Dangerous scene generation method, device, equipment and readable storage medium

Abstract

The present invention provides a method, device, equipment and readable storage medium for generating a dangerous scene. The method comprises: extracting a target scene segment based on target information; generating a target object list in the target scene segment; generating a target vehicle driving trajectory and a high-precision road map in the target scene segment based on the target object list; weighting based on a weighting factor to generate a first preset number of driving behavior parameters of a host vehicle and a target vehicle driving behavior parameter; combining to obtain a second preset number of driving behavior combinations; mapping the driving behavior combinations to the high-precision road map to generate a target scene; and performing simulation re-play on the target scene, with the target scene in which an accident occurs in the simulation re-play and the target scene in which an accident is dangerous but no accident occurs as the dangerous scene. The present invention makes it possible to obtain dangerous scenes without artificially creating dangerous driving, and reduces costs while improving the efficiency of dangerous scene generation.

Description

Technical Field

[0001] The present invention relates to the technical field of dangerous scenario testing for automatic driving of automobiles, and in particular to a dangerous scenario generation method, device, equipment and readable storage medium. Background Art

[0002] Dangerous scenario testing is of great value in improving the safety of autonomous driving. By collecting dangerous driving data, reproducing and executing tests in a simulation environment or a closed test environment, and verifying the performance of autonomous driving vehicles in dangerous scenarios, the performance of autonomous driving vehicle algorithms can be improved and enhanced. The current methods of collecting dangerous scenario data include: data from existing accidents collected in the National Vehicle Accident In-depth Investigation System, and qualitative descriptions of dangerous scenarios through logical reasoning or expert experience. The above methods of collecting accident data have limitations, such as: (1) the process of dangerous scenarios and the driver's status information before the accident cannot be accurately recorded; (2) the National Vehicle Accident In-depth Investigation System does not include scenarios that do not cause traffic accidents but are dangerous driving; (3) dangerous scenarios are based on the location, behavior relationship and expert experience between traffic participants, and cannot quantitatively describe the process of danger or accident occurrence, resulting in the use of generalized test cases for testing, and the test results are not reliable. Summary of the invention

[0003] The main purpose of the present invention is to provide a dangerous scene generation method, device, equipment and readable storage medium, aiming to solve the technical problem that the dangerous scene data obtained by the existing solutions have limitations.

[0004] In a first aspect, the present invention provides a method for generating a dangerous scene, the method comprising the following steps:

[0005] Extracting target scene segments based on target information, wherein the target information includes behavior parameters of the driver of the host vehicle and driving scene parameters;

[0006] Generating a list of target objects in the target scene segment based on driving scene parameters;

[0007] Based on the target object list, generating a target vehicle driving trajectory and a high-precision road map in a target scene segment;

[0008] Weighting the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle;

[0009] Weighting the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters;

[0010] Combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations;

[0011] Mapping the second preset number of driving behavior combinations into the high-precision road map to generate a target scene;

[0012] The target scenario is simulated and replayed, and the target scenario where an accident occurs and the target scenario that is dangerous but no accident occurs in the simulation and replay are regarded as dangerous scenarios.

[0013] Optionally, before the step of extracting the target scene segment based on the target information, the step includes:

[0014] Acquiring preset information, wherein the preset information includes behavior parameters of the driver of the host vehicle and driving scenario parameters;

[0015] The preset information without the main vehicle positioning information or with abnormal sensor signals is filtered out to obtain the target information.

[0016] Optionally, the step of extracting the target scene segment based on the target information includes:

[0017] Screening out the behavior parameters of the driver of the host vehicle that meet the first preset threshold condition and the time segments of the driving scene parameters that meet the second preset threshold condition;

[0018] The time segment is taken as the target scene segment.

[0019] Optionally, the step of generating a list of target objects in the target scene segment based on the driving scene parameters includes:

[0020] Based on the driving scene parameters, the target objects in the target scene segment are annotated;

[0021] The same target object monitored by multiple sensors is fused, and based on the fusion result, a target object list is output, wherein the target object list includes: a timestamp of the detected target object, a target object type, and a state parameter of the target object.

[0022] Optionally, the step of generating a target vehicle driving trajectory and a high-precision road map in a target scene segment based on the target object list includes:

[0023] Determine a first target object in the target object list whose target object type is a traffic lane line or a road edge, combine the first target object with the map data, and generate a high-precision road map in the target scene segment;

[0024] A second target object whose target object type is a vehicle in the target object list is determined, and a target vehicle driving trajectory in the target scene segment is generated by combining the second target object and a state parameter corresponding to the second target object.

[0025] Optionally, the step of combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations includes:

[0026] By traversing the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner, a third preset number of driving behavior combinations is obtained;

[0027] From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not satisfy the third preset threshold condition, or whose distance between the host vehicle and the target vehicle does not satisfy the fourth preset threshold condition are screened out to obtain the second preset number of driving behavior combinations.

