A parking path planning method based on ackerman model motion curve

Abstract

The application discloses the technical fields of parking path planning and a parking path planning method based on Ackerman model motion curves, and comprises the following steps: obtaining the door opening angle of a vehicle based on the adjacent obstacles of a target parking space, and establishing a vehicle passenger and load getting on and off comfort representation model, wherein the door opening angle of the vehicle is the angle at which the door of the adjacent obstacle is allowed to open and the angle required for the passenger and load to get on and off the vehicle. The application improves the getting on and off comfort of the passenger and load and the parking efficiency by taking the comfort of the vehicle passenger and load getting on and off and the parking time into the optimization design of the parking path, reduces the dimension of the path parameters by presetting the vehicle motion parameters and the parking path mode, improves the efficiency of the parking path planning, and meets the real-time requirement.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of parking path planning, and particularly relates to a parking path planning method based on an Ackermann model motion curve. BACKGROUND

[0002] In recent years, with the rapid development of intelligent transportation and automatic driving fields, automatic parking path planning and vehicle motion planning problems have attracted widespread attention from the academic and industrial circles. The automatic parking technology shows extremely broad application prospects, which can significantly reduce the operating pressure of the driver in the parking process, effectively simplify the parking operation process, and greatly reduce various accidents related to parking, and improve the safety and efficiency of parking. At present, the existing parking planning methods are mainly based on curve fitting methods, which usually use polynomial curves, segmented curves and B-spline curves for parking path design. Such methods are simple in design, short in operation time and strong in numerical optimization calculation capacity. In addition, methods based on graph search, sampling and intelligent algorithms, such as hybrid A* algorithm, RRT algorithm, genetic algorithm and neural network algorithm, are used for parking path design, so as to realize automatic parking operation.

[0003] The existing parking path planning methods all set the final parking state of the vehicle as an ideal situation, that is, the vehicle is accurately parked at the center of the target parking space, and the side line of the vehicle body is parallel to the edge line of the parking space. But the actual application scenarios are much more complex. First, manual and automatic parking modes coexist. Because different drivers have different levels of manual parking operation, it is often difficult to achieve the ideal parking position of the adjacent vehicles in the target parking space. For example, in the parking lot of an old community, the parking space resources are scarce and lack clear guiding signs, and some car owners park their vehicles randomly to increase the number of parking spaces, which causes irregular parking of vehicles and seriously interferes with the normal parking of subsequent vehicles in the adjacent parking spaces. Second, due to the current development level of automatic driving technology, at least one safety officer needs to be equipped in the vehicle during automatic parking to supervise. At the same time, the complex environmental conditions in the parking lot bring many challenges to parking. In the parking lot of a large shopping mall, a logistics park or a port, the opening angle of the vehicle door is limited by facilities such as load-bearing columns and walls. When the vehicle drives between containers in the park, it faces the situation of dense operation equipment and small operation space, and not only needs to avoid collision with containers and operation equipment, but also needs to consider the problem of loading and unloading of the vehicle. For example, in the container parking lot of a port, the containers are arranged compactly, and when the vehicle drives between them, the path needs to be accurately planned to ensure that the door can be normally opened and the goods can be smoothly loaded and unloaded. In the vertical parking scenario, the influence of the above factors is particularly significant. It not only reduces the comfort of passengers getting on and off the vehicle, but also may even cause the passengers to be unable to normally get on and off the vehicle. For example, in the vertical parking space of an underground parking lot, if the adjacent vehicles are parked improperly or there are load-bearing columns beside the parking space, the opening angle of the door will be limited, which will bring great inconvenience to the passengers, especially those carrying large luggage or having difficulty moving. In addition, the path generated by the existing parking path planning method usually needs additional smoothing processing and curvature continuity optimization, which increases the risk of vehicle collision to some extent and reduces the safety of parking. In complex scenarios such as airport parking lots where vehicles and personnel flow frequently, any deviation in the path processing process may cause collision between the vehicle and the surrounding objects or pedestrians.

[0004] In summary, the existing automatic parking method is not suitable for complex scenarios where obstacles exist on both sides of the target parking space, and cannot comprehensively consider the comfort of passengers getting on and off the vehicle, the convenience of loading and unloading of the vehicle, the safety of parking, and the efficiency of parking during the parking process, and the use effect is not good. SUMMARY

[0005] The purpose of the present application is to provide a parking path planning method based on the Ackermann model motion curve to solve the problem of not being suitable for complex scenarios where obstacles exist on both sides of the target parking space, and not being able to comprehensively consider the comfort of passengers getting on and off the vehicle, the convenience of loading and unloading of the vehicle, the safety of parking, and the efficiency of parking during the parking process.

[0006] To achieve the above object, the present application provides the following technical solutions.

[0007] A parking path planning method based on Ackermann model motion curve, comprising the following steps:

[0008] S1: obtaining the door opening angle of the vehicle based on the adjacent obstacles of the target parking space, and establishing a vehicle passenger load loading and unloading comfort representation model, wherein the door opening angle is the angle of the door opening of the adjacent obstacles and the angle required for the passenger load loading and unloading;

[0009] S2: constructing a vehicle basic motion curve library based on the Ackermann model, wherein the vehicle basic motion curve library comprises a plurality of vehicle basic motion curves, and the required time of the vehicle passing through is obtained according to the vehicle basic motion curve;

[0010] S3: establishing the basic motion curve to include a transition path, a maneuvering path, a driving-in path and an adjusting path, and generating a parking path near the target parking space connecting any starting state and target point state, so as to reduce the parameter dimension of the vehicle basic motion curve;

[0011] S4: establishing a parking path multi-objective optimization model, wherein the parking path multi-objective optimization model takes the vehicle completely in the target parking space, the parking path parameters, the passenger load loading and unloading, the passenger load loading and unloading into the parking space and the collision avoidance as constraint conditions;

[0012] S5: converting and solving the parking path multi-objective optimization model based on the penalty function to obtain the automatic parking path.

