Automatic parking trajectory planning method and electronic equipment for narrow vertical parking spaces

Through the automatic parking trajectory planning method, the trajectory planning combined with five-degree polynomial and geometric method is used to solve the problem of inefficient automatic parking in narrow vertical parking lots, and the efficient and smooth parking of vehicles in narrow parking spaces is achieved, which improves the efficiency and smooth parking of vehicles in narrow parking spaces. Comfort and control accuracy of the parking process.

CN115489515BActive Publication Date: 2024-11-19WUHAN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211040726.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-08-29
Publication Date
2024-11-19
Estimated Expiration
2042-08-29

AI Technical Summary

Technical Problem

The existing automatic parking technology is inefficient in narrow vertical parking lots, with multiple front and rear pivoting and gear switching, and it is difficult to meet the comfort requirements in narrow parking lot environments.

Method used

Using an automatic parking trajectory planning method, the coordinate system is established through perceived information, the one-time parking position is calculated, and the trajectory planning combined with five-degree polynomial and geometric method can be used to achieve smooth parking of the vehicle from the initial position to the parking space, reducing gear position Switch and maneuver.

Benefits of technology

It improves parking efficiency, reduces the demand for front passage space of parking spaces, enhances the comfort and control accuracy of the parking process, and adapts to narrow parking space environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115489515B_ABST
    Figure CN115489515B_ABST
Patent Text Reader

Abstract

The present invention discloses an automatic parking trajectory planning method and electronic device for narrow vertical parking spaces, including the following steps: according to the vehicle perception information, delineate the initial range that can take over the vehicle to complete automatic parking; according to the relative position between the vehicle and the parking space, establish a coordinate system, and import the parameter model of the vehicle's surrounding environment; complete the first segment of trajectory planning based on the quintic polynomial, so that the vehicle drives to a position that can be parked once; based on the geometric method and the cubic polynomial coupling method, the second segment of trajectory planning and the third segment of planning are completed in sequence, so that the vehicle can park once from the position that can be parked once into the parking space, and finally park in the middle of the parking space; finally, the calculated trajectory information is output in the form of a matrix to provide data for the control of the vehicle. Avoid multiple forward and backward maneuvers and gear switching, save the parking trajectory of the front channel space of the parking space to the greatest extent, and meet the comfort requirements of Du Yu's parking process in a narrow parking lot environment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of automatic driving in automobile-related artificial intelligence technology, and in particular relates to an automatic parking trajectory planning method and electronic equipment for narrow vertical parking spaces. Background Art

[0002] Nowadays, the automobile industry is developing rapidly, automobile-related technologies are improving rapidly, and people's demand for travel tools is also increasing; many automobile products have begun to launch intelligent assisted driving and unmanned driving functions to help drivers drive the vehicle, improve the driving safety of the vehicle, and reduce the fatigue caused by long-term driving.

[0003] In parking lots or parks, the environment is relatively fixed and the types of obstacles are relatively simple. Therefore, autonomous driving in parking lots is the most widely used technical environment. At the same time, the number of motor vehicles is increasing, resulting in increasingly tight parking spaces, and the area of ​​many parking lots is very limited. Faced with the vertical parking lot environment with narrow passages, a few existing automatic parking strategies will not be able to complete parking. Most will use the traditional multi-step parking method, which completes parking by switching gears multiple times in front of the parking space and moving back and forth. This method is inefficient and has the possibility of failure. At the same time, multiple parking and gear switching place high demands on the accuracy of tracking control. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide an automatic parking trajectory planning method and electronic equipment for narrow vertical parking spaces in response to the above-mentioned problems, so as to avoid multiple forward and backward maneuvers and gear switching, save the parking trajectory of the front channel space of the parking space to the greatest extent, and at the same time meet the comfort requirements of Du Yu's parking process in a narrow parking lot environment.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A method for automatic parking trajectory planning for narrow vertical parking spaces, characterized by comprising the following steps:

[0007] S1: Based on the vehicle perception information, a coordinate system is established to define the initial range in which the vehicle can be taken over from the initial position to complete automatic parking;

[0008] S2: According to the relative position between the initial range of the vehicle and the parking space, a parameter model of the vehicle's surrounding environment is imported into the coordinate system to calculate the position that can be parked at one time;

[0009] S3: Complete the first segment trajectory planning based on the quintic polynomial, so that the vehicle moves from the initial position to a position where it can be parked in one go;

[0010] S4: completing the second segment trajectory planning based on the geometric method and the cubic polynomial coupling method, so that the vehicle can be parked in the parking space at one time from the position that can be parked at one time;

[0011] S5: The third trajectory is planned based on the fifth-order polynomial, so that the vehicle finally stops in the middle of the parking space;

[0012] S6: Output the calculated trajectory information in the form of a matrix to provide data for vehicle control.