[0028] Optionally, the step of mapping the second preset number of driving behavior combinations into the high-precision road map to generate a target scene includes:

[0029] The driving behavior parameters of the host vehicle and the target vehicle corresponding to each driving behavior combination are aligned with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene;

[0030] Based on the starting coordinates, each driving behavior combination is mapped into the high-precision road map to generate a target scene.

[0031] In a second aspect, the present invention further provides a dangerous scene generating device, the dangerous scene generating device comprising:

[0032] An extraction module, used to extract a target scene segment based on target information, wherein the target information includes a behavior parameter of a driver of a host vehicle and a driving scene parameter;

[0033] A first generating module, configured to generate a list of target objects in the target scene segment based on driving scene parameters;

[0034] A second generating module, for generating a target vehicle driving trajectory and a high-precision road map in a target scene segment based on the target object list;

[0035] A first weighting module, used for weighting the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle;

[0036] A second weighting module, used to weight the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters;

[0037] A combination module, used for combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations;

[0038] A third generating module, used for mapping the second preset number of driving behavior combinations into the high-precision road map to generate a target scene;

[0039] The simulation module is used to simulate and replay the target scene, with the target scene where an accident occurs and the target scene that is dangerous but no accident occurs in the simulation and replay as dangerous scenes.

[0040] Optionally, the dangerous scenario generating device further includes an acquisition module, which is used to:

[0041] Acquiring preset information, wherein the preset information includes behavior parameters of the driver of the host vehicle and driving scenario parameters;

[0042] The preset information without the main vehicle positioning information or with abnormal sensor signals is filtered out to obtain the target information.

[0043] Optionally, the extraction module is specifically used to:

[0044] Screening out the behavior parameters of the driver of the host vehicle that meet the first preset threshold condition and the time segments of the driving scene parameters that meet the second preset threshold condition;

[0045] The time segment is taken as the target scene segment.

[0046] Optionally, the first generating module is specifically used to:

[0047] Based on the driving scene parameters, the target objects in the target scene segment are annotated;

[0048] The same target object monitored by multiple sensors is fused, and based on the fusion result, a target object list is output, wherein the target object list includes: a timestamp of the detected target object, a target object type, and a state parameter of the target object.

[0049] Optionally, the second generating module is specifically used to:

[0050] Determine a first target object in the target object list whose target object type is a traffic lane line or a road edge, combine the first target object with the map data, and generate a high-precision road map in the target scene segment;

[0051] A second target object whose target object type is a vehicle in the target object list is determined, and a target vehicle driving trajectory in the target scene segment is generated by combining the second target object and a state parameter corresponding to the second target object.

[0052] Optionally, the combined module is specifically used for:

[0053] By traversing the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner, a third preset number of driving behavior combinations is obtained;

[0054] From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not satisfy the third preset threshold condition, or whose distance between the host vehicle and the target vehicle does not satisfy the fourth preset threshold condition are screened out to obtain the second preset number of driving behavior combinations.

[0055] Optionally, the third generating module is specifically used to:

[0056] The driving behavior parameters of the host vehicle and the target vehicle corresponding to each driving behavior combination are aligned with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene;

[0057] Based on the starting coordinates, each driving behavior combination is mapped into the high-precision road map to generate a target scene.

[0058] In a third aspect, the present invention also provides a dangerous scenario generating device, comprising a processor, a memory, and a dangerous scenario generating program stored in the memory and executable by the processor, wherein when the dangerous scenario generating program is executed by the processor, the steps of the dangerous scenario generating method as described above are implemented.

[0059] In a fourth aspect, the present invention further provides a readable storage medium, on which a dangerous scenario generation program is stored, wherein when the dangerous scenario generation program is executed by a processor, the steps of the dangerous scenario generation method as described above are implemented.

[0060] The present invention provides a method, device, equipment and readable storage medium for generating a dangerous scene. The method comprises: extracting a target scene segment based on target information, wherein the target information comprises a behavior parameter of a driver of a main vehicle and a driving scene parameter; generating a list of target objects in the target scene segment based on the driving scene parameter; generating a driving trajectory of a target vehicle in the target scene segment and a high-precision road map based on the list of target objects; weighting the behavior parameter of the driver of the main vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the main vehicle; weighting the parameters corresponding to the driving trajectory of the target vehicle based on the first preset number of weighting factors to generate a first preset number of driving behavior parameters of the target vehicle; combining the first preset number of driving behavior parameters of the main vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations; mapping the second preset number of driving behavior combinations to the high-precision road map to generate a target scene; and re-simulating the target scene to re-play a simulation, and taking the target scene in which an accident occurs and the target scene that is dangerous but does not occur in the simulation as the dangerous scene. The present invention superimposes weighted factors of different driving aggressiveness on the actually collected driving scene data, thereby making it possible to obtain dangerous scenes without artificially creating dangerous driving, enriching the sources of accident scenes, and solving the technical problem that the dangerous scene data obtained by the existing solutions have limitations, thereby improving the efficiency of dangerous scene generation and reducing costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A schematic diagram of the hardware structure of a dangerous scene generating device involved in an embodiment of the present invention;