[0013] As a further scheme of the present application, in S1, the adjacent obstacles are divided into a plurality of obstacle line segments, and the door opening angle a is calculated according to the plurality of obstacle line segments;

[0014]

[0015] wherein a is the allowable opening angle of the door, a M is the maximum opening angle of the door without interference, a0, a m , a B , a U and respectively represent the end point Q0 when the door is not started, the end point Q m when the door is opened to the maximum angle, the lower end point O LB of the obstacle line segment, the upper end point O LU of the obstacle line segment, the angle between the line connecting the foot P ⊥ and the door rotation point P and the positive direction of the X axis, a s0 is the safety angle, and a INTo determine the permissible opening angle of the vehicle door under obstruction line interference, l D d is the length of the car door. m The distance from the door pivot point to the obstacle line segment O LU O LB The shortest distance, d m The calculation formula is as follows;

[0016]

[0017] Where Pm is the line segment O between the door and the obstacle. LU O L The intersection, Perpendicular feet P ⊥ The angle between the line connecting the door pivot point P and the positive X-axis;

[0018] Allowable door opening angle α under obstacle line interference IN The calculation depends on whether the interference of the obstacle segment on the car door is a line constraint or a point constraint. IN The calculation formula is as follows:

[0019]

[0020] Where, α P and α L The allowable opening angles of the car door are defined by point constraints and line constraints, respectively.

[0021] As a further aspect of the present invention: the calculation of the required door opening angle α for passengers and cargo to get on and off the vehicle based on the preset space required for passengers and cargo to disembark. r α r The calculation formula is as follows:

[0022] α r = k·arcsin[(2d r +W t ) / l D ]

[0023] Where, α r The required door opening angle for passengers to get on and off the vehicle, d r For redundant distance, W t is the passenger and cargo dimensions, and k is a logical variable indicating whether there are passengers or cargo at the door.

[0024] As a further aspect of the present invention: an angle margin is obtained based on the allowable opening angle of the vehicle door and the opening angle of the vehicle door required for passengers and cargo to get on and off the vehicle;

[0025] Δα j =(α j -α rj ) / α rj× 100% j = 1, 2, 3, 4...

[0026] wherein, Δα j is the door j angle allowance, i.e. the passenger load vehicle up and down comfort representation variable, α j is the door j door opening angle allowed, α rj is the door j door opening angle required when the passenger load vehicle can get on and off the vehicle;

[0027] The passenger load vehicle up and down comfort representation model is obtained based on the angle allowance;

[0028] Δα = min{Δα ij} = min{(α j - α rj ) / α rj × 100%, i = 1, 2,..., n O ; j = 1, 2,..., n D}

[0029] wherein, Δα is the passenger load vehicle up and down comfort, Δα ij is the angle allowance of the door j under the influence of the obstacle line segment i, n O is the number of obstacle line segments, n D is the number of doors.

[0030] As a further scheme of the application: in S2, the speed, acceleration, equivalent front wheel angle and equivalent front wheel angle acceleration of the vehicle are preset, a coordinate system is established based on the rear axle center of the vehicle, and the vehicle Ackerman model about time is established based on the Ackerman model;

[0031]

[0032] wherein, x E is the horizontal coordinate of the rear axle center of the vehicle, y E is the longitudinal coordinate of the rear axle center of the vehicle, θ is the vehicle heading angle, t is the time, is the equivalent front wheel angle, l A is the wheelbase of the vehicle, and v is the vehicle speed;

[0033] A plurality of vehicle basic motion curves are obtained according to the vehicle Ackerman model, a smooth path satisfying vehicle kinematics and continuous curvature is obtained, the vehicle basic motion curves include cycloid-circular arc-deceleration curve, acceleration curve-circular arc-cycloid curve, cycloid-circular arc-cycloid curve and acceleration curve-circular arc-deceleration curve, uniform deceleration straight line and uniform speed straight line;

[0034] The central angle δ of the spiral curve-circular arc-deceleration curve, acceleration curve-circular arc-spiral curve, spiral curve-circular arc-spiral curve, and acceleration curve-circular arc-deceleration curve is the difference between the vehicle's heading angles θ1 and θ2 at the two endpoints of the curve. It is positively correlated with the travel time. The time to traverse the curve with a central angle δ is obtained through discrete calculation and fitting, and the corresponding time function is...

[0035] t m =f CAM (δ),t m =f MAC (δ),t m =f CAC (δ),t m =f MAM (δ)

[0036] The required time for a vehicle to pass through is obtained based on the time function.

[0037] As a further aspect of the present invention: in step S3, the transition path includes two segments of spiral curve-circular arc-spiral curve and a uniform straight line, wherein the parameters of the spiral curve-circular arc-spiral curve are passed through the intermediate node S. T heading angle θ T It is determined that the formula for calculating the spiral curve-circular arc-spiral curve is:

[0038]

[0039] Among them, H CAC (θ1, θ2) represents the vertical height of the spiral-circular-spiral curve with an initial heading angle of θ1 and an ending heading angle of θ2. cr The critical angle is θ, and the initial heading angle is θ. cr The vertical height y of the spiral-circular-spiral curve with a final heading angle of 0 n+2 -y s θ n+2 θ is the heading angle at the starting point of the maneuver path. s Let y be the heading angle at the starting point. n+2 The y-coordinate of the starting point of the maneuver path. s Let n be the ordinate of the starting point and n be the number of maneuvers.