[0013] In the above technical solution, step S1 establishes a coordinate system according to the initial position of the parking space based on the perception information data, and defines an initial parking range on one side in front of or above the parking space, wherein the initial range is as follows: the longitudinal boundary of the vehicle close to one side of the parking space is the lower vertical coordinate of the boundary, and the length of the frame beam spaced from the boundary vertical coordinate is set as the upper vertical coordinate of the boundary; the boundary point on the one side in front of or above the parking space is the proximal horizontal coordinate of the initial range, and the center point of the frame beam of the rear axle of the vehicle located at the upper part of the parking space is set as the far horizontal coordinate of the initial range.

[0014] In the above technical solution, in the initial range of S1, the center point of the frame crossbeam of the rear axle of the vehicle is located at the upper part of the parking space, 1m to 4m away from the upper edge line of the parking space; the length of the frame crossbeam is set to 8m according to the size parameters of the position of the frame crossbeam of a general vehicle and the position of the longitudinal beams on both sides; the initial range is defined on the front right side of the parking space as follows:

[0015] b_4=b+1

[0016] b_5=b+4

[0017] a_5=a+8;

[0018] Among them, (a, b) is the right boundary point of the parking space, a is the horizontal coordinate of the right boundary of the parking space; b is the vertical coordinate of the right boundary of the parking space;

[0019] (a_4,b_4) are the coordinates of the lower left boundary point of the initial position;

[0020] (a_5,b_5) are the coordinates of the upper right boundary point of the initial position;

[0021] The steps for defining the initial range on the left side in front of the parking space are similar to those above.

[0022] In the above technical solution, step S2 calculates the center of the trajectory (center_x, center_y) of the vehicle that will not collide with the parking space during the parking process based on the kinematic model of the vehicle, the right corner collision constraint, and the left corner collision constraint.

[0023] Among them, the right corner collision constraint:

[0024]

[0025] Where:

[0026] Rmin: minimum turning radius of the car;

[0027] W: width of the car;

[0028] safe_dis: The minimum safe distance that the car needs to maintain from the parking space boundary;

[0029] a is the horizontal coordinate of the right corner of the boundary point of the parking space, and b is the vertical coordinate of the right corner of the boundary point of the parking space;

[0030] Left corner collision constraint:

[0031] center_x-R min≥a_2+W / 2+safe_dis

[0032] Among them; a-2 is the horizontal coordinate of the left corner, b-2 is the vertical coordinate of the right corner;

[0033] We obtain: center_x=a_2+safe_dis+W / 2+R min

[0034]

[0035] In the above technical solution, step S3 obtains the position or initial position that can be parked once from the center of the trajectory: the ordinate of the position that can be parked once is directly obtained according to the ordinate of the center trajectory:

[0036] y_2=center_y+Rmin

[0037] The distance between the horizontal coordinates x_2 and x_1 is set to be positively correlated with the distance between the vertical coordinates y_2 and y_1, which is:

[0038] x_2-x_1=5*(y_2-y_1).

[0039] In the above technical solution, the first segment of trajectory f in step S3 is the segment from the initial position to the position where the vehicle can be parked once. The first segment of trajectory planning is performed using a five-order polynomial to ensure that the jump of this segment of trajectory is minimal, that is, Take the minimum value;

[0040] The time t taken by the vehicle from the initial position to the position where it can be parked once is as follows:

[0041] t0=0;

[0042] t1=(a_2-a_1+1) / v_target;

[0043] The target speed is set to v_target = 0.8 m / s.

[0044] In the above technical solution, in step S4, the second segment trajectory is planned by combining the geometric method and the cubic polynomial. The path is planned by the geometric method of connecting straight lines and curves, and the speed is planned by the cubic polynomial. Finally, the two are combined to plan the trajectory:

[0045] The vehicle's path s(t) = a0 + a1(t-t2) + a2(t-t2) 2 +a3(t-t2) 3 ,t2≤t≤t3;

[0046]

[0047] Where t2 is the initial time point of the second trajectory, t3 is the end time point, a0, a1, a2, a3 are polynomial algebras, s0 represents the initial position, v0 represents the initial velocity, h represents the displacement of the second trajectory, and T represents the time consumed by the second trajectory.

