Automatic parking path planning method, device, system and readable storage medium

Abstract

The application provides an automatic parking path planning method, device and system and a readable storage medium. In the automatic parking path planning method, when a parking path of a vehicle is planned, an error caused by parameters of an initial node pose of the vehicle is considered, a parameter interval of the initial node pose is obtained according to an error interval of the parameters of the initial node pose, a parameter interval of a next node is solved by using a path planning function and the parameter interval of the initial node pose, and the next node with no interference with obstacles and a minimum cost function value in the parameter interval thereof is taken as a planning node, so that the planning node has no interference with the obstacles in the parameter interval thereof, and thus when the vehicle parks according to the planning node, the probability of interference between the vehicle and the obstacles is greatly reduced. Therefore, the automatic parking path planning method can improve the success rate of parking path planning and increase the feasibility of parking path planning.

Description

[0001] This application is a divisional application of the invention patent application with application number 202210513104.7, application date May 11, 2022, entitled "Automatic Parking Path Planning Method, Apparatus, System and Readable Storage Medium". Technical Field

[0002] This invention relates to the field of route planning, and more particularly to an automatic parking route planning method, apparatus, system, and readable storage medium. Background Technology

[0003] In recent years, with the rapid growth of the total number of cars in China, parking spaces in cities have become increasingly scarce and limited, significantly increasing the likelihood of parking-related accidents. Automatic parking technology can help people park safely and reliably. It acquires the vehicle's and parking space's positions, plans a feasible path, and then automatically controls the vehicle to follow that path to complete the parking maneuver. Path planning is a crucial step in automatic parking; its basic requirement is that the planned path is feasible and collision-free.

[0004] The path planning algorithms used in automatic vehicle parking mainly include: ① methods that represent the vehicle's trajectory through functions, such as circular arcs, straight lines, curved lines, polynomial curves, etc.; ② sampling planning methods, which perform random sampling planning between the starting point and the ending point, such as fast random tree methods and probabilistic route graph methods; ③ graph search methods, which build a grid map of the search space and then search and select grids, such as Dijktra's algorithm, A* algorithm, and D* algorithm; ④ optimization prediction-based algorithms, which construct an objective function of optimization indicators and add constraints to optimize the optimal trajectory.

[0005] Hybrid A* algorithm is often used for path planning in automatic parking. A* is a heuristic search algorithm that can evaluate the direction of path search at each step through a cost function. By adding vehicle kinematics to the A* algorithm, the search path is transformed from discrete points into a line with a certain degree of continuity.

[0006] Currently, most path planning for automatic parking is based on deterministic parameters or models. However, in actual automatic parking, due to differences and changes in the external environment and the performance of various vehicle modules (perception and localization module, kinematic model, control module, etc.), the vehicle's parameters are uncertain. For example, the vehicle's initial pose, wheel angle, and the position coordinates of obstacles often lead to a certain error between the parking path and the actual path, or require multiple path planning attempts, and may even cause automatic parking to fail. Summary of the Invention

[0007] The purpose of this invention is to provide an automatic parking path planning method, apparatus, system, and readable storage medium to address the shortcomings of the aforementioned technical problems, thereby improving the success rate of automatic vehicle parking.

[0008] To achieve the above objectives, the present invention provides an automatic parking path planning method, comprising:

[0009] Collect the initial node pose parameters of the vehicle and the location of the target parking space;

[0010] The parameter range of the initial node pose is obtained, and the parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose.

[0011] The parameter range of the next node pose is solved using the path planning function and the parameter range of the initial node pose;

[0012] Obstacle interference detection is performed based on the parameter range of the next node pose, and the cost function value of the next node is solved.

[0013] The next node that passes obstacle interference detection and has the minimum cost function value is selected as the planning node;

[0014] The parameter range of the initial node pose is:

[0015]

[0016] in, The initial pose parameters to be acquired, δ a The acquired rotation angle, and the parameters of the acquired initial node pose include the initial node pose parameters. and the collected wheel rotation angle δ a ; The initial pose parameters The error range, δ e The wheel rotation angle δ collected a The error range;

[0017] The parameter range for solving the next node pose using the path planning function and the parameter range of the initial node pose includes:

[0018] Within the range of each parameter of the initial node pose, select m interpolation points, and arrange and combine the m interpolation points of each parameter to obtain multiple input sample points of the initial node pose.