[0040] The uniform straight line in the transition path includes the length L of the curve. T The length L of the curve T The calculation formula is as follows:

[0041] L T =x n+2 -x s -[L CAC (θ s ,θ T )+LCAC (θ T ,θ n+2 )]

[0042] wherein, L CAC (θ1,θ2) is the horizontal length of the clothoid-circular arc-clothoid with the initial heading angle θ1 and the terminal heading angle θ2, x n+2 is the horizontal coordinate of the starting point of the maneuver path, x s is the horizontal coordinate of the starting point;

[0043] the maneuver path comprises a clothoid-circular arc-deceleration curve / constant deceleration straight line, a plurality of acceleration motion-circular arc-deceleration motion curves and an acceleration motion-circular arc-clothoid;

[0044] the entry path comprises a constant speed straight line;

[0045] the adjustment path comprises a clothoid-circular arc-deceleration curve;

[0046] the central angle δ A of the clothoid-circular arc-deceleration curve of the adjustment path is;

[0047] δ A = θ eT - π / 2, wherein θ et is the vehicle heading angle when the vehicle passengers get on or off the vehicle;

[0048] based on the initial state and the final state of the vehicle, the parameters of the four-segment parking path mode are the central angle vector δ of the maneuver path and the length L S of the entry path; and a parking path connecting any initial state and a target point state near the target parking space is generated, so as to reduce the parameter dimension of the basic motion curve of the vehicle.

[0049] As a further scheme of the present application: in the S4, the multi-objective optimization model of the parking path comprises optimization variables, a comfort objective function and constraint conditions, and the optimization variables X are;

[0050] X = [x eT , y eT , θ eT , δ, L s ]

[0051] wherein, X is the optimization variable, [x eT , y eT , θ eT ] is the final parking state of the target vehicle, (x eT , y eT ) is the final coordinate of the rear axle center of the target vehicle; and θ eTis the final heading angle of the target vehicle, and δ is the center angle vector of the maneuver path curve S is the length of the entry path segment

[0052] The comfort objective function includes the comfort of the target vehicle and the comfort of the adjacent vehicles

[0053] f1(X) = - [Δα T + η A · (Δα L + Δα R )]

[0054] Wherein, f1(X) is the comfort objective function, Δα T is the comfort of the target vehicle, Δα L is the comfort of the left adjacent vehicle, Δα R is the comfort of the right adjacent vehicle, and η A is the comfort weight of the adjacent vehicle

[0055] The target function of parking time is constructed based on the time function of the basic path

[0056] f2(X) = T M + L T / v m + f CAC (|θ T - θ s |) + f CAC (|θ n+2 - θ T |) + L S / v m + f CAM (δ A )

[0057]

[0058] Wherein, f2(X) is the parking time objective function, T M is the time through the maneuver path, and δ i is the center angle of each curve in the maneuver path

[0059] The constraint conditions include that the vehicle is completely in the target parking space in the final parking state, the parking path parameters, the passengers and cargos can get on and off the vehicle smoothly, the passengers and cargos can get in and out of the parking space smoothly, and collision is avoided. The constraint condition that the passengers and cargos can get on and off the vehicle smoothly is:

[0060]

[0061] When Δα T ≥ 30%, the target vehicle meets the requirement that the passengers and cargos can get on and off the vehicle smoothly

[0062] obtain a parking path multi-objective optimization model;

[0063] min f ob (X)=[f1(X),f2(X)] T

[0064] s.t.g i (X)≤0i=1,2,…,N

[0065] Wherein, f ob (X) is an optimization objective function, g i (X) is a constraint condition, and N is the number of constraint conditions.

[0066] As a further scheme of the application: in S5, the parking path multi-objective optimization model is converted into an unconstrained optimization of optimization variables in a preset interval based on a penalty function, to obtain a comfort optimization objective function and a parking time optimization objective function, and the comfort optimization objective function and the parking time optimization objective function are

[0067]

[0068] Wherein, f 1C (X) is a comfort optimization objective function with a penalty function, f 2C (X) is a parking time optimization objective function with a penalty function, C1 is a comfort penalty constant, C2 is a parking time penalty constant, and J is a nonlinear constraint set including passenger load and cargo can smoothly get on and off the vehicle, passenger load and cargo can smoothly enter and exit the parking space, and collision avoidance constraints.

[0069] Based on the comfort optimization objective function and the parking time optimization objective function, a parking path multi-objective optimization model with a penalty function is obtained, and the parking path multi-objective optimization model with a penalty function is

[0070] min f obC (X)=[f 1C (X),f 2C (X)] T

[0071] s.t.X∈X D

[0072] Wherein, f obC (X) is an optimization objective function with a penalty function, and X D is an optimization variable feasible region under other constraints.

[0073] An automatic parking path is obtained, wherein the automatic parking path is an optimal parking path with high passenger load and cargo on and off the vehicle comfort and short time.

[0074] Compared with the prior art, the present application has the following beneficial effects:

[0075] 1、 The present application improves the comfort of passengers and cargo getting on and off the vehicle and the parking time, and improves the comfort of passengers and cargo getting on and off the vehicle and the parking efficiency, reduces the dimension of the path parameters by presetting the vehicle motion parameters and the parking path mode, improves the efficiency of the parking path planning, and meets the real-time requirement.

[0076] 2、 The present application directly obtains a smooth path meeting the kinematics of the vehicle and the continuity of the curvature by adopting the vehicle basic motion curve based on the Ackermann model, avoids the increased collision risk of path smoothing processing, improves the smoothness of parking, and reduces the tire wear and the steering assist pressure.

