A vehicle drivable boundary generation method and device, electronic equipment and storage medium

Abstract

The application provides a vehicle drivable boundary generation method and device, electronic equipment and storage medium. The method obtains obstacle information of a vehicle in a vehicle camera coordinate system through a vehicle camera, the vehicle has at least three obstacles, and the obstacle information at least includes lateral and longitudinal distances of the obstacles. The obstacle information of the vehicle in the vehicle camera coordinate system is subjected to coordinate system conversion to obtain an obstacle information coordinate system conversion result of the vehicle, which at least includes obstacle information of the vehicle in a self-coordinate system. The drivable boundary of the vehicle is generated according to the obstacle information coordinate system conversion result, without relying on map data, and the drivable boundary can be generated according to actual perception data, solving the problem that the existing technical solution of generating a drivable boundary using a high-precision map relies heavily on map data, which easily leads to the problem that an automatic driving system cannot accurately identify road conditions, thereby affecting the safety of driving.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, specifically to a method, apparatus, electronic device, and storage medium for generating drivable vehicle boundaries. Background Technology

[0002] Driving boundaries typically refer to the range within which a vehicle can safely travel in road traffic; that is, the maximum allowable boundary range for a vehicle to travel on the road.

[0003] Currently, in the field of autonomous driving, the technology for generating drivable boundaries of vehicles mainly uses high-precision map data, such as information on roads and radar signs, to help vehicles with path planning.

[0004] The inventors discovered that the technical solution of using high-precision maps to generate drivable boundaries is highly dependent on map data. If the map information is outdated, especially in the case of road reconstruction or new road construction, the autonomous driving system may be unable to accurately identify road conditions, thereby affecting driving safety. Summary of the Invention

[0005] In response to this, this application provides a method, apparatus, electronic device, and storage medium for generating drivable boundaries of a vehicle, in order to solve the problem that existing technical solutions for generating drivable boundaries using high-precision maps are highly dependent on map data. If the map information is outdated, especially in the case of road reconstruction or new road construction, it can easily lead to the autonomous driving system being unable to accurately identify road conditions, thereby affecting driving safety.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of this application discloses a method for generating a vehicle's drivable boundary, including:

[0008] Obstacle information of the vehicle in the vehicle camera coordinate system is obtained. The vehicle has at least three obstacles. The obstacle information includes at least the horizontal and vertical distances of the obstacles.

[0009] The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system.

[0010] Based on the coordinate system transformation result of the obstacle information, the drivable boundary of the vehicle is generated.

[0011] Optionally, in the above-described method for generating drivable vehicle boundaries, the obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed to obtain the obstacle information coordinate transformation result of the vehicle, including:

[0012] The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed into the vehicle coordinate system to obtain the obstacle information of the vehicle in the vehicle coordinate system.

[0013] Determine whether the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation;

[0014] If the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into both lane line prediction trajectory coordinate system and kinematic prediction trajectory coordinate system, respectively, to obtain the obstacle information of the vehicle in the lane line prediction trajectory coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system, the obstacle information of the vehicle in the lane line prediction trajectory coordinate system, and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation result of the vehicle.

[0015] If the vehicle does not meet the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into the kinematic prediction trajectory coordinate system to obtain the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation result of the vehicle.

[0016] Optionally, in the above-described method for generating drivable vehicle boundaries, the vehicle satisfies the lane line prediction trajectory coordinate system transformation requirements, including:

[0017] The lane quality of the vehicle is greater than the preset lane quality.

[0018] Optionally, in the above method for generating drivable vehicle boundaries, the process of generating the lane line prediction trajectory coordinate system includes:

[0019] Determine the lane line information of the vehicle in the vehicle coordinate system. The lane line information includes at least: lane line lateral distance, lane line slope, lane line curvature, and lane line curvature change rate.

[0020] The lane line information of the vehicle in the vehicle coordinate system is fitted to obtain the lane line prediction trajectory curve of the vehicle, and the lane line prediction trajectory coordinate system of the vehicle is obtained based on the lane line prediction trajectory curve of the vehicle.

[0021] The process of generating the kinematic trajectory coordinate system includes:

[0022] Obtain the kinematic information of the vehicle, the kinematic information including at least: vehicle angular velocity, steering wheel angle, and steering gear ratio;

[0023] Based on the kinematic information, the vehicle's turning radius is obtained;

[0024] The kinematic prediction trajectory curve of the vehicle is obtained by fitting the vehicle's turning radius, and the kinematic prediction trajectory coordinate system of the vehicle is obtained based on the kinematic prediction trajectory curve.

[0025] Optionally, in the above-described method for generating the drivable boundary of a vehicle, generating the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information includes:

[0026] Determine whether the number of obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result is equal;

[0027] If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, then the drivable boundary of the vehicle is generated based on the obstacle seed points in the coordinate system with the largest number of obstacle seed points in the obstacle information coordinate system transformation result.

[0028] If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then the drivable boundary of the vehicle is generated according to the priority order of the coordinate systems and the obstacle seed points in the coordinate system with the highest priority in the obstacle information coordinate system transformation result.

[0029] The priority order of the coordinate systems is as follows: lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system.

[0030] Optionally, in the above-described method for generating the drivable boundary of a vehicle, generating the drivable boundary of the vehicle based on obstacle seed points in the corresponding coordinate system includes:

[0031] Based on the obstacle seed points in the corresponding coordinate system, determine the vehicle's interior point within the obstacle in the corresponding coordinate system;

[0032] Determine whether the number of points inside the obstacle in the corresponding coordinate system is greater than or equal to a preset number and whether the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than a preset quadratic term coefficient.

