An automatic driving vehicle parking method, system and vehicle

Abstract

The present invention relates to a method, system and vehicle for an autonomous vehicle to park in a parking space. It is determined whether the vehicle type of the vehicle to be parked corresponds to the target parking space. If they correspond, a trajectory planning coordinate system is established with the center of the assumed vehicle in the target parking space as the coordinate origin, and the vehicle coordinate system of the assumed vehicle is transformed with reference to the trajectory planning coordinate system. A parking point after the assumed vehicle exits the warehouse is selected according to a set rule and an optimal trajectory is generated. It is determined whether the current position of the vehicle to be parked is on any of the trajectories. If it is, the parking-in-position step is executed; if not, an adjustment step is executed to adjust the vehicle to a point on any of the trajectories, and then the parking-in-position is executed. In the present invention, the vehicle pose deflection angle of the corresponding point can be calculated for each coordinate, so as to achieve accurate vehicle pose adjustment, so that the vehicle after adjusting the pose can be parked in the parking space without stopping once.

Description

Technical Field

[0001] The present invention relates to the field of autonomous vehicles, and in particular to a method, a system and a vehicle for an autonomous vehicle to park in a parking space. Background Art

[0002] An autonomous vehicle (also known as a driverless vehicle) can rely on the collaborative cooperation of artificial intelligence, visual computing, radar, monitoring devices, etc., so that the in-vehicle computer can automatically and safely control the autonomous vehicle without any human operation. According to the "Specification for the Setting of On-street Parking Spaces in Urban Roads", the standard size design of parking spaces is mainly divided into two sizes, large and small. The large parking space is 15.6 meters long and 3.25 meters wide, suitable for medium and large vehicles; the small parking space is 6 meters long and 2.5 meters wide, suitable for small vehicles.

[0003] The arrangement of parking spaces can be divided into parallel, inclined and perpendicular: the standard length of the parallel parking space is 6 meters and the width is 2.5 meters; the inclined length of the inclined parking space reaches 6 meters and the width is 2.8 meters, and the perpendicular distance between the two diagonal lines should be kept at the standard of 2.5 meters; the length of the perpendicular parking space is greater than or equal to 5 meters, usually set to 6 meters, and the width is 2.5 meters. Generally, 2.5x5.3m is the best standard parking space size. Inclined and perpendicular parking spaces should not be used for large vehicles. Different parking spaces are matched with vehicles of different external dimensions (mini cars, small cars, light vehicles, medium vehicles, large buses, articulated buses, large trucks and articulated trucks). If parked in a parking space of the wrong type, it will cause waste of public resources or prevent other vehicles from parking.

[0004] During the driving process of an autonomous vehicle, there are significant differences between the parking stage and the general driving stage. The space for parking is often relatively narrow, so the control accuracy requirements for autonomous vehicles in the parking stage are very high. Autonomous vehicles should consider the reasonable accuracy of the parking position more when parking, so as to better solve the problems of non-standard parking and occupying other resources. Summary of the Invention

[0005] The present invention provides a method for an autonomous vehicle to park in a parking space, which ensures the accuracy of the autonomous vehicle parking in the parking space and can park in the parking space without interruption at one time.

[0006] The object of the present invention is achieved by the following technical solutions.

[0007] A method for an autonomous vehicle to park in a parking space includes the following steps:

[0008] Judge whether the type of the vehicle to be parked corresponds to the target parking space. If it corresponds, execute the parking step:

[0009] A trajectory planning coordinate system is established with the center of the assumed vehicle in the target parking space as the coordinate origin. The front and back of the driving direction of the assumed vehicle are the x-axis, and the front of the vehicle is positive. The left and right sides of the driving direction of the assumed vehicle are the y-axis, and the right side is positive. The up and down directions of the driving direction of the assumed vehicle are the z-axis, and the up direction is positive;

[0010] The vehicle coordinate system of the assumed vehicle is transformed with reference to the trajectory planning coordinate system;

[0011] Select the parking point after the assumed vehicle exits the warehouse according to the set rules and generate the optimal trajectory;

