Autonomous parking path planning method and device based on skeleton-assisted Reeds-Shepp curve

Abstract

The invention provides an autonomous parking path planning method and device based on a skeleton-assisted Reeds-Shepp curve, and the method comprises the steps: determining a grid map corresponding to a real environment map of a parking lot; according to the parking information in the grid map and the minimum turning radius of a to-be-parked vehicle, determining a target parking point of the to-be-parked vehicle; the parking information comprises a vehicle parking mode, a parking starting point and a parking ending point; according to the target parking point, a curve skeleton corresponding to the grid map and the small turning radius, a feasible path of the to-be-parked vehicle during parking is determined, and the curve skeleton is obtained by performing skeletonization processing on the grid map based on a preset skeletonization algorithm; and testing and optimizing the feasible path to obtain a target parking path of the to-be-parked vehicle. According to the scheme, the safety of the to-be-parked vehicle can be ensured, and meanwhile, the parking path planning efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to an autonomous parking path planning method and device based on a skeleton-assisted Reeds-Shepp curve. Background Art

[0002] With the development of autonomous driving technology, the Autonomous Valet Parking (AVP) system has become an effective solution to the parking problem in urban environments. The AVP system improves the utilization rate of parking lots and the user's parking experience in an automated way. In the AVP system, path planning algorithms are the core components. They are responsible for generating a safe and efficient path from the parking lot entrance to the designated parking space while complying with the kinematic constraints of the vehicle. This challenge involves not only point-to-point navigation, but also the ability to handle complex parking environments and ensure reliable execution in the real world.

[0003] Path planning faces many challenges in parking environments, such as how to ensure that the vehicle can achieve accurate obstacle avoidance, how to complete parking quickly and efficiently, and how to solve parking path planning in complex "dead-end" scenarios. These challenges make existing path planning technologies often difficult to meet requirements when faced with scenarios that require multiple motion sequences or precise position control. In addition, the computational complexity of generating optimal trajectories while maintaining real-time performance further exacerbates these challenges.

[0004] Scholars have conducted extensive research on path planning algorithms and proposed a variety of methods. These methods mainly include graph-based search algorithms and geometric curve-based algorithms. The A* algorithm is one of the most typical graph-based search algorithms and is widely used in path planning problems in static environments due to its high efficiency. The A* algorithm finds the optimal path by calculating the path cost from the starting point to the target point, ensuring the efficiency and optimality of the path search, and can quickly approach the target position under the guidance of the heuristic function.

[0005] However, the standard A* algorithm consumes large computational resources, has high time complexity, and has poor path smoothness. To solve these problems, researchers introduced an improved heuristic function to reduce unnecessary node searches and combined geometric curves to make the path smoother. The Reeds-Shepp curve is a classic curve that can satisfy vehicle kinematics. However, when dealing with complex environments, such as urban parking scenarios, the traditional RS curve faces the problems of large computational complexity and difficult path optimization. At the same time, due to the complexity of the Voronoi diagram calculation, this will generate high time costs and a large amount of redundant information. Using the geometric curve method to combine different curves to generate paths can adapt to a variety of parking scenarios, but may require further optimization in terms of kinematic feasibility and computational efficiency.

[0006] Although these existing parking technology solutions have made progress in some aspects, most of them use inefficient algorithms and cannot quickly and accurately obtain an optimal path, or obtain a shorter path with poor smoothness or even an unfeasible path. There is still no good solution to the path planning problem in complex parking scenarios, especially dead-end road scenarios. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide an autonomous parking path planning method and device based on a skeleton-assisted Reeds-Shepp curve, which is used in an autonomous parking system, especially in complex environments and dead-end road scenarios, to solve the technical problems of the prior art, such as large consumption of computing resources, high time complexity, and poor path smoothness, which result in the inability to quickly and accurately provide an optimal parking path.