[0077] 3、 The present application reduces the dimension of the path parameters by presetting the vehicle motion parameters and the four-section parking path mode of the transition path, the maneuvering path, the driving-in path and the adjusting path, and converts the multiple nonlinear constraint parking path multi-objective optimization model into an unconstrained optimization problem of the optimization variable in the preset range by using the penalty function method, reduces the parking path planning time, and meets the real-time requirement. BRIEF DESCRIPTION OF DRAWINGS

[0078] Figure 1 Schematic diagram of the allowable and required door opening angle of the vehicle under the influence of adjacent obstacles;

[0079] Figure 2 Schematic diagram of the motion parameters corresponding to the vehicle basic motion curve based on the Ackermann model;

[0080] Figure 3 Schematic diagram of the center angle of the vehicle basic motion curve;

[0081] Figure 4 Schematic diagram of the four-section parking path mode of the transition, maneuvering, driving-in and adjusting;

[0082] Figure 5 Schematic diagram of the final scheme selection of the parking path obtained by multi-objective optimization. DETAILED DESCRIPTION

[0083] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0084] Embodiment:

[0085] Please refer toFigures 1-5 A parking path planning method based on Ackermann model motion curve, comprising the following steps:

[0086] S1: obtaining the door opening angle of the vehicle based on the adjacent obstacles of the target parking space, and establishing a comfort representation model of the vehicle passengers and the load getting on and off the vehicle, wherein the door opening angle is the angle of the door opening and the angle of the passengers and the load getting on and off the vehicle;

[0087] S2: constructing a vehicle basic motion curve library based on the Ackermann model, wherein the vehicle basic motion curve library comprises a plurality of vehicle basic motion curves, and the required time of the vehicle passing through the corresponding vehicle basic motion curve is obtained;

[0088] S3: establishing a four-section parking path mode including a transition path, a maneuvering path, a driving-in path and an adjusting path based on the basic motion curve, and generating a parking path connecting any starting state and target point state near the target parking space, so as to reduce the parameter dimension of the vehicle basic motion curve;

[0089] S4: establishing a multi-objective optimization model of the parking path, wherein the multi-objective optimization model of the parking path takes the vehicle completely in the target parking space, the parking path parameters, the passengers and the load getting on and off the vehicle smoothly, the passengers and the load getting in and out of the parking space smoothly and the avoidance of collision as constraint conditions;

[0090] S5: converting and solving the multi-objective optimization model of the parking path based on the penalty function, and obtaining the automatic parking path.

[0091] Specifically, the comfort of the passengers and the load getting on and off the vehicle and the parking time are taken into the optimization design of the parking path, the comfort of the passengers and the load getting on and off the vehicle and the parking efficiency are improved, the four-section parking path mode including the transition path, the maneuvering path, the driving-in path and the adjusting path is established based on the vehicle basic motion curve of the Ackermann model, the smooth path meeting the kinematic requirements of the vehicle and the curvature continuity is directly generated, the collision risk caused by the smooth processing of the path is effectively avoided, the smoothness of the parking is improved, the tire wear and the pressure of the steering assist system are reduced, the dimension of the path parameters is reduced through the preset vehicle motion parameters and the parking path mode, the efficiency of the parking path planning is improved, and the real-time requirement is met.

[0092] Preferably, as Figure 1As shown in S1, based on the allowable opening angle of the target vehicle door under the influence of obstacles near the target parking space and the opening angle required to ensure smooth entry and exit for passengers and cargo, the nearby obstacles are divided into several obstacle segments. The allowable opening angle of the door under the influence of all obstacle segments is calculated. First, it is determined whether the obstacle segments will interfere with the opening of the door; if not, the door angle can be opened to its maximum. The nearby obstacles are divided into several obstacle segments, and the allowable opening angle α of the door is calculated based on these obstacle segments.

[0093]

[0094] Where α is the permissible opening angle of the car door, α M The maximum opening angle of the car door without interference, α0, α m α B α U and These represent the endpoint Q0 when the door is not open and the endpoint Q when the door is open to its maximum angle, respectively. m The lower endpoint O of the obstacle segment LB Obstacle segment upper endpoint O LU Perpendicular P ⊥ The angle α between the line connecting the pivot point P of the car door and the positive direction of the X-axis. s0 From a safety perspective, α IN To determine the permissible opening angle of the vehicle door under obstruction line interference, l D d is the length of the car door. m The distance from the door pivot point to the obstacle line segment O LU O LB The shortest distance, d m The calculation formula is as follows;

[0095]

[0096] Where Pm is the line segment O between the door and the obstacle. LU O L The intersection, Perpendicular feet P ⊥ The angle between the line connecting the door pivot point P and the positive X-axis;

[0097] Allowable door opening angle α under obstacle line interference IN The calculation depends on whether the interference of the obstacle segment on the car door is a line constraint or a point constraint. IN The calculation formula is as follows:

[0098]

[0099] Where, α P and α L The allowable opening angles of the car door are defined by point constraints and line constraints, respectively.

[0100] As a further aspect of the present invention: the required door opening angle α for passengers and cargo to get on and off the vehicle is calculated based on the preset space required for passengers and cargo to disembark. r α r The calculation formula is as follows:

[0101] α r = k·arcsin[(2d r +W t ) / l D ]

[0102] Where, α r The required door opening angle for passengers to get on and off the vehicle, d r For redundant distance, W t is the passenger and cargo dimensions, and k is a logical variable indicating whether there are passengers or cargo at the door.

[0103] Preferred, such as Figure 1 As shown, the angle margin is obtained based on the allowable opening angle of the door and the door opening angle required for passengers to get on and off the vehicle.

[0104] Δα j =(α j -α rj ) / α rj ×100%j=1,2,3,4……

[0105] Where, Δα j Let α be the door angle margin, which corresponds to the variable representing passenger and cargo comfort when getting on and off the vehicle. j Let α be the allowable opening angle of door j. rj The required door opening angle for passengers to be able to get on and off the vehicle;

[0106] A passenger comfort model for getting on and off the vehicle with cargo is derived based on the angle margin.

[0107] Δα=min{Δα ij}=min{(α j -α rj ) / α rj ×100%, i=1,2,...,n O j = 1, 2, ..., n D}

[0108] Where Δα represents the passenger comfort when getting on and off the vehicle, Δα ij Let n be the angular margin of door j under the influence of obstacle segment i. O Let n be the number of obstacle segments. D This refers to the number of car doors.