[0012] Judge whether the current position of the vehicle to be parked is on any trajectory. If it is, execute the parking-in-position step. If not, execute the adjustment step to adjust the vehicle to a point on any trajectory, and then execute the parking-in-position;

[0013] The adjustment step is to establish a coordinate system with the center of the current position of the vehicle to be parked as the coordinate origin. The front and back of the driving direction of the vehicle to be parked are the x-axis, and the front of the vehicle is positive. The left and right sides of the driving direction of the vehicle to be parked are the y-axis, and the right side is positive. The up and down directions of the driving direction of the vehicle to be parked are the z-axis, and the up direction is positive. The coordinate system surrounds the current position and the adjusted position of the vehicle to be parked. The adjusted position is located on any trajectory. The coordinate system is equally divided into n*m to establish a Map table. The Map table parameter is defined as the vehicle pose deflection angle φ. The grid division of the Map table is where a, b, c, and d are input limit values, and n and m are the number of grids divided in the coordinate system; there is a one-to-one functional relationship φ between the Map table parameter and the input i,j = f(x i , y j ); Use the piecewise bilinear interpolation algorithm to search for the parameter φ in the Map table;

[0014]

[0015] As a better solution of the present invention, the judgment of whether the vehicle type of the vehicle to be parked corresponds to the target parking space is specifically: the sensor collects and obtains the current information of the vehicle to be parked, including the image information of the vehicle surroundings obtained by the camera. The surrounding image information includes parking space information, and judges whether the vehicle type of the vehicle to be parked meets the size of the parking space; the vehicle types are minicar, small car, light vehicle, medium vehicle, large bus, small bus, large truck, and articulated truck.

[0016] As a better solution of the present invention, the transformation of the vehicle coordinate system of the assumed vehicle with reference to the trajectory planning coordinate system is specifically: let the position information of the autonomous vehicle be (x t , y t , θ t) T , the expression of the vehicle pose change per unit time is: (x t+1 , y t+1 , θ t+1 ) T = (x t , y t , θ t ) T + (△x, △y, △θ) T ; where (△x, △y, △θ) T represents the moving distance of the vehicle per unit time; the vehicle motion model is: where L is the wheelbase of the front and rear wheels, φ is the steering angle of the vehicle's front wheels; x represents the trajectory of the vehicle moving in the x-axis direction, y represents the trajectory of the vehicle moving in the y-axis direction, S represents the linear velocity of the vehicle, and θ represents the deflection angle of the vehicle at the current position; the linear velocity S of the vehicle is obtained through the IMU inertial measurement unit, and the deflection angle θ at the current position is obtained through the vehicle bus.

[0017] Another object of the present invention is to provide an automatic driving vehicle parking system, including:

[0018] A scene recognition module, which is used to receive the perception map information, identify the scene where the vehicle is located and the target parking space, and output the scene information to the scene trajectory generation module;

[0019] A scene trajectory generation module, which is used to generate an optimal trajectory starting from the hypothetical vehicle in the target parking space and ending at the parking point after the hypothetical vehicle exits according to the set rules based on the output scene information, and output it to the judgment module;

[0020] A judgment module, which is used to judge whether the current position of the vehicle to be parked is on any one of the trajectories. If it is, the parking-in-position step is executed; if not, the adjustment step is executed;

[0021] An adjustment module, which is used to adjust the vehicle to a point on any one of the trajectories, and then execute the parking-in-position step.