[0008] To solve the above technical problems, an embodiment of the present invention provides an autonomous parking path planning method based on a skeleton-assisted Reeds-Shepp curve, comprising:

[0009] Determine the grid map corresponding to the real environment map of the parking lot;

[0010] Determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking information includes a vehicle parking method, a parking start point, and a parking end point; determining a feasible path for the vehicle to be parked according to the target parking point, a curve skeleton corresponding to the grid map, and the minimum turning radius, wherein the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm;

[0011] The feasible path is tested and optimized to obtain a target parking path for the vehicle to be parked.

[0012] In one embodiment, determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked includes:

[0013] Determining a target position of the vehicle to be parked according to the parking method and the parking end point;

[0014] The target parking point is determined according to the target posture, the parking starting point and the minimum turning radius.

[0015] In one embodiment, determining a feasible path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius includes:

[0016] Determining a guidance path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius;

[0017] A feasible path for parking the vehicle to be parked is determined according to the guiding nodes on the guiding path and the first preset path curve.

[0018] In one embodiment, determining a guidance path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius includes:

[0019] Extracting target skeleton nodes on the curve skeleton;

[0020] Based on a first preset A* algorithm, searching and processing the target skeleton node and the target parking point to determine an intermediate path for the vehicle to be parked;

[0021] A guiding path for the vehicle to be parked is determined according to the intermediate path and the minimum turning radius.

[0022] In one embodiment, determining the guidance path for the vehicle to be parked according to the intermediate path and the minimum turning radius includes:

[0023] Determine a plurality of intersection points between a circle having each intermediate node on the intermediate path as a circle center and the minimum turning radius as a drawing radius and the intermediate path;

[0024] The guide path is determined according to the plurality of intersection points.

[0025] In one embodiment, determining a feasible path for the vehicle to be parked based on the guidance path includes:

[0026] Each guiding node on the guiding path is used as a target node, and the target node is expanded to obtain a feasible path for the vehicle to be parked.

[0027] In one embodiment, the feasible path is tested and optimized to obtain a target parking path for the vehicle to be parked, including:

[0028] A collision test is performed on the feasible nodes in the feasible path according to a preset collision strategy. If a test result obtained by the collision test indicates that the current feasible path can avoid all obstacles, the current feasible path is determined as the target parking path; otherwise, a subsequent feasible path to which the feasible node including the collision location belongs is optimized to obtain the target parking path.

[0029] In one embodiment, optimizing the subsequent feasible paths including the feasible nodes where the collision occurs to obtain the target parking path includes:

[0030] When the feasible node at the collision location is located at the last segment of the feasible path and a first body posture angle of the vehicle to be parked at the feasible node at the collision location is opposite to a second body posture angle at the end point of the last segment of the feasible path, a previous feasible node of the feasible node at the collision location is determined, and an extension process is performed on the path between the previous feasible node and the parking end point to obtain a first subsequent target path.

[0031] In one embodiment, optimizing the subsequent feasible paths including the feasible nodes where the collision occurs to obtain the target parking path includes:

[0032] When the feasible node at the collision location is located at the last segment of the feasible path and the first body posture angle of the vehicle to be parked at the feasible node at the collision location is not opposite to the second body posture angle at the end point of the last segment of the feasible path, the feasible node at the collision location, the nodes between the feasible node at the collision location and the parking end point, and the parking end point are searched based on a second preset A* algorithm, and the nodes in the searched path are expanded to obtain a second subsequent target path.

[0033] Otherwise, based on a second preset A* algorithm, search processing is performed on the feasible nodes at the collision location, the nodes between the feasible nodes at the collision location and the parking end point, and the parking end point, and expansion processing is performed on the nodes in the searched path to obtain a second subsequent target path.