[0033] If it is determined that the number of obstacle interior points of the vehicle in the corresponding coordinate system is greater than or equal to a preset number, and the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient, then the obstacle interior points of the vehicle in the corresponding coordinate system are taken as target obstacle interior points, and the drivable boundary of the vehicle is generated based on the target obstacle interior points.

[0034] Optionally, in the above method for generating the drivable boundary of a vehicle, determining the interior points of the vehicle within the obstacle in the corresponding coordinate system based on the obstacle seed points in the corresponding coordinate system includes:

[0035] Based on the obstacle seed points in the corresponding coordinate system, generate the obstacle seed point fitting curve of the vehicle in the corresponding coordinate system;

[0036] Based on the fitted curve of the obstacle seed point, determine the abscissa of the fitted curve of each obstacle seed point of the vehicle in the corresponding coordinate system;

[0037] For each obstacle seed point of the vehicle in the corresponding coordinate system, its lateral distance error is obtained based on its coordinate system abscissa and the fitted curve abscissa.

[0038] Obstacle seed points with lateral distance errors less than a preset lateral distance error are taken as interior points of the vehicle in the corresponding coordinate system.

[0039] The second aspect of this application discloses a vehicle drivable boundary generation device, comprising:

[0040] The acquisition unit is used to acquire obstacle information of the vehicle in the vehicle camera coordinate system through the vehicle's onboard camera. The vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles.

[0041] The transformation unit is used to perform coordinate system transformation on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system.

[0042] The generation unit is used to generate the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information.

[0043] A third aspect of this application discloses an electronic device, including: a memory and a processor;

[0044] The memory is used to store computer programs;

[0045] The processor is used to execute the computer program, specifically to implement the vehicle drivable boundary generation method as disclosed in any of the first aspects.

[0046] The fourth aspect of this application discloses a computer storage medium for storing a computer program, which, when executed, is specifically used to implement the vehicle drivable boundary generation method as disclosed in any of the first aspects.

[0047] This invention provides a method for generating a vehicle's drivable boundary, comprising: acquiring obstacle information of the vehicle in the vehicle's coordinate system using an onboard camera, wherein the vehicle has at least three obstacles, and the obstacle information includes at least the lateral and longitudinal distances of the obstacles; performing coordinate system transformation on the obstacle information in the vehicle's coordinate system to obtain the obstacle information coordinate system transformation result, the obstacle information coordinate system transformation result including at least the obstacle information of the vehicle in its own coordinate system; and generating the vehicle's drivable boundary based on the obstacle information coordinate system transformation result. This method does not rely on map data and can generate the drivable boundary based on actual perception data, solving the problem that existing technical solutions for generating drivable boundaries using high-precision maps are highly dependent on map data. If the map information is outdated, especially in the case of road reconstruction or new road construction, it can easily lead to the autonomous driving system being unable to accurately identify road conditions, thus affecting driving safety. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0049] Figure 1 A flowchart of a method for generating a vehicle drivable boundary provided in this application;

[0050] Figure 2 A flowchart of a coordinate system transformation provided in an embodiment of this application;

[0051] Figure 3 A flowchart illustrating the generation of a kinematic prediction trajectory coordinate system for a vehicle, as provided in an embodiment of this application;

[0052] Figure 4 A schematic diagram of the structure of a vehicle lane line prediction trajectory coordinate system provided in an embodiment of this application;

[0053] Figure 5 A schematic diagram of a lane line prediction trajectory and a kinematic prediction trajectory provided for an embodiment of this application;

[0054] Figure 6 An example diagram illustrating the generation of a vehicle's drivable boundary, provided as an embodiment of this application;

[0055] Figure 7 This is a schematic diagram of a vehicle drivable boundary generation device provided in an embodiment of this application;

[0056] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] First, it should be noted that in this application, "obstacle" refers to any object that may or could obstruct the normal driving path of a vehicle.

[0059] This application provides a method, apparatus, electronic device, and storage medium for generating drivable boundaries of a vehicle, in order to solve the problem that existing technical solutions for generating drivable boundaries using high-precision maps are highly dependent on map data. If the map information is outdated, especially in the case of road reconstruction or new road construction, it can easily lead to the autonomous driving system being unable to accurately identify road conditions, thereby affecting driving safety.

[0060] Please see Figure 1 The method for generating the drivable boundary of this vehicle mainly includes the following steps:

[0061] S101. Obtain obstacle information of the vehicle in the vehicle camera coordinate system through the vehicle's onboard camera; the vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles.

[0062] In practice, the vehicle's onboard forward-facing camera can be used to obtain obstacle information of the vehicle under the coordinates of the onboard camera; of course, it is not limited to this, and it can also be obtained through other cameras of the vehicle, all of which are within the protection scope of this application.

[0063] Specifically, each vehicle is generally equipped with 1 to 3 onboard forward-facing cameras, which are usually installed inside the windshield, near the top or rearview mirror.

[0064] It should be noted that the vehicle's onboard camera can obtain obstacle information such as the vehicle's lateral and longitudinal distances, lateral and longitudinal velocities, length, width, and orientation angle. Since this obstacle information is obtained through the onboard camera, the coordinate system of the onboard camera is used as a reference point. The origin of the onboard camera coordinate system is the location where the onboard camera is installed. shaft and The axial direction can be customized by the vehicle-mounted camera.

[0065] To accurately represent multi-vehicle stationary scenarios, after acquiring obstacle information of the vehicle in the vehicle's coordinate system using the vehicle's onboard camera, the information can be filtered and used to generate the vehicle's drivable boundary. Filtering criteria can include obstacles having lateral and longitudinal distances less than preset lateral and longitudinal distances and being stationary. Obstacles with lateral and longitudinal distances less than preset lateral and longitudinal distances and being stationary are then used for generating the vehicle's drivable boundary. Of course, filtering criteria are not limited to these and can be determined based on the application environment and user needs, all of which are within the scope of this application.

