Vehicle simulation scene modeling method and device, computer equipment and storage medium

Abstract

The invention provides a vehicle simulation scene modeling method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining static scene data, and constructing a static scene model according to the static scene data; loading the static scene model and the vehicle kinematics model to construct a simulation initial model; obtaining dynamic environment data, and performing coordinate conversion on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model; obtaining dynamic main vehicle data, and driving the vehicle through the dynamic main vehicle data based on the simulation initial model; detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle reaching the preset speed; and generating a dynamic scene model according to the dynamic environment model and the dynamic main vehicle model. And dynamic scene model reconstruction based on data playback realizes dynamic traffic scene reconstruction and fault scene reproduction, thereby realizing data-driven modeling of a vehicle intelligent driving key scene.

Description

Technical Field

[0001] The present invention relates to the technical field of simulation modeling, and in particular to a vehicle simulation scene modeling method, device, computer equipment and storage medium. Background Art

[0002] With the development of the economy, especially the development of the automobile industry, the number of vehicles in society is increasing. With the improvement of people's living standards, it is very common to use cars as a means of transportation, and the configuration requirements for vehicles are getting higher and higher. At present, the driving mode of vehicles has shifted from manned driving to unmanned driving, mainly through unmanned driving algorithms. If you want to reach the level of human drivers, you need to improve the unmanned driving algorithm through a large amount of data. At least 17.7 billion kilometers of driving data are needed to improve the algorithm. If a fleet of 100 self-driving test vehicles is configured, and road tests are carried out 24 hours a day without stopping, with an average speed of 25 miles (40 kilometers) per hour, it will take more than 500 years to complete the target mileage. From the above, it can be seen that the time and cost spent during this period are unbearable.

[0003] Among the related technologies, simulation testing based on virtual scenarios is an important way to solve the high cost, long cycle and low coverage of unmanned driving tests. By constructing virtual scenarios, closed-loop simulation testing of autonomous driving perception, decision-making planning, control and other algorithms can be achieved. Among them, the simulation test of unmanned vehicles first needs to simulate and build a vehicle operation scenario that is consistent with the real world, and operate the unmanned vehicle in the vehicle operation scenario. However, when key scenarios such as being taken over by a safety officer and abnormal parking of autonomous driving appear during the autonomous driving process, it is particularly important to realize the dynamic and static scene modeling of these key scenarios and apply them to simulation testing of data closed loops. Summary of the invention

[0004] Based on this, it is necessary to provide a vehicle simulation scenario modeling method, device, computer equipment and storage medium to address the above technical problems.

[0005] A vehicle simulation scene modeling method, comprising:

[0006] Acquire static scene data, and construct a static scene model according to the static scene data;

[0007] Loading the static scene model and the vehicle kinematics model to construct an initial simulation model;

[0008] Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model;

[0009] Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model;

[0010] Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed;

[0011] A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

[0012] In one embodiment, the step of acquiring static scene data and constructing a static scene model according to the static scene data includes:

[0013] Collecting characteristic parameter data of preset road marking lines, performing segmented fitting on the characteristic parameter data to obtain multiple segments of preview paths, wherein the preview paths have multiple discrete points;

[0014] Integrating the discrete points of each preview path to form a discrete point sample set;

[0015] The discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve.

[0016] In one embodiment, the step of acquiring static scene data and constructing a static scene model according to the static scene data further includes:

[0017] Preprocessing the characteristic parameter data;

[0018] Detecting whether the number of the characteristic parameter data is equal to a preset number within a preset time, and when the number of the characteristic parameter data is equal to the preset number within the preset time, performing global fitting on the discrete point sample set to obtain a road curve;

[0019] A static scene model is constructed based on the road curve.

[0020] In one embodiment, the step of acquiring dynamic environment data, performing coordinate transformation on the dynamic environment data based on the simulation initial model, and obtaining a dynamic environment model comprises:

[0021] Acquire the dynamic environment data, arrange the dynamic environment data according to the acquisition time, and obtain dynamic environment sequence data;

[0022] Performing coordinate transformation on the dynamic environment sequence data to generate dynamic environment element data;

[0023] Inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes;

[0024] sequentially detecting whether each of the first simulated dynamic scenes is consistent with a preset scene, and when the first simulated dynamic scene is consistent with the preset scene, determining that the first simulated dynamic scene is a first prepared dynamic scene;

[0025] The dynamic environment model is constructed based on the plurality of the first prepared dynamic scenes.

[0026] In one embodiment, the step of inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes includes:

[0027] inputting the initial values ​​of the dynamic environment element data into the simulation initial model, and operating the vehicle in the simulation initial model;

[0028] Detect whether the vehicle state reaches a preset state, and when the vehicle state reaches the preset state, obtain the state data of the vehicle in the simulation initial model, thereby generating a plurality of first simulation dynamic scenes.

[0029] In one embodiment, after the step of sequentially detecting whether each of the first simulated dynamic scenes is consistent with the preset scene, the method further includes:

[0030] When the first simulation dynamic scene does not match the preset scene, determining that the dynamic environment element data is element data to be supplemented, and supplementing the element data to be supplemented;

[0031] Using the supplemented dynamic environment element data to input into the simulation initial model to generate a plurality of second simulation dynamic scenes;

[0032] It is detected in turn whether each of the second simulated dynamic scenes is consistent with the preset scene. When the second simulated dynamic scene is consistent with the preset scene, the dynamic environment model is constructed based on multiple second simulated dynamic scenes and the first prepared dynamic scene.

[0033] In one embodiment, the step of obtaining dynamic main vehicle data and driving the vehicle by the dynamic main vehicle data based on the simulation initial model comprises:

[0034] Acquire the dynamic main vehicle data, arrange the dynamic main vehicle data according to the acquisition time, and obtain dynamic main vehicle sequence data;

[0035] Establishing a vehicle acceleration model according to the dynamic main vehicle sequence data;

[0036] The vehicle is driven by the vehicle acceleration model.

