Automatic parking path planning method based on target planning

Abstract

The invention belongs to the technical field of automatic driving, and particularly relates to an automatic parking path planning method based on target planning, which comprises the following steps: presetting a plurality of performance indexes # imgabs0 # for automatic parking path planning; according to the specific parameters of the path planning vehicle or robot, calculating the theoretical optimal value # imgabs1 # of each performance index; a target function is constructed, the deviation # imgabs4 # between each performance index # imgabs2 # and a theoretical optimal value # imgabs3 # is minimized, a path planning trajectory equation is generated and specifically is # imgabs5 #, # imgabs6 # is a weight coefficient and meets # imgabs7 #, and a path planning trajectory equation is generated; and solving the objective function by using a solver to obtain an automatic parking driving path. According to the path planning method taking the plurality of performance indexes as the objective function, the optimization of the plurality of performance indexes is realized by utilizing the objective planning method, and the path planning method has a better effect on solving the path planning problem with the requirements of the plurality of performance indexes.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic driving, and in particular relates to an automatic parking path planning method based on target planning. Background Art

[0002] In recent years, self-driving cars, which integrate advanced sensing technology, artificial intelligence technology and the latest control technology in the automotive industry, have become increasingly popular and gradually come into people's view. The path planning of self-driving cars can automatically generate a feasible path for the vehicle from a given starting point to the desired destination, and plays a very important role in the autonomous control module of self-driving cars.

[0003] In the actual process of autonomous driving path planning, a certain performance indicator is often optimized as an objective function, which is difficult to meet the needs of decision makers. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to address the deficiencies in the prior art and to provide an automatic parking path planning method based on target planning, a path planning method that takes multiple performance indicators as objective functions, and uses the target planning method to optimize multiple performance indicators, which has a better effect on solving path planning problems with multiple performance indicators.

[0005] Technical solution: The automatic parking path planning method based on target planning of the present invention comprises the following steps: (1) Preset multiple performance indicators , used for automatic parking path planning; (2) Calculate the theoretical optimal value of each performance indicator based on the specific parameters of the automatic parking vehicle ; (3) Construct the objective function by minimizing each performance indicator Theoretical optimal value Deviation , generate the path planning trajectory equation, specifically: in, is the weight coefficient, satisfying ; (4) Use the solver to solve the objective function and obtain the automatic parking driving path.

[0006] To further improve the above technical solution, the multiple performance indicators include but are not limited to at least one of the following: total path planning time, total change of front wheel steering angle, total change of front wheel steering angular velocity, path smoothness, and smoothness of acceleration change.

[0007] Furthermore, the path planning trajectory equation is based on the vehicle's motion characteristic constraints and the constraints in the parking process, including but not limited to at least one of the following: constraints on the vehicle parking in a parking space; vehicle's motion characteristic constraints, including maximum limits on acceleration, angular velocity, angular acceleration and velocity; path smoothness constraints; acceleration smoothness constraints.

[0008] Furthermore, the specific parameters of the automatic parking vehicle include a parking model and a parking space model. The parking model includes: the instantaneous center of velocity of the vehicle; , the positions of the four vertices of the vehicle, the position of the midpoint of the vehicle's rear axle , front wheel steering angle , vehicle heading angle , vehicle front and rear suspension distance ; The parking space model includes: is the horizontal width of the parking space, is the longitudinal width of the parking space, is the road width; The constraint of the vehicle parking into the parking space during the parking process is expressed by a step function: , is the horizontal width of the parking space, when the vehicle coordinates satisfy , the vehicle is considered to be in the parking space.

[0009] Furthermore, the constructed path planning trajectory equation includes the following based on the vehicle's motion characteristic constraints and the constraints during parking: The constraints for a vehicle to park in a parking space include that the lateral coordinates of the four vertices of the vehicle satisfy the lateral constraints of the parking space and the longitudinal coordinates of the four vertices of the vehicle satisfy the longitudinal constraints of the parking space, which can be expressed as: in, are the lateral coordinates of the four vertices of the vehicle, are the longitudinal coordinates of the four vertices of the vehicle, is the road width, It is used to constrain the lateral coordinates of the four vertices of the vehicle to meet the lateral constraints of the parking space. It is used to constrain the longitudinal coordinates of the four vertices of the vehicle to meet the longitudinal constraint conditions of the parking space; The vehicle's kinematic constraints are expressed as: in, , , , They represent vehicle acceleration, angular velocity, angular acceleration, and speed respectively. Used to constrain the vehicle's acceleration, angular velocity, angular acceleration, and speed to ensure that they do not exceed the vehicle's maximum allowable acceleration. , maximum angular velocity , maximum angular acceleration , Maximum speed ; Smoothness constraints for paths: in, is the curvature of the path, is the maximum permissible curvature, Used to limit the curvature of the path; Smoothness constraint on acceleration: in, is the acceleration when the velocity is 0, Used to constrain the smoothness of the acceleration profile.