[0066] In practice, the specific values ​​of the preset horizontal and vertical distances can be determined based on the detection range of the vehicle-mounted camera. For example, the farther the detection range of the vehicle-mounted camera, the larger the preset horizontal and vertical distances. When the horizontal and vertical velocities of the obstacle are both 0, the obstacle is considered to be in a stationary state.

[0067] S102. Perform coordinate system transformation on the obstacle information of the vehicle in the coordinate system of the on-board camera to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result shall include at least the obstacle information of the vehicle in the vehicle coordinate system.

[0068] In some embodiments, the specific process of performing coordinate system transformation on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system in step S102 to obtain the coordinate system transformation result of the obstacle information of the vehicle can be as follows: Figure 2 As shown, it mainly includes steps S201 to S204:

[0069] S201. Perform a self-vehicle coordinate system transformation on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system to obtain the obstacle information of the vehicle in the self-vehicle coordinate system.

[0070] In this application, the vehicle coordinate system is a coordinate system established with the vehicle itself as the reference point, and the origin is usually set at the center of the rear axle or the center of mass of the vehicle. The axis points forward of the vehicle. The axis points to the left. The axis is vertically upward.

[0071] Since obstacle information obtained through the vehicle's onboard camera is referenced to the onboard camera's coordinate system, it is necessary to align the obstacle information to the vehicle's own coordinate system.

[0072] In practical applications, the lateral and longitudinal distances of obstacles on the vehicle can be converted from the vehicle-mounted camera coordinates to the vehicle's coordinate system.

[0073] S202. Determine whether the vehicle meets the requirements for coordinate system transformation of the lane line prediction trajectory.

[0074] For a vehicle to meet the requirements for coordinate system transformation of the lane prediction trajectory, the lane quality must be greater than the preset lane quality. The lane quality generally includes the lane quality on the left and right sides of the vehicle.

[0075] If the quality of the vehicle's left lane line and / or right lane line is greater than the preset lane line quality, that is, the quality of the vehicle's left lane line and / or right lane line is relatively high, the vehicle is considered to meet the requirements for lane line prediction trajectory coordinate system transformation; if the quality of the vehicle's left lane line and right lane line is not greater than the preset lane line quality, that is, the quality of the vehicle's left lane line and right lane line is relatively low, the vehicle is considered to not meet the requirements for lane line prediction trajectory coordinate system transformation.

[0076] It should be noted that the lane quality of a vehicle can be determined by parameters such as lane length and lane confidence level; the specific value of the preset lane quality can be determined according to the application environment and user needs, and this application does not make specific limitations, all of which are within the protection scope of this application.

[0077] If the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation, then proceed to step S202; if the vehicle does not meet the requirements for lane line prediction trajectory coordinate system transformation, then proceed to step S203.

[0078] S203. Perform lane line prediction trajectory coordinate system transformation and kinematic prediction trajectory coordinate system transformation on the obstacle information of the vehicle in the vehicle coordinate system, respectively, to obtain the obstacle information of the vehicle in the lane line prediction trajectory coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. Use the obstacle information of the vehicle in the vehicle coordinate system, the obstacle information of the vehicle in the lane line prediction trajectory coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system as the obstacle information coordinate transformation results of the vehicle.

[0079] In practice, if the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in its own coordinate system is transformed into both lane line prediction trajectory coordinate system and kinematic prediction trajectory coordinate system.

[0080] In practical applications, the lateral and longitudinal distances of the vehicle to obstacles can be transformed from the vehicle coordinate system to the lane line prediction trajectory coordinate system and the kinematic prediction trajectory coordinate system, respectively.

[0081] S204. Perform kinematic prediction trajectory coordinate system transformation on the obstacle information of the vehicle in the vehicle coordinate system to obtain the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system, and use the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system as the obstacle information coordinate transformation result of the vehicle.

[0082] In practice, if the vehicle does not meet the requirements for lane line prediction trajectory coordinate system transformation, then only the obstacle information in the vehicle's own coordinate system is transformed into the kinematic prediction trajectory coordinate system.

[0083] In this application, the kinematic prediction trajectory coordinate system is a coordinate system used to describe the vehicle's motion state, including the changes of parameters such as position, velocity, and acceleration over time. In some embodiments, the process of generating the vehicle's kinematic prediction trajectory coordinate system is as follows: Figure 3 As shown, the main steps include S301 to S303:

[0084] S301. Obtain the vehicle's kinematic information, which includes at least: vehicle angular velocity, steering wheel angle, and steering ratio.

[0085] In practice, vehicle kinematic information can be obtained through the vehicle's onboard system.

[0086] S302. Based on the kinematic information, obtain the vehicle's turning radius.

[0087] The turning radius of the vehicle, also known as the vehicle's turning radius, is calculated based on the vehicle's angular velocity, steering wheel angle, and steering transmission ratio.

[0088] Specifically, the vehicle's turning radius can be used express.

[0089] S303. Fit the vehicle's turning radius to obtain the vehicle's kinematic prediction trajectory curve, and obtain the vehicle's kinematic prediction trajectory coordinate system based on the vehicle's kinematic prediction trajectory curve.

[0090] In this application, the lane line prediction trajectory coordinate system specifically refers to the coordinate system related to the vehicle and the lane line, which is used to accurately describe the positional relationship of the vehicle relative to the lane line.