[0037] A vehicle simulation scene modeling device, comprising:

[0038] Static data acquisition module, used to acquire static scene data;

[0039] A static model building module, used for building a static scene model according to static scene data;

[0040] An initial model building module, used to load the static scene model and the vehicle kinematics model to build an initial simulation model;

[0041] A dynamic data acquisition module is used to acquire dynamic environment data and dynamic main vehicle data;

[0042] A dynamic model building module is used to build a dynamic scene model based on the simulation initial model according to the dynamic environment data and the dynamic main vehicle data.

[0043] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the following steps when executing the computer program:

[0044] Acquire static scene data, and construct a static scene model according to the static scene data;

[0045] Loading the static scene model and the vehicle kinematics model to construct an initial simulation model;

[0046] Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model;

[0047] Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model;

[0048] Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed;

[0049] A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

[0050] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0051] Acquire static scene data, and construct a static scene model according to the static scene data;

[0052] Loading the static scene model and the vehicle kinematics model to construct an initial simulation model;

[0053] Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model;

[0054] Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model;

[0055] Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed;

[0056] A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

[0057] The above-mentioned vehicle simulation scene modeling method, device, computer equipment and storage medium obtain static scene data, construct a static scene model based on the static scene data, and load the static scene model and the vehicle kinematic model into the simulation platform to construct an initial simulation model. In addition, dynamic environment data and dynamic main vehicle data are obtained, and on the basis of the initial simulation model, a dynamic environment model is constructed according to the dynamic environment data, a dynamic main vehicle model is constructed according to the dynamic main vehicle data, and then a dynamic scene model is generated according to the dynamic environment model and the dynamic main vehicle model. The dynamic scene model reconstruction based on the playback of radar recognition result data realizes the reconstruction of dynamic traffic scenes and the reproduction of fault scenes, thereby realizing data-driven modeling of key scenes of vehicle intelligent driving, improving the authenticity of scene modeling, and expanding the effective data source of intelligent driving simulation test scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of a process flow of a vehicle simulation scenario modeling method in one embodiment;

[0059] Figure 2 is a structural block diagram of a vehicle simulation scene modeling device in one embodiment;

[0060] Figure 3 is a diagram of the internal structure of a computer device in one embodiment;

[0061] Figure 4 A schematic diagram of a process for constructing a static scene model in an embodiment;

[0062] Figure 5 A schematic diagram of a process for constructing an initial simulation model in an embodiment;

[0063] Figure 6 A schematic diagram of a process for constructing a dynamic environment model in an embodiment;

[0064] Figure 7 A schematic diagram of a process for constructing a dynamic main vehicle model in an embodiment. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] Embodiment 1

[0067] In this embodiment, Figure 1 As shown, a vehicle simulation scenario modeling method is provided, which includes:

[0068] Step 110: Acquire static scene data, and construct a static scene model according to the static scene data.

[0069] In this embodiment, static scene data refers to environmental data in the scene where the vehicle is traveling, for example, the static scene data includes road width, road marking lines, road length, curvature, etc. Static scene data is mainly obtained from road marking lines collected during vehicle driving, including parameters such as the position, length, width, curvature, etc. of road marking lines. In practical applications, by installing monitoring equipment such as a camera in front of the vehicle and cooperating with a positioning sensor, the camera continuously collects characteristic parameter data of different positions of the road marking lines when the vehicle is traveling on the road to generate static scene data, and the static scene data is input into the simulation platform to construct a static scene model. In some embodiments, static scene data can be obtained through sensors such as high-precision maps and laser radars. In this embodiment, static scene data is collected based on the camera and a static scene model is constructed to reconstruct the road marking lines on the simulation platform. The static scene model reconstruction based on the playback of the camera recognition result data reduces the reliance on high-precision maps or preset static scene models in the prior art, and realizes the reproduction of static scenes.

[0070] Step 120, loading the static scene model and the vehicle kinematic model to construct an initial simulation model.

[0071] After the static scene model is constructed, the static scene model and the vehicle kinematic model are loaded on the simulation platform, so that the simulation platform enters initialization to construct the initial simulation model. In this embodiment, the vehicle kinematic model describes the motion characteristics of the vehicle in a specific coordinate system. Different coordinate systems can be selected according to the application scenario and requirements, such as the global coordinate system or the road coordinate system. In the field of autonomous driving, the vehicle kinematic model is used to plan and predict the motion trajectory of the vehicle, as well as to control the driving of the vehicle under various road and traffic conditions. This is a prior art, so it is not described in detail here.

[0072] Step 130 , acquiring dynamic environment data, and performing coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model.

[0073] In this embodiment, the main vehicle is a vehicle in an unmanned driving state for which a simulation model needs to be established at present, and the slave vehicle refers to other vehicles on the same road or in the same traffic flow as the main vehicle. It can be understood that the dynamic environment data refers to the vehicle parameter data collected from the slave vehicle during driving, and the dynamic environment data is obtained from the slave vehicle during driving. Among them, the dynamic environment data includes real-time data such as the position, speed, acceleration, steering angle, roll angle, etc. of the slave vehicle, as well as vehicle state estimation data (such as perception and judgment of the environment). The dynamic environment data can be obtained through monitoring equipment. When obtaining the dynamic environment data, the slave vehicle during driving can be detected by laser radar, GPS positioning system, camera, sensor and other equipment to obtain dynamic environment data. For example, the main vehicle and the taxi have a scratch, which is regarded as a fault scene. Among them, the taxi and other vehicles in the traffic flow are slave vehicles, and the parameter data of the slave vehicle is obtained. The parameter data of the slave vehicle is the dynamic environment data to be obtained. After obtaining the dynamic environment data, the dynamic environment data is coordinate-converted on the simulation platform based on the constructed initial simulation model, and the dynamic environment data after coordinate conversion is input into the slave vehicle in the simulation platform, so that the slave vehicle runs on the simulation platform. At this time, the dynamic environment model is obtained according to the parameters of the slave vehicle.