[0022] As a more preferable technical solution of the present invention, the adjustment module is specifically:

[0023] Taking the center of the current position of the vehicle to be parked as the coordinate origin, establish a coordinate system. The front and back in the driving direction of the vehicle to be parked are the x-axis with the head direction being positive, the left and right sides in the driving direction of the vehicle to be parked are the y-axis with the right direction being positive, and the up and down directions in the driving direction of the vehicle to be parked are the z-axis with the up direction being positive. The coordinate system encloses the current position and the adjusted position of the vehicle to be parked. The adjusted position is located on any one of the trajectories. Divide the coordinate system into n*m equal parts to establish a Map table. Define the Map table parameter as the vehicle pose deflection angle φ, and define the grid division of the Map table as where a, b, c, and d are the input limit values, and n and m are the number of grids divided in the coordinate system; there is a one-to-one functional relationship φ i,j = f(x i , y j ); Use the piecewise bilinear interpolation algorithm to look up the parameter φ in the Map table;

[0024]

[0025] Another object of the present invention is to provide a vehicle, including the automatic driving vehicle parking-in-position system.

[0026] Another object of the present invention is to provide a computer-readable storage medium, on which a computer program is stored, and the program can be loaded and run by a processor to execute the automatic driving vehicle parking-in-position method.

[0027] When the adjustment module provided by the present invention adjusts the pose, the vehicle pose deflection angle of the corresponding point can be calculated for each coordinate, so as to achieve accurate vehicle pose adjustment, so that the vehicle after adjusting the pose can be parked in the parking space without stopping once. Description of the Drawings

[0028] Figure 1 is a flowchart of the automatic driving vehicle parking-in-position method of the present invention;

[0029] Figure 2 is a schematic diagram of parking in a balanced parking space for the automatic driving vehicle parking-in-position method of the present invention;

[0030] Figure 3 is a diagram for defining vehicle model parameters of the present invention;

[0031] Figure 4 is a schematic diagram of adjusting the vehicle pose by the bilinear interpolation method of the present invention;

[0032] Figure 5 is a schematic diagram of parking in a vertical parking space for the automatic driving vehicle parking-in-position method of the present invention;

[0033] Figure 6This is a parking schematic diagram of the method for an autonomous vehicle to park in a tilted parking space according to the present invention. Detailed implementation manners

[0034] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0035] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in the present invention should be of the ordinary meaning understood by those skilled in the art to which the present invention pertains.

[0036] In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, or devices.

[0037] The present invention provides a method for an autonomous vehicle to park in a parking space, including the following steps:

[0038] Judge whether the model of the vehicle to be parked corresponds to the target parking space. If it corresponds, execute the parking step:

[0039] Establish a trajectory planning coordinate system with the center of the assumed vehicle in the target parking space as the coordinate origin. The front and back of the assumed vehicle's driving direction is the x-axis with the head direction being positive, the left and right sides of the assumed vehicle's driving direction are the y-axis with the right side being positive, and the up and down direction of the assumed vehicle's driving direction is the z-axis with the up direction being positive;

[0040] Convert the assumed vehicle's coordinate system with reference to the trajectory planning coordinate system;

[0041] Select the parking point after the assumed vehicle exits the warehouse according to the set rules and generate the optimal trajectory; the set rules involve how to avoid the influence of obstacles.

[0042] Judge whether the current position of the vehicle to be parked is on any one of the trajectories. If it is, execute the parking-in-position step. If not, execute the adjustment step to adjust the vehicle to a point on any one of the trajectories, and then execute the parking-in-position;

[0043] The present invention also provides an autonomous vehicle parking-in-position system, including:

[0044] A scene recognition module, configured to receive perception map information, recognize the scene where the vehicle is located and the target parking space, and output the scene information to the scene trajectory generation module;

[0045] A scene trajectory generation module, configured to generate an optimal trajectory with a hypothetical vehicle in the target parking space as the starting point and a parking point selected according to a set rule after the hypothetical vehicle exits the warehouse as the end point based on the output scene information, and output it to the judgment module;

[0046] A judgment module, configured to judge whether the current position of the vehicle to be parked is on any one of the trajectories. If so, execute the parking-in-position step; if not, execute the adjustment step;