[0062] Figure 2 A schematic diagram of a flow chart of an embodiment of a method for generating a dangerous scene according to the present invention;

[0063] Figure 3 A schematic diagram of a flow chart of another embodiment of a method for generating a dangerous scene according to the present invention;

[0064] Figure 4 A schematic diagram of a flow chart of another embodiment of a method for generating a dangerous scene according to the present invention;

[0065] Figure 5 Schematic diagram of functional modules of an embodiment of a dangerous scene generating device of the present invention.

[0066] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0067] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0068] In a first aspect, an embodiment of the present invention provides a dangerous scenario generating device.

[0069] Reference Figure 1 , Figure 1 It is a schematic diagram of the hardware structure of the dangerous scene generation device involved in the embodiment of the present invention. In the embodiment of the present invention, the dangerous scene generation device may include a processor 1001 (such as a central processing unit Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components; the user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard); the network interface 1004 may optionally include a standard wired interface, a wireless interface (such as wireless fidelity WIreless-FIdelity, WI-FI interface); the memory 1005 may be a high-speed random access memory (random access memory, RAM), or a stable memory (non-volatile memory), such as a disk storage, and the memory 1005 may optionally be a storage device independent of the aforementioned processor 1001. Those skilled in the art will understand that Figure 1 The hardware structure shown in the figure does not constitute a limitation of the present invention, and may include more or less components than those shown in the figure, or combine certain components, or arrange the components differently.

[0070] Continue to refer to Figure 1 , Figure 1 The memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a dangerous scene generation program. The processor 1001 may call the dangerous scene generation program stored in the memory 1005 and execute the dangerous scene generation method provided in the embodiment of the present invention.

[0071] In a second aspect, an embodiment of the present invention provides a method for generating a dangerous scenario.

[0072] Reference Figure 2 , Figure 2 It is a flowchart of an embodiment of a method for generating a dangerous scene according to the present invention.

[0073] In an embodiment of a dangerous scene generation method of the present invention, the dangerous scene generation method includes:

[0074] Step S10, extracting a target scene segment based on target information, wherein the target information includes a behavior parameter of a driver of a host vehicle and a driving scene parameter;

[0075] In this embodiment, the actual natural driving scene is first collected based on the scene collection vehicle, wherein the scene collection is for the main vehicle. In the process of the main vehicle collecting the natural driving scene: first, the actual natural driving scene parameters, such as static obstacles, dynamic vehicles, vehicle navigation positioning, traffic signs and markings, etc., will be collected and recorded based on sensor units such as laser radar, camera, GPS / IMU, collection and recording unit, etc.; second, the behavior parameters of the driver of the main vehicle will be collected and recorded based on sensor units such as brake pedal stroke, pedal force sensor, steering wheel angle sensor, etc., such as the steering wheel angle operated by the driver, wheel cylinder pressure, brake pedal stroke, throttle opening, etc.; third, the state parameters of the driver of the main vehicle will be collected and recorded based on human body information collection instruments such as eye tracker, electroencephalogram, electrocardiogram, etc., such as eye movement trajectory, electroencephalogram, electrocardiogram, etc.; fourth, the vehicle state parameters of the main vehicle itself will be recorded, such as the vehicle's speed, acceleration, yaw angular acceleration, lateral acceleration, etc. The above-obtained driving scene parameters, the behavior parameters of the driver of the main vehicle, the state parameters of the driver of the main vehicle, and the vehicle state parameters of the main vehicle are target information. The above collection can accurately record the occurrence process of dangerous scenes and the driver's status information before the accident.

[0076] Based on the driving scene parameters in the above target information and the behavior parameters of the main vehicle driver, the required target scene segments can be extracted. The target scene segments include: (1) Single vehicle control: high-speed cloverleaf ramp, high-speed long bridge deck, high-speed long tunnel, high-speed tire blowout and other dangerous scenes; (2) Single vehicle passing: passing through intersections with signal lights (including crossroads and T-junctions), passing through intersections without signal lights (including crossroads and T-junctions), city navigation (no other vehicles or pedestrians involved, testing vehicle navigation positioning, traffic sign and marking recognition, normal passage without deviation and other dangerous scene capabilities), obstacle recognition (static vehicles, traffic facilities such as road closure cones, guardrails), passing through isolation belts, pedestrians crossing in front and other dangerous scenes; (3) Multi-vehicle interaction: dangerous scenes such as the vehicle in front is stationary, the vehicle in front brakes urgently, the vehicle in front cuts in, the vehicle in front crosses and other dangerous scenes; (4) The main vehicle overtakes.