[0074] Step 140 , obtaining dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model.

[0075] In this embodiment, the dynamic main vehicle data refers to the vehicle parameter data collected from the main vehicle in the unmanned driving state during the driving process. The dynamic main vehicle data is obtained from the main vehicle during the driving process. The dynamic main vehicle data includes the real-time data such as the position, speed, acceleration, steering angle, roll angle, and vehicle state estimation data (such as the perception and judgment of the environment). The dynamic main vehicle data can be obtained through monitoring equipment. When obtaining the dynamic main vehicle data, based on the fusion data of the real vehicle global navigation satellite system (GNSS) and the inertial navigation system, the vehicle state parameters are collected to obtain the dynamic main vehicle data. It should be noted that the fusion data of the global navigation satellite system (GNSS) and the inertial navigation system can provide high-precision positioning and navigation information. In the field of unmanned driving, the fusion of GNSS and inertial navigation systems can provide more accurate parking space information, helping vehicles to perform high-precision navigation in parking lots or in complex urban environments. In addition, during the driving process of the unmanned vehicle, this fusion data can also provide real-time information such as the speed, direction, acceleration, etc. of the vehicle to help the vehicle make autonomous decisions and control.

[0076] In this embodiment, after obtaining the dynamic main vehicle data, based on the constructed initial simulation model, the vehicle is driven according to the dynamic main vehicle data so that the vehicle performs kinematic simulation in the simulation platform. It should be noted that the vehicle here is the main vehicle for simulation.

[0077] Step 150 , detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed.

[0078] In this embodiment, after obtaining the dynamic main vehicle data, the main vehicle is driven according to the dynamic main vehicle data, so that the main vehicle performs kinematic simulation in the simulation platform. It should be understood that in this process, the speed of the main vehicle needs to be detected. When the speed of the main vehicle reaches the preset speed, the dynamic main vehicle model can be constructed according to the parameters of the main vehicle that reaches the preset speed. When the dynamic main vehicle data is collected from the main vehicle, when the actual collected data has data truncation, the initial state of some dynamic information represented by speed may not be reached instantly at the initialization start time, and cannot be directly input into the simulation platform. Therefore, it is necessary to detect whether the main vehicle reaches the preset speed required by the simulation platform. When the speed of the main vehicle reaches the preset speed, the dynamic main vehicle model can be constructed according to the parameters of the main vehicle that reaches the preset speed. When the speed of the main vehicle does not reach the preset speed, it is necessary to drive the main vehicle through the dynamic main vehicle data based on the simulation initial model in step 140 until the speed of the main vehicle reaches the preset speed, and then construct the dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed.

[0079] Step 160: Generate a dynamic scene model according to the dynamic environment model and the dynamic host vehicle model.

[0080] In this embodiment, after obtaining the dynamic environment model and the dynamic main vehicle model, the dynamic environment model and the dynamic main vehicle model are loaded on the simulation platform to generate a dynamic scene model. The initial simulation model is constructed according to the static scene model and the vehicle kinematic model. On the basis of the initial simulation model, the dynamic environment data is collected by the laser radar and a dynamic environment model is constructed to reconstruct the slave vehicle on the simulation platform, and the dynamic main vehicle data is collected to construct a dynamic main vehicle model. The dynamic scene model is reconstructed based on the playback of radar recognition result data to achieve dynamic traffic scene reconstruction and fault scene reproduction, thereby realizing data-driven modeling of key scenes of vehicle intelligent driving, improving the authenticity of scene modeling, and expanding the effective data source of intelligent driving simulation test scenes.

[0081] In the above embodiment, by acquiring static scene data, a static scene model is constructed according to the static scene data. After obtaining the static scene model, the static scene model and the vehicle kinematic model are loaded into the simulation platform to construct the initial simulation model. The dynamic environment data and the dynamic main vehicle data are acquired. On the basis of the initial simulation model, the dynamic environment data is converted into coordinates to obtain the dynamic environment model. On the basis of the initial simulation model, the main vehicle is driven according to the dynamic main vehicle data, and it is detected whether the main vehicle reaches the preset speed required by the simulation platform. When the speed of the main vehicle reaches the preset speed, the dynamic main vehicle model can be constructed according to the parameters of the main vehicle that reaches the preset speed. Finally, a dynamic scene model is generated according to the constructed dynamic environment model and the dynamic main vehicle model. The dynamic scene model reconstruction based on the playback of radar recognition result data realizes the reconstruction of dynamic traffic scenes and the reproduction of fault scenes, thereby realizing data-driven modeling of key scenes of vehicle intelligent driving, improving the authenticity of scene modeling, and expanding the effective data source of intelligent driving simulation test scenes.

[0082] In one embodiment, the step of acquiring static scene data and constructing a static scene model according to the static scene data includes:

[0083] Collecting characteristic parameter data of preset road marking lines, performing segmented fitting on the characteristic parameter data to obtain multiple segments of preview paths, wherein the preview paths have multiple discrete points;

[0084] Integrating the discrete points of each preview path to form a discrete point sample set;

[0085] The discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve.