[0034] An embodiment of the present invention further provides an autonomous parking path planning device based on a skeleton-assisted Reeds-Shepp curve, comprising:

[0035] A map construction module, used to determine a grid map corresponding to a real environment map of the parking lot;

[0036] a path planning module, for determining a target parking point for the vehicle to be parked according to parking information in the grid map and a minimum turning radius of the vehicle to be parked; the parking information includes a vehicle parking method, a parking start point, and a parking end point; determining a feasible path for the vehicle to be parked according to the target parking point, a curve skeleton corresponding to the grid map, and the minimum turning radius, the curve skeleton being obtained by skeletonizing the grid map based on a preset skeletonization algorithm; and testing and optimizing the feasible path to obtain a target parking path for the vehicle to be parked.

[0037] The above solution of the present invention includes at least the following beneficial effects:

[0038] The above scheme of the present invention provides an autonomous parking path planning method and device based on a skeleton-assisted Reeds-Shepp curve, wherein the method includes: determining a grid map corresponding to a real environment map of a parking lot; determining a target parking point for the vehicle to be parked according to parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking space information includes a vehicle parking method, a parking start point, and a parking end point; determining a feasible path for the vehicle to be parked according to the target parking point, a curve skeleton corresponding to the grid map, and the minimum turning radius, wherein the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm; testing and optimizing the feasible path to obtain the target parking path for the vehicle to be parked. The scheme of the present invention can ensure the timeliness, accuracy, and safety of parking path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of a method for determining a parking path provided by an embodiment of the present invention;

[0040] Figure 2 is a grid map corresponding to a parking lot scene provided by an optional embodiment of the present invention;

[0041] Figure 3 is a curve skeleton corresponding to a grid map provided by an optional embodiment of the present invention;

[0042] Figure 4 is a simplified curve skeleton provided by an optional embodiment of the present invention;

[0043] Figure 5 is a guiding path effect diagram provided by an optional embodiment of the present invention;

[0044] Figure 6 It is a feasible path effect diagram obtained by extending a preset RS curve with each node of the guide path as the target point provided by an optional embodiment of the present invention;

[0045] Figure 7 is a target parking path effect diagram corresponding to a "dead-end road" scenario provided by an optional embodiment of the present invention;

[0046] Figure 8 It is a module block diagram of a device for determining a parking path provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0048] In the following description, certain specific details are set forth for the purpose of illustrating various disclosed embodiments to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the relevant art will recognize that the embodiments may be practiced without one or more of these specific details. In other cases, well-known devices, structures, and techniques associated with the present application may not be shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.

[0049] References throughout the specification to "one embodiment" or "an embodiment" indicate that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any manner in one or more embodiments.

[0050] In the following description, in order to clearly show the structure and working mode of the present invention, many directional words will be used for description, but the words "front", "back", "left", "right", "outside", "inside", "outward", "inward", "up", "down", etc. should be understood as convenient terms and should not be understood as restrictive terms.

[0051] like Figure 1 As shown, an embodiment of the present invention provides an autonomous parking path planning method 10 based on a skeleton-assisted Reeds-Shepp curve, comprising:

[0052] Step 11, determining the grid map corresponding to the real environment map of the parking lot;

[0053] Step 12, determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking information includes the vehicle parking method and the parking starting point;

[0054] Step 13, determining a feasible path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius, wherein the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm;

[0055] Step 14: Test and optimize the feasible paths to obtain a target parking path for the vehicle to be parked.

[0056] In this embodiment, the grid map is a scene map that satisfies the AVP system of the vehicle to be parked. The grid map can be modeled based on the real environment map of the parking lot, so that the vehicle to be parked can obtain parking information on the grid map; here, a feature map model, a topological map model or a grid map model can be used for modeling;

[0057] In an achievable example, based on the grid map corresponding to the real environment map of the vertical parking lot, a parking lot with a size of 102m*102m is set with a resolution of 0.1m. Figure 2 As shown, each grid C corresponds to a parking space; A represents the parking lot entrance and B represents the parking lot exit. A total of 236 parking spaces are set (corresponding to 236 grids C). The size of each parking space is set to 3m*6m. There are five "dead-end road" parking space areas as shown in the red circles in the figure. The detailed experimental data are shown in Table 1.