[0109] Preferred, such asFigure 2 As shown in S2, the speed, acceleration, equivalent front wheel angle and equivalent front wheel angle acceleration of the preset vehicle are established, a coordinate system is established based on the Ackermann model, and the Ackermann model of the vehicle about time is established;

[0110]

[0111] wherein x E is the horizontal coordinate of the rear axle center of the vehicle, y E is the longitudinal coordinate of the rear axle center of the vehicle, θ is the heading angle of the vehicle, and t is time, is the equivalent front wheel angle, l A is the wheelbase of the vehicle, and v is the speed;

[0112] According to the vehicle Ackermann model, a plurality of vehicle basic motion curves are obtained, and the plurality of vehicle basic motion curves form a vehicle basic motion curve library, which is used for parking path design to obtain a smooth path that satisfies vehicle kinematics and curvature continuity, avoids the increase of collision risk caused by path smoothing processing, presets the speed, acceleration, equivalent front wheel angle and equivalent front wheel angle acceleration of the vehicle, reduces the parameter dimension of the parking path, and directly obtains the time required to pass each basic path for evaluating the parking efficiency,

[0113] The motion parameters of the vehicle are preset, and the parameter dimension of the path is reduced. The maximum speed of the vehicle is v m , the maximum equivalent front wheel angle is φ m , the time required for the vehicle to uniformly accelerate from rest to v m is t v , and the time required for the equivalent front wheel angle to uniformly increase from 0 to φ m is also set to t v . As shown in Figure 2 , different motion time t m , speed v and equivalent front wheel angle φ change mode are set to obtain the basic motion curve. As shown in Figure 2 (a), the speed is set to be unchanged at v m in [0, t m -t v ] time, and uniformly reduced to 0 in [t m -t v , t m ] time; the equivalent front wheel angle is uniformly increased from 0 to φ m in [0, t v ] time, and is unchanged at φ m in [t v , t m -t v ] time; [t m -t v , t mThe motion time t decreases uniformly to 0; when the motion time t m ≥2t v When t is reached, the cyclotron-circular-deceleration motion curve (CAM) is obtained. m <2t v At that time, proceed as follows Figure 2 The adjustments are shown in (a). The motion parameter settings for Accelerated Motion-Circular Arc-Rotation Curve (MAC), Accelerated Motion-Circular Arc-Decelerated Motion Curve (MAM), Rotation-Circular Arc-Rotation Curve (CAC), Uniformly Decelerated Linear Motion (DL), and Uniform Linear Motion (UL) are as follows: Figure 2 (b) Figure 2 (c) Figure 2 (d) Figure 2 (e) and Figure 4 As shown in (f);

[0114] The basic motion curves of a vehicle include the spiral curve-circular arc-deceleration curve (CAM), the acceleration curve-circular arc-spiral curve (MAC), the spiral curve-circular arc-spiral curve (CAC), and the acceleration curve-circular arc-deceleration curve (MAM), as well as the uniform deceleration straight line (DL) and the uniform velocity straight line (UL).

[0115] The center angle δ of the curve-circular-deceleration curve (CAM), curve-circular-curve-cyclone (MAC), curve-circular-cyclone (CAC), and curve-circular-deceleration curve (MAM) is the difference between the vehicle's heading angles θ1 and θ2 at the two endpoints of the curve. It is positively correlated with the travel time. The time to traverse the curve with a center angle δ is obtained through discrete calculation and fitting, and the corresponding time function is:

[0116] t m =f CAM (δ),t m =f MAC (δ),t m =f CAC (δ),t m =f MAM (δ)

[0117] The required time for a vehicle to pass through is obtained based on the time function.

[0118] Preferably, in S3, a parking path mode is set to generate a parking path connecting any starting state and target point state near the target parking space, further reducing the dimensionality of path parameters, such as... Figure 4 As shown in (a), a four-segment parking path pattern is established based on the basic motion curve, including a transition path, a maneuvering path, an entry path, and an adjustment path. Figure 4As shown in (b), the transition path includes two segments of spiral curve-circular arc-spiral curve and a uniform straight line. Its parameters are determined by the starting point state and the starting point state of the maneuver path. The parameters of the spiral curve-circular arc-spiral curve are obtained through the intermediate node S. T heading angle θ T Confirmed, with Figure 4 In the embodiment of (b), the formula for calculating the spiral curve-circular arc-spiral curve is:

[0119]

[0120] Among them, H CAC (θ1, θ2) represents the vertical height of the CAC curve with an initial heading angle of θ1 and an ending heading angle of θ2. cr The critical angle is θ, and the initial heading angle is θ. cr The vertical height y of the CAC curve with a final heading angle of 0 n+2 -y s θ n+2 θ is the heading angle at the starting point of the maneuver path. s Let y be the heading angle at the starting point. n+2 The y-coordinate of the starting point of the maneuver path. s Let n be the ordinate of the starting point and n be the number of maneuvers.

[0121] The length L of the uniform straight line in the transition path, including the curve. T The length L of the curve T The calculation formula is as follows:

[0122] L T =x n+2 -x s -[L CAC (θ s ,θ T )+L CAC (θ T ,θ n+2 )]

[0123] Among them, L CAC (θ1, θ2) represents the horizontal length of the spiral-circular-spiral curve with an initial heading angle of θ1 and an ending heading angle of θ2. n+2 x is the x-coordinate of the starting point of the maneuver path. s The x-coordinate of the starting point;

[0124] like Figure 4As shown in (e), the maneuver path includes a section of spiral curve-circular arc-deceleration curve / uniform deceleration straight line, several sections of acceleration motion-circular arc-deceleration motion curves, and a section of acceleration motion-circular arc-spiral curve. The maneuver path is determined by the number of maneuvers and the central angle of each curve, that is, by the central angle vector δ=[δ1,δ2,...,δn]. In particular, the maneuver path of a single maneuver (n=1) consists of a section of DL curve and a section of MAC curve, and the maneuver path of a second maneuver (n=2) consists of a section of CAM curve and a section of MAC curve.