[0019] The next node pose is obtained by using multiple input sample points and the path planning function.

[0020] The parameter range of the next node pose is solved based on the input sample points, the output sample points, and the polynomial surrogate model.

[0021] Optionally, the step of solving the parameter range of the next node pose based on the input sample points, the output sample points, and the polynomial surrogate model includes:

[0022] By using multiple input sample points as independent variables of the polynomial surrogate model and multiple output sample points as dependent variables of the polynomial surrogate model, the coefficient relationship between the parameters of the initial node pose and the parameters of the next node pose is solved.

[0023] The parameter range of the next node pose is derived based on the coefficient relationship and the parameter range of the initial node pose.

[0024] Optionally, the polynomial proxy model is the Chebyshev proxy model:

[0025]

[0026] Where g is the number of parameters in the initial node pose; u is the number of parameters i1, i2, ..., i in the Chebyshev series. g The number of zeros. The pose of the input sample points is used as the independent variable (x1, x2, ..., x) of the ChebysheV surrogate model. g The output sample points are used as the dependent variables of the Chebyshev surrogate model.

[0027] The coefficient relationship between the parameters of the initial node pose and the parameters of the next node pose is as follows:

[0028]

[0029] Optionally, the multiple input sample points are obtained based on the tensor product of m interpolation points for each parameter of the initial node pose.

[0030] Optionally, the path planning function includes the HybridA* algorithm.

[0031] To achieve the above objectives, the present invention also provides an automatic parking path planning device, comprising:

[0032] The positioning module is used to collect the parameters of the vehicle's initial node pose and the location of the target parking space;

[0033] The acquisition module is used to acquire the parameter range of the initial node pose, wherein the parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose.

[0034] The solution module is used to solve for the parameter range of the next node pose using the path planning function and the parameter range of the initial node pose;

[0035] The detection module is used to perform obstacle interference detection based on the parameter range of the pose of the next node, and to solve for the cost function value of the next node;

[0036] The determination module is used to select the next node that passes obstacle interference detection and has the minimum cost function value as the planning node;

[0037] The parameter range of the initial node pose is:

[0038]

[0039] in, The initial pose parameters to be acquired, δ a The acquired rotation angle, and the parameters of the acquired initial node pose include the initial node pose parameters. and the collected wheel rotation angle δ a ; The initial pose parameters The error range, δ e The wheel rotation angle δ collected a The error range;

[0040] The parameter range for solving the next node pose using the path planning function and the parameter range of the initial node pose includes:

[0041] Within the range of each parameter of the initial node pose, select m interpolation points, and arrange and combine the m interpolation points of each parameter to obtain multiple input sample points of the initial node pose.

[0042] The next node pose is obtained by using multiple input sample points and the path planning function.

[0043] The parameter range of the next node pose is solved based on the input sample points, the output sample points, and the polynomial surrogate model.

[0044] To achieve the above objectives, the present invention also provides an automatic parking path planning system, comprising:

[0045] Processor; and

[0046] Memory for storing the executable instructions of the processor;

[0047] The processor is configured to execute the automated parking path planning method as described above by executing the executable instructions.

[0048] To achieve the above objectives, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the automatic parking path planning method as described above.

[0049] The present invention also provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the automatic parking path planning method as described above.

[0050] In the automatic parking path planning method provided by this invention, when planning a parking path for a vehicle, the error generated by the parameters of the initial node pose is taken into account. The parameter range of the initial node pose is obtained based on the error range of the initial node pose parameters. The parameter range of the next node is solved using the path planning function and the parameter range of the initial node pose. The next node with no interference with obstacles and the minimum cost function value within its parameter range is selected as the planning node. This ensures that the planning node will not interfere with obstacles within its parameter range, thereby significantly reducing the probability of interference between the vehicle and obstacles when parking according to the planning node. Therefore, the automatic parking path planning method of this invention can improve the success rate of parking path planning and increase the feasibility of parking path planning. Attached Figure Description

[0051] Figure 1 This is a flowchart of the automatic parking path planning method according to an embodiment of the present invention.