[0058] Table 1. Vehicle dynamics and parking simulation parameters

[0059]

[0060] Here, the parking information includes but is not limited to the vehicle parking method, parking starting point, parking end point, and parking space information (parking space size, coordinates, direction, and type (whether it is a "dead-end road" parking space)); wherein the parking starting point indicates the position corresponding to the parking lot entrance, and the parking end point indicates the parking space coordinates of the final parking space for the vehicle to be parked (one parking space corresponds to one parking end point), and the vehicle parking method includes vertical parking space parking, parallel parking space parking, and diagonal parking space parking. In this application, since the grid map is constructed based on the vertical parking lot, the subsequent vehicle parking methods are all vertical parking space parking;

[0061] The target parking point indicates the position point where the vehicle to be parked moves from the parking starting point to the point where it starts to park in the current parking space; the minimum turning radius is the basic attribute information of the vehicle to be parked; here, the target parking point can be determined based on the parking information and the minimum turning radius of the vehicle to be parked and through a preset RS curve algorithm to ensure the accuracy and efficiency of subsequent target parking path planning and reduce computing resource consumption;

[0062] Furthermore, the grid map can be skeletonized according to a preset skeletonization algorithm to obtain a curve skeleton of a one-dimensional structure to simplify a complex parking environment and reduce the complexity of calculations during subsequent path planning; based on the simplified curve skeleton, the target parking point and the minimum turning radius, a feasible path from the parking position to the target parking point is determined, and the feasible path is tested and optimized to ensure that the target parking path ultimately obtained avoids all obstacles, ensures safety and reliability when parking using the target parking path, and ensures the actual feasibility of the target parking path.

[0063] In an optional embodiment of the present invention, the above step 12 may include:

[0064] Step 121, determining a target position of the vehicle to be parked according to the parking mode and the parking end point of the vehicle;

[0065] Step 122, determining a target parking point according to the target posture and minimum turning radius of the vehicle to be parked.

[0066] In this embodiment, the target position and posture of the vehicle to be parked includes the parking end point and posture corresponding to the vehicle to be parked; here, the parking end point corresponds to the parking end point (parking space coordinates), and the posture of the vehicle to be parked corresponds to the parking space direction;

[0067] When the vehicle to be parked enters the parking lot, firstly, the parking type of the parking space can be determined according to the parking mode and parking end point corresponding to the parking space in the parking lot, so as to determine the parking end point and posture of the vehicle to be parked;

[0068] Furthermore, according to the initial posture and the minimum turning radius, the specific process of determining the target parking point is as follows:

[0069] Step 1221, taking the center point of the parking end point as the initial position, taking the posture of the vehicle to be parked as the initial direction, and using the preset RS curve algorithm to calculate the target parking point, the pseudo code is shown in Table 2. Specifically: starting from the center point, draw a straight line segment of a preset length along the initial direction as the initial path; taking the minimum turning radius of the vehicle to be parked as the radius, the center of the parking arc segment is calculated based on the parking side and the parking end point of the vehicle parking method; at the same time, the RS curve is used to calculate the target parking point, and collision detection is performed. If a collision occurs, the distance of the straight line out of the warehouse is increased until no collision occurs, and finally the target parking point is obtained.

[0070] Table 2. Parking posture algorithm

[0071]

[0072]

[0073] In an achievable example of the present invention, skeletonization of the grid map is performed based on a preset skeletonization algorithm to obtain a curve skeleton, which may specifically include the following steps:

[0074] Step 123, extracting key feature information in the grid map, including but not limited to walls, obstacles and parking space boundaries, and performing spatial decomposition based on the key feature information to obtain a two-dimensional layout map corresponding to the walls, obstacles and parking space boundaries;

[0075] Step 124, skeletonizing the two-dimensional layout images by using a preset skeletonization algorithm, specifically: converting each of the two-dimensional layout images into a skeleton line of the image, extracting all key geometric feature points (turning points and intersections) on the skeleton line, and connecting them in the order of the key geometric feature points to form a curve, and determining the connected curve as the curve skeleton of the grid map;

