, the trajectory data is converted into driving behaviors along fixed time path points (timepoints).

[0129] It should be understood that after determining the driving behavior corresponding to each trajectory segment, corresponding trigger conditions and termination conditions can also be set for each driving behavior. For example, the trigger condition for lane change behavior can be that the current vehicle deviates from the starting point of the lane, and the termination condition is that the current vehicle enters the new lane.

[0130] In addition, you can refer to Figure 5 The traditional dynamic scene reconstruction based on driving trajectory and the dynamic scene reconstruction based on driving behavior of this application are compared and explained. Figure 5 Comparison graph of the source scene required for dynamic scene reconstruction.

[0131] exist Figure 5 Among them, 5-a is a scene diagram required for traditional dynamic scene reconstruction based on driving trajectory, and 5-b is a scene diagram required for dynamic scene reconstruction based on driving behavior corresponding to trajectory segments in this application.

[0132] From the comparison between 5-a and 5-b above, it can be seen that the scene graph required for the traditional dynamic scene reconstruction based on the driving trajectory contains a lot of unnecessary detail information, and fails to effectively abstract the core elements of driving behavior in complex scenes. Therefore, when faced with complex scenes, its interpretability and generalization capabilities are poor, and it is difficult to support the comprehensive testing and optimization of intelligent driving systems. In the scene graph required for the dynamic scene reconstruction of driving behaviors corresponding to trajectory segments in this application, the driving trajectory is divided into trajectory segments and includes advanced driving behaviors such as changing lanes, accelerating, decelerating, and turning, thereby effectively simplifying the driving trajectory, removing redundant information in the driving trajectory, and helping to grasp the core characteristics of driving behavior.

[0133] In a specific implementation, the generating device can convert the complete driving trajectory into a series of driving behaviors and their corresponding triggering conditions and termination conditions by extracting the driving behaviors corresponding to each of the trajectory segments based on the above-mentioned preset driving behavior judgment conditions, so as to facilitate the subsequent construction of a dynamic scene.

[0134] In addition, this can also be combined with Figure 6 The extraction process of driving behavior in this embodiment is described. Figure 6 This is a schematic diagram of the driving behavior extraction process in the third embodiment of the driving simulation scene generation method of the present application.

[0135] based on Figure 6After obtaining the complete driving trajectory of the current vehicle, the driving trajectory can be sliced ​​first and divided into short and continuous trajectory segments. Then, combined with the road information, lane information and the current vehicle speed provided by the high-precision map, the lateral behavior, longitudinal behavior and trajectory behavior of the trajectory segment are extracted to determine the driving behavior corresponding to each trajectory segment and obtain a driving behavior set.

[0136] This embodiment slices the complete driving trajectory of the current vehicle into short, continuous trajectory segments. Then, combined with the road information and lane information in the high-precision map, driving behavior is extracted from these trajectory segments, thereby converting the trajectory data into a set of lateral behavior, longitudinal behavior, and trajectory behavior. This embodiment can effectively reduce the complex redundant information in the original driving data by extracting a higher level of driving behavior of the current vehicle, making the driving scene of the current vehicle more concise and easy to analyze. This not only reduces the complexity of the scene, but also enhances the interpretability of the driving scene, making the understanding of the driving scene more focused and accurate, and helping to grasp the core characteristics of driving behavior.

[0137] Based on the first to third embodiments of the present application, in the fourth embodiment of the present application, the same or similar contents as those of the first to third embodiments can be referred to the above description, and will not be repeated in the following. Figure 7 , Figure 7 This is a flow chart of the fourth embodiment of the driving simulation scene generation method of the present application.

[0138] In this embodiment, in order to further illustrate how to reconstruct the simulation scene, step S30 specifically includes: steps S301 to S303:

[0139] Step S301: Initializing a static road structure of a simulation scene according to the road map information and each of the trajectory segments.

[0140] It should be understood that the generating device can determine the road map information corresponding to each track segment, and construct the static road structure of the simulation scene according to the road type, attribute and road topology relationship in the road map information. When there is a high-precision map provided by a map provider, the high-precision map is preferentially used as the road map information to construct the static road structure and generate the OpenDrive file.