[0052] Figure 2 yes Figure 1 The flowchart for step 300.

[0053] Figure 3 yes Figure 2 The flowchart for step 340.

[0054] Figure 4 yes Figure 1 The flowchart for step 600.

[0055] Figure 5 yes Figure 1 The flowchart after step 600.

[0056] Figure 6 This is a block diagram of the automatic parking path planning device according to an embodiment of the present invention.

[0057] Figure 7 This is a block diagram of the automatic parking path planning system according to an embodiment of the present invention.

[0058] Figure 8 This is a search graph in an embodiment of the present invention that uses a path planning function and a parameter range of planned node poses to search for the next node pose. Detailed Implementation

[0059] To explain in detail the technical content, structural features, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0060] Please see Figure 1 This invention discloses an automatic parking path planning method, which includes:

[0061] 100. Collect the parameters of the initial node pose of the vehicle and the position of the target parking space.

[0062] The initial node pose and the target parking space location can be acquired using the vehicle's positioning module. The vehicle's positioning module can acquire its initial node pose and the target parking space location from the map. The map can be a local map including the initial node and the target parking space, containing enough space to facilitate path planning, or it can be a global map of the parking lot, etc.

[0063] 200. Obtain the parameter range of the initial node pose. The parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose.

[0064] Because the vehicle's perception and control modules may contain errors, the initial node pose acquired will also have some inaccuracies. To improve the accuracy of the initial node pose, a parameter range for the initial node pose is derived based on the error ranges of each parameter. Specifically, the parameter range for the initial node pose is:

[0065]

[0066] in, The initial node pose is used for acquisition, δa is the acquired wheel rotation angle, and the acquired parameters of the vehicle's initial node pose include the vehicle's initial node pose parameters. and the rotation angle δ of the collected wheel a ; Initial node pose parameters The error range, δ e The collected rotation angle δ a The error range.

[0067] In this embodiment of the invention, the error range of the initial node pose parameters δ can be obtained from experimental measurements and statistics based on certain data. a It can be obtained from the vehicle control module.

[0068] In some specific examples, the error range of the initial node pose parameters can be expressed as:

[0069]

[0070] The error range for each parameter can be expressed as:

[0071]

[0072] in, a m and These represent the lower and upper bounds of the parameter's error, respectively. For example, the error range of the initial pose parameter x0 can be expressed as... The error range of the initial pose parameter y0 can be expressed as: Initial pose parameters The error range can be expressed as The error range of the initial pose parameter δ can be expressed as:

[0073] 300. Solve for the parameter range of the next node pose using the path planning function and the parameter range of the initial node pose.

[0074] The path planning function is used to plan the path from the initial node of the vehicle to the target parking space. The parameter range of the pose of the next node can be solved according to the path planning function known to those skilled in the art. For example, the path planning function can be the Hybrid A* algorithm, Dijkstra's algorithm, etc.

[0075] 400. Perform obstacle interference detection based on the parameter range of the pose of the next node, and solve for the cost function value of the next node.

[0076] When performing obstacle interference detection on nodes, vehicles can be simplified into polygons and then expanded. Obstacle interference detection can then be performed on the expanded vehicles at the nodes. Of course, other obstacle interference detection methods known to those skilled in the art can also be used. The cost function value of a node can be solved using any existing cost value algorithm known to those skilled in the art.

[0077] It is understandable that since the parameter range of the initial node pose is derived from the error range of the parameters, and since the errors generated by the parameters in each module of the vehicle are not too large, the span of the parameter range of the initial node pose will not be very large. Therefore, the span of the parameter range of the next node calculated based on the parameter range of the initial node pose will also not be very large. Thus, when calculating the cost function value of the next node, a value can be taken from its parameter range for calculation, thereby reducing the amount of computation.

[0078] 500. The next node that passes the obstacle interference detection and has the minimum cost function value is selected as the planning node.

[0079] Obstacle interference detection means that the solved next node pose will not interfere with obstacles within its parameter range. Since solving for the parameters of the next node pose based on the path planning function and the initial node pose generates multiple nodes corresponding to multiple search directions, each node corresponding to a parameter range, to optimize the planned path, the next node corresponding to the search direction that minimizes interference with obstacles within the parameter range and has the lowest cost function value is selected as the planning node.