[0125] like Figure 4 As shown in (c), the driving path includes a straight section with a constant velocity, whose length L S Sure;

[0126] The adjustment path includes a spiral curve-circular arc-deceleration curve, and the adjustment path is a CAM curve whose parameters are determined by the final state of the vehicle.

[0127] like Figure 5 As shown in (d), adjust the center angle δ of the spiral curve-circular arc-deceleration curve of the path. A for;

[0128] δ A =θ eT -π / 2, where θ et The vehicle heading angle for the final loading and unloading of passengers and goods;

[0129] Based on the vehicle's initial and final states, the parameters of the four parking path patterns are the center angle vector δ of the maneuvering path and the entry path length L. S It also generates parking paths near the target parking space that connect any starting state and the target point state, in order to reduce the parameter dimension of the vehicle's basic motion curve.

[0130] Preferably, in S4, a multi-objective optimization model for parking paths is established, with the comfort of passengers getting on and off the target vehicle and adjacent vehicles, as well as parking time, as optimization objectives. The final parking state of the vehicle and the parameters of the four parking paths are used as optimization variables. Constraints include the vehicle being completely within the target parking space in the final parking state, the parking path parameters being within a certain range, passengers being able to get on and off smoothly, passengers being able to smoothly enter and exit the parking space, and collision avoidance. By incorporating passenger comfort into parking path planning and considering parking time, an optimal parking path with high passenger comfort and short time is obtained. The multi-objective optimization model for parking paths includes optimization variables, a comfort objective function, and constraints. The optimization variable X is...

[0131] X = [x eT ,y eT ,θ eT ,δ,Ls ]

[0132] wherein X is an optimization variable, [x eT ,y eT ,θ eT ] is the final parking state of the target vehicle, (x eT ,y eT ) is the final coordinate of the center of the rear axle of the target vehicle; θ eT is the final heading angle of the target vehicle, δ is the central angle vector of the maneuver path curve, L S is the length of the path segment for entering;

[0133] The comfort objective function includes the comfort of getting on and off the target vehicle and the comfort of the adjacent vehicles;

[0134] f1(X) = - [Δα T + η A · (Δα L + Δα R )]

[0135] wherein f1(X) is the comfort objective function, Δα T is the comfort of getting on and off the target vehicle, Δα L is the comfort of getting on and off the left adjacent vehicle, Δα R is the comfort of getting on and off the right adjacent vehicle, and η A is the comfort weight of getting on and off the adjacent vehicles;

[0136] The time function based on the basic path is used to construct the objective function of parking time

[0137] f2(X) = T M + L T / v m + f CAC (|θ T - θ s |) + f CAC (|θ n+2 - θ T |) + L S / v m + f CAM (δ A )

[0138]

[0139] wherein f2(X) is the parking time objective function, T M is the time for passing the maneuver path, and δ i is the central angle of each curve in the maneuver path;

[0140] The constraints include that the vehicle is completely in the target parking space in the final parking state, the parking path parameters, the passenger load can smoothly get on and off the vehicle, the passenger load can smoothly enter and exit the parking space, and collision is avoided, and the constraint that the passenger load can smoothly get on and off the vehicle is;

[0141]

[0142] When Δα T ≥ 30%, the target vehicle meets the requirement that the passenger load can smoothly get on and off the vehicle;

[0143] The parking path multi-objective optimization model is obtained;

[0144] min f ob (X) = [f1(X), f2(X)] T

[0145] s.t.g i (X) ≤ 0 i = 1, 2, …, N

[0146] Wherein, f ob (X) is an optimization objective function, g i (X) is a constraint condition, and N is the number of constraint conditions.

[0147] Preferably, in S5, the input vehicle, the parking space parameters, the adjacent obstacle parameters, the target vehicle initial state and the target parking space are input, the parking path multi-objective optimization model is solved, the parking path multi-objective optimization model is converted into an unconstrained optimization of optimization variables in a preset interval based on a penalty function, the comfort optimization objective function and the parking time optimization objective function are obtained, and the comfort optimization objective function and the parking time optimization objective function are;

[0148]

[0149] Wherein, f 1C (X) is a comfort optimization objective function with a penalty function, f 2C (X) is a parking time optimization objective function with a penalty function, C1 is a comfort penalty constant, C2 is a parking time penalty constant, and J is a nonlinear constraint set including the constraints that the passenger load can smoothly get on and off the vehicle, the passenger load can smoothly enter and exit the parking space, and collision is avoided;

[0150] The comfort optimization objective function and the parking time optimization objective function are used to obtain a parking path multi-objective optimization model with a penalty function, and the parking path multi-objective optimization model with a penalty function is;

[0151] min f obC (X) = [f 1C (X), f 2C (X)] T

[0152] s.t.X∈X D

[0153] wherein f obC (X) is an optimization objective function with a penalty function, X D is an optimization variable feasible region under other constraints;

[0154] The automatic parking path is obtained by solving, wherein the automatic parking path is an optimal parking path with high passenger load comfort and short time;

[0155] The method of penalty function is used to convert the multiple nonlinear constraint parking path multi-objective optimization model into an unconstrained optimization problem of optimization variables in a certain range, and a multi-objective optimization algorithm is used for solving, so as to reduce the parking path planning time, and the multi-objective optimization model is solved to obtain a set of optimal solutions, and the final scheme is determined by using the following principles: as shown in the formula (1), if the optimal solution set contains a scheme meeting the comfortable getting-on and off conditions (Δα≥30%), the scheme meeting the comfortable getting-on and off requirements and the shortest parking time is selected, if the optimal solution set does not contain a scheme meeting the comfortable getting-on and off requirements, the scheme with the highest getting-on and off comfort index is selected as the final scheme, and the optimal parking path with high passenger load comfort and short time is obtained based on the final scheme. ​