[0141] Specifically, the generating device may use the driving origin of the vehicle of interest as the map origin of the static road structure, and record the roads passed by the vehicle of interest and the target vehicle one by one in the XML format into the OpenDrive file.

[0142] For example, the description of each road may include the following information:

[0143] 1. Road topology: record the length, reference line, curvature and other information of the road, as well as the connection relationship between the road and the roads before and after.

[0144] 2. Lane line topology: including lane attributes (such as lane type, whether there is an emergency lane), lane width, lane line type (dashed line, solid line, etc.), and the front and rear connection relationship of the lane.

[0145] Specifically, the generating device may obtain an OpenDrive file containing information of multiple lanes, and the OpenDrive file may define a static road structure of the entire simulation scene.

[0146] Step S302: Based on the static road structure, a scene entity of a simulation scene is initialized in combination with the current vehicle.

[0147] Step S303: Associating the driving behaviors corresponding to the trajectory segments with the scene entities, and generating an initial simulation scene file and scene parameters according to the association results.

[0148] It should be noted that the generating device can first initialize the scene entities on the corresponding static road structure according to the current vehicles (focus vehicle and target vehicle) involved in each trajectory segment, and determine the vehicle information corresponding to each scene entity (including three-dimensional detection frame, vehicle type, vehicle initial speed, vehicle initial position and vehicle direction).

[0149] Next, you can use the OpenScenario syntax to define the driving behavior of each scenario entity and the trigger / termination conditions of the driving behavior, and add them to the OpenScenario file in XML format. In the OpenScenario file, each driving behavior includes detailed parameters related to it. Taking lane change as an example, the corresponding detailed parameters include:

[0150] a. Lane changing entity: specifies the scene entity on which the lane change is to be performed.

[0151] b. Lane change direction: Determine whether the lane change is to the left or right.

[0152] c. Lane change duration or distance: Describes the time or distance required to change lanes.

[0153] d. Trigger condition: defines when the behavior is triggered, such as starting a lane change when a scene entity reaches a certain time point, position, or satisfies a certain sensor status.

[0154] Finally, in addition to the trigger conditions of a single driving behavior, the scenario trigger / termination conditions of each simulation scenario can also be stored in the OpenScenario file. For example, common scenario termination conditions may include:

[0155] a. Arrive at the designated location: The vehicle reaches the predetermined destination and the scene ends.

[0156] b. Timeout: When the simulation time exceeds the preset time limit, the scenario ends.

[0157] In addition, the scene triggering conditions may also include the arrival of the focus vehicle or target vehicle at the predetermined destination, the time point when the focus vehicle starts driving, etc.; the scene termination conditions may also include the triggering of specific driving behaviors, etc.

[0158] It should be understood that by adding driving behaviors in XML format in the OpenScenario file and setting the trigger / termination conditions of the prevention scenario, all driving behaviors and trigger / termination conditions of the scenario can be combined to ultimately generate a complete OpenScenario file (initial simulation scenario file) to describe the dynamic content in the simulation scenario.

[0159] You can also refer to Figure 8 The simulation scene initialization process of this embodiment is described. Figure 8 This is a schematic diagram of simulation scene reconstruction in the fourth embodiment of the driving simulation scene generation method of the present application.

[0160] exist Figure 8 In the simulation scene reconstruction process, the reconstruction process can be divided into static scene reconstruction and dynamic scene reconstruction. The static road structure of the simulation scene can be defined by the OpenDrive file, that is, the OpenDrive file contains the road information in the simulation scene; the dynamic content in the simulation scene can be defined by the OpenScenario file, and the dynamic content comes from the focus vehicle and the target vehicle, that is, the OpenScenario file contains the self-vehicle information corresponding to the focus vehicle and the target vehicle information corresponding to the target vehicle.

[0161] based on Figure 8 In the simulation scene reconstruction process, the obtained simulation scene can be Fig. 9 As shown, Fig. 9 This is an example diagram of a simulation scene in the fourth embodiment of the driving simulation scene generation method of the present application.

[0162] exist Fig. 9 In the above figure, ego identifies the vehicle of interest in the simulation scene, and c1~c4 respectively identify the target vehicles in the simulation scene.