[0048] In the above technical solution, the planning of the third segment of the trajectory in step S5 uses the same method as S4 for trajectory planning, and the end position is limited to the center position of the parking space to ensure the parking standardization of the vehicle.

[0049] In the above technical solution, the speed of the end point in the second segment planning in S4 is the same as the speed of the initial point in the third segment planning in S5, and the steering angles at the two points are set the same. This ensures the smooth connection of the two segments of the trajectory and enables a smooth transition.

[0050] Preferably, the sum of the speeds of the end points in the second planning section in S4 is set to 0.6 m / s, and the speed of the initial point of the third planning section is also 0.6 m / s, so as to ensure the smooth connection of the two trajectories. At the same time, the steering angles of the two are also 0, which can fully achieve a smooth transition.

[0051] An electronic device, comprising: one or more processors; a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include instructions for performing the following steps:

[0052] S1: a device for defining an initial range capable of taking over the vehicle to complete automatic parking according to vehicle perception information;

[0053] S2: According to the relative position between the vehicle and the parking space, a coordinate system is established and a parameter model of the vehicle's surrounding environment is imported;

[0054] S3: Complete the first segment trajectory planning based on the quintic polynomial, so that the vehicle drives to a position where it can be parked in one go;

[0055] S4: The second segment trajectory planning is completed based on the geometric method and the cubic polynomial coupling method, so that the vehicle can be parked in one time from the position where it can be parked in one time;

[0056] S5: The third trajectory is planned based on the fifth-order polynomial, so that the vehicle finally stops in the middle of the parking space;

[0057] S6: Output the calculated trajectory information in the form of a matrix to provide data for vehicle control.

[0058] Therefore, the present invention provides an automatic parking trajectory planning method and device for narrow vertical parking spaces. After re-trajectory planning using the strategy mentioned above, the vehicle saves 15% of the longitudinal space during parking, which reduces the demand for longitudinal space during parking and can better adapt to the increasingly narrow parking spaces nowadays.

[0059] In step S2, the center position of the circle during direct parking is first determined, and then the first step of adjusting the trajectory is determined in S3, so that the vehicle can park without shifting gears during the parking process, which improves comfort and reduces control pressure. In S5, a lateral movement process based on a fifth-order polynomial is added after the vehicle has been initially parked in the parking space, so that the vehicle can be parked in the center of the parking space and the demand for the longitudinal space of the front passage of the parking space can be minimized.

[0060] The beneficial effects of the present invention are:

[0061] Under the planning strategy of the present invention, the vehicle only switches gears once when it is adjusted to a parking position. Compared with the traditional parking method, the number of maneuvers is reduced, and the connection of the curvature is smoother, without the process of turning on the spot.

[0062] The geometric method is combined with the relative position of the vehicle and the parking space and the vehicle's kinematic model to determine the starting and ending points of the vehicle's movement. Then the trajectory planning of the vehicle's movement process is completed using a quintic polynomial to obtain a parking trajectory that can save the space in the aisle in front of the parking space to the greatest extent, while meeting Du Yu's requirements for comfort during the parking process.

[0063] The longitudinal space of the front passage of the parking space can be used less. The program reduces the demand for longitudinal space by increasing the lateral movement of the last section of the path. At the same time, when taking over the parking, the vehicle is adjusted to a position where it can be parked in one go, avoiding the repeated gear switching near the parking space, improving comfort and reducing the psychological pressure on the driver. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The initial range of the center point of the rear axle of the vehicle specified in the present invention is

[0065] Figure 2 It is a right collision constraint diagram of an embodiment of the automatic parking trajectory planning method for a narrow vertical parking space of the present invention.

[0066] Figure 3 This is a left collision constraint diagram according to an embodiment of the present invention.

[0067] Figure 4 This is the first step trajectory target position of an embodiment of the present invention.

[0068] Figure 5 The first step trajectory of an embodiment of the present invention is shown.

[0069] FIG. 6( a ) shows the change of the first step coordinate x over time in one embodiment of the present invention.

[0070] FIG6( b ) is a first-step speed planning output curve according to an embodiment of the present invention.