[0077] Furthermore, in one embodiment, before step S10, the following steps are included:

[0078] Acquiring preset information, wherein the preset information includes behavior parameters of the driver of the host vehicle and driving scenario parameters;

[0079] The preset information without the main vehicle positioning information or with abnormal sensor signals is filtered out to obtain the target information.

[0080] In this embodiment, before the step of extracting the target scene fragment based on the target information, it is also necessary to obtain preset information, which includes the behavioral parameters and driving scene parameters of the main vehicle driver collected during the entire process of the main vehicle collecting the natural driving scene. Since it is necessary to collect data parameters through sensors such as laser radar, video, positioning navigation, etc. during the entire process of the main vehicle collecting the natural driving scene, if the collection process encounters scenes such as multi-path effects of urban high-rise buildings, shade shading, tunnel shielding, etc., the positioning information will be lost, and the data parameters recorded in this section do not have positioning information. The subsequent scene cannot be accurately reproduced based on this section of data parameters, so it is necessary to exclude the above invalid data parameters. In addition, if the signal output by a certain sensor has abnormal recording such as noise or power failure, if this section of data parameters is used for the subsequent extraction of target scene fragments, corresponding errors will be generated. Therefore, after obtaining the above preset information, it is necessary to filter the preset information that does not have the main vehicle positioning information or sensor signal abnormality to obtain the target information, so as to facilitate the subsequent extraction of the target scene fragment based on the filtered target information.

[0081] Further, in one embodiment, the step S10 includes:

[0082] Screening out the behavior parameters of the driver of the host vehicle that meet the first preset threshold condition and the time segments of the driving scene parameters that meet the second preset threshold condition;

[0083] The time segment is taken as the target scene segment.

[0084] In this embodiment, specifically, after obtaining the target information, the step of extracting the scene segment using the target information includes: screening out the time segment of the driver's behavior parameters that meet the first preset threshold condition and the driving scene parameters that meet the second preset threshold condition, and taking the time segment as the target scene segment. Wherein, for different target scene segments, the first preset threshold condition and the second preset threshold condition are different.

[0085] Among them, for single-vehicle control: (1) High-speed up and down cloverleaf ramps: the first preset threshold condition is that the vehicle speed is greater than the threshold V0, the steering wheel angle is greater than the threshold θ0, and the yaw angular velocity threshold is greater than φ. All time segments that meet the above conditions are extracted, and the above time segments are used as the target scene segments of the high-speed on- and off-ramps; (2) High-speed long bridge deck passage: the first preset threshold condition is that the vehicle speed is greater than the threshold V0 and the lateral wind speed is greater than the rule V1. All time segments that meet the above conditions are extracted, and the above time segments are used as the target scene segments of the high-speed long bridge deck passage; (3) High-speed long tunnel: the second preset threshold condition is that the tunnel length is greater than the threshold L0. This scene mainly relies on the visual camera to identify the location of entering and exiting the tunnel and calculates the mileage of entering and exiting the tunnel by the vehicle speed to obtain the length of the tunnel. All time segments that meet the above conditions are extracted, and the above time segments are used as the target scene segments of the high-speed long tunnel; (4) High-speed tire blowout: the first preset threshold condition is that the tire pressure exceeds the preset threshold and there is a tire pressure alarm signal. All time segments that meet the above conditions are extracted, and the above time segments are used as the target scene segments of the high-speed tire blowout.

[0086] For single-vehicle traffic: (1) Intersection traffic: First, the vehicle positioning information that meets the second preset threshold condition in the driving scenario parameters is matched with the navigation map information to determine the approximate position of the vehicle. At the same time, the first preset threshold condition is that the vehicle speed is greater than the threshold V1 and the steering wheel angle is greater than the threshold θ1. All time segments that meet the above conditions are cut out, and the above time segments are used as target scene segments for intersection traffic; (2) Obstacle recognition: If it is determined that there are time segments that meet the second preset threshold conditions of the static traffic facility label in the driving scene parameters, all time segments that meet the above conditions are cut out, and the above time segments are used as target scene segments for ice cream cones and guardrails in front of the vehicle; (3) Passing through isolation belts: If it is determined that there are time segments that meet the second preset threshold conditions of the static isolation belt label in the driving scene parameters, all time segments that meet the above conditions are cut out, and the above time segments are used as target scene segments for intersection traffic; (4) Pedestrians crossing the front: If it is determined that there are time segments that meet the second preset threshold conditions of the pedestrian label and carry deceleration decision information in the driving scene parameters, all time segments that meet the above conditions are cut out, and the above time segments are used as target scene segments for pedestrians crossing the front of the vehicle.