[0047] An adjustment module, configured to adjust the vehicle to a point on any one of the trajectories, and then execute the parking-in-position step. Specifically, the adjustment module is: establish a coordinate system with the center of the current position of the vehicle to be parked as the coordinate origin, the front and back of the driving direction of the vehicle to be parked as the x-axis and the head direction as positive, the left and right sides of the driving direction of the vehicle to be parked as the y-axis and the right as positive, and the up and down directions of the driving direction of the vehicle to be parked as the z-axis and the up as positive. The coordinate system encloses the current position and the adjusted position of the vehicle to be parked, and the adjusted position is on any one of the trajectories. Divide the coordinate system into n*m equal parts to establish a Map table, define the Map table parameter as the vehicle pose deflection angle φ, and define the grid division of the Map table as where a, b, c, and d are input limit values, and n and m are the number of grids divided in the coordinate system; there is a one-to-one functional relationship φ between the Map table parameter and the input i,j =f(x i ,y j ); use the piecewise bilinear interpolation algorithm to look up the parameter φ in the Map table;

[0048]

[0049] The present invention also provides a vehicle, which is equipped with the automatic driving vehicle parking-in-position system.

[0050] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and the program can be loaded and run by a processor to execute the automatic driving vehicle parking-in-position method.

[0051] Embodiment 1

[0052] This embodiment relates to an example of an autonomous vehicle pulling over in front of a roadside parallel parking space. The specific position of the vehicle is as Figure 2 shown, and the specific steps are as follows:

[0053] S1. Based on the information collected by the camera, perform parking space detection using the YOLOv3 method, detect the size of the parking space, determine that the parking space type is a small car parking space, and this vehicle is also a small car. After judgment, this vehicle meets the detected parking space.

[0054] S2. Establish an environment model, a vehicle model, define parameters, and establish a motion model. Search backward for the vehicle's motion trajectory starting from the coordinate origin, and determine a series of model parameters corresponding to the respective trajectories. The coordinate origin is the center point of the vehicle's parking position in the parking space.

[0055] First, determine the coordinate system according to the established environment model, and describe the vehicle's motion trajectory on this coordinate system. Let the current vehicle position information be (x t , y t , θ t ). T , then the expression for the vehicle pose change within a unit time is: (x t+1 , y t+1 , θ t+1 ) T = (x t , y t , θ t ) T + (△x, △y, △θ) T . Where (△x, △y, △θ) T represents the vehicle's movement distance within a unit time, where where is the vehicle model.

[0056] The established vehicle motion model:

[0057] L represents the wheelbase of the front and rear wheels, φ represents the steering angle of the vehicle's front wheels, as Figure 3 shown.

[0058] S3. If the current position of the vehicle is not on a reasonable motion trajectory, then the starting point of parking needs to be adjusted, and the bilinear interpolation algorithm is used to adjust the initial pose of the vehicle; when using the bilinear interpolation algorithm of the present invention for pose adjustment, as shown in Figure 2 , if the current vehicle position is at point A (or point B), it can be determined that the current vehicle is on the vehicle motion trajectory searched backward, and no pose adjustment is required. However, in this embodiment, the current vehicle is at point C, and the vehicle pose is relatively deviated and not on the vehicle motion trajectory searched backward, so the vehicle pose needs to be adjusted. According to different actual requirements (such as avoiding obstacles), the vehicle is adjusted from point C to point A (or point B).

[0059] According to the improved bilinear interpolation method, the vehicle pose can be accurately adjusted to the appropriate position. The specific operation method is as follows:

[0060] First, taking point C of the current parking space as the coordinate origin, a coordinate system is established, which is characterized in that the coordinate system encloses point A (or B). That is, the coordinate system should enclose the current vehicle position and the position after adjusting the pose. The local part of the coordinate system is equally divided into n*m parts, a map table is established, and its map table parameter is defined as the vehicle pose deflection angle φ. Define the grid division of the map table.

[0061] Among them, a, b and c, d are the input limit values, n and m are the number of grids divided in the coordinate system, and there is a one-to-one functional relationship between the map table parameter and the input: φ i,j = f(x i , y j ). Use the improved piecewise bilinear interpolation algorithm to look up the parameter φ of the map table:

[0062]

[0063] As Figure 4 shown, when adjusting the pose, the vehicle pose deflection angle φ of the corresponding point can be calculated for each coordinate (x, y). Thus, the accurate adjustment of the vehicle pose is achieved. So that the vehicle after adjusting the pose can be parked in the parking space without stopping once.