[0086] In this embodiment, the characteristic parameter data of the road marking line can be collected by installing a camera in front of the vehicle and integrating a positioning sensor, so that the camera collects different positions of the road marking line when the vehicle is driving on the road, and the characteristic parameter data of the road marking line is calculated after the camera recognizes it. It can also be considered that during the driving process of the automatic driving, the road marking line is scanned and detected by using sensors and algorithms to obtain the characteristic parameter data of the road marking line and predict the vehicle driving path. For each coordinate position of the main vehicle and the characteristic parameter data of the road marking line recognized by the corresponding camera, the road marking line is segmented and fitted according to the preset distance to obtain multiple segments of the preview path. The segmented fitting is to divide the characteristic parameter data into multiple different segments, and then fit each segment. Its main purpose is to fit a set of discrete data points in the form of a piecewise function, and make the fitting result conform to the characteristics of the original data to the greatest extent. Therefore, each segment of the preview path has multiple discrete data points (discrete points), and then the discrete points of all the preview paths are integrated to form a discrete point sample set of the static scene model, and then the discrete point sample set is globally fitted to obtain the road curve. Global fitting refers to a data fitting technology that can perform overall fitting in a set of data at the same time. Specifically, global fitting first requires merging multiple data sets into one data set, and then using certain mathematical methods to perform overall fitting on this data set to obtain a fitting function or model. Specifically, global fitting of discrete point sample sets can be performed by fitting the discrete sample point sets with a smooth spline curve fitting method to obtain a road curve. It should be noted that before performing segmented fitting on the feature parameter data, the feature parameter data needs to be converted into coordinates. When the camera collects feature parameter data, it is collected based on the camera's coordinate system, that is, with the camera's position as the coordinate origin. The collected data is the relative position of the road marking line with the camera's position as the coordinate origin. Therefore, before performing segmented fitting, the feature parameter data collected by the camera needs to be converted into coordinates, that is, converted into the vehicle coordinate system, that is, with the vehicle's front axle or rear axle as the coordinate origin. The conversion can be performed using existing mathematical formulas.

[0087] In this embodiment, after obtaining the road curve, a static scene model is constructed based on the road curve according to the OPENDRIVE format. The OPENDRIVE format is a format for describing a static road traffic network required for an autonomous driving simulation application, which mainly covers the description of modeling contents such as roads, lanes, and intersections.

[0088] In one embodiment, the step of acquiring static scene data and constructing a static scene model according to the static scene data further includes:

[0089] Preprocessing the characteristic parameter data;

[0090] Detecting whether the number of the characteristic parameter data is equal to a preset number within a preset time, and when the number of the characteristic parameter data is equal to the preset number within the preset time, performing global fitting on the discrete point sample set to obtain a road curve;

[0091] A static scene model is constructed based on the road curve.

[0092] In this embodiment, after the characteristic parameter data of the preset road marking line is collected by the camera, the characteristic parameter data needs to be preprocessed. The preprocessing includes: filtering the characteristic parameter data, performing cleaning processes such as deduplication, outlier removal, and downsampling on the filtered characteristic parameter data to eliminate data errors and noise and improve the accuracy and reliability of the data.

[0093] It should be noted that the filtering process needs to filter the feature parameter data according to the effective flag of the camera. Before filtering the feature parameter data according to the effective flag of the camera, the data credibility of the feature parameter data needs to be evaluated. When evaluating the credibility of the data, first, verify the data source: it is necessary to confirm whether the source of the data is reliable; second, check the data quality: it is necessary to evaluate the quality and clarity of the data; then, consider the data deviation: it is necessary to evaluate whether there is a deviation that may affect the credibility of the data. Next, compare with other data sources: if possible, the credibility can be evaluated by comparing data from other sources. Finally, apply statistical methods: some statistical methods can be used to evaluate the credibility of the data. It should be understood that the process of filtering the feature parameter data according to the effective flag of the camera can include the following steps: first, collect data: obtain image or video data from the camera, and record the effective flag information of each data point. Secondly, determine the filtering criteria: according to the application scenario and target, determine the criteria for filtering the feature parameters, which may include the range, change trend, correlation, etc. of the parameters. Then, extract the feature parameters: pre-process each data point to extract the feature parameters related to the target, which may include color, texture, shape, etc. Next, filter the feature parameters: filter each feature parameter according to the valid flag and filtering criteria. For example, if the valid flag indicates that the camera view corresponding to the data point is partially blocked, then some feature parameters of the data point can be ignored. Finally, process the filtered data: process the filtered feature parameters to achieve the application goal. For example, the filtered feature parameters can be used for object recognition, behavior analysis, scene understanding, etc.

[0094] In this embodiment, during the preprocessing of the characteristic parameter data, if it is found that the characteristic parameter data is partially missing due to factors such as backlighting and blurred lane lines, it is necessary to detect whether the preprocessed characteristic parameter data is missing. Specifically, by detecting whether the number of characteristic parameter data within the preset time is equal to the preset number, when the number of characteristic parameter data within the preset time is equal to the preset number, the discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve according to the OPENDRIVE format. However, when the number of characteristic parameter data within the preset time is not equal to the preset number, multi-sensor data interpolation filling and abnormal scene reconstruction are performed. First, the characteristic parameter data within the preset time is interpolated and supplemented according to the existing characteristic parameter data, and then the positioning sensor data is fused. For the road marking line with missing data, the vehicle driving path coordinate data is used as the reference path, and the interpolation data is corrected to ensure that the path filling fitting error is within the allowable range. After that, the road marking line with missing data can also refer to the on-site cause analysis, and abnormal scene models such as blurred or lost lane lines and rutting interference can be constructed in the simulation scene to restore the actual static scene in the static scene model. After interpolation, supplementation and correction processing are performed on the characteristic parameter data, global fitting is performed using the supplemented characteristic parameter data to obtain the road curve, and then a static scene model is constructed based on the road curve in accordance with the OPENDRIVE format.

[0095] like Figure 6 As shown, in one embodiment, the steps of acquiring dynamic environment data, performing coordinate transformation on the dynamic environment data based on the simulation initial model, and obtaining a dynamic environment model include:

[0096] Acquire the dynamic environment data, arrange the dynamic environment data according to the acquisition time, and obtain dynamic environment sequence data;

[0097] Performing coordinate transformation on the dynamic environment sequence data to generate dynamic environment element data;

[0098] Inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes;

[0099] sequentially detecting whether each of the first simulated dynamic scenes is consistent with a preset scene, and when the first simulated dynamic scene is consistent with the preset scene, determining that the first simulated dynamic scene is a first prepared dynamic scene;

[0100] The dynamic environment model is constructed based on the plurality of the first prepared dynamic scenes.