[0080] In the automatic parking path planning method provided by this invention, when planning a parking path for a vehicle, the error generated by the parameters of the initial node pose is taken into account. The parameter range of the initial node pose is obtained based on the error range of the initial node pose parameters. The parameter range of the next node is solved using the path planning function and the parameter range of the initial node pose. The next node with the smallest cost function value and no interference with obstacles within its parameter range is selected as the planning node. This ensures that the planning node will not interfere with obstacles within its parameter range, thereby significantly reducing the probability of interference between the vehicle and obstacles when parking according to the planning node. Therefore, the automatic parking path planning method of this invention can improve the success rate of parking path planning and increase the feasibility of parking path planning.

[0081] In a specific example, after solving for the planned nodes, the automatic parking path planning method further includes: using the path planning function and the interval parameters of the planned node poses to solve for the parameter interval of the next node pose corresponding to the planned node; performing obstacle detection based on the parameter interval of the next node pose corresponding to the planned node and solving for the cost function value of the next node; taking the next node that passes obstacle interference detection and has the smallest cost function value as the new planned node; repeating the above steps until the target parking space is planned, thereby obtaining the parking path formed by the initial node and multiple planned nodes. Of course, the automatic parking path planning method of this embodiment is not limited to this.

[0082] like Figure 2 As shown, step 300 includes:

[0083] 310. Select m interpolation points within the interval of each parameter of the initial node pose, and arrange and combine the m interpolation points of each parameter to obtain multiple input sample points of the current planned node pose.

[0084] Specifically, multiple input sample points are obtained based on the tensor product of m interpolation points for each parameter. For example, the m interpolation points selected within the interval of parameter x0 of the initial node pose are (x...01 x 02 x 03 , ..., x 0m ), where m interpolation points are selected within the interval of the parameter y0 of the initial node pose (y 01 y 02 y 03 , ..., y 0m ), parameters of the initial node pose The m interpolation points selected within the interval are The m interpolation points selected within the interval of the parameter δ of the initial node pose are (δ... 01 δ 02 δ 03 , …, δ 0m The permutation and combination of m interpolation points for each parameter can be based on the tensor product of the m interpolation points for each parameter: In other words, multiple input sample points can be obtained from the tensor product of m interpolation points of each parameter of the initial node pose. For example, the pose of some input sample points is... etc.

[0085] 320. Based on multiple input sample points and the path planning function, obtain multiple output sample points for the pose of the next node.

[0086] The path planning function calculates the pose of the next node based on the pose of the previous node. Therefore, by substituting the poses of multiple input sample points of the initial node obtained in step 310 into the path planning function, the poses of multiple output sample points of the next node can be calculated.

[0087] 330. Solve the parameter range of the vehicle's next node pose based on the input sample points, output sample points, and the polynomial surrogate model.

[0088] Furthermore, such as Figure 3 As shown, step 330 includes:

[0089] 331. Using multiple input sample points as independent variables of a polynomial surrogate model and multiple output sample points as dependent variables of a polynomial surrogate model, solve for the coefficient relationship between the initial node pose and the next node pose.

[0090] 332. Based on the coefficient relationship and the parameter range of the initial node pose, the parameter range of the next node pose is obtained.

[0091] Specifically, the polynomial proxy model can be the Chebyshev proxy model, whose general formula is:

[0092]

[0093] Where g is the number of parameters in the initial node pose; u is the number of parameters i1, i2, ..., i in the Chebyshev series. g The number of zeros. The pose of the input sample points is used as the independent variable (x1, x2, ..., x) of the ChebysheV surrogate model. g The output sample points are used as the dependent variables of the Chebyshev surrogate model.