[0087] For multi-vehicle interaction: (1) The vehicle ahead is stationary: The first preset threshold condition is that the braking speed and pedal pressure of the main vehicle are less than the corresponding thresholds. All time segments that meet the above conditions are cut out, and the above time segments are used as the target scene segments for the vehicle ahead to be stationary; (2) The vehicle ahead cuts in: The second preset threshold condition is to determine based on the driving scene parameters that the vehicle ahead has changed lanes (left lane change, the left front wheel and the right rear wheel cross the right lane line of the main vehicle). All time segments that meet the above conditions are cut out, and the above time segments are used as the target scene segments for the vehicle ahead to cut in; (3) The vehicle ahead crosses: The second preset threshold condition is that the speed vector of the vehicle ahead is perpendicular to the speed vector of the main vehicle. All time segments that meet the above conditions are cut out, and the above time segments are used as the target scene segments for the vehicle ahead to cross;

[0088] For the main vehicle overtaking: the second preset condition is that the main vehicle speed is greater than the first preset threshold and the steering wheel angle is greater than the second preset threshold. All time segments that meet the above conditions are cut out, and the above time segments are used as target scene segments for the main vehicle overtaking.

[0089] Step S20, generating a list of target objects in the target scene segment based on the driving scene parameters;

[0090] In this embodiment, after extracting the target scene segment based on the target information, a list of target objects in the target scene segment is generated based on the driving scene parameters in the target information. The driving scene parameters include parameters of target objects that need attention and have a corresponding impact on the driving of the main vehicle in the driving scene. The target objects include: lane lines, traffic signs, traffic lights, vehicles, pedestrians, and other traffic targets involved in traffic.

[0091] Further, in one embodiment, the step S20 includes:

[0092] Based on the driving scene parameters, the target objects in the target scene segment are annotated;

[0093] The same target object monitored by multiple sensors is fused, and based on the fusion result, a target object list is output, wherein the target object list includes: a timestamp of the detected target object, a target object type, and a state parameter of the target object.

[0094] In this embodiment, the driving scene parameters correspond to the images acquired by the camera and the point cloud data acquired by the laser radar sensor. Therefore, specifically, the step of generating the target object list in the target scene segment based on the driving scene parameters includes: based on the driving scene parameters, the target objects and semantic information in the target scene segment are annotated by offline automatic annotation + manual verification. Then, based on the visual data parameters of the semantic information and depth information, the same target object monitored by multiple sensors such as laser radar, camera, GPS / IMU, etc. is determined. Among them, the semantic information corresponds to the semantic features of the target object, and the depth information corresponds to the real three-dimensional coordinates corresponding to the pixel position of the target object in the image. The parameters of different dimensions of the same target object are fused. Then, based on the fusion result, the target object list is output. Among them, the target object list includes: the timestamp of the target object being monitored, the target object type and the state parameters of the target object.

[0095] Step S30, generating a target vehicle driving trajectory and a high-precision road map in a target scene segment based on the target object list;

[0096] In this embodiment, after the target object list in the target scene segment is generated based on the driving scene parameters, the target vehicle driving trajectory and high-precision road map in the target scene segment are generated based on the target object list. The generated target vehicle driving trajectory is used for subsequent analysis of the interaction risk between the host vehicle and the target vehicle in each target scene, and the generated high-precision map is used for subsequent mapping of the corresponding host vehicle and the target vehicle in the corresponding target scene for scene reproduction simulation deduction.

[0097] Further, in one embodiment, referring to Figure 3 , the step S30 comprises:

[0098] Step S301, determining a first target object in the target object list whose target object type is a traffic lane line or a road edge, combining the first target object with map data to generate a high-precision road map in the target scene segment;

[0099] Step S302: determine a second target object whose target object type is a vehicle in the target object list, and generate a target vehicle driving trajectory in the target scene segment by combining the second target object and a state parameter corresponding to the second target object.

[0100] In this embodiment, specifically, based on the target object list, the step of generating the target vehicle driving trajectory and the high-precision road map in the target scene segment includes: firstly determining the first target object whose target object type is a traffic lane line or a road edge based on the target object list, the first target object can be determined based on the image data parameters acquired by the camera, and then combining the first target object with the map data, that is, generating the high-precision road map in the target scene segment. At the same time, based on the target object list, determining the second target object whose target object type is a vehicle, combining the state parameters corresponding to the second target object, such as the position of the target vehicle from the main vehicle within a period of time, fitting the position points of each frame of the target object from the main vehicle, and generating the target vehicle driving trajectory in the target scene segment.