[0163] It should also be noted that the final generated OpenScenario file contains the scene entities, initialization states, driving behaviors, and trigger conditions in the simulation scene, which ensures the logical integrity and accuracy of each simulation scene, and is conducive to comprehensive testing and evaluation of intelligent driving systems in various scenarios.

[0164] Furthermore, in order to generate more diverse simulation scenarios to cover different driving behaviors and scenario environment conditions, the generated OpenScenario file can be generalized by updating the scenario parameters. Therefore, step S40 specifically includes: steps S401 to S403:

[0165] Step S401: determining a preset generalization method according to a current generalization requirement, wherein the preset generalization method includes at least one of a road parameter adjustment method, an environment parameter adjustment method, and a vehicle state parameter adjustment method.

[0166] Step S402: determining the scene parameters to be adjusted in each of the scene parameters according to the preset generalization method, and updating the scene parameters to be adjusted, wherein the scene parameters include: road parameters, environmental parameters and vehicle state parameters.

[0167] The road parameter adjustment method can generate different simulation test maps by modifying multiple road parameters in the simulation scene. The specific road parameter adjustment methods include:

[0168] a. Changes in road width: Different lane widths are set to simulate narrow urban roads or spacious highways to test the adaptability of the intelligent driving system to different spatial environments.

[0169] b. Adjustment of road curvature: By adjusting the curvature of the road, various curve scenarios are generated to evaluate the steering control ability of the intelligent driving system under different curve conditions.

[0170] c. Lane type and lane line attributes: Switch a single lane to multiple lanes or add bus lanes and emergency lanes, and change the form of lane lines (dashed lines, solid lines, double yellow lines, etc.) to test the driving decision-making ability of the intelligent driving system under different road types and rules.

[0171] d. Lane connection relationship: Change the way lanes are connected, such as adding forks, merges, and U-turn areas, to simulate complex road conditions and intersections, and test the intelligent driving system's understanding and navigation capabilities of the road network.

[0172] The environmental parameter adjustment method can simulate a variety of complex environments by adjusting the initial environmental conditions in the simulation scene. The specific environmental parameter adjustment methods include:

[0173] a. Ambient weather conditions: Change weather conditions such as lighting, visibility, precipitation (rain, snow) to simulate various weather scenarios, such as foggy days, night driving, and slippery roads in rainy days. This allows the intelligent driving system to be tested for perception, decision-making, and control performance under different weather conditions.

[0174] b. Road conditions: By changing the road friction coefficient, the slippery degree of different road surfaces (such as ice and snow, sandy and gravel roads, etc.) is simulated to test the driving stability and emergency braking ability of the intelligent driving system under different grip conditions.

[0175] c. Initial position and state of the vehicle: Adjust the initial position, speed, direction, etc. of the vehicle and other vehicles to simulate different driving scenarios, such as waiting at an intersection, dynamic rear-end collision, etc., to examine the intelligent driving system's ability to respond to emergencies.

[0176] The vehicle state parameter adjustment method can generate a variety of driving behavior scenarios by modifying the relevant parameters of the driving behavior. The specific vehicle state parameter adjustment methods include:

[0177] a. Behavior parameter adjustment: For example, modify the following distance of the vehicle-to-vehicle distance keeping behavior and the lane-changing action time / distance of the lane-changing behavior to generate different driving behavior change patterns. By adjusting these behavior parameters, the adaptability of the intelligent driving system to different driving habits and behavior patterns can be tested.

[0178] b. Behavior trigger / termination conditions: Modify the trigger conditions of driving behavior, such as advancing or delaying the behavior trigger conditions for a period of time to make the simulation scenarios more diverse and challenging. The termination conditions can also be adjusted according to the simulation scenarios.

[0179] Step S403: updating the initial simulation scene file according to the scene parameter update result to obtain a target driving simulation scene.

[0180] It should be understood that by combining the above-mentioned road parameter adjustment method, environmental parameter adjustment method, and vehicle state parameter adjustment method, and combining or changing the road parameters, environmental parameters, and vehicle state parameters, a rich variety of simulation test scenarios can be generated, ensuring that the obtained target driving simulation scenario effectively covers various complex, borderline, and rare real driving scenarios. Thus, the target driving simulation scenario is added to the simulation scenario library, so that the target driving simulation scenario can be loaded in the Carla simulation engine to realize the simulation test of the intelligent driving system, thereby ensuring the comprehensive performance test of the intelligent driving system and improving its robustness and reliability in different driving scenarios.