[0091] In some embodiments, the process of generating the lane line prediction trajectory coordinate system is as follows: Figure 4 As shown, it mainly includes steps S401 and S402:

[0092] S401. Determine the lane line information of the vehicle in the vehicle coordinate system. The lane line information shall include at least: lane line lateral distance, lane line slope, lane line curvature and lane line curvature change rate.

[0093] In the vehicle coordinate system, the lateral distance of the road line represents the distance between the road line and the installation position of the vehicle-mounted camera in the vehicle coordinate system. The absolute distance in direction; the slope of the road line represents the road line relative to the vehicle in the vehicle's coordinate system. The slope of the direction, the curvature of the track indicates the degree of curvature of the track at a certain point, and the rate of change of track curvature indicates the change of curvature of the track at different positions.

[0094] In practice, lane line information of the vehicle in the vehicle's coordinate system can be obtained through the vehicle's onboard camera. Then, the lane line information in the vehicle's coordinate system can be transformed to obtain the lane line information in the vehicle's own coordinate system.

[0095] Specifically, the lateral distance of the lane can be used This indicates that the slope of the track can be expressed as... This indicates that the curvature of the track line can be used. This indicates that the rate of change of track curvature can be expressed as... express.

[0096] S402. Fit the lane line information of the vehicle to obtain the lane line prediction trajectory curve of the vehicle, and obtain the lane line prediction trajectory coordinate system of the vehicle based on the lane line prediction trajectory curve of the vehicle.

[0097] In practical applications, the lane line prediction trajectory curve of a vehicle is generally a cubic polynomial curve, used to represent lateral distance. With longitudinal distance Changes:

[0098] ;

[0099] The coefficient in the curve , , , The initial value is set to 0, corresponding to the above. , , , .

[0100] When the quality of the left and right lane lines detected by the vehicle camera is very low, it is considered that the lane lines in the current frame do not meet the requirements, that is, the vehicle does not meet the requirements for the coordinate system transformation of the lane line prediction trajectory. Therefore, no values ​​are assigned to the four coefficients in the curve, that is, the vehicle has no lane line trajectory.

[0101] When the vehicle-mounted camera detects high quality for both the left and right lane markings, then... , , , The value is assigned as the average of the corresponding coefficients for the lane markings on both sides, and these coefficients are provided by the vehicle-mounted camera. For example, Left lane line , Right lane line , .

[0102] If the vehicle-mounted camera detects that the left lane marking quality is high and the right lane marking quality is low, then... , , , Assign the corresponding coefficient to the left lane line; if the vehicle camera detects that the right lane line quality is high and the left lane line quality is low, then... , , , Assign it as the coefficient corresponding to the right track line.

[0103] In summary, the coordinate system transformation relationship in this application is as follows: In the vehicle coordinate system, the longitudinal coordinate of the obstacle can be expressed as... The horizontal axis is represented as In the lane prediction trajectory coordinate system, the obstacle's vertical coordinate can be expressed as: The horizontal axis is represented as . , and , The conversion relationship is as follows:

[0104] ;

[0105] .

[0106] In the kinematic prediction trajectory coordinate system, the obstacle's ordinate can be expressed as: The horizontal axis is represented as The turning radius of a vehicle is expressed as . , and , The conversion relationship is as follows:

[0107] ;

[0108] ;

[0109] .

[0110] It is worth noting that, combined with Figure 5 When a vehicle is in motion, there are generally two predicted trajectory curves: the kinematic predicted trajectory curve (the blue dashed line in the figure) and the lane line predicted trajectory curve (the red dashed line in the figure).

[0111] S103. Based on the coordinate system transformation results of the obstacle information, generate the drivable boundary of the vehicle.

[0112] In practical applications, since the obstacle information coordinate system transformation result includes obstacle information in at least one coordinate system, determining which coordinate system's obstacle information to use as the basis for generating the vehicle's drivable boundary is a crucial step in the process of generating the vehicle's drivable boundary. Specifically, the vehicle's drivable boundary can be generated based on the obstacle information coordinate system transformation result, combined with the RANSAC (Random Sample Consensus) algorithm. The detailed process is as follows:

[0113] S501. Determine whether the number of obstacle seed points in each coordinate system is equal in the result of the obstacle information coordinate system transformation.

[0114] In practice, the obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result can be determined by the RANSAC algorithm; the seed points represent the observation data that can be used for the RANSAC algorithm, namely the horizontal and vertical coordinates of the obstacle.

[0115] Taking a multi-vehicle stationary scenario as an example, if there is only a stationary target on the left side of the vehicle, then the stationary target on the left side of the vehicle is used as the seed point, that is, the obstacle on the left side of the coordinate system is used as the obstacle seed point of the coordinate system; if there is only a stationary target on the right side of the vehicle, then the stationary targets on both sides of the vehicle are used as the seed points, that is, the obstacle on the right side of the coordinate system is used as the obstacle seed point of the coordinate system; if there are stationary targets on both sides of the vehicle, then the stationary target on the side with more stationary targets is used as the seed point, that is, the obstacle on the side with more obstacles in the coordinate system is used as the obstacle seed point of the coordinate system.

[0116] It should be noted that after determining the obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result, the number of obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result can be judged to determine whether the number of obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result is equal, based on the quantitative relationship of the obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result.

[0117] If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, then step S502 can be executed; if it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then step S503 can be executed.

[0118] S502. Based on the obstacle sub-points in the coordinate system with the largest number of obstacle sub-points in the obstacle information coordinate system transformation result, generate the drivable boundary of the vehicle.

[0119] In practical applications, if it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, the drivable boundary of the vehicle can be generated based on the obstacle seed points in the coordinate system with the largest number of obstacle seed points in the obstacle information coordinate system transformation result.

[0120] S503. According to the coordinate system priority order, generate the vehicle's drivable boundary based on the obstacle seed points in the highest priority coordinate system of the obstacle information coordinate system transformation result.