[0071] Figure 7 This is the second path planning of an embodiment of the present invention.

[0072] Figure 8 This is the second speed planning of an embodiment of the present invention.

[0073] Fig. 9 It is the target point of the third trajectory of an embodiment of the present invention.

[0074] FIG. 10( a ) is a display of the third trajectory of an embodiment of the present invention.

[0075] FIG10( b ) is a third-stage speed planning according to an embodiment of the present invention.

[0076] FIG. 11( a ) shows the result of traditional planning in the environment of an embodiment of the present invention.

[0077] FIG. 11( b ) is a comparison between an embodiment of the present invention and traditional planning results. DETAILED DESCRIPTION

[0078] In order to better understand the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0079] The automatic parking trajectory planning method for narrow vertical parking spaces of the present invention takes a vertical parking lot with a front passage width of 4.5m as an example environment. Figure 1-Figure 10(b) , the technical route and key points of the present invention are as follows:

[0080] Step 1:

[0081] First, after the vehicle recognizes the parking space, in order to improve the driver's comfort during parking and reduce the number of times the vehicle needs to be moved during parking, in the first preparation stage, the vehicle is first adjusted from the initial position to a parking position that can be parked in one go, avoiding multiple steps of parking near the parking space.

[0082] The parking position is calculated using the following method. First, according to the vehicle kinematic model, the low-speed parking process of the car can be regarded as the movement of the center point of the rear axle of the vehicle. When the vehicle maintains the minimum turning radius, the position of the rear wheels will be the position of the vehicle's minimum movement radius.

[0083] Therefore, by calculating the center of the trajectory (center_x, center_y) of the vehicle that will not collide with the parking space during the parking process, the first parking position or initial position of the vehicle can be obtained.

[0084]

[0085] In the formula,

[0086] Rmin: minimum turning radius of the car;

[0087] W: width of the car;

[0088] safe_dis: The minimum safe distance that the car needs to maintain from the parking space boundary, which is 0.2m in the calculation;

[0089] The calculation of the parking position must not only meet the constraint of no collision with the minimum radius, but also meet the constraint of no collision with the parking space boundary with the maximum movement radius of the vehicle, such as Figure 1 and 2 ,From the vehicle kinematic model, it can be known that during the parking process, the vehicle movement radius is located at the rear corner of the vehicle, that is, the rear corner of the vehicle is the first to collide with the parking space boundary.

[0090] In order to avoid collision between the rear corner of the vehicle and the parking space, the maximum movement radius of the vehicle should maintain a certain distance from the boundary of the parking space. Therefore, the horizontal coordinate of the center of the movement circle can be obtained:

[0091] center_x-Rmin≥a_2+W / 2+safe_dis (2)

[0092] Combined with the image, we can see that when center_y is required to be small, center_x should also be small. Therefore, in order to save longitudinal space, we directly move the initial position of the vehicle entering the parking space as close to the edge of the parking space as possible, that is, determine:

[0093] center_x=a_2+safe_dis+W / 2+Rmin

[0094] By inverting the value of center_x and substituting it into formula 1, the target value of center_y can be calculated:

[0095]

[0096] Step 2:

[0097] After obtaining the center point of the arc when parking, you can start parking. In the initial state, when the driver stops the vehicle and turns on the automatic parking function, the car first automatically adjusts to a one-step parking position. Figure 3 and 4 The purpose of adjusting to the single parking position first is to reduce the number of gear shifting and forward and backward maneuvers during the parking process, improve the comfort of the parking process, and reduce the pressure of the tracking control process.

[0098] The initial x_1 and y_1 are obtained by the driver based on the relative position between the vehicle and the parking space when the car is first parked. Since it is a one-step parking state, the value of y_2 is the ordinate of the center of the arc when parking:

[0099] y_2=center_y+Rmin

[0100] The value of X_2 is related to the driver's parking position. If the distance between x_2 and x_1 is too close, the steering angle planned by the quintic polynomial path will be too large, causing the vehicle's maximum steering angle to be unable to meet the planning requirements, thereby causing the tracking control to fail. Therefore, the distance between x_2 and x_1 is set to be positively correlated with the distance from y_2 to y_1.

[0101] x_2-x_1=5*(y_2-y_1)

[0102] After obtaining the coordinates of the starting and ending positions of the adjustment phase, we can start the first step of trajectory planning. When the parameters of two points are known, we optimize the trajectory between them. What we need to consider is to optimize the rate of change of the slope to minimize its value, that is, to solve the process of minimizing Jerk. In the motion planning of the vehicle, comfort is a very important indicator. In physics, the physical quantity that measures comfort is jump.