[0101] Step S40, weighting the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle;

[0102] In this embodiment, since the above-mentioned target scene segment is obtained based on the target information in the actual scene collection, in order to ensure driving safety, it can only approach the more dangerous driving state during the test operation, but it is impossible to obtain a more detailed and real target scene segment of the main vehicle in a more dangerous driving state based on the actual collection. Therefore, different weighting factors are superimposed on the original collected behavior parameters of the main vehicle driver, and more driving behavior parameters under the target scene segment are expanded with the weighted main vehicle driving behavior parameters. Among them, different weighting factors correspond to the aggressiveness of different drivers in braking, overtaking, and lane changing conditions. For example, they can be divided into soft type, general type, aggressive type, and abnormal aggressive type, and each aggressiveness corresponds to a weighting factor. That is, the behavior parameters of the main vehicle driver can be weighted based on the first preset number of weighting factors to generate the first preset number of main vehicle driving behavior parameters, thereby expanding to obtain more driving behavior parameters of the main vehicle under the target scene segment, which includes the driving behavior parameters of the main vehicle in a dangerous driving state. By superimposing weighting factors of different driving aggressiveness on the real collected driving scene data, it is possible to obtain dangerous scenes without artificially creating dangerous driving.

[0103] Step S50, weighting the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters;

[0104] In this embodiment, similarly, in the actual driving collection scenario, the driving trajectory of the target vehicle is also variable. In order to obtain more driving behavior parameters of the target vehicle in the target scene segment, it is also necessary to weight the parameters corresponding to the driving trajectory of the target vehicle based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters, thereby expanding to obtain more driving behavior parameters of the target vehicle in the target scene segment, including the driving behavior parameters of the target vehicle in a dangerous driving state.

[0105] Step S60, combining the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations;

[0106] In this embodiment, after obtaining the first preset number of driving behavior parameters of the host vehicle and the target vehicle in different target scene segments, it is necessary to combine them to obtain a driving behavior combination of multiple vehicles interacting in different target scene segments. The above driving behavior combination can be used to map to obtain a specific target scene. The above combination enriches the source of accident scenes, improves the efficiency of dangerous scene generation, and reduces costs.

[0107] Further, in one embodiment, the step S60 includes:

[0108] By traversing the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner, a third preset number of driving behavior combinations is obtained;

[0109] From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not satisfy the third preset threshold condition, or whose distance between the host vehicle and the target vehicle does not satisfy the fourth preset threshold condition are screened out to obtain the second preset number of driving behavior combinations.

[0110] In this embodiment, specifically, the step of combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain the second preset number of driving behavior combinations includes: traversing the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner to obtain a third preset number of driving behavior combinations. From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not meet the third preset threshold condition (such as the speed of the host vehicle does not exceed the speed of the target vehicle by 20 km / h) or whose distance between the host vehicle and the target vehicle does not meet the fourth preset threshold condition (the longitudinal distance between the two vehicles does not exceed 50m, and the lateral distance does not exceed 5m) are screened out to obtain the second preset number of driving behavior combinations.

[0111] Step S70, mapping the second preset number of driving behavior combinations to the high-precision road map to generate a target scene;

[0112] In this embodiment, the second preset number of driving behavior combinations obtained above is mapped to the high-precision road map, so that target scenarios with different driving aggressiveness levels can be generated.

[0113] Further, in one embodiment, referring to Figure 4 , the step S70 comprises:

[0114] Step S701, aligning the host vehicle driving behavior parameters and the target vehicle driving behavior parameters corresponding to each driving behavior combination with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene;

[0115] Step S702: Based on the starting coordinates, each driving behavior combination is mapped to the high-precision road map to generate a target scene.

[0116] In this embodiment, specifically, the step of mapping the second preset number of driving behavior combinations to the high-precision road map to generate the target scene includes: aligning the host vehicle driving behavior parameters and the target vehicle driving behavior parameters corresponding to each driving behavior combination with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene. Based on the starting coordinates, each driving behavior combination is mapped to the high-precision road map to generate the target scene.

[0117] Step S80, performing simulation replay on the target scene, with the target scene where an accident occurs and the target scene that is dangerous but no accident occurs in the simulation replay as the dangerous scene.

[0118] In this embodiment, after the target scenario is generated, the target scenario is simulated and replayed, and the target scenario where an accident occurs and the target scenario that is dangerous but does not cause an accident are regarded as dangerous scenarios. The dangerous scenarios include scenarios that do not cause traffic accidents but are dangerous driving scenarios.