[0121] The priority order of the coordinate systems is as follows: lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system.

[0122] In practical applications, if it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then the drivable boundary of the vehicle can be generated according to the priority order of lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system, based on the obstacle seed points in the coordinate system with the highest priority in the obstacle information coordinate system transformation result.

[0123] For example:

[0124] ① Assume that the obstacle information coordinate system transformation result includes: obstacle information in the vehicle coordinate system, obstacle information in the lane line prediction trajectory coordinate system, and obstacle information in the kinematics prediction trajectory coordinate system. The number of obstacle seed points in each coordinate system is equal. According to the coordinate system priority order of lane line prediction trajectory coordinate system > kinematics prediction trajectory coordinate system > vehicle coordinate system, the coordinate system with the highest priority is the lane line prediction trajectory coordinate system. The drivable boundary of the vehicle can be generated based on the obstacle seed points of the vehicle in the lane line prediction trajectory coordinate system.

[0125] ② Assuming that the obstacle information coordinate system transformation result includes: obstacle information in the vehicle coordinate system and obstacle information in the lane line prediction trajectory coordinate system, and that the number of obstacle seed points in each coordinate system is equal, the coordinate system priority order is lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system, with the highest priority being the lane line prediction trajectory coordinate system. The drivable boundary of the vehicle can be generated based on the obstacle seed points in the lane line prediction trajectory coordinate system.

[0126] ③ Assuming the obstacle information coordinate system transformation result includes: obstacle information in the vehicle coordinate system and obstacle information in the kinematic prediction trajectory coordinate system, and the number of obstacle seed points in each coordinate system is equal, the coordinate system priority order is lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system, with the kinematic prediction trajectory coordinate system having the highest priority. The drivable boundary of the vehicle can be generated based on the obstacle seed points in the kinematic prediction trajectory coordinate system.

[0127] In some embodiments, the specific process of generating the drivable boundary of the vehicle based on obstacle seed points in the corresponding coordinate system is as follows, mainly including steps S601 to S603:

[0128] In practice, the corresponding coordinate system is the coordinate system used to generate the drivable boundary of the vehicle. When the number of obstacle seed points in each coordinate system is not equal in the obstacle information coordinate system transformation result, the corresponding coordinate system is the coordinate system with the most obstacle seed points in the obstacle information coordinate system transformation result. When the number of obstacle seed points in each coordinate system is equal in the obstacle information coordinate system transformation result, the corresponding coordinate system is the coordinate system with the highest priority in the obstacle information coordinate system transformation result.

[0129] S601. Determine the vehicle's interior point within the obstacle in the corresponding coordinate system based on the obstacle seed points in the corresponding coordinate system.

[0130] In this application, interior points of obstacles can be used as interior points in the RANSAC algorithm.

[0131] In some embodiments, the specific execution process of step S601, determining the vehicle's interior point in the obstacle according to the obstacle seed points in the corresponding coordinate system, is as follows, mainly including steps S701 to S704:

[0132] S701. Generate a fitting curve of the vehicle based on the obstacle seed points in the corresponding coordinate system.

[0133] In practice, obstacle seed points in the corresponding coordinate system can be fitted to generate a vehicle obstacle seed point fitting curve in the corresponding coordinate system.

[0134] For specific fitting methods, please refer to existing technologies, such as fitting using the least squares method, which are all within the scope of protection of this application.

[0135] S702. Based on the fitted curve of the obstacle seed point, determine the abscissa of the fitted curve of each obstacle seed point of the vehicle in the corresponding coordinate system.

[0136] In practice, the ordinate of each obstacle sub-point in the corresponding coordinate system can be input into the obstacle sub-point fitting curve to obtain the abscissa of the fitting curve of that obstacle sub-point.

[0137] S703. For each obstacle seed point of the vehicle in the corresponding coordinate system, obtain its lateral distance error based on its coordinate system abscissa and the abscissa of the fitted curve.

[0138] The x-coordinate of an obstacle seed point refers to the x-coordinate of that obstacle seed point in the corresponding coordinate system.

[0139] In practice, the lateral distance error of each obstacle seed point can be obtained by subtracting its coordinate system x-coordinate from the x-coordinate of the fitted curve.

[0140] S704. Obstacle seed points with lateral distance errors less than the preset lateral distance error are taken as obstacle interior points of the vehicle in the corresponding coordinate system.

[0141] In practice, after determining the lateral distance error of each obstacle seed point under the corresponding coordinates, obstacle seed points with lateral distance errors less than the preset lateral distance error can be used as the interior points of the vehicle in the corresponding coordinate system.

[0142] The specific value of the preset lateral distance error can be determined based on the actual problem scenario and big data screening, and all of these are within the protection scope of this application.

[0143] S602. Determine whether the number of points inside obstacles in the corresponding coordinate system is greater than or equal to the preset number and whether the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient.

[0144] If it is determined that the number of points inside the obstacle in the corresponding coordinate system is greater than or equal to the preset number, and the quadratic coefficient of the corresponding cubic polynomial curve is less than the preset quadratic coefficient, then step S603 can be executed. Otherwise, the RANSAC algorithm is used to continue iterative calculation until the number of points inside the obstacle in the corresponding coordinate system is greater than or equal to the preset number and the quadratic coefficient of the corresponding cubic polynomial curve is less than the preset quadratic coefficient.

[0145] The specific value of the preset quantity can be determined according to the application environment and user needs. For example, it can be any positive integer, which is within the protection scope of this application. The specific value of the preset quadratic term coefficient can be determined based on the actual scenario and big data screening.