[0064] S4. Park the vehicle with the adjusted pose at point A (or B) into the target parking space at one time.

[0065] Embodiment 2

[0066] This embodiment proposes an example of a driverless vehicle pulling over before a vertical parking space. As Figure 5 shown, the specific steps include the following steps:

[0067] S1. According to the information collected by the camera, use the YOLOv3 method for parking space detection, detect the size of the parking space, determine that the parking space type is a small parking space, and judge whether the vehicle conforms to the detected parking space. The judgment result is conforming.

[0068] S2. Establish an environment model, a vehicle model, define parameters and establish a motion model. Starting from the coordinate origin (the parking position of the vehicle in the parking space is the coordinate origin), search for the vehicle motion trajectory backward and determine a series of model parameters corresponding to the corresponding trajectory.

[0069] S3. Judge whether it is necessary to adjust the starting point of parking; in this embodiment, it is judged that there is no need to adjust the vehicle pose. As can be seen from Figure 5 , the vehicle does not need to adjust the vehicle pose at point D (or E), and can directly park in the garage at one time, that is, directly perform S4.

[0070] S4. Park the vehicle (such as point D or E in the figure) into the target parking space in one go. Figure 5

[0071] Embodiment 3

[0072] This embodiment presents an example of a driverless vehicle pulling over to the side before an inclined parking space. As shown in the figure, the specific steps include the following: Figure 6

[0073] S1. According to the information collected by the camera, use the YOLOv3-based method to detect the parking space, detect the size of the parking space, determine that the type of the parking space is a small parking space, and judge whether the vehicle conforms to the detected parking space. The judgment result is affirmative.

[0074] S2. Establish an environment model, a vehicle model, define parameters, and establish a motion model. Start from the origin of coordinates (the parking position of the vehicle in the parking space is the origin of coordinates) and search backward for the vehicle's motion trajectory, and determine a series of model parameters corresponding to the corresponding trajectory.

[0075] S3. Judge whether it is necessary to adjust the starting point of parking; in this embodiment, it is judged that it is necessary to adjust the vehicle pose. As can be seen from the figure, it is necessary to adjust the pose from point G to point F through the bilinear interpolation method. The specific operation method can be seen in Embodiment 1. Figure 6

[0076] S4. Park the vehicle with the adjusted pose (such as point F in the figure) into the target parking space in one go. Figure 4

[0077] The above embodiments prove that the parking method provided by the present invention enables the vehicle after adjustment to park into the parking space in one go without pausing.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent substitutions. Any modification or equivalent substitution without departing from the spirit and scope of the present invention shall be covered by the scope of the claims of the present invention.