[0076] Step 125, optimizing the terminal points of the skeleton curve, removing unnecessary skeleton nodes to reduce the computational complexity, while ensuring the connectivity and integrity of the path; here, the terminal points of the curve represent the skeleton nodes in the four directions of the curve skeleton;

[0077] The grid map is reduced to the central axis representation through the preset skeletonization algorithm, while retaining the original topological and geometric properties to ensure the accuracy of subsequent operations; here the two-dimensional layout diagram is simplified into a curve skeleton of a one-dimensional structure, such as Figure 3 As shown in the figure, the complex parking environment is simplified; the terminal points of the curve skeleton are further optimized and simplified to obtain the following Figure 4 As shown in the figure, the optimized curve skeleton shows that the number of skeleton nodes on the curve skeleton is further reduced. At the same time, the simplified skeleton nodes can effectively retain the core information of the passable area and eliminate unnecessary redundant data, so as to reduce the computational complexity and consumption of computing resources in subsequent parking path planning, and ensure the accuracy and real-time performance of parking path planning.

[0078] In an optional embodiment of the present invention, the above step 13 may include:

[0079] Step 131, determining a guidance path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius;

[0080] In this embodiment, the guidance path is a path between a skeleton node on the curve skeleton and a target parking point, which provides a simplified but information-rich environment representation for the vehicle, helping the vehicle to navigate in a complex parking environment;

[0081] Specifically, the above step 131 may include:

[0082] Step 1311, extracting target skeleton nodes on the curve skeleton;

[0083] Step 1312, based on the first preset A* algorithm, searching and processing the skeleton nodes and the target parking point to determine the intermediate path when the vehicle to be parked is parked;

[0084] Step 1313, determining a guidance path for the vehicle to be parked according to the intermediate path and the minimum turning radius.

[0085] In this embodiment, the curve skeleton includes multiple skeleton nodes, where the target skeleton node is a skeleton node between the target parking point and the parking start point and located on the curve skeleton; the target skeleton node may include a turning node, an intersection node, and a terminal point that maintains path connectivity; the standard for extracting the target skeleton node is based on path connectivity, spatial decomposition efficiency, safety and obstacle avoidance, and vehicle kinematic constraints.

[0086] By extracting and screening the skeleton nodes on the curve skeleton, and using the screened target skeleton nodes as intermediate nodes, these intermediate nodes and the target parking point are used as intermediate path construction nodes of the first preset A* algorithm to construct a preliminary path (intermediate path) from the current position of the vehicle to the target parking point, thereby reducing the calculation complexity and ensuring the accuracy and real-time performance of the path planning;

[0087] Here, based on the first preset A* algorithm, searching and processing the skeleton nodes and the target parking point to determine the intermediate path when the vehicle to be parked is parked may specifically include the following steps:

[0088] Step 13121, by creating two lists: an open list (Open List) and a closed list (ClosedList); the open list is used to store the target skeleton nodes to be explored, and the closed list is used to store the target skeleton nodes that have been explored, and the parking start point (the current position of the vehicle) and the target parking entry point are also added to the open list; the cost function f(n) of the open list is set, and g(n) is set to represent the actual cost from the parking start point to the current target skeleton node, and h(n) is set to be the estimated cost (heuristic function) from the current target skeleton node to the target parking entry point;

[0089] Step 13122, when the open list is not empty, perform the following process: select the target skeleton node with the lowest f(n) value from the open list, called the current target skeleton node (as the parent node); if the current target skeleton node is the target docking point, the algorithm is completed and a path is found; move the current target skeleton node from the open list to the closed list. For each neighbor target skeleton node of the current target skeleton node: if the neighbor target skeleton node is not in the closed list, calculate the value of its cost function f(n). If the neighbor target skeleton node is in the open list and the cost of reaching the neighbor target skeleton node through the current target skeleton node is lower, update its cost and parent node. According to the calculated value of the cost function f(n), update the order of the target skeleton nodes in the open list to ensure that the target skeleton node with the lowest f(n) value is explored first; repeat the above steps until the target docking point in the open list is found or the open list is empty (indicating that the target docking point cannot be reached).