[0094] Based on the coefficients of the Chebyshev proxy model, the relationship between the coefficients of the initial node pose and the parameters of the next node pose can be derived as follows:

[0095]

[0096] Where i1, i2, ..., i g =0, 1, 2, 3, ... To solve the coefficients of the ChebysheV surrogate model. The multiple integrals of the general formula of the Chebyshev surrogate model are transformed into numerical integrals. Using Gauss-Chebyshev numerical integration, the complex multiple integrals are solved numerically to obtain the coefficients of the Chebyshev surrogate model. This approach ensures both accuracy and significantly improves the solution speed. Specifically, it is assumed that:

[0097]

[0098] It can be concluded that:

[0099]

[0100] Where (x1, x2, ..., x h Let x be the Gaussian points of the Gaussian integral, with h being the number of Gaussian points in each dimension, and let X be a one-dimensional Gaussian point x1, x2, ..., xn. h The tensor product, since all parameters of the initial node pose have been normalized, has Gaussian points for each parameter as (x1, x2, ..., x...). h After normalization, each interpolation point corresponding to the same parameter is the same, therefore X is:

[0101]

[0102] This allows us to solve for the coefficients of the Chebyshev surrogate model, and further derive the coefficient relationship between the initial node pose and the next node pose in the Chebyshev surrogate model. Since the parameter range of the initial node pose is... Based on the coefficient relationship between the initial node pose and the next node pose, the parameter range of the corresponding next node pose can be obtained. in,

[0103]

[0104]

[0105]

[0106] In this embodiment of the invention, the use of the Chebyshev proxy model in the polynomial proxy model can improve computational accuracy and efficiency. Of course, this embodiment of the invention does not limit the specific form of the polynomial proxy model; it can also be other proxy models known to those skilled in the art, such as the Kriging proxy model, etc.

[0107] In some specific examples, after solving for the parameter range of the first planned node pose based on the parameter range of the initial node pose, the parameter range of the second planned node pose is solved based on the parameter range of the first planned node pose, and so on, iteratively solving for each planned node pose. The process of solving for the parameter range of each planned node pose is the same as the process of solving for the parameter range of the planned node pose based on the parameter range of the initial node pose. For example, the parameter range of the nth planned node pose is... The m interpolation points selected for each parameter interval of the pose of the nth planning node are respectively (x n1 x n2 x n3 , ..., x nm ), (y n1 y n2 y n3 , ..., y nm ), (δ n1 δ n2 δ n3 , …, δ nm The multiple input sample points obtained are:

[0108]

[0109] Multiple output sample points are obtained by solving the next node pose using the path planning function. The input sample points are used as independent variables, and the output sample points as dependent variables, and substituted into the Chebyshev surrogate model to solve the coefficient relationship between the nth planned node pose and its corresponding next node pose. Based on the coefficient relationship and the parameter range of the nth planned node pose, the following is... Solve for the parameter range of the next node pose. in,

[0110]

[0111]

[0112]

[0113] like Figure 8 As shown, based on the parameter range of the planned node pose and the path planning function, the parameter ranges of the next node pose in multiple different search directions can be solved. Therefore, the optimal next node is selected as the planned node through obstacle interference detection and cost function value solving.

[0114] It is understood that in this embodiment of the invention, the parameter range of the next node pose corresponding to the planned node is solved by iteratively solving the parameter range of the initial node pose. Since the parameter range of the initial node pose is related to the error range of the parameters, the parameter range of each node pose solved iteratively will be related to the error range of the parameters.

[0115] Furthermore, since the specific representation of the vehicle's nodal pose is generally as follows: Therefore, the parameter range of the solved next node pose can be expressed simply as However, since the wheel rotation angle δ also affects the vehicle's nodal pose, when solving for the parameter range of the next nodal pose, it is also necessary to solve for the range of its wheel rotation angle δ. Therefore, the representation of the parameter range of the nodal pose is not limited to... It could also be

[0116] In some specific examples, such as Figure 4 As shown, the automatic parking path planning method may also include:

[0117] 600. Determine if there is a path from the planned node to the target parking space.

[0118] When a path exists from the planned node to the target parking space, it indicates that a Reed-Shepp curve exists between the planned node and the target parking space. It is understood that the planned node in this example is not limited to the first planned node obtained from the initial node; it can be any planned node obtained through iterative solving.

[0119] 610. When there is a path from the planning node to the target parking space, select n interpolation points in the interval of each parameter of the planning node pose, and obtain multiple input sample points of the planning node pose based on the n interpolation points of each parameter of the planning node pose.