[0010] Furthermore, the optimization process of the objective function further includes adding multiple sub-objective functions to improve the path planning quality, and the sub-objective functions include but are not limited to: time optimization objective function; steering angle change smoothing objective function; steering angular velocity change smoothing objective function.

[0011] Furthermore, the objective function The construction method is: in, represents the time optimization objective function, represents the steering angle change smoothing objective function, represents the objective function for smoothing the change of steering angular velocity, is the theoretical optimal value corresponding to each performance indicator, and the weight coefficient of each objective function satisfy .

[0012] Beneficial effect: Compared with the prior art, the advantages of the present invention are: the present invention adopts a target planning-based method, incorporating multiple performance indicators (such as path smoothness, vehicle steering angle change, acceleration change, etc.) into the optimization target, making the path planning result more comprehensive and better able to meet the diverse needs in actual driving scenarios.

[0013] By introducing performance indicators related to path smoothness (such as the smoothness of steering angle changes and acceleration changes), this method significantly reduces sharp turns, sudden accelerations, and sudden decelerations in the path, making the path smoother. This not only improves the stability of autonomous vehicles, but also improves the comfort of passengers, especially in high-precision application scenarios such as automatic parking. By incorporating the mechanical characteristics of the vehicle (such as steering angle, steering angular velocity, acceleration) into constraints and performing multi-objective optimization of the objective function in the path planning process, this method significantly improves the quality of trajectory planning and avoids unreasonable trajectory problems caused by over-simplification, thereby ensuring the vehicle's high-precision execution capabilities in the actual path.

[0014] In the automatic parking scenario, the combination of parking model and parking space model for constraints can effectively avoid collisions between vehicles and obstacles or parking space edges during parking. In addition, by optimizing time and path smoothness, this method improves parking efficiency while ensuring parking safety.

[0015] The present invention introduces a weight coefficient, which can flexibly adjust the weight of each performance indicator in the objective function to meet the requirements of different driving tasks (such as prioritizing optimal time for highway driving and prioritizing path smoothness for complex road conditions), making the algorithm more adaptable and extensible.

[0016] This method reduces the computational complexity of path planning by simplifying the optimization problem of multiple objectives into a deviation minimization problem through the goal planning method. Compared with the traditional path planning method based on large-scale multi-objective optimization, this method significantly improves the real-time performance of the algorithm while achieving similar optimization results, thus meeting the real-time requirements of the autonomous driving system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of vehicle parking model analysis in Example 2 of the present invention; Figure 2 It is a schematic diagram of the parking space model analysis in Example 2 of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the embodiments.

[0019] Embodiment 1: The automatic parking path planning method based on target planning provided in this embodiment includes the following steps: (1) Preset multiple performance indicators , used for path planning; (2) Calculate the theoretical optimal value of each performance indicator based on the specific parameters of the path planning vehicle or robot ; (3) Construct the objective function by minimizing each performance indicator Theoretical optimal value The deviation of the path planning trajectory equation is generated, which is: in, is the weight coefficient, satisfying , Represents the deviation of the performance indicator from the theoretical optimal value.

[0020] (4) Use the solver to solve the path planning problem and obtain the autonomous driving path.

[0021] Embodiment 2: This embodiment is specifically described using automatic parking as an application scenario.

[0022] Parking model Figure 1 As shown, Point is the instantaneous center of the vehicle's velocity. Indicates the midpoint of the rear axle. represents the front wheel steering angle, represents the vehicle heading angle, It represents the distance between the front and rear suspension beams of the vehicle, and A, B, C, and D represent the positions of the four vertices of the vehicle.

[0023] Parking space model Figure 2 As shown, is the horizontal width of the parking space, is the longitudinal width of the parking space, is the road width, then the parking space constraint is expressed by a step function: , when the vehicle coordinates satisfy , the vehicle is considered to be in the parking space.

[0024] First, the parking space constraints of the vehicle are sorted out. Its nonlinear inequality constraint matrix is ​​as follows: in, are the lateral coordinates of the four vertices of the vehicle, are the longitudinal coordinates of the four vertices of the vehicle, is the road width, It is used to constrain the lateral coordinates of the four vertices of the vehicle to meet the lateral constraints of the parking space. It is used to constrain the longitudinal positions of the four vertices of the vehicle not to exceed the longitudinal range of the parking space.

[0025] Then, the mechanical characteristic constraints of the vehicle are arranged and the expressions are as follows: in, , , , They represent vehicle acceleration, angular velocity, angular acceleration, and speed respectively. Used to constrain the vehicle's motion characteristics (such as acceleration, angular velocity, angular acceleration, speed) to not exceed the vehicle's maximum allowed acceleration , maximum angular velocity , maximum angular acceleration , Maximum speed .

[0026] Then add constraints on the smoothness of the path, the expression is as follows: in, is the curvature of the path, is the maximum permissible curvature, It is used to limit the curvature of the path to ensure that the path is not too curved, so as to avoid drastic dynamic changes of the vehicle when turning and ensure smooth driving.