[0090] The above process can be implemented in the Autonomous Valet Parking (AVP) system through the following code 2:

[0091] Table 3: First preset A* algorithm

[0092]

[0093] Since the first preset A* algorithm does not consider the turning radius and collision volume of the vehicle to be parked, directly guiding the kinematic path extension will cause the path of the vehicle to be parked to become longer and increase the risk of collision; the intermediate path at this time is not the optimal path for guiding parking. In order to make the planned kinematic path trajectory conform to the actual operation scenario, the intermediate path needs to be processed according to the minimum turning radius of the vehicle to be parked to obtain a guiding path that meets the requirements;

[0094] In an optional embodiment of the present invention, the above step 1313 may specifically include:

[0095] Step 13131, determine multiple intersection points of a circle with each intermediate node on the intermediate path as the center and the minimum turning radius as the drawing radius and the intermediate path, such as Figure 5 As shown, the red dots in the figure indicate the intersection points;

[0096] Step 13132, determine a guidance path based on multiple intersecting intersection points.

[0097] In this embodiment, the starting point of the vehicle's optimal turning path is often at the minimum turning radius from the lane turning point (that is, the minimum turning radius of the vehicle). Therefore, when performing kinematic path planning, each node of the middle path is used as the center of the circle, and a circle is drawn with the minimum turning radius of the vehicle, and a guiding path is constructed by finding the points that intersect with the middle path;

[0098] Specifically: take each intermediate node on the intermediate path as the center of the circle, take the minimum turning radius of the vehicle to be parked as the drawing radius, and draw a circle corresponding to each intermediate node; after drawing the circle corresponding to each intermediate node, determine the two intersecting nodes of the circle and the intermediate path, and then obtain multiple intersecting nodes; and connect the intersecting nodes in sequence to obtain a guiding path.

[0099] In an optional embodiment of the present invention, the above step 13 may further include:

[0100] Step 132, determining a feasible path for the vehicle to be parked according to the guidance path.

[0101] Specifically, the above step 132 may include:

[0102] Step 1321, taking each guiding node on the guiding path as a target node, and performing expansion processing on the target node to obtain a feasible path for the vehicle to be parked.

[0103] In this embodiment, the target node can be expanded by a preset RS curve algorithm to ensure that the path not only meets the kinematic constraints of the vehicle, but also can navigate safely and effectively in a complex parking environment; the specific expansion process is as follows:

[0104] Step 13211, starting from the first target node of the guidance path, insert an RS curve path segment between two adjacent target nodes in sequence (corresponding to the order in which the vehicles to be parked pass the target nodes); here, each RS curve path segment is obtained by connecting the coordinates and angles of two adjacent target nodes, and each RS curve path segment is obtained and added to the final feasible path in sequence until all target nodes of the guidance path are processed, and finally a smooth feasible path that meets the vehicle motion constraints is obtained.

[0105] In an optional embodiment of the present invention, the above step 14 may include:

[0106] Step 141, performing collision test processing on feasible nodes in the feasible path according to a preset collision strategy, if the test result obtained by the collision test processing indicates that the current feasible path can avoid all obstacles, then determining the current feasible path as the target parking path;

[0107] Step 142: Otherwise, optimize the subsequent feasible paths to which the feasible nodes including the collision location belong to obtain a target parking path.