Claims

1. A method for a self-driving vehicle to park in a parking space, characterized in that: The method includes the following steps: determining whether the vehicle type of the vehicle to be parked corresponds to the target parking space, and if so, executing the parking step: A trajectory planning coordinate system is established with the center of the assumed vehicle in the target parking space as the coordinate origin. The front and back of the driving direction of the assumed vehicle are the axis, and the front of the vehicle is the positive direction. The left and right sides of the driving direction of the assumed vehicle are the axis, and the right is the positive direction. The up and down directions of the driving direction of the assumed vehicle are the axis, and the up is the positive direction; Converting the coordinate system of the hypothetical vehicle into a vehicle coordinate system with reference to the trajectory planning coordinate system; Selecting the parking point of the hypothetical vehicle after leaving the garage according to the set rules and generating an optimal trajectory; Determining whether the current position of the vehicle to be parked is on any of the trajectories. If it is, execute the parking-in-place step. If not, execute the adjustment step to adjust the vehicle to a point on any of the trajectories, and then execute the parking-in-place; The adjustment step is to establish a coordinate system with the center of the current position of the vehicle to be parked as the coordinate origin. The front and back in the driving direction of the vehicle to be parked are axis, and the front of the vehicle is positive. The left and right sides in the driving direction of the vehicle to be parked are axis, and the right side is positive. The up and down directions in the driving direction of the vehicle to be parked are axis, and the up side is positive. The coordinate system encloses the current position and the adjusted position of the vehicle to be parked. The adjusted position is located on any one of the trajectories. Divide the coordinate system n * m equally into parts to establish a Map table. Define the Map table parameter as the vehicle pose deflection angle , and define the grid division of the Map table as ， Among them a 、 b and c 、 d are the input limit values, n and m are the number of grids divided within the coordinate system; there is a one-to-one functional relationship between the Map table parameters and the input ; use the piecewise bilinear interpolation algorithm to find the Map table parameters ; 。 2. The method for a self-driving vehicle to park in a parking space according to claim 1, characterized in that: The determination of whether the vehicle type of the vehicle to be parked corresponds to the target parking space is specifically as follows: The sensor collects the current information of the vehicle to be parked, including the image information around the vehicle obtained by the camera. The surrounding image information includes the parking space information, and determines whether the vehicle type of the vehicle to be parked meets the size of the parking space; the vehicle types are mini cars, small cars, light vehicles, medium-sized vehicles, large buses, small buses, large trucks, and articulated trucks.

3. The method for a self-driving vehicle to park in a parking space according to claim 1, characterized in that: The vehicle coordinate system transformation of the assumed vehicle with reference to the trajectory planning coordinate system is specifically as follows: Let the position information of the autonomous vehicle be , and the expression of the vehicle pose change within a unit time is: ; Among them ; It represents the moving distance of the vehicle per unit time; the vehicle motion model is: Among them L is the wheelbase of the front and rear wheels, is the steering angle of the front wheel of the vehicle; represents the trajectory of the vehicle moving in the axis direction, represents the moving trajectory of the vehicle in the axis direction, S represents the linear velocity of the vehicle, represents the deflection angle of the vehicle at the current position; the linear velocity of the vehicle is obtained through the IMU inertial measurement unit , and the deflection angle at the current position is obtained through the vehicle bus .

4. A system for a self-driving vehicle to park in a parking space, characterized in that: It includes: A scene recognition module, configured to receive the perception map information, recognize the scene where the vehicle is located and the target parking space, and output the scene information to the scene trajectory generation module; A scene trajectory generation module, configured to generate an optimal trajectory with a hypothetical vehicle in the target parking space as the starting point and the parking point of the hypothetical vehicle after leaving the garage selected according to the set rules as the end point according to the output scene information, and output it to the judgment module; A judgment module, configured to determine whether the current position of the vehicle to be parked is on any of the trajectories. If it is, execute the parking-in-place step. If not, execute the adjustment step; An adjustment module, configured to adjust the vehicle to a point on any of the trajectories, and then execute the parking-in-place step; specifically, the adjustment module is: Taking the center of the current position of the vehicle to be parked as the coordinate origin, establish a coordinate system. The front and back in the driving direction of the vehicle to be parked are axis, and the front of the vehicle is positive. The two sides in the driving direction of the vehicle to be parked are axis, and the right is positive. The up and down directions in the driving direction of the vehicle to be parked are axis, and the up is positive. The coordinate system encloses the current position and the adjusted position of the vehicle to be parked. The adjusted position is located on any one of the trajectories. Perform n * m equal division on the coordinate system to establish a Map table. Define the Map table parameter as the vehicle pose deflection angle . Define the grid division of the Map table as, where a 、 b and c 、 d are the input limit values, n and m are the number of grids divided in the coordinate system ; There is a one-to-one functional relationship between the Map table parameter and the input ; Use the piecewise bilinear interpolation algorithm to perform Map table lookup for parameters ; 。 5. A vehicle, characterized in that: An automatic driving vehicle parking-in-place system as described in claim 4.

6. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and the program can be loaded and run by a processor to execute the automatic driving vehicle parking-in-place method as described in any one of claims 1 to 3.

Images