[0146] In practice, a cubic polynomial curve can be fitted first based on the vehicle's interior points in the corresponding coordinate system. Then, it can be determined whether the quadratic coefficient of the fitted cubic polynomial curve is less than a preset quadratic coefficient. If the result is yes, it can be determined whether the number of the vehicle's interior points in the corresponding coordinate system is greater than or equal to a preset number. If so, the vehicle's interior points in the corresponding coordinate system are determined as the target obstacle interior points. Alternatively, it can be determined whether the number of the vehicle's interior points in the corresponding coordinate system is greater than or equal to a preset number. If the result is yes, then the determination of whether the quadratic coefficient of the fitted cubic polynomial curve is less than a preset quadratic coefficient can be performed.

[0147] S603. Take the interior point of the vehicle in the corresponding coordinate system as the interior point of the target obstacle, and generate the drivable boundary of the vehicle based on the interior point of the target obstacle.

[0148] In practice, if the number of obstacle interior points of the vehicle in the corresponding coordinate system is greater than or equal to the preset number and the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient, the obstacle interior points of the vehicle in the corresponding coordinate system can be used as target obstacle interior points, and the drivable boundary of the vehicle can be generated based on the target obstacle interior points.

[0149] Specifically, the drivable boundary of a vehicle can be generated based on the horizontal and vertical coordinates of points within the target obstacle; alternatively, it can be generated based on the IDs of points within the target obstacle. The ID is used to distinguish different objects; in a multi-vehicle stationary scene, the ID remains unchanged from the time a target appears until the time it disappears.

[0150] For example, such as Figure 6 As shown, if the point inside the target obstacle is the red dot in the figure, the red dashed line is the drivable boundary of the vehicle.

[0151] Based on the above principles, the vehicle drivable boundary generation method provided in this embodiment includes: acquiring obstacle information of the vehicle in the vehicle-mounted camera coordinate system through the vehicle's onboard camera, wherein the vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles; performing coordinate system transformation on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system to obtain the obstacle information coordinate system transformation result of the vehicle, and the obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle's own coordinate system; generating the vehicle's drivable boundary according to the obstacle information coordinate system transformation result, without relying on map data, and generating the drivable boundary based on actual perception data. This solves the problem that existing technical solutions for generating drivable boundaries using high-precision maps are highly dependent on map data. If the map information is outdated, especially in the case of road reconstruction or new road construction, it can easily lead to the autonomous driving system being unable to accurately identify road conditions, thereby affecting driving safety.

[0152] It is worth noting that, through research, the inventors discovered that existing methods use sensors such as cameras, radar, and lidar to collect environmental information and generate drivable boundaries or areas by training neural network models to help vehicles perceive road conditions in real time. However, the scheme of using sensor data and training neural network models to generate drivable boundaries requires high system computing power and a large amount of data for model training to achieve a relatively accurate result. In contrast, this application uses obstacle information perceived and output by an onboard camera and the RANSAC algorithm to effectively reduce the amount of computation required to generate drivable boundaries, thereby improving the response speed and processing efficiency of the autonomous driving system.

[0153] It should be noted that, since the lateral distance difference between the inner point of the target obstacle and the drivable boundary of the vehicle is small, the obstacle corresponding to the inner point of the target obstacle can be processed accordingly, such as not identifying it as the nearest vehicle on the path. This can effectively solve the problem of misselection of the target causing the vehicle to brake unnecessarily. Obstacles corresponding to the inner points of other obstacles can be left unfiltered and processed for subsequent target selection. In other words, this application can also improve the vehicle's perception of obstacles in multi-vehicle stationary scenarios, avoiding the misselection of obstacles with no collision risk as the nearest vehicle on the path, thus preventing the vehicle from braking unnecessarily and improving driving safety.

[0154] Optionally, another embodiment of this application also provides a vehicle drivable boundary generation device, see [link to relevant documentation]. Figure 7 The device includes:

[0155] The acquisition unit 101 is used to acquire obstacle information of the vehicle in the vehicle camera coordinate system through the vehicle's on-board camera. The vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles.

[0156] The transformation unit 102 is used to perform coordinate system transformation on the obstacle information of the vehicle in the vehicle camera coordinate system to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system.

[0157] The generation unit 103 is used to generate the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information.

[0158] In some embodiments, when the transformation unit 102 performs coordinate system transformation on the obstacle information of the vehicle in the coordinate system of the on-board camera to obtain the coordinate system transformation result of the obstacle information of the vehicle, it is specifically used for:

[0159] The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed into the vehicle coordinate system to obtain the obstacle information of the vehicle in the vehicle coordinate system.

[0160] Determine whether the vehicle meets the requirements for coordinate system transformation of the lane line prediction trajectory;

[0161] If the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into both lane line prediction trajectory coordinate system and kinematic prediction trajectory coordinate system, respectively, to obtain the obstacle information of the vehicle in the lane line prediction trajectory coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system, the obstacle information of the vehicle in the lane line prediction trajectory coordinate system, and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation results of the vehicle.

[0162] If the vehicle does not meet the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into the kinematic prediction trajectory coordinate system to obtain the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation result of the vehicle.

[0163] In some embodiments, the vehicle meets the lane line prediction trajectory coordinate system transformation requirements, including:

[0164] The vehicle's lane quality is greater than the preset lane quality.

[0165] In some embodiments, the vehicle drivable boundary generation apparatus further includes:

[0166] The lane line information determination unit is used to determine the lane line information of the vehicle in the vehicle coordinate system. The lane line information includes at least: lane line lateral distance, lane line slope, lane line curvature and lane line curvature change rate.