[0101] In this embodiment, after acquiring dynamic environment data through laser radar collection, the dynamic environment data is arranged according to the time series based on the simulation platform model, and the time series here can be the acquisition time, and the dynamic environment data is preprocessed using the nearest neighbor data association algorithm to generate dynamic environment sequence data. After obtaining the dynamic environment sequence data, the dynamic environment sequence data is subjected to coordinate transformation according to the global coordinate system, and dynamic environment element data is generated according to the OpenScenario standard, and the global coordinate system here is the coordinate system of the vehicle. Then, the dynamic environment element data is input into the simulation initial model to generate the first simulation dynamic scene, and then each first simulation dynamic scene is detected in turn to be consistent with the preset scene. When the first simulation dynamic scene is consistent with the preset scene, the first simulation dynamic scene is determined to be the first preparatory dynamic scene, and a dynamic environment model is constructed based on multiple first preparatory dynamic scenes to restore the actual dynamic scene in the dynamic environment model. For example, in practice, an unmanned vehicle and a taxi have a scratch on the road, the unmanned vehicle is the main vehicle, and the taxi is the slave vehicle. The parameter data of the slave vehicle is collected as the dynamic environment data, and the dynamic environment data is input into the simulation initial model to generate the first simulation dynamic scene. During the simulation on the simulation platform, if it is found that the first simulated dynamic scene is consistent with the preset scene (a collision between an unmanned vehicle and a taxi on the road), the first simulated dynamic scene is determined to be the first prepared dynamic scene.

[0102] In one embodiment, the step of inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes includes:

[0103] inputting the initial values ​​of the dynamic environment element data into the simulation initial model, and operating the vehicle in the simulation initial model;

[0104] Detect whether the vehicle state reaches a preset state, and when the vehicle state reaches the preset state, obtain the state data of the vehicle in the simulation initial model, thereby generating a plurality of first simulation dynamic scenes.

[0105] In this embodiment, in the collection of dynamic environment data, when there is data truncation in the actual collected data, the initial state of some dynamic information such as speed as a typical representative may not be reached instantly at the beginning of initialization, and cannot be directly input into the simulation platform. Therefore, the dynamic environment element data needs to be input into the simulation initial model. Specifically, a scene initialization section can be added to the static scene model, and the initial value of the dynamic environment element data is first input into the simulation initial model, and the initial value of the dynamic environment element data is used as the control target to pre-prepare the slave vehicle in the initialization section. The pre-preparation process here is to set the starting point position and posture of the slave vehicle in the simulation scene. After the slave vehicle is pre-prepared, it is necessary to detect whether the state of the slave vehicle reaches the preset state. When the state of the slave vehicle reaches the preset state, switch back to control the slave vehicle with the dynamic environment element data as the control target, and obtain the state data of the slave vehicle in the simulation initial model, thereby generating multiple first simulation dynamic scenes, and realizing the restoration of the actual dynamic scene in the dynamic environment model.

[0106] In one embodiment, after the step of sequentially detecting whether each of the first simulated dynamic scenes is consistent with the preset scene, the method further includes:

[0107] When the first simulation dynamic scene does not match the preset scene, determining that the dynamic environment element data is element data to be supplemented, and supplementing the element data to be supplemented;

[0108] Using the supplemented dynamic environment element data to input into the simulation initial model to generate a plurality of second simulation dynamic scenes;

[0109] It is detected in turn whether each of the second simulated dynamic scenes is consistent with the preset scene. When the second simulated dynamic scene is consistent with the preset scene, the dynamic environment model is constructed based on multiple second simulated dynamic scenes and the first prepared dynamic scene.

[0110] In this embodiment, when the first simulated dynamic scene does not match the preset scene, it is necessary to perform multi-source data supplementation on the dynamic environment element data. First, determine that the dynamic environment element data is the element data to be supplemented, perform data supplementation on the element data to be supplemented, and then input the supplemented dynamic environment element data into the simulated initial model to generate multiple second simulated dynamic scenes, and then sequentially detect whether each second simulated dynamic scene matches the preset scene. When the second simulated dynamic scene matches the preset scene, a dynamic environment model is constructed based on the multiple second simulated dynamic scenes and the first prepared dynamic scene to restore the actual dynamic scene in the dynamic environment model.

[0111] like Figure 7As shown, in one embodiment, the step of acquiring dynamic main vehicle data and driving the vehicle through the dynamic main vehicle data based on the simulation initial model includes:

[0112] Acquire the dynamic main vehicle data, arrange the dynamic main vehicle data according to the acquisition time, and obtain dynamic main vehicle sequence data;

[0113] Establishing a vehicle acceleration model according to the dynamic main vehicle sequence data;

[0114] The vehicle is driven by the vehicle acceleration model.

[0115] In this embodiment, after obtaining the dynamic main vehicle data, the dynamic main vehicle data is arranged according to the acquisition time based on the simulation initial model, and the dynamic main vehicle data is pre-processed by data cleaning, data format consistency processing, missing data fitting interpolation and other pre-processing work, so as to generate dynamic main vehicle sequence data. A vehicle acceleration model is established according to the dynamic main vehicle sequence data to drive the vehicle kinematic model in the simulation initial model. Based on the simulation initial model, the main vehicle is switched to the manual driving mode, and the main vehicle is driven by the vehicle acceleration model to realize the kinematic simulation of the main vehicle in the simulation platform. Subsequently, it is detected whether the speed of the main vehicle reaches the preset speed. When the speed of the main vehicle reaches the preset speed, the main vehicle is switched to the automatic driving mode, and the dynamic main vehicle model is constructed according to the parameters of the main vehicle switched to the automatic driving mode. Finally, the dynamic scene model is constructed according to the dynamic environment model and the dynamic main vehicle model. It should be noted that when the speed of the main vehicle does not reach the preset speed, the main vehicle is driven by the vehicle acceleration model to perform kinematic simulation until the speed of the main vehicle reaches the preset speed.