[0103] In order to ensure comfort, the acceleration needs to change smoothly. For a trajectory s = f(t), when When it is the smallest, it can meet our needs. The following boundary conditions exist in this trajectory:

[0104] ls(0)=s0 s(T)=s n

[0105] s′(0)=v0 s′(T)=v n

[0106] s″(0)=a0 s″(T)=a n

[0107] To satisfy the six boundary conditions at the same time, it is necessary to introduce a functional problem with boundary condition constraints. To solve this functional extreme value problem with 6 constraints, it is necessary to introduce a quintic polynomial. Construct the matrix:

[0108]

[0109] Bx=(x0vx0ax0x1vx1ax1)

[0110] Multiply the two together:

[0111] a=Ax*Bx

[0112] The curve of coordinate x changing with time t can be obtained:

[0113] xt=a(1)+a(2)*dt+a(3)*dt 2 +a(4)*dt 3 +a(5)*dt 4 +a(6)*dt 5

[0114] Next, we need to solve the relationship between the coordinate y and x:

[0115]

[0116] By = (y0vy0ay0y1vy1ay1)

[0117] Multiply the two together:

[0118] b=Ay*By

[0119] You can get the curve of the change of coordinate y with x, which is the path of the vehicle.

[0120] yt=b(1)+b(2)*dt+b(3)*dt 2 +b(4)*dt 3 +b(5)*dt 4 +b(6)*dt 5

[0121] In this equation, we know the values ​​of the x-matrix and the y-matrix. x0, vx0, ax0 represent the initial position and the horizontal coordinate, velocity, and acceleration respectively; y0, vy0, ay0 represent the vertical coordinate, velocity, and acceleration of the initial position respectively. Secondly, the time t0 and t1 are also known, representing the time of the initial position and the final position respectively. We use the target velocity to solve the two time points.

[0122] v_target=0.8m / s

[0123] t0=0 t1=(a_2-a_1+1) / v_target

[0124] In this way, the X, Y, and T matrices are all known, and we can easily find the two matrices.

[0125] After that, we set the time interval △t and substitute it into the T matrix, and we can use the A matrix and the T matrix to calculate the position, velocity and acceleration of the point between the initial position and the end position.

[0126] Substituting the obtained initial position (x_1, y_1) and end position (x_2, y_2) values ​​into the above formula, the trajectory point can be obtained. Figure 6 (a) shows the change of the coordinate x over time in the first step of the present invention.

[0127] By taking the derivative of x and y, we can get the curve of velocity changing with time.

[0128] dx_dt=a(2)+2*a(3)*dt+3*a(4)*dt 2 +4*a(5)*dt 3 +5*a(6)*dt 4

[0129] dy_dt=b(2)+2*b(3)*dt+3*b(4)*dt 2 +4*b(5)*dt 3 +5*b(6)*dt 4

[0130]

[0131] That is, the velocity is obtained and coupled with the path obtained above to obtain the trajectory planning, as shown in Figure 6(b).

[0132] Step 3:

[0133] After completing the first trajectory, the second step of planning is performed, such as Figure 6(a) , 6(b)The second step consists of a reversing process and an arc driving process. After completing the second step, the vehicle will drive to a vertical position in the parking space. The vehicle's straight-line backward process keeps the steering wheel angle at 0, and the arc backward process is processed according to the minimum turning radius (maximum turning angle).

[0134] As can be seen from the figure, the waypoints of the second path are:

[0135] x_3=center_x y_3=center_y+Rmin

[0136] x_4=center_x-Rmin y_4=center_y

[0137] This section uses the speed and path decoupling method, uses a cubic polynomial for speed planning, outputs the change data of speed v relative to displacement s, and finally couples the speed and path to form a planned trajectory. Because the second and third trajectories do not shift gears, and the curvature at the junction of the two end paths does not change suddenly, there is no need to stop at (x_4, y_4). During the speed planning process, the speed of point (x_4, y_4) is set to 1m / s.