[0167] The lane line prediction trajectory coordinate system generation unit is used to fit the lane line information of the vehicle in the vehicle coordinate system to obtain the lane line prediction trajectory curve of the vehicle, and to obtain the lane line prediction trajectory coordinate system of the vehicle based on the lane line prediction trajectory curve of the vehicle.

[0168] The acquisition unit is used to acquire the vehicle's kinematic information, which includes at least: vehicle angular velocity, steering wheel angle, and steering ratio;

[0169] The turning radius determination unit is used to obtain the vehicle's turning radius based on kinematic information;

[0170] The kinematics prediction trajectory coordinate system generation unit is used to fit the vehicle's turning radius to obtain the vehicle's kinematics prediction trajectory curve, and to obtain the vehicle's kinematics prediction trajectory coordinate system based on the vehicle's kinematics prediction trajectory curve.

[0171] In some embodiments, the generation unit 103 is specifically used to generate the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information as follows:

[0172] Determine whether the number of obstacle seed points in each coordinate system is equal in the result of the obstacle information coordinate system transformation;

[0173] If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, then the drivable boundary of the vehicle is generated based on the obstacle seed points in the coordinate system with the largest number of obstacle seed points in the obstacle information coordinate system transformation result.

[0174] If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then the drivable boundary of the vehicle is generated according to the priority order of the coordinate systems and the obstacle seed points in the coordinate system with the highest priority in the obstacle information coordinate system transformation result.

[0175] The priority order of the coordinate systems is as follows: lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system.

[0176] In some embodiments, generating the drivable boundary of the vehicle based on obstacle seed points in the corresponding coordinate system includes:

[0177] Based on the obstacle seed points in the corresponding coordinate system, determine the vehicle's point inside the obstacle in the corresponding coordinate system;

[0178] Determine whether the number of points inside the obstacle in the corresponding coordinate system is greater than or equal to the preset number and whether the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient.

[0179] If it is determined that the number of obstacle interior points of the vehicle in the corresponding coordinate system is greater than or equal to the preset number, and the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient, then the obstacle interior points of the vehicle in the corresponding coordinate system are taken as the target obstacle interior points, and the drivable boundary of the vehicle is generated based on the target obstacle interior points.

[0180] In some embodiments, determining the vehicle's interior point within the obstacle in the corresponding coordinate system based on obstacle seed points in the corresponding coordinate system includes:

[0181] Based on the obstacle seed points in the corresponding coordinate system, generate the vehicle's obstacle seed point fitting curve in the corresponding coordinate system;

[0182] Based on the fitted curve of the obstacle sub-point, determine the abscissa of the fitted curve of each obstacle sub-point of the vehicle in the corresponding coordinate system;

[0183] For each obstacle seed point of the vehicle in the corresponding coordinate system, its lateral distance error is obtained based on its coordinate system x-coordinate and the fitted curve x-coordinate.

[0184] Obstacle seed points with lateral distance errors less than the preset lateral distance error are taken as the vehicle's interior points in the corresponding coordinate system.

[0185] In the vehicle drivable boundary generation device provided in this embodiment, the acquisition unit 101 is used to acquire obstacle information of the vehicle in the vehicle-mounted camera coordinate system through the vehicle's onboard camera. The vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles. The conversion unit 102 is used to perform coordinate system conversion on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system to obtain the obstacle information coordinate system conversion result of the vehicle. The obstacle information coordinate system conversion result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The generation unit 103 is used to generate the drivable boundary of the vehicle based on the obstacle information coordinate system conversion result. It does not rely on map data and can generate the drivable boundary based on actual perception data. This solves the problem that the existing technical solutions for generating drivable boundaries using high-precision maps are heavily dependent on map data. If the map information is lagging, especially in the case of road reconstruction or new road construction, it is easy for the autonomous driving system to be unable to accurately identify the road conditions, thereby affecting driving safety.

[0186] It should be noted that for relevant descriptions of each unit in the vehicle drivable boundary generation device, please refer to the corresponding method embodiments mentioned above, and will not be repeated here.

[0187] Optionally, another embodiment of this application also provides a computer storage medium for storing a computer program, which, when executed, is specifically used to implement the vehicle drivable boundary generation method provided in any embodiment of this application.

[0188] It should be noted that the relevant explanations regarding the method for generating the drivable boundary of a vehicle can be found in the above embodiments, and will not be repeated here.

[0189] Optionally, another embodiment of this application also provides an electronic device, such as... Figure 7 As shown, the electronic device includes a memory 601 and a processor 602.

[0190] Among them, memory 601 is used to store computer programs;

[0191] The processor 602 is used to execute computer programs, specifically to implement the vehicle drivable boundary generation method provided in any embodiment of this application.

[0192] It should be noted that the relevant explanations regarding the method for generating the drivable boundary of a vehicle can also be found in the above embodiments, and will not be repeated here.

[0193] The features described in the various embodiments of this specification can be substituted for or combined with each other. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the description of the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort. Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0194] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0195] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A vehicle drivable boundary generation method characterized by, include: Obstacle information of the vehicle in the vehicle camera coordinate system is obtained. The vehicle has at least three obstacles. The obstacle information includes at least the horizontal and vertical distances of the obstacles. The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. Based on the coordinate system transformation result of the obstacle information, the drivable boundary of the vehicle is generated; The step of generating the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information includes: Determine whether the number of obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result is equal; If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, then the drivable boundary of the vehicle is generated based on the obstacle seed points in the coordinate system with the largest number of obstacle seed points in the obstacle information coordinate system transformation result. If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then the drivable boundary of the vehicle is generated according to the priority order of the coordinate systems and the obstacle seed points in the coordinate system with the highest priority in the obstacle information coordinate system transformation result. The priority order of the coordinate systems is as follows: lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system.