[0116] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0117] Embodiment 2

[0118] In this embodiment, a vehicle simulation scene modeling method is provided, including static scene restoration and dynamic scene reconstruction.

[0119] like Figure 4As shown, first, static scene restoration: In order to solve the authenticity problem of static scene restoration in the absence of high-precision maps, the present invention proposes a static scene restoration modeling method based on camera data playback, and constructs a simulation platform initialization model on this basis.

[0120] For static scene restoration, it is mainly divided into the following steps:

[0121] Data collection and preprocessing: A camera sensor is installed in front of the vehicle, which is integrated with a positioning sensor. When the vehicle is driving on an open road, the characteristic parameter data of the camera at different positions is collected to identify specific road marking lines; the characteristic parameter data is filtered according to the effective flag of the camera, and then the characteristic parameter data is cleaned by deduplication, outlier removal, downsampling, etc. to eliminate data errors and noise and improve data accuracy and reliability. In the data preprocessing process, the credibility of the data needs to be evaluated.

[0122] Local path segmented fitting and global fitting: for each coordinate position of the vehicle and the corresponding characteristic parameter data of the road marking line recognized by the camera, the road marking line is segmentedly fitted at a preset distance to obtain multiple preview paths, and the discrete points of each preview path are obtained. All the segmented fitting results are integrated to establish a discrete point sample set of the static scene model; the discrete point sample set is fitted with the smoothing spline curve fitting method to obtain the road curve, and then the static scene model is constructed according to the road curve in the OPENDRIVE format.

[0123] Multi-source data supplementation: In particular, during the data cleaning process, it was found that due to factors such as backlighting and blurred lane lines, some feature parameter data was missing. The following solutions are proposed for this type of problem: fusion of multi-sensor data interpolation filling and abnormal scene reconstruction. First, interpolation filling is performed based on the existing feature parameter data, and then the positioning sensor data is fused. For sections with missing data, the vehicle's driving path coordinate data is used as the reference path to correct the interpolation data to ensure that the path filling fitting error is within the allowable range. Then, for sections with missing data, refer to the on-site cause analysis, and construct abnormal scene models such as blurred or missing lane lines and rutting interference in the simulation scene to restore the actual static scene.

[0124] like Figure 5 As shown, the construction of the simulation platform initialization model is mainly based on the existing simulation platform, loading the reconstructed static scene model and vehicle motion / dynamics model, so that the simulation platform enters the initialization.

[0125] Then, the dynamic scene is restored:

[0126] Dynamic scene restoration includes two parts: dynamic environment scene restoration and dynamic main vehicle scene restoration.

[0127] like Figure 6 As shown in the figure, for the dynamic environment scene restoration part, the above-mentioned simulation platform initialization scene construction is used as the basis. First, the radar recognition dynamic environment data is collected, and each frame of dynamic environment data is arranged according to the time series. The dynamic environment data is preprocessed using the improved nearest neighbor data association algorithm to generate dynamic environment sequence data. After the dynamic environment sequence data is converted according to the global coordinate system, the dynamic environment element data is generated according to the OpenScenario standard.

[0128] In particular, the actual collected data is truncated, and the initial state of some dynamic information, typically represented by speed, may not be reached instantly at the beginning of initialization and cannot be directly input into the simulation system. It is necessary to establish an initialization construction plan. Specifically, a scene initialization section can be added to the static road restoration. Before the actual data is input, the initial value of the dynamic environment element data needs to be used as the control target, and the simulated vehicle is pre-prepared in the initialization section; when the simulated vehicle state meets the initialization requirements of the dynamic environment element data, the simulated vehicle switches to the dynamic environment element data as the control target to achieve dynamic scene restoration.

[0129] like Figure 7 As shown, for the reconstruction of the dynamic main vehicle model, it is also based on the construction of the initialization scene of the above-mentioned simulation platform, with data-driven simulation as the main technical means.

[0130] Firstly, based on the fusion data of the real vehicle's global navigation satellite system (GNSS) and inertial navigation system, the vehicle state parameters are collected as the dynamic main vehicle data, and the dynamic main vehicle data is preprocessed such as data cleaning, data format consistency processing, missing data fitting interpolation, etc., to obtain the dynamic main vehicle sequence data in the scene coordinate system, and establish a data-driven vehicle acceleration model, which is used to drive the vehicle kinematic model in the scene model of the simulation platform initialization.

[0131] Then, the scenario model is initialized based on the simulation platform, the vehicle is switched to manual driving mode, and data-driven simulation is realized by importing external data, similar to manual driving. Then, it is determined whether the vehicle speed at the target point of main concern meets the requirements. If it meets the requirements, the vehicle is switched to automatic driving mode. It is considered that the simulation platform has achieved dynamic and static scenario modeling when key events occur.

[0132] Embodiment 3

[0133] In this embodiment, Figure 2 As shown, a vehicle simulation scene modeling device is provided, comprising:

[0134] Static data acquisition module 310, used to acquire static scene data;

[0135] A static model building module 320, used to build a static scene model according to static scene data;

[0136] An initial model building module 330 is used to load the static scene model and the vehicle kinematics model to build a simulation initial model;

[0137] A dynamic data acquisition module 340 is used to acquire dynamic environment data and dynamic host vehicle data;

[0138] The dynamic model building module 350 is used to build a dynamic scene model based on the simulation initial model according to the dynamic environment data and the dynamic host vehicle data.