[0120] Since the pose of the planning node is not deterministic but rather has a parameter range, the Reed-Shepp curve between the planning node and the target node is also not a deterministic trajectory but rather a trajectory within a certain range. To solve for the Reed-Shepp curve between the planning node and the target node, n interpolation points need to be selected within the interval of each parameter of the planning node's pose. Multiple input sample points are obtained based on the permutation and combination of the n interpolation points for each parameter.

[0121] 620. Based on multiple input sample points of the planned node pose and the path planning function, obtain multiple planned paths and the parameter range of each node pose of each planned path.

[0122] Since each input sample point corresponds to a planned path, the input sample point is substituted into the path planning function to solve for its corresponding planned path. Then, based on the error range of the initial node pose parameters, the parameter range of each node pose in the planned path is solved. Those skilled in the art can use the known path planning function for solving Reed-Shepp curves to solve for the planned path corresponding to each input sample point.

[0123] Understandably, to more accurately determine the parameter range of each node pose in each planned path, when planning a node using the path planning function, the parameter range of the current node pose can be calculated based on the parameter range of the previous node pose. The calculation method can be the same as the method described above for determining the parameter range of the next node pose: n interpolation points are selected within the parameter range of each parameter of the previous node pose; multiple input sample points of the previous node pose are obtained based on these interpolation points; these multiple input sample points are substituted into the path planning function to obtain multiple output sample points of the current node pose; the input and output sample points are substituted into the polynomial surrogate model to obtain the coefficient relationship between the previous and current node poses; and the parameter range of the current node pose is then determined based on the parameter range of the previous node pose and the coefficient relationship. In essence, the determination of the parameter range of each node pose in the planned path iteratively progresses from the parameter range of the planned node to the target parking space.

[0124] 630. Perform obstacle interference detection based on the parameter range of the pose of each node in each planned path.

[0125] Obstacle interference detection can be performed either after the complete planning path has been planned, or it can be performed on each node after the parameter range of the pose has been solved.

[0126] 640. The planned path obtained through obstacle interference detection is taken as the target path.

[0127] 650. Obtain the parking path based on the initial node, the planned node, and the target path.

[0128] In the above example, determining whether a path exists from the planned node to the target parking space simplifies the calculation of automatic parking path planning and improves calculation speed. Furthermore, since the pose of each node in both the planned node and the target path is free from interference with obstacles within its parameter range, it ensures that the parking path planned in this embodiment of the invention will not collide with obstacles, effectively preventing path planning failure or parking failure.

[0129] It is understandable that when planning parking paths for vehicles, for each planning node that is solved, it is necessary to determine whether there is a path to the target parking space. If there is, the planning path is solved based on the planning node. If there is no path, the parameter range of the pose of the next node is solved based on the planning node, the path planning function, and the polynomial surrogate model.

[0130] Furthermore, such as Figure 5 As shown, in order to meet the actual movement needs of the vehicle, the automatic parking path planning method provided in this embodiment of the invention further includes:

[0131] 700. Smooth the parking path.

[0132] 800. Output the smoothed parking path.

[0133] like Figure 6 As shown, this embodiment of the invention also provides an automatic parking path planning device, comprising:

[0134] The positioning module 10 is used to collect the parameters of the vehicle's initial node pose and the position of the target parking space;

[0135] The acquisition module 11 is used to acquire the parameter range of the initial node pose, wherein the parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose.

[0136] Solver module 12 is used to solve for the parameter range of the next node pose using the path planning function and the parameter range of the initial node pose;

[0137] The detection module 13 is used to perform obstacle interference detection based on the parameter range of the pose of the next node, and to solve for the cost function value of the next node;

[0138] The determination module 14 is used to select the next node that passes obstacle interference detection and has the minimum cost function value as the planning node.

[0139] It should be noted that the various modules in the automatic parking path planning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0140] like Figure 7 As shown, this embodiment of the invention also discloses an automatic parking path planning system, comprising:

[0141] Processor 20; and

[0142] Memory 30 is used to store executable instructions of processor 20;

[0143] The processor 20 is configured to execute the automated parking path planning method as described above by executing executable instructions.

[0144] This invention also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the automatic parking path planning method as described above.

[0145] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the automatic parking path planning method as described above.