2. The vehicle drivable boundary generation method according to claim 1, characterized by, The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed to obtain the obstacle information coordinate transformation result of the vehicle, including: The obstacle information of the vehicle in the vehicle-mounted camera coordinate system is transformed into the vehicle coordinate system to obtain the obstacle information of the vehicle in the vehicle coordinate system. Determine whether the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation; If the vehicle meets the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into both lane line prediction trajectory coordinate system and kinematic prediction trajectory coordinate system, respectively, to obtain the obstacle information of the vehicle in the lane line prediction trajectory coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system, the obstacle information of the vehicle in the lane line prediction trajectory coordinate system, and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation result of the vehicle. If the vehicle does not meet the requirements for lane line prediction trajectory coordinate system transformation, then the obstacle information of the vehicle in the vehicle coordinate system is transformed into the kinematic prediction trajectory coordinate system to obtain the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system are used as the obstacle information coordinate transformation result of the vehicle.

3. The vehicle drivable boundary generation method according to claim 2, characterized by, The vehicle meets the requirements for coordinate system transformation of lane line prediction trajectory, including: The lane quality of the vehicle is greater than the preset lane quality.

4. The vehicle drivable boundary generation method according to claim 2, characterized by, The process of generating the lane line prediction trajectory coordinate system includes: Determine the lane line information of the vehicle in the vehicle coordinate system. The lane line information includes at least: lane line lateral distance, lane line slope, lane line curvature, and lane line curvature change rate. The lane line information of the vehicle in the vehicle coordinate system is fitted to obtain the lane line prediction trajectory curve of the vehicle, and the lane line prediction trajectory coordinate system of the vehicle is obtained based on the lane line prediction trajectory curve of the vehicle. The process of generating the kinematic trajectory coordinate system includes: Obtain the kinematic information of the vehicle, the kinematic information including at least: vehicle angular velocity, steering wheel angle, and steering gear ratio; Based on the kinematic information, the vehicle's turning radius is obtained; The kinematic prediction trajectory curve of the vehicle is obtained by fitting the vehicle's turning radius, and the kinematic prediction trajectory coordinate system of the vehicle is obtained based on the kinematic prediction trajectory curve.

5. The vehicle drivable boundary generation method according to claim 1, characterized by, Based on the obstacle seed points in the corresponding coordinate system, the drivable boundary of the vehicle is generated, including: Based on the obstacle seed points in the corresponding coordinate system, determine the vehicle's interior point within the obstacle in the corresponding coordinate system; Determine whether the number of points inside the obstacle in the corresponding coordinate system is greater than or equal to a preset number and whether the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than a preset quadratic term coefficient. If it is determined that the number of obstacle interior points of the vehicle in the corresponding coordinate system is greater than or equal to a preset number, and the coefficient of the quadratic term of the corresponding cubic polynomial curve is less than the preset quadratic term coefficient, then the obstacle interior points of the vehicle in the corresponding coordinate system are taken as target obstacle interior points, and the drivable boundary of the vehicle is generated based on the target obstacle interior points.

6. The vehicle drivable boundary generation method according to claim 5, characterized by, Based on the obstacle seed points in the corresponding coordinate system, determine the vehicle's interior points within the obstacle in the corresponding coordinate system, including: Based on the obstacle seed points in the corresponding coordinate system, generate the obstacle seed point fitting curve of the vehicle in the corresponding coordinate system; Based on the fitted curve of the obstacle seed point, determine the abscissa of the fitted curve of each obstacle seed point of the vehicle in the corresponding coordinate system; For each obstacle seed point of the vehicle in the corresponding coordinate system, its lateral distance error is obtained based on its coordinate system abscissa and the fitted curve abscissa. Obstacle seed points with lateral distance errors less than a preset lateral distance error are taken as interior points of the vehicle in the corresponding coordinate system.

7. A vehicle drivable boundary generating apparatus characterized by comprising: include: The acquisition unit is used to acquire obstacle information of the vehicle in the vehicle camera coordinate system through the vehicle's onboard camera. The vehicle has at least three obstacles, and the obstacle information includes at least the horizontal and vertical distances of the obstacles. The transformation unit is used to perform coordinate system transformation on the obstacle information of the vehicle in the vehicle-mounted camera coordinate system to obtain the obstacle information coordinate system transformation result of the vehicle. The obstacle information coordinate system transformation result includes at least the obstacle information of the vehicle in the vehicle coordinate system and the obstacle information of the vehicle in the kinematic prediction trajectory coordinate system. The generation unit is used to generate the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information; The generation unit, when generating the drivable boundary of the vehicle based on the coordinate system transformation result of the obstacle information, includes: Determine whether the number of obstacle seed points in each coordinate system of the obstacle information coordinate system transformation result is equal; If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is not equal, then the drivable boundary of the vehicle is generated based on the obstacle seed points in the coordinate system with the largest number of obstacle seed points in the obstacle information coordinate system transformation result. If it is determined that the number of obstacle seed points in each coordinate system in the obstacle information coordinate system transformation result is equal, then the drivable boundary of the vehicle is generated according to the priority order of the coordinate systems and the obstacle seed points in the coordinate system with the highest priority in the obstacle information coordinate system transformation result. The priority order of the coordinate systems is as follows: lane line prediction trajectory coordinate system > kinematic prediction trajectory coordinate system > vehicle coordinate system.

8. An electronic device, comprising: include: Memory and processor; The memory is used to store computer programs; The processor is used to execute the computer program, specifically to implement the vehicle drivable boundary generation method as described in any one of claims 1-6.

9. A computer storage medium, characterized in that Used to store a computer program, which, when executed, is specifically used to implement the vehicle drivable boundary generation method as described in any one of claims 1-6.

Images