[0181] After generating the initial simulation scenario file, this embodiment generalizes the initial simulation scenario based on the road parameter adjustment method, the environmental parameter adjustment method, and the vehicle state parameter adjustment method, and can generate more diverse simulation scenarios, thereby providing rich test conditions for the comprehensive testing and optimization of the intelligent driving system. Compared with the limitation of only generalizing the vehicle trajectory in the existing method, the target driving simulation scenario obtained based on this embodiment can more comprehensively test the performance of the intelligent driving system in different scenarios, thereby improving the reliability and safety of the system.

[0182] Furthermore, this can also be combined with Fig.10 The whole process of the driving simulation scene generation method of this application is explained. Fig.10 This is a schematic diagram of the entire process of the driving simulation scenario generation method of this application.

[0183] exist Fig.10 In the process of driving simulation scene generation, the driving behavior analysis stage and the simulation scene generation stage can be divided into two stages. The driving behavior analysis stage includes: driving trajectory generation and driving behavior analysis, and the simulation scene generation stage includes: simulation scene reconstruction and simulation scene generalization.

[0184] In the driving trajectory generation stage, the generating device extracts camera images and millimeter-wave radar point clouds from the driving data of the vehicle of interest that triggers the shadow mode, and inputs them into the preset target detection model to detect and output each target vehicle; then, combined with the GPS / IMU positioning data of the vehicle of interest, the target vehicle is tracked in real time through a multi-sensor fusion algorithm, and finally the driving trajectories of the vehicle of interest and the target vehicle are obtained.

[0185] In the driving behavior analysis stage, the driving trajectories of the concerned vehicles and target vehicles are analyzed in detail by combining the road map information. Specifically, the driving trajectories can be sliced ​​first and divided into several trajectory segments, and then the driving behaviors (lateral behaviors, longitudinal behaviors, and trajectory behaviors) corresponding to each trajectory segment are extracted.

[0186] In the simulation scene reconstruction stage, the road map information, trajectory fragments and corresponding driving behaviors are converted into simulation scene files, which mainly include OpenDrive files used to describe static content in the simulation scene and OpenScenario files used to describe dynamic content in the simulation scene.

[0187] In the simulation scenario generalization stage, the generated OpenScenario file is generalized, and a rich variety of simulation test scenarios are generated by modifying the scenario parameters, ensuring that the target driving simulation scenario obtained effectively covers various complex, borderline and rare real driving scenarios.

[0188] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the driving simulation scene generation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0189] In addition, the present application also provides a driving simulation scene generation device, referring to Fig.11 , Fig.11 This is a structural block diagram of the first embodiment of the driving simulation scene generation device of the present application; Fig.11 As shown, the device comprises:

[0190] The driving trajectory generation module 1101 is used to obtain driving data and determine the driving trajectory of the current vehicle according to the driving data;

[0191] A driving behavior extraction module 1102, for dividing the driving trajectory into a plurality of trajectory segments based on road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition;

[0192] A simulation scene reconstruction module 1103 is used to initialize the simulation scene and scene parameters according to the road map information and the driving behavior corresponding to each of the trajectory segments;

[0193] The simulation scenario generalization module 1104 is used to update the scenario parameters according to a preset generalization method, and generate a target driving simulation scenario based on the simulation scenario according to the update result.

[0194] The current vehicles include: a focused vehicle and a target vehicle;

[0195] Furthermore, the driving trajectory generation module 1101 is also used to: obtain driving data of the vehicle of interest in the shadow mode through vehicle sensors, and input the driving data into a preset target detection model to determine the target vehicle; based on the driving data, determine the trajectory data corresponding to the target vehicle according to a preset target tracking algorithm; generate the driving trajectories of the vehicle of interest and the target vehicle according to the trajectory data.

[0196] Furthermore, the driving trajectory generation module 1101 is also used to: obtain the driving trajectory of the vehicle of interest based on the driving data, and determine the relative driving trajectory corresponding to the target vehicle based on the trajectory data; match the relative driving trajectory with the driving trajectory of the vehicle of interest according to the absolute position and timestamp of the vehicle of interest, and determine the driving trajectory of the target vehicle based on the matching result.