[0119] The present embodiment provides a method for generating a dangerous scene, including: extracting a target scene segment based on target information, wherein the target information includes behavior parameters of a main vehicle driver and driving scene parameters; generating a list of target objects in the target scene segment based on the driving scene parameters; generating a target vehicle driving trajectory and a high-precision road map in the target scene segment based on the target object list; weighting the behavior parameters of the main vehicle driver based on a first preset number of weighting factors to generate a first preset number of main vehicle driving behavior parameters; weighting parameters corresponding to the target vehicle driving trajectory based on the first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters; combining the first preset number of main vehicle driving behavior parameters with the first preset number of target vehicle driving behavior parameters to obtain a second preset number of driving behavior combinations; mapping the second preset number of driving behavior combinations to the high-precision road map to generate a target scene; and performing simulation re-play on the target scene, with the target scene in which an accident occurs and the target scene in which an accident is dangerous but no accident occurs in the simulation re-play as the dangerous scene. The present invention superimposes weighted factors of different driving aggressiveness on the actually collected driving scene data, thereby making it possible to obtain dangerous scenes without artificially creating dangerous driving, enriching the sources of accident scenes, and solving the technical problem that the dangerous scene data obtained by the existing solutions have limitations, thereby improving the efficiency of dangerous scene generation and reducing costs.

[0120] In a third aspect, an embodiment of the present invention further provides a dangerous scene generating device.

[0121] Reference Figure 5 , a schematic diagram of functional modules of an embodiment of a dangerous scene generating device.

[0122] In this embodiment, the dangerous scene generating device includes:

[0123] An extraction module 10 is used to extract a target scene segment based on target information, wherein the target information includes a behavior parameter of a driver of a host vehicle and a driving scene parameter;

[0124] A first generating module 20, configured to generate a list of target objects in the target scene segment based on driving scene parameters;

[0125] A second generating module 30, for generating a target vehicle driving trajectory and a high-precision road map in a target scene segment based on the target object list;

[0126] A first weighting module 40, configured to weight the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle;

[0127] A second weighting module 50 is used to weight the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters;

[0128] A combination module 60, configured to combine the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations;

[0129] A third generating module 70, configured to map the second preset number of driving behavior combinations to the high-precision road map to generate a target scene;

[0130] The simulation module 80 is used to simulate and replay the target scene, with the target scene where an accident occurs and the target scene that is dangerous but no accident occurs in the simulation and replay being the dangerous scene.

[0131] Optionally, the dangerous scenario generating device further includes an acquisition module, which is used to:

[0132] Acquiring preset information, wherein the preset information includes behavior parameters of the driver of the host vehicle and driving scenario parameters;

[0133] The preset information without the main vehicle positioning information or with abnormal sensor signals is filtered out to obtain the target information.

[0134] Optionally, the extraction module 10 is specifically used for:

[0135] Screening out the behavior parameters of the driver of the host vehicle that meet the first preset threshold condition and the time segments of the driving scene parameters that meet the second preset threshold condition;

[0136] The time segment is taken as the target scene segment.

[0137] Optionally, the first generating module 20 is specifically configured to:

[0138] Based on the driving scene parameters, the target objects in the target scene segment are annotated;

[0139] The same target object monitored by multiple sensors is fused, and based on the fusion result, a target object list is output, wherein the target object list includes: a timestamp of the detected target object, a target object type, and a state parameter of the target object.

[0140] Optionally, the second generating module 30 is specifically configured to:

[0141] Determine a first target object in the target object list whose target object type is a traffic lane line or a road edge, combine the first target object with the map data, and generate a high-precision road map in the target scene segment;

[0142] A second target object whose target object type is a vehicle in the target object list is determined, and a target vehicle driving trajectory in the target scene segment is generated by combining the second target object and a state parameter corresponding to the second target object.

[0143] Optionally, the combination module 60 is specifically used for:

[0144] By traversing the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner, a third preset number of driving behavior combinations is obtained;

[0145] From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not satisfy the third preset threshold condition, or whose distance between the host vehicle and the target vehicle does not satisfy the fourth preset threshold condition are screened out to obtain the second preset number of driving behavior combinations.

[0146] Optionally, the third generating module 70 is specifically configured to:

[0147] The driving behavior parameters of the host vehicle and the target vehicle corresponding to each driving behavior combination are aligned with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene;

[0148] Based on the starting coordinates, each driving behavior combination is mapped into the high-precision road map to generate a target scene.

[0149] Among them, the functional implementation of each module in the above-mentioned dangerous scene generation device corresponds to the various steps in the above-mentioned dangerous scene generation method embodiment, and its functions and implementation processes will not be repeated here one by one.

[0150] In a fourth aspect, an embodiment of the present invention further provides a readable storage medium.

[0151] The readable storage medium of the present invention stores a dangerous scenario generation program, wherein when the dangerous scenario generation program is executed by a processor, the steps of the dangerous scenario generation method as described above are implemented.

[0152] Among them, the method implemented when the dangerous scene generation program is executed can refer to the various embodiments of the dangerous scene generation method of the present invention, and will not be repeated here.

[0153] 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 existence of other identical elements in the process, method, article or system including the element.

[0154] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0155] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present invention.