[0156] The above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited to this, any person skilled in the art can make equivalent replacement or change according to the technical scheme and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered in the protection scope of the present application.​

Claims

1. A parking path planning method based on Ackerman model motion curves, characterized in that, Includes the following steps: S1: Based on the obstacles near the target parking space, obtain the allowable opening angle of the vehicle door and establish a vehicle passenger loading and unloading comfort model, wherein the allowable opening angle of the door is the allowable opening angle of the door of the nearby obstacle and the opening angle required for passengers to load and unload the vehicle. S2: Construct a basic vehicle motion curve library based on the Ackerman model. The basic vehicle motion curve library includes several basic vehicle motion curves. The required time for a vehicle to pass through the corresponding basic vehicle motion curves is obtained from the basic vehicle motion curves. S3: Establish a four-segment parking path pattern including transition path, maneuver path, entry path and adjustment path for the basic motion curve, and generate a parking path near the target parking space that connects any starting state and target point state, in order to reduce the parameter dimension of the vehicle's basic motion curve. S4: Establish a multi-objective optimization model for parking paths. The multi-objective optimization model for parking paths is subject to the following constraints: the vehicle is completely within the target parking space in the final parking state; the parking path parameters are met; passengers and cargo can get in and out of the vehicle smoothly; passengers and cargo can enter and exit the parking space smoothly; and collisions are avoided. S5: Based on the penalty function, the multi-objective optimization model of the parking path is transformed and solved to obtain the automatic parking path; In step S1, the adjacent obstacle is divided into several obstacle segments, and the allowable opening angle α of the door is calculated based on the several obstacle segments. Where α is the permissible opening angle of the car door, α M The maximum opening angle of the car door without interference, α0, α m α B α U and α P⊥ These represent the endpoint Q0 when the door is not open and the endpoint Q when the door is open to its maximum angle, respectively. m The lower endpoint O of the obstacle segment LB Obstacle segment upper endpoint O LU Perpendicular P ⊥ The angle α between the line connecting the pivot point P of the car door and the positive direction of the X-axis. s0 From a safety perspective, α IN To determine the permissible opening angle of the vehicle door under obstruction line interference, l D d is the length of the car door. m The distance from the door pivot point to the obstacle line segment O LU O LB The shortest distance, d m The calculation formula is as follows; Where Pm is the line segment O between the door and the obstacle. LU O L The intersection point, α P⊥ Perpendicular feet P ⊥ The angle between the line connecting the door pivot point P and the positive X-axis; Allowable door opening angle α under obstacle line interference IN The calculation depends on whether the interference of the obstacle segment on the car door is a line constraint or a point constraint. IN The calculation formula is as follows: Where, α P and α L The allowable opening angles of the car door are defined by point constraints and line constraints, respectively.

2. The parking path planning method based on the Ackerman model motion curve according to claim 1, characterized in that: Calculate the required door opening angle α based on the preset space needed for passengers and cargo to get on and off the vehicle. r α r The calculation formula is as follows: α r =k·arcsin[(2d r +W t ) / l D ] Where, α r The required door opening angle for passengers to get on and off the vehicle, d r For redundant distance, W t is the passenger and cargo dimensions, and k is a logical variable indicating whether there are passengers or cargo at the door.

3. The parking path planning method based on the Ackerman model motion curve according to claim 2, characterized in that: Angle margin is obtained based on the allowable opening angle of the vehicle door and the door opening angle required for passengers and cargo to get on and off the vehicle. Da j =(a j -a rj ) / a rj ×100%j=1,2,3,4…… Where, Δα j Let α be the door angle margin, which corresponds to the variable representing passenger and cargo comfort when getting on and off the vehicle. j Let α be the allowable opening angle of door j. rj The required door opening angle for passengers to be able to get on and off the vehicle; Based on the aforementioned angular margin, a passenger comfort model for getting on and off the vehicle is derived. Δα=min{Δα ij }=min{(a j -a rj ) / a rj ×100%,i=1,2,…,n O ;j=1,2,...,n D } Where Δα represents the passenger comfort when getting on and off the vehicle, Δα ij Let n be the angular margin of door j under the influence of obstacle segment i. O Let n be the number of obstacle segments. D This refers to the number of car doors.

4. The parking path planning method based on the Ackerman model motion curve according to claim 3, characterized in that: In S2, the vehicle speed, acceleration, equivalent front wheel angle and equivalent front wheel angle acceleration are preset, a coordinate system is established with the rear axle center of the vehicle, and an Ackerman model of the vehicle with respect to time is established based on the Ackerman model. Where, x E Let y be the x-coordinate of the rear axle center of the vehicle. E Let θ be the ordinate of the rear axle center of the vehicle, θ be the vehicle's heading angle, and t be time. For the equivalent front wheel steering angle, l A v represents the vehicle's wheelbase and the vehicle speed. Based on the vehicle Ackerman model, several basic vehicle motion curves are obtained, resulting in a smooth path that satisfies vehicle kinematics and has continuous curvature. The basic vehicle motion curves include a spiral curve-circular arc-deceleration curve, an acceleration curve-circular arc-spiral curve, a spiral curve-circular arc-spiral curve, and an acceleration curve-circular arc-deceleration curve, as well as uniformly decelerated straight lines and uniform straight lines. The central angle δ of the spiral curve-circular arc-deceleration curve, acceleration curve-circular arc-spiral curve, spiral curve-circular arc-spiral curve, and acceleration curve-circular arc-deceleration curve is the difference between the vehicle's heading angles θ1 and θ2 at the two endpoints of the curve. It is positively correlated with the travel time. The time to traverse the curve with a central angle δ is obtained through discrete calculation and fitting, and the corresponding time function is... t m =f CAM (d),t m =f MAC (d),t m =f CAC (d),t m =f MAM (d) The required time for a vehicle to pass through is obtained based on the time function.