[0197] Furthermore, the driving behavior extraction module 1102 is also used to determine the road information and lane information related to the driving trajectory based on the road map information, and determine the driving speed of the current vehicle based on the road information and lane information; split the driving trajectory into a plurality of trajectory segments based on the road information, the lane information and the driving speed; and determine the driving behavior corresponding to each of the trajectory segments through preset driving behavior judgment conditions.

[0198] The preset driving behavior determination conditions include: lateral behavior determination conditions, longitudinal behavior determination conditions and trajectory behavior determination conditions;

[0199] Furthermore, the driving behavior extraction module 1102 is also used to obtain trajectory data corresponding to each of the trajectory segments; determine whether the trajectory data meets the lateral behavior determination condition and / or the longitudinal behavior determination condition; if so, determine the driving behavior corresponding to each of the trajectory segments according to the determination result; if not, convert the trajectory data into a driving behavior of traveling along a preset fixed point according to the trajectory behavior determination condition, and determine it as the driving behavior corresponding to the trajectory segment.

[0200] Furthermore, the simulation scene reconstruction module 1103 is also used to initialize the static road structure of the simulation scene according to the road map information and each of the trajectory segments; based on the static road structure, initialize the scene entity of the simulation scene in combination with the current vehicle; associate the driving behavior corresponding to each of the trajectory segments to each of the scene entities, and generate an initial simulation scene file and scene parameters according to the association results.

[0201] Furthermore, the simulation scenario generalization module 1104 is also used to determine a preset generalization method according to a current generalization requirement, the preset generalization method including at least one of a road parameter adjustment method, an environmental parameter adjustment method, and a vehicle state parameter adjustment method; determine the scene parameters to be adjusted from each of the scene parameters according to the preset generalization method, and update the scene parameters to be adjusted, the scene parameters including road parameters, environmental parameters, and vehicle state parameters; update the initial simulation scenario file according to the scene parameter update result to obtain a target driving simulation scenario.

[0202] This embodiment can divide the current vehicle's driving trajectory into trajectory segments according to the road map information, so that the high-level driving behavior of the vehicle can be captured based on the trajectory segments. Compared with the existing limitation of only generalizing the vehicle trajectory, the scene parameters of the simulation scene generated based on the driving behavior can be updated according to the preset generalization method, which is conducive to obtaining a target driving simulation scene with higher adaptability and authenticity.

[0203] The present application also provides a driving simulation scene generation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the driving simulation scene generation method in the above-mentioned embodiment one.

[0204] Reference below Fig.12 , Fig.12 The schematic diagram of the structure of the driving simulation scene generation device of the present application. The driving simulation scene generation device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.12 The driving simulation scene generation device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0205] like Fig.12As shown, the driving simulation scene generation device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the driving simulation scene generation device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the driving simulation scene generation device to communicate wirelessly or wired with other devices to exchange data. Although the driving simulation scene generation device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or provided instead.

[0206] In addition, the present application also provides a computer-readable storage medium having computer-readable program instructions (ie, computer program) stored thereon, and the computer-readable program instructions are used to execute the driving simulation scene generation method in the above-mentioned embodiment.

[0207] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0208] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system 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 system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other elements in the process, method, article or system including the element.

[0209] The serial numbers of the above-mentioned embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. They are only some embodiments of the present application and do not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the description and drawings of the present application under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A driving simulation scene generation method, characterized in that: The method comprises: Acquiring driving data, and determining a current driving trajectory of the vehicle based on the driving data; Dividing the driving trajectory into a plurality of trajectory segments based on the road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition; Initialize simulation scenes and scene parameters according to the road map information and the driving behaviors corresponding to each of the trajectory segments; The scenario parameters are updated in a preset generalized manner, and a target driving simulation scenario is generated based on the simulation scenario according to the update result.