[0027] In order to make the acceleration change process traceable, it is also necessary to impose smooth constraints on the acceleration profile, as shown in the following formula: in, is the acceleration when the velocity is 0, It is used to constrain the smoothness of the acceleration profile, that is, the acceleration change will not be too drastic. Especially at the end of parking, the acceleration of the vehicle needs to be smoothly reduced to zero to avoid drastic impact.

[0028] The above is the constraint equation of the parking process. The traditional path planning method usually takes time optimization as the only objective function, as shown in the following formula: in, is the number of pose points, i.e., the number of iterations, is a single sampling time.

[0029] This method constructs multiple sub-objective functions, as shown in the following formula: in, .

[0030] The objective function of this method is as follows: in , is the weight coefficient, .

[0031] In the addition of the objective function and After that, the vehicle's front wheel steering angular velocity and front wheel steering angle The change of is obviously smoothed, which proves that increasing the number of objective functions of the trajectory optimization problem is effective in improving the trajectory quality.

[0032] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. An automatic parking path planning method based on target planning, characterized in that: The steps include: (1) Preset multiple performance indicators , used for automatic parking path planning; (2) Calculate the theoretical optimal value of each performance indicator based on the specific parameters of the automatic parking vehicle ; (3) Construct the objective function by minimizing each performance indicator Theoretical optimal value Deviation , generate the path planning trajectory equation, specifically: in, is the weight coefficient, satisfying ; (4) Use the solver to solve the objective function and obtain the automatic parking driving path.

2. The automatic parking path planning method based on target planning according to claim 1, characterized in that: The multiple performance indicators include but are not limited to at least one of the following: total path planning time, total amount of change in front wheel steering angle, total amount of change in front wheel steering angular velocity, smoothness of the path, and smoothness of acceleration change.

3. The automatic parking path planning method based on target planning according to claim 2 is characterized in that: The path planning trajectory equation is based on the vehicle's motion characteristic constraints and the constraints during the parking process, including but not limited to at least one of the following: constraints on the vehicle parking in a parking space; vehicle's motion characteristic constraints, including maximum limits on acceleration, angular velocity, angular acceleration and velocity; path smoothness constraints; acceleration smoothness constraints.

4. The automatic parking path planning method based on target planning according to claim 3 is characterized in that: The specific parameters of the automatic parking vehicle include a parking model and a parking space model. The parking model includes: the instantaneous center of velocity of the vehicle , the positions of the four vertices of the vehicle, the position of the midpoint of the vehicle's rear axle , front wheel steering angle , vehicle heading angle , vehicle front and rear suspension distance ; The parking space model includes: is the horizontal width of the parking space, is the longitudinal width of the parking space, is the road width; The constraint of the vehicle parking into the parking space during the parking process is expressed by a step function: , is the horizontal width of the parking space, when the vehicle coordinates satisfy , the vehicle is considered to be in the parking space.

5. The automatic parking path planning method based on target planning according to claim 4 is characterized in that: The constructed path planning trajectory equation is based on the vehicle's motion characteristic constraints and the constraints during parking, including: The constraints for a vehicle to park in a parking space include that the lateral coordinates of the four vertices of the vehicle satisfy the lateral constraints of the parking space and the longitudinal coordinates of the four vertices of the vehicle satisfy the longitudinal constraints of the parking space, which can be expressed as: in, are the lateral coordinates of the four vertices of the vehicle, are the longitudinal coordinates of the four vertices of the vehicle, is the road width, It is used to constrain the lateral coordinates of the four vertices of the vehicle to meet the lateral constraints of the parking space. It is used to constrain the longitudinal coordinates of the four vertices of the vehicle to meet the longitudinal constraint conditions of the parking space; The vehicle's kinematic constraints are expressed as: in, , , , They represent vehicle acceleration, angular velocity, angular acceleration, and speed respectively. Used to constrain the vehicle's acceleration, angular velocity, angular acceleration, and speed to ensure that they do not exceed the vehicle's maximum allowable acceleration. , maximum angular velocity , maximum angular acceleration , Maximum speed ; Smoothness constraints for paths: in, is the curvature of the path, is the maximum permissible curvature, Used to limit the curvature of the path; Smoothness constraint on acceleration: in, is the acceleration when the velocity is 0, Used to constrain the smoothness of the acceleration profile.

6. The automatic parking path planning method based on target planning according to claim 5, characterized in that: The optimization process of the objective function further includes adding multiple sub-objective functions to improve the path planning quality, and the sub-objective functions include but are not limited to: time optimization objective function; steering angle change smoothing objective function; steering angular velocity change smoothing objective function.

7. The automatic parking path planning method based on target planning according to claim 6, characterized in that: The objective function The construction method is: in, represents the time optimization objective function, represents the steering angle change smoothing objective function, represents the objective function for smoothing the change of steering angular velocity, is the theoretical optimal value corresponding to each performance indicator, and the weight coefficient of each objective function satisfy .

Images