[0139] In this embodiment, the static data acquisition module 310 is used to obtain the characteristic parameter data of the road marking line. The characteristic parameter data can be obtained through monitoring equipment such as radar and camera. After obtaining the static scene data, the static scene data is filtered and cleaned to eliminate data errors and noise to improve the accuracy and reliability of the data. The static data acquisition module 310 inputs the static scene data into the static model construction module 320, and then constructs the static scene model through the static model construction module 320. Subsequently, the static scene model and the vehicle kinematic model are loaded into the simulation platform through the initial model construction module 330, so as to construct the simulation initial model.

[0140] The dynamic data acquisition module 340 acquires dynamic environment data, arranges the dynamic environment data according to a time series, and pre-processes the dynamic environment data, thereby generating dynamic environment sequence data. The dynamic environment sequence data is subjected to coordinate conversion according to a global coordinate system to generate dynamic environment element data. In addition, the dynamic data acquisition module 340 acquires dynamic main vehicle data, arranges the dynamic main vehicle data according to the acquisition time, and pre-processes the dynamic main vehicle data, thereby generating dynamic main vehicle sequence data. Then, the dynamic data acquisition module 340 inputs the dynamic environment element data and the dynamic main vehicle sequence data into the dynamic model construction module 350, and then the dynamic scene model is constructed by the dynamic model construction module 350.

[0141] The specific definition of the vehicle simulation scene modeling device can be found in the definition of the vehicle simulation scene modeling method above, which will not be repeated here. Each unit in the above-mentioned vehicle simulation scene modeling device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units.

[0142] Embodiment 4

[0143] In this embodiment, a computer device is provided. Its internal structure diagram can be shown as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and the non-volatile storage medium is deployed with a database, which is used to store static scene data, dynamic environment data and dynamic main vehicle data. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices deployed with application software. When the computer program is executed by the processor, a vehicle simulation scene modeling method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse, etc.

[0144] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0145] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0146] Acquire static scene data, and construct a static scene model according to the static scene data;

[0147] Loading the static scene model and the vehicle kinematics model to construct an initial simulation model;

[0148] Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model;

[0149] Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model;

[0150] Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed;

[0151] A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

[0152] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0153] Collecting characteristic parameter data of preset road marking lines, performing segmented fitting on the characteristic parameter data to obtain multiple segments of preview paths, wherein the preview paths have multiple discrete points;

[0154] Integrating the discrete points of each preview path to form a discrete point sample set;

[0155] The discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve.

[0156] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0157] Preprocessing the characteristic parameter data;

[0158] Detecting whether the number of the characteristic parameter data is equal to a preset number within a preset time, and when the number of the characteristic parameter data is equal to the preset number within the preset time, performing global fitting on the discrete point sample set to obtain a road curve;

[0159] A static scene model is constructed based on the road curve.

[0160] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0161] Acquire the dynamic environment data, arrange the dynamic environment data according to the acquisition time, and obtain dynamic environment sequence data;

[0162] Performing coordinate transformation on the dynamic environment sequence data to generate dynamic environment element data;

[0163] Inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes;

[0164] sequentially detecting whether each of the first simulated dynamic scenes is consistent with a preset scene, and when the first simulated dynamic scene is consistent with the preset scene, determining that the first simulated dynamic scene is a first prepared dynamic scene;

[0165] The dynamic environment model is constructed based on the plurality of the first prepared dynamic scenes.

[0166] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0167] inputting the initial values ​​of the dynamic environment element data into the simulation initial model, and operating the vehicle in the simulation initial model;

[0168] Detect whether the vehicle state reaches a preset state, and when the vehicle state reaches the preset state, obtain the state data of the vehicle in the simulation initial model, thereby generating a plurality of first simulation dynamic scenes.

[0169] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0170] When the first simulation dynamic scene does not match the preset scene, determining that the dynamic environment element data is element data to be supplemented, and supplementing the element data to be supplemented;

[0171] Using the supplemented dynamic environment element data to input into the simulation initial model to generate a plurality of second simulation dynamic scenes;

[0172] It is detected in turn whether each of the second simulated dynamic scenes is consistent with the preset scene. When the second simulated dynamic scene is consistent with the preset scene, the dynamic environment model is constructed based on multiple second simulated dynamic scenes and the first prepared dynamic scene.

[0173] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0174] Acquire the dynamic main vehicle data, arrange the dynamic main vehicle data according to the acquisition time, and obtain dynamic main vehicle sequence data;

[0175] Establishing a vehicle acceleration model according to the dynamic main vehicle sequence data;

[0176] The vehicle is driven by the vehicle acceleration model.

[0177] Embodiment 5

[0178] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0179] Acquire static scene data, and construct a static scene model according to the static scene data;

[0180] Loading the static scene model and the vehicle kinematics model to construct an initial simulation model;

[0181] Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model;

[0182] Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model;

[0183] Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed;

[0184] A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

[0185] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0186] Collecting characteristic parameter data of preset road marking lines, performing segmented fitting on the characteristic parameter data to obtain multiple segments of preview paths, wherein the preview paths have multiple discrete points;

[0187] Integrating the discrete points of each preview path to form a discrete point sample set;

[0188] The discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve.

[0189] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0190] Preprocessing the characteristic parameter data;

[0191] Detecting whether the number of the characteristic parameter data is equal to a preset number within a preset time, and when the number of the characteristic parameter data is equal to the preset number within the preset time, performing global fitting on the discrete point sample set to obtain a road curve;

[0192] A static scene model is constructed based on the road curve.

[0193] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0194] Acquire the dynamic environment data, arrange the dynamic environment data according to the acquisition time, and obtain dynamic environment sequence data;

[0195] Performing coordinate transformation on the dynamic environment sequence data to generate dynamic environment element data;

[0196] Inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes;

[0197] sequentially detecting whether each of the first simulated dynamic scenes is consistent with a preset scene, and when the first simulated dynamic scene is consistent with the preset scene, determining that the first simulated dynamic scene is a first prepared dynamic scene;

[0198] The dynamic environment model is constructed based on the plurality of the first prepared dynamic scenes.