[0138] Time required for the process:

[0139] t3=(a_2-a_3) / v_target t4=(2*π*Rmin / 4) / v_target

[0140] The velocity planning of this segment of trajectory uses a cubic polynomial. After the time of the initial position and the end position is determined, the coordinates of each point are known. By converting the coordinates into displacement, the relationship between displacement and time can be obtained, and then the velocity planning of this segment of trajectory is obtained:

[0141] s(t)=a0+a1(t-t2)+a2(t-t2) 2 +a3(t-t2) 3 ,t2≤t≤t3

[0142]

[0143] The process from (x_2, y_2) to (x_4, y_4) consists of two sections. Since the vehicle does not stop at (x_3, y_3), the speed here is planned to be 1m / s. Substituting the parameters into the formula, we get the planned speed of the second section.

[0144] Step 4:

[0145] In order to save longitudinal space at the beginning of planning, the vehicle is not in the middle of the parking space in this state. Therefore, it is necessary to add a track at the last time to make the vehicle as central as possible to facilitate the driver of the vehicle and the driver of the adjacent vehicle to get on and off. Figure 8-Figure 10(b) .

[0146] The initial position of the vehicle is known, that is, (x_4, y_4), and the final parking position is Figure 8 As shown:

[0147] x_5=(a_2+a) / 2

[0148] y_5=b_3+safe_dis+Lr

[0149] During the planning process, attention should be paid to the collision between the front edge of the vehicle and the parking space boundary. The goal of saving longitudinal space should be met under the premise of ensuring that the vehicle does not collide. The planning of the last segment of the trajectory is the same as the first segment, using the quintic polynomial trajectory planning method. It is connected with the second segment of the trajectory planning above, so the speed at the initial position (x_4, y_4) is set to 1m / s.

[0150] Combined with the output results of the vehicle parking process trajectory planning, it can be seen that:

[0151] The vehicle's trajectory is very smooth, and the slope of the curve has no sudden changes, which can meet the needs of smoothness and comfort and provide a basis for precise trajectory tracking control.

[0152] Using traditional planning strategies, for the parking environment shown in the adopted example, Figure 11(a) and 11(b) As shown, during the parking process, the coordinate position of the left front corner of the vehicle is 12.93, which has collided with the boundary and cannot meet the environment of the target parking lot. In the whole process, there were two gear switches. In the middle of the adjustment process, there were two stages that required on-site special operations, which had an adverse effect on the vehicle's steering gear. After re-planning using the optimization strategy disclosed in this patent, in the optimized trajectory, the highest point of the vehicle was reduced to 12.15, and the longitudinal coordinate was reduced by 15%. There was no boundary collision. At the same time, the number of gear switches was reduced to 1, and the path curvature changed more smoothly. At the same time, there was no on-site turning process, which improved the parking efficiency and occupant comfort. It solves the problems existing in the original planning strategy very well, and can better adapt to the current more narrow parking lot environment.

[0153] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. An automatic parking trajectory planning method for narrow vertical parking spaces, characterized in that The steps include: S1: Based on the vehicle perception information, a coordinate system is established according to the initial position of the parking space, and an initial range that can take over the vehicle from the initial position to complete automatic parking is defined on one side of the front or upper part of the parking space; the initial range is: the longitudinal boundary of the vehicle on one side of the front or upper part of the parking space is the lower boundary ordinate, and the length of the frame beam from the boundary ordinate is set as the upper boundary ordinate; the boundary point on one side of the front or upper part of the parking space is the near-end horizontal coordinate of the initial range, and the center point of the frame beam of the rear axle of the vehicle located at the upper part of the parking space is set as the far-end horizontal coordinate of the initial range; S2: According to the relative position between the initial range of the vehicle and the parking space, a parameter model of the vehicle's surrounding environment is imported into the coordinate system; and the position that can be parked at one time is calculated; Step S2 calculates the center of the trajectory (center_x, center_y) of the vehicle that will not collide with the parking space during the parking process based on the vehicle's kinematic model, the right corner collision constraint, and the left corner collision constraint. Among them, the right corner collision constraint: Where: Rmin: minimum turning radius of the car; W: width of the car; safe_dis: The minimum safe distance that the car needs to maintain between itself and the parking space boundary; a is the horizontal coordinate of the right corner of the boundary point of the parking space, and b is the vertical coordinate of the right corner of the boundary point of the parking space; left corner collision constraint: center_x-Rmin≥a_2+W / 2+safe_dis Wherein: a-2 is the horizontal coordinate of the left corner of the boundary point of the parking space, and b-2 is the vertical coordinate of the left corner of the boundary point of the parking space; We obtain: center_x = a_2 + safe_dis + W / 2 + Rmin S3: Complete the first segment trajectory planning based on the quintic polynomial, so that the vehicle moves from the initial position to a position where it can be parked in one go; S4: completing the second segment trajectory planning based on the geometric method and the cubic polynomial coupling method, so that the vehicle can be parked in the parking space at one time from the position that can be parked at one time; S5: The third trajectory is planned based on the fifth-order polynomial, so that the vehicle finally stops in the middle of the parking space; S6: Output the calculated trajectory information in the form of a matrix to provide data for vehicle control.

2. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1 is characterized in that In the initial range in step S1, the center point of the frame crossbeam of the rear axle of the vehicle is located at the upper part of the parking space, 1m to 4m away from the upper edge line of the parking space; the length of the frame crossbeam is set to 8m according to the size parameters of the position of the frame crossbeam of a general vehicle and the position of the longitudinal beams on both sides; the initial range is defined on the front right side of the parking space as follows: b_4=b+1 b_5=b+4 a_5=a+8; Among them, (a, b) is the right boundary point of the parking space, a is the horizontal coordinate of the right boundary of the parking space; b is the vertical coordinate of the right boundary of the parking space; (a_4,b_4) are the coordinates of the lower left boundary point of the initial position; (a_5,b_5) are the coordinates of the upper right boundary point of the initial position.

3. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1 is characterized in that Step S3 obtains the position that can be parked once from the center of the trajectory: The ordinate of the position that can be parked in one time can be directly obtained based on the ordinate of the center trajectory: y_2 = center_y + Rmin; The distance between the horizontal coordinates x_2 and x_1 is set to be positively correlated with the distance between the vertical coordinates y_2 and y_1, which is: x_2-x_1=5*(y_2-y_1).

4. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1 is characterized in that In step S3, the first segment of trajectory f is the segment from the initial position to the position where the vehicle can be parked once. The first segment of trajectory planning is performed using a five-order polynomial to ensure that the jump of this segment of trajectory is minimal, that is, Take the minimum value; The time t taken by the vehicle from the initial position to the position where it can be parked once is as follows: t0=0; t1=(a_2-a_1+1) / v_target; The target speed is set to v_target = 0.8 m / s.

5. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1 is characterized in that In step S4, the second segment trajectory is planned by combining the geometric method and the cubic polynomial. The path is planned by the geometric method of connecting straight lines and curves, and the speed is planned by the cubic polynomial. Finally, the two are combined to plan the trajectory: The vehicle's path s(t) = a0 + a1(t-t2) + a2(t-t2) 2 +a3(t-t2) 3 ,t2≤t≤t3; Where t2 is the initial time point of the second trajectory, t3 is the end time point, a0, a1, a2, a3, are polynomial algebras, s0 represents the initial position, v0 represents the initial velocity, h represents the displacement of the second trajectory, and T represents the time consumed by the second trajectory.

6. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1, characterized in that The third trajectory planning in step S5 is performed by combining the geometric method and the fifth-order polynomial, and the end position is limited to the center position of the parking space to ensure the parking standardization of the vehicle.

7. The automatic parking trajectory planning method for narrow vertical parking spaces according to claim 1 is characterized in that The speed of the end point in the second segment trajectory planning in S4 is the same as the speed of the initial point in the third segment trajectory planning in S5, and the steering angles at the two points are set the same.

8. An electronic device, comprising: one or more processors; A memory, one or more programs, characterized in that the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include instructions for executing the automatic parking trajectory planning method for narrow vertical parking spaces according to any one of claims 1 to 7: S1: a device for establishing a coordinate system based on vehicle perception information and defining an initial range capable of taking over the vehicle to complete automatic parking; S2: According to the relative position between the vehicle and the parking space, a parameter model of the vehicle's surrounding environment is imported to calculate the position that can be parked at one time; S3: Complete the first segment trajectory planning based on the quintic polynomial, so that the vehicle travels from the initial position to a position where it can be parked in one go; S4: The second segment trajectory planning is completed based on the geometric method and the cubic polynomial coupling method, so that the vehicle can be parked in one time from the position where it can be parked in one time; S5: The third trajectory is planned based on the fifth-order polynomial, so that the vehicle finally stops in the middle of the parking space; S6: Output the calculated trajectory information in the form of a matrix to provide data for vehicle control.