5. The parking path planning method based on the Ackerman model motion curve according to claim 4, characterized in that: In step S3, the transition path includes two segments of spiral curve-circular arc-spiral curve and a uniform straight line. The parameters of the spiral curve-circular arc-spiral curve are passed through the intermediate node S. T heading angle θ T It is determined that the formula for calculating the spiral curve-circular arc-spiral curve is: Among them, H CAC (θ1, θ2) represents the vertical height of the spiral-circular-spiral curve with an initial heading angle of θ1 and an ending heading angle of θ2. cr The critical angle is θ, and the initial heading angle is θ. cr The vertical height y of the spiral-circular-spiral curve with a final heading angle of 0 n+2 -y s θ n+2 θ is the heading angle at the starting point of the maneuver path. s Let y be the heading angle at the starting point. n+2 The y-coordinate of the starting point of the maneuver path. s Let n be the ordinate of the starting point and n be the number of maneuvers. The uniform straight line in the transition path includes the length L of the curve. T The length L of the curve T The calculation formula is as follows: L T =x n+2 -x s -[L CAC (i s ,i T )+L CAC (i T ,i n+2 )] Among them, L CAC (θ1, θ2) represents the horizontal length of the spiral-circular-spiral curve with an initial heading angle of θ1 and an ending heading angle of θ2. n+2 x is the x-coordinate of the starting point of the maneuver path. s The x-coordinate of the starting point; The motion path includes a section of spiral curve-circular arc-deceleration curve / uniform deceleration straight line, several sections of acceleration motion-circular arc-deceleration motion curve, and a section of acceleration motion-circular arc-spiral curve; The driving path includes a straight section with uniform speed; The adjustment path includes a spiral curve-circular arc-deceleration curve; The central angle δ of the spiral curve-circular arc-deceleration curve of the adjustment path. A for; δ A =θ eT -π / 2, where θ et The vehicle heading angle for the final loading and unloading of passengers and goods; Based on the vehicle's initial and final states, the parameters of the four parking path patterns are the center angle vector δ of the maneuvering path and the entry path length L. S It also generates parking paths near the target parking space that connect any starting state and the target point state, in order to reduce the parameter dimension of the vehicle's basic motion curve.

6. The parking path planning method based on the Ackerman model motion curve according to claim 5, characterized in that: In S4, the multi-objective optimization model for parking paths includes optimization variables, a comfort objective function, and constraints, wherein the optimization variable X is; X=[x eT ,y eT ,i eT ,d,L s ] Where X is the optimization variable, [x eT ,y eT ,θ eT [x] represents the final parking state of the target vehicle. eT ,y eT θ represents the final coordinates of the rear axle center of the target vehicle; eT Let L be the target vehicle's final heading angle, δ be the center angle vector of the maneuver path curve, and L be the center angle vector of the target vehicle. S The length of the entry path segment; The comfort objective function includes the comfort of getting in and out of the target vehicle and the comfort of adjacent vehicles; f1(X)=-[Dα T +n A ·(Da L +Da R )] Where f1(X) is the comfort objective function, Δα T For the target vehicle's comfort when getting in and out, Δα L For the comfort of getting on and off the adjacent vehicle on the left, Δα R For the comfort of getting on and off the adjacent vehicle on the right, η A Weighting of comfort when getting on and off adjacent vehicles; Constructing an objective function for parking time based on the time function of the basic path. f2(X)=T M +L T / v m +f CAC (|θ T -θ s |)+f CAC (|θ n+2 -θ T |)+L S / v m +f CAM (δ A ) Where f2(X) is the objective function for parking time, and T M δ is the time to travel the maneuver path. i The center angle of each curve in the aforementioned maneuvering path; The constraints include the vehicle being completely within the target parking space in the final parking state, parking path parameters, passengers and cargo being able to get in and out of the vehicle smoothly, passengers and cargo being able to enter and exit the parking space smoothly, and avoiding collisions. The constraint that passengers and cargo can get in and out of the vehicle smoothly is as follows: When Δα T When the percentage is ≥30%, the target vehicle meets the requirement that passengers and cargo can get on and off the vehicle smoothly. A multi-objective optimization model for parking paths was obtained; min f ob (X)=[f1(X),f2(X)] T s.t.g i (X)≤0i=1,2,…,N Among them, f ob (X) is the optimization objective function, g i (X) represents the constraint, and N represents the number of constraints.

7. The parking path planning method based on the Ackerman model motion curve according to claim 6, characterized in that: In step S5, the multi-objective optimization model of the parking path is transformed into unconstrained optimization of the optimization variables within a preset interval based on the penalty function, resulting in the comfort optimization objective function and the parking time optimization objective function. The comfort optimization objective function and the parking time optimization objective function are as follows: Among them, f 1C (X) is the comfort optimization objective function with a penalty function, f 2C (X) is the objective function for optimizing parking time with a penalty function, C1 is the comfort penalty constant, C2 is the parking time penalty constant, and J is a set of nonlinear constraints including the ability of passengers and cargo to get on and off the vehicle smoothly, the ability of passengers and cargo to enter and exit the parking space smoothly, and collision avoidance constraints. Based on the aforementioned comfort optimization objective function and parking time optimization objective function, a multi-objective optimization model for parking paths with a penalty function is obtained. The multi-objective optimization model for parking paths with a penalty function is as follows: min f obC (X)=[f 1C (X),f 2C (X)] T s.t.X∈X D Among them, f obC (X) is the optimization objective function with a penalty function, X D For the feasible region of optimization variables under other constraints; The automatic parking path is obtained by solving the problem. The automatic parking path is the optimal parking path that provides high comfort for passengers getting on and off the vehicle and takes the shortest time.

Images