[0146] It should be understood that, in the embodiments of the present invention, the processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DP), application-specific integrated circuits (AIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0147] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by hardware related to computer program instructions. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0148] The above-disclosed examples are merely preferred embodiments of the present invention, intended to facilitate understanding and implementation by those skilled in the art. They should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. An automatic parking path planning method, characterized in that, include: Collect the initial node pose parameters of the vehicle and the position of the target parking space; The parameter range of the initial node pose is obtained, and the parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose. The parameter range of the next node pose is solved using the path planning function and the parameter range of the initial node pose; Obstacle interference detection is performed based on the parameter range of the next node pose, and the cost function value of the next node is solved. The next node that passes obstacle interference detection and has the minimum cost function value is selected as the planning node; The parameter range of the initial node pose is: in, The initial pose parameters to be acquired, δ a The acquired rotation angle, and the parameters of the acquired initial node pose include the initial node pose parameters. and the collected wheel rotation angle δ a ; The initial pose parameters The error range, δ e The wheel rotation angle δ collected a The error range; The parameter range for solving the next node pose using the path planning function and the parameter range of the initial node pose includes: Within the range of each parameter of the initial node pose, select m interpolation points, and arrange and combine the m interpolation points of each parameter to obtain multiple input sample points of the initial node pose. The next node pose is obtained by using multiple input sample points and the path planning function. The parameter range of the next node pose is solved based on the input sample points, the output sample points, and the polynomial surrogate model.

2. The automatic parking path planning method according to claim 1, characterized in that, The parameter range for solving the pose of the next node based on the input sample points, the output sample points, and the polynomial surrogate model includes: By using multiple input sample points as independent variables of the polynomial surrogate model and multiple output sample points as dependent variables of the polynomial surrogate model, the coefficient relationship between the parameters of the initial node pose and the parameters of the next node pose is solved. The parameter range of the next node pose is derived based on the coefficient relationship and the parameter range of the initial node pose.

3. The automatic parking path planning method according to claim 2, characterized in that, The polynomial proxy model is the Chebyshev proxy model: Where g is the number of parameters in the initial node pose; u is the number of parameters i1, i2, ..., i in the Chebyshev series. g The number of zeros. The pose of the input sample points is used as the independent variable (x1, x2, ..., x) of the Chebyshev surrogate model. g The output sample points are used as the dependent variable of the Chebyshev surrogate model. The coefficient relationship between the parameters of the initial node pose and the parameters of the next node pose is as follows:

4. The automatic parking path planning method according to claim 1, characterized in that, The multiple input sample points are obtained by the tensor product of m interpolation points for each parameter of the initial node pose.

5. The automatic parking path planning method according to claim 1, characterized in that, The path planning function includes the Hybrid A* algorithm.

6. An automatic parking path planning device, characterized in that, include: The positioning module is used to collect the parameters of the vehicle's initial node pose and the location of the target parking space; The acquisition module is used to acquire the parameter range of the initial node pose, wherein the parameter range of the initial node pose is obtained based on the error range of the parameters of the initial node pose. The solution module is used to solve for the parameter range of the next node pose using the path planning function and the parameter range of the initial node pose; The detection module is used to perform obstacle interference detection based on the parameter range of the pose of the next node, and to solve for the cost function value of the next node; The determination module is used to select the next node that passes obstacle interference detection and has the minimum cost function value as the planning node; The parameter range of the initial node pose is: in, The initial pose parameters to be acquired, δ a The acquired rotation angle, and the parameters of the acquired initial node pose include the initial node pose parameters. and the collected wheel rotation angle δ a ; The initial pose parameters The error range, δ e The wheel rotation angle δ collected a The error range; The parameter range for solving the next node pose using the path planning function and the parameter range of the initial node pose includes: Within the range of each parameter of the initial node pose, select m interpolation points, and arrange and combine the m interpolation points of each parameter to obtain multiple input sample points of the initial node pose. The next node pose is obtained by using multiple input sample points and the path planning function. The parameter range of the next node pose is solved based on the input sample points, the output sample points, and the polynomial surrogate model.

7. An automatic parking path planning system, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the automatic parking path planning method according to any one of claims 1 to 5 by executing the executable instructions.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the automatic parking path planning method according to any one of claims 1 to 5.

Images