[0199] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0200] inputting the initial values ​​of the dynamic environment element data into the simulation initial model, and operating the vehicle in the simulation initial model;

[0201] Detect whether the vehicle state reaches a preset state, and when the vehicle state reaches the preset state, obtain the state data of the vehicle in the simulation initial model, thereby generating a plurality of first simulation dynamic scenes.

[0202] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0203] When the first simulation dynamic scene does not match the preset scene, determining that the dynamic environment element data is element data to be supplemented, and supplementing the element data to be supplemented;

[0204] Using the supplemented dynamic environment element data to input into the simulation initial model to generate a plurality of second simulation dynamic scenes;

[0205] It is detected in turn whether each of the second simulated dynamic scenes is consistent with the preset scene. When the second simulated dynamic scene is consistent with the preset scene, the dynamic environment model is constructed based on multiple second simulated dynamic scenes and the first prepared dynamic scene.

[0206] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0207] Acquire the dynamic main vehicle data, arrange the dynamic main vehicle data according to the acquisition time, and obtain dynamic main vehicle sequence data;

[0208] Establishing a vehicle acceleration model according to the dynamic main vehicle sequence data;

[0209] The vehicle is driven by the vehicle acceleration model.

[0210] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0211] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0212] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A vehicle simulation scene modeling method, characterized in that: include: Acquire static scene data, and construct a static scene model according to the static scene data; Loading the static scene model and the vehicle kinematics model to construct an initial simulation model; Acquire dynamic environment data, and perform coordinate transformation on the dynamic environment data based on the simulation initial model to obtain a dynamic environment model; Acquiring dynamic main vehicle data, and driving the vehicle using the dynamic main vehicle data based on the simulation initial model; Detecting whether the speed of the vehicle reaches a preset speed, and when the speed of the vehicle reaches the preset speed, constructing a dynamic main vehicle model according to the parameters of the vehicle that reaches the preset speed; A dynamic scene model is generated according to the dynamic environment model and the dynamic host vehicle model.

2. The vehicle simulation scene modeling method according to claim 1, characterized in that: The step of acquiring static scene data and constructing a static scene model according to the static scene data comprises: Collecting characteristic parameter data of preset road marking lines, performing segmented fitting on the characteristic parameter data to obtain multiple segments of preview paths, wherein the preview paths have multiple discrete points; Integrating the discrete points of each preview path to form a discrete point sample set; The discrete point sample set is globally fitted to obtain a road curve, and a static scene model is constructed based on the road curve.

3. The vehicle simulation scenario modeling method according to claim 2, characterized in that: The step of acquiring static scene data and constructing a static scene model according to the static scene data further comprises: Preprocessing the characteristic parameter data; Detecting whether the number of the characteristic parameter data is equal to a preset number within a preset time, and when the number of the characteristic parameter data is equal to the preset number within the preset time, performing global fitting on the discrete point sample set to obtain a road curve; A static scene model is constructed based on the road curve.

4. The vehicle simulation scene modeling method according to claim 1, characterized in that: The step of acquiring dynamic environment data, performing coordinate transformation on the dynamic environment data based on the simulation initial model, and obtaining a dynamic environment model comprises: Acquire the dynamic environment data, arrange the dynamic environment data according to the acquisition time, and obtain dynamic environment sequence data; Performing coordinate transformation on the dynamic environment sequence data to generate dynamic environment element data; Inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes; sequentially detecting whether each of the first simulated dynamic scenes is consistent with a preset scene, and when the first simulated dynamic scene is consistent with the preset scene, determining that the first simulated dynamic scene is a first prepared dynamic scene; The dynamic environment model is constructed based on the plurality of the first prepared dynamic scenes.

5. The vehicle simulation scene modeling method according to claim 4, characterized in that: The step of inputting the dynamic environment element data into the simulation initial model to generate a plurality of first simulation dynamic scenes comprises: inputting the initial values ​​of the dynamic environment element data into the simulation initial model, and operating the vehicle in the simulation initial model; Detect whether the vehicle state reaches a preset state, and when the vehicle state reaches the preset state, obtain the state data of the vehicle in the simulation initial model, thereby generating a plurality of first simulation dynamic scenes.

6. The vehicle simulation scene modeling method according to claim 5, characterized in that: After the step of sequentially detecting whether each of the first simulation dynamic scenes is consistent with the preset scene, the method further includes: When the first simulation dynamic scene does not match the preset scene, determining that the dynamic environment element data is element data to be supplemented, and supplementing the element data to be supplemented; Using the supplemented dynamic environment element data to input into the simulation initial model to generate a plurality of second simulation dynamic scenes; It is detected in turn whether each of the second simulated dynamic scenes is consistent with the preset scene. When the second simulated dynamic scene is consistent with the preset scene, the dynamic environment model is constructed based on multiple second simulated dynamic scenes and the first prepared dynamic scene.

7. The vehicle simulation scene modeling method according to claim 5, characterized in that: The step of acquiring dynamic main vehicle data and driving the vehicle by the dynamic main vehicle data based on the simulation initial model comprises: Acquire the dynamic main vehicle data, arrange the dynamic main vehicle data according to the acquisition time, and obtain dynamic main vehicle sequence data; Establishing a vehicle acceleration model according to the dynamic main vehicle sequence data; The vehicle is driven by the vehicle acceleration model.

8. A vehicle simulation scene modeling device, characterized in that: include: Static data acquisition module, used to acquire static scene data; A static model building module, used for building a static scene model according to static scene data; An initial model building module, used to load the static scene model and the vehicle kinematics model to build an initial simulation model; A dynamic data acquisition module is used to acquire dynamic environment data and dynamic main vehicle data; A dynamic model building module is used to build a dynamic scene model based on the simulation initial model according to the dynamic environment data and the dynamic main vehicle data.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images