[0156] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for generating a dangerous scene, characterized in that: The dangerous scene generation method comprises: Extracting target scene segments based on target information, wherein the target information includes behavior parameters of the driver of the host vehicle and driving scene parameters; Generating a list of target objects in the target scene segment based on driving scene parameters; Based on the target object list, generating a target vehicle driving trajectory and a high-precision road map in a target scene segment; Weighting the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle; Weighting the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters; Combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations; Mapping the second preset number of driving behavior combinations into the high-precision road map to generate a target scene; The target scenario is simulated and replayed, and the target scenario where an accident occurs and the target scenario that is dangerous but no accident occurs in the simulation and replay are regarded as dangerous scenarios.

2. The method for generating a dangerous scene according to claim 1, characterized in that: Before the step of extracting the target scene fragment based on the target information, the method includes: Acquiring preset information, wherein the preset information includes behavior parameters of the driver of the host vehicle and driving scenario parameters; The preset information without the main vehicle positioning information or with abnormal sensor signals is filtered out to obtain the target information.

3. The method for generating a dangerous scene according to claim 1, characterized in that: The step of extracting the target scene fragment based on the target information comprises: Screening out the behavior parameters of the driver of the host vehicle that meet the first preset threshold condition and the time segments of the driving scene parameters that meet the second preset threshold condition; The time segment is taken as the target scene segment.

4. The method for generating a dangerous scene according to claim 1, characterized in that: The step of generating a list of target objects in the target scene segment based on the driving scene parameters comprises: Based on the driving scene parameters, the target objects in the target scene segment are annotated; The same target object monitored by multiple sensors is fused, and based on the fusion result, a target object list is output, wherein the target object list includes: a timestamp of the detected target object, a target object type, and a state parameter of the target object.

5. The method for generating a dangerous scene according to claim 4, characterized in that: The step of generating the target vehicle driving trajectory and the high-precision road map in the target scene segment based on the target object list includes: Determine a first target object in the target object list whose target object type is a traffic lane line or a road edge, combine the first target object with the map data, and generate a high-precision road map in the target scene segment; A second target object whose target object type is a vehicle in the target object list is determined, and a target vehicle driving trajectory in the target scene segment is generated by combining the second target object and a state parameter corresponding to the second target object.

6. The method for generating a dangerous scene according to claim 1, characterized in that: The step of combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations includes: By traversing the first preset number of driving behavior parameters of the host vehicle and the first preset number of driving behavior parameters of the target vehicle in an evenly spaced arrangement and combination manner, a third preset number of driving behavior combinations is obtained; From the third preset number of driving behavior combinations, the driving behavior combinations whose speed difference between the host vehicle and the target vehicle does not satisfy the third preset threshold condition, or whose distance between the host vehicle and the target vehicle does not satisfy the fourth preset threshold condition are screened out to obtain the second preset number of driving behavior combinations.

7. The method for generating a dangerous scene according to claim 1, characterized in that: The step of mapping the second preset number of driving behavior combinations to the high-precision road map to generate a target scene includes: The driving behavior parameters of the host vehicle and the target vehicle corresponding to each driving behavior combination are aligned with the coordinates of the target scene segment to obtain the starting coordinates of the corresponding scene; Based on the starting coordinates, each driving behavior combination is mapped into the high-precision road map to generate a target scene.

8. A dangerous scene generating device, characterized in that: The dangerous scene generating device comprises: An extraction module, used to extract a target scene segment based on target information, wherein the target information includes a behavior parameter of a driver of a host vehicle and a driving scene parameter; A first generating module, configured to generate a list of target objects in the target scene segment based on driving scene parameters; A second generating module, for generating a target vehicle driving trajectory and a high-precision road map in a target scene segment based on the target object list; A first weighting module, used for weighting the behavior parameters of the driver of the host vehicle based on a first preset number of weighting factors to generate a first preset number of driving behavior parameters of the host vehicle; A second weighting module, used to weight the parameters corresponding to the target vehicle driving trajectory based on a first preset number of weighting factors to generate a first preset number of target vehicle driving behavior parameters; A combination module, used for combining the first preset number of driving behavior parameters of the host vehicle with the first preset number of driving behavior parameters of the target vehicle to obtain a second preset number of driving behavior combinations; A third generating module, used for mapping the second preset number of driving behavior combinations into the high-precision road map to generate a target scene; The simulation module is used to simulate and replay the target scene, with the target scene where an accident occurs and the target scene that is dangerous but no accident occurs in the simulation and replay as dangerous scenes.

9. A dangerous scene generation device, characterized in that: The dangerous scenario generation device includes a processor, a memory, and a dangerous scenario generation program stored in the memory and executable by the processor, wherein when the dangerous scenario generation program is executed by the processor, the steps of the dangerous scenario generation method as described in any one of claims 1 to 7 are implemented.

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

Images