2. The method according to claim 1, characterized in that The current vehicle includes: a focused vehicle and a target vehicle. The step of acquiring driving data and determining the driving trajectory of the current vehicle according to the driving data includes: Acquire driving data of the vehicle of interest in the shadow mode through a vehicle sensor, and input the driving data into a preset target detection model to determine the target vehicle; Based on the driving data, determining the trajectory data corresponding to the target vehicle according to a preset target tracking algorithm; The driving trajectories of the vehicle of interest and the target vehicle are generated according to the trajectory data.

3. The method according to claim 2, characterized in that The step of generating the driving trajectories of the vehicle of interest and the target vehicle according to the trajectory data comprises: Acquire the driving trajectory of the vehicle of interest according to the driving data, and determine the relative driving trajectory corresponding to the target vehicle according to the trajectory data; The relative driving trajectory is matched with the driving trajectory of the vehicle of interest according to the absolute position and timestamp of the vehicle of interest, and the driving trajectory of the target vehicle is determined according to the matching result.

4. The method according to claim 1, characterized in that The step of dividing the driving trajectory into a plurality of trajectory segments based on the road map information, and determining the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition, comprises: Determining road information and lane information related to the driving trajectory according to the road map information, and determining the current driving speed of the vehicle according to the road information and lane information; Splitting the driving trajectory into a plurality of trajectory segments by combining the road information, the lane information and the driving speed; The driving behavior corresponding to each of the trajectory segments is determined by presetting the driving behavior determination condition.

5. The method according to claim 4, characterized in that The preset driving behavior determination conditions include: a lateral behavior determination condition, a longitudinal behavior determination condition and a trajectory behavior determination condition; the step of determining the driving behavior corresponding to each trajectory segment by the preset driving behavior determination conditions includes: Acquire trajectory data corresponding to each of the trajectory segments; Determining whether the trajectory data satisfies the lateral behavior determination condition and / or the longitudinal behavior determination condition; If yes, determining the driving behavior corresponding to each of the trajectory segments according to the judgment result; If not, the trajectory data is converted into a driving behavior of traveling along a preset fixed point according to the trajectory behavior determination condition, and is determined as the driving behavior corresponding to the trajectory segment.

6. The method according to claim 1, characterized in that The step of initializing the simulation scene and scene parameters according to the road map information and the driving behavior corresponding to each of the trajectory segments comprises: Initialize a static road structure of a simulation scene according to the road map information and each of the trajectory segments; Initialize a scene entity of a simulation scene based on the static road structure and in combination with the current vehicle; The driving behaviors corresponding to the trajectory segments are associated with the scene entities, and an initial simulation scene file and scene parameters are generated according to the association results.

7. The method according to claim 6, characterized in that The step of updating the scenario parameters in a preset generalization manner and generating a target driving simulation scenario based on the simulation scenario according to the update result includes: Determining a preset generalization method according to the current generalization requirement, wherein the preset generalization method includes: at least one of a road parameter adjustment method, an environment parameter adjustment method, and a vehicle state parameter adjustment method; Determining a scene parameter to be adjusted from each of the scene parameters according to the preset generalization method, and updating the scene parameter to be adjusted, wherein the scene parameter includes: a road parameter, an environmental parameter, and a vehicle state parameter; The initial simulation scene file is updated according to the scene parameter update result to obtain the target driving simulation scene.

8. A driving simulation scene generation device, characterized in that: The device comprises: A driving trajectory generation module, used to obtain driving data and determine the current vehicle driving trajectory according to the driving data; A driving behavior extraction module, used to divide the driving trajectory into a plurality of trajectory segments based on road map information, and determine the driving behavior corresponding to each of the trajectory segments according to a preset driving behavior determination condition; A simulation scene reconstruction module, used to initialize the simulation scene and scene parameters according to the road map information and the driving behavior corresponding to each of the trajectory segments; The simulation scenario generalization module is used to update the scenario parameters according to a preset generalization method, and generate a target driving simulation scenario based on the simulation scenario according to the update result.

9. A driving simulation scene generation device, characterized in that: The device comprises: a memory, a processor, and a driving simulation scenario generation program stored in the memory and executable on the processor, wherein the driving simulation scenario generation program is configured to implement the steps of the driving simulation scenario generation method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a driving simulation scenario generation program, and when the driving simulation scenario generation program is executed by the processor, the steps of the driving simulation scenario generation method according to any one of claims 1 to 7 are implemented.

Images