[0108] In this embodiment, the preset collision strategy may be a two-circle detection method; by performing obstacle detection on each feasible node in the feasible path, it is ensured that the generated path can be correctly executed by the vehicle to be parked, and the feasible path that succeeds the test is determined as the final target parking path;

[0109] Here, obstacle detection is performed using the two-circle detection method. Specifically, obstacle detection can be performed at each feasible node based on the vehicle to be parked. The specific detection process is as follows:

[0110] Step 1411, defining two centers of the vehicle to be parked at each feasible node: a front circle and a rear circle, and simulating the position and shape of the vehicle to be parked in space;

[0111] Step 1412, determining a first position corresponding to a front circle of the vehicle to be parked and a second position corresponding to a rear circle in the global coordinate system; here, the first position of the front circle is the center of the front axle of the vehicle to be parked, and the second position of the rear circle is the center of the rear axle of the vehicle to be parked;

[0112] Step 1413, checking whether the first position corresponding to the front circle or the second position corresponding to the rear circle collides with an obstacle, so as to determine whether the vehicle to be parked is in a safe state; specifically: when the first position corresponding to the front circle or the second position corresponding to the rear circle overlaps with the position of the obstacle, the vehicle to be parked collides with the obstacle; otherwise, the vehicle to be parked will not collide.

[0113] When the detection indicates that the current feasible path cannot avoid the obstacle (collision occurs), the state of the current feasible path is identified to facilitate subsequent optimization of the feasible path to obtain the final target parking path.

[0114] In an optional embodiment of the present invention, the above step 142 may include:

[0115] Step 1421, when the feasible node at the collision location is located at the last segment of the feasible path and the first body posture angle of the vehicle to be parked at the feasible node at the collision location is opposite to the second body posture angle at the parking end point of the last segment of the feasible path, determine the previous feasible node of the feasible node at the collision location, and extend the path between the previous feasible node and the parking end point to obtain a first subsequent target path;

[0116] Step 1422, otherwise, based on the second preset A* algorithm, search and process the feasible nodes at the collision location, the nodes between the feasible nodes at the collision location and the parking end point, and the parking end point, and expand the nodes in the searched path to obtain a second subsequent target path.

[0117] In this embodiment, step 1421 is mainly used to identify a "dead-end road" parking space, because the vehicle to be parked needs to perform a complex reversing operation to exit the "dead-end road" parking space; when the collision detection fails, firstly, the feasible node at the collision location is located at the last segment of the feasible path. If the first body posture angle at the feasible node at the collision location is opposite to the second body posture angle at the parking end of the last segment of the feasible path, then the parking space corresponding to the parking end of the last segment is determined to be the "dead-end road" parking space; here, the feasible node at the collision location is the target parking point; here, the first body posture angle and the second body posture angle may also be 180 degrees apart;

[0118] After the current parking space is identified as a "dead-end" parking space, the last segment of the path is deleted, and the previous feasible node of the feasible node where the collision occurs is redirected, and the path between the previous feasible node and the parking end point is extended to obtain the first subsequent target path; here, the specific process of extending the path between the previous feasible node and the parking end point is the same as the above-mentioned process of generating a feasible path, which will not be repeated here;

[0119] Here, the first subsequent target path represents the path between the feasible node at the collision location and the parking end point; further, the first subsequent target path is combined with the path before the collision location, and the final path is obtained as follows: Figure 7 The target parking path shown is to enable the vehicle to be parked to complete the operation of reversing into the "dead-end" parking space, so as to effectively avoid the risk of the vehicle to be parked being unable to enter the parking space in the narrow "dead-end" parking space scenario, thereby improving the practicality and safety of path planning.

[0120] When the obstacle detection fails and it is determined that it is not a "dead-end" parking space, it is necessary to re-plan the path between the feasible node at the collision site and the parking end point; at this time, return to the path where the collision occurred, and use the second preset A* algorithm to search and process the feasible node at the collision site, the node between the feasible node at the collision site and the parking end point, and the parking end point, and perform obstacle detection based on the two-circle detection method to obtain a path away from the obstacle, and expand the nodes in the searched, i.e., detected, path to obtain a second subsequent target path; here, the second subsequent target path represents the path between the feasible node at the collision site and the parking end point; further, the second subsequent target path is combined with the path before the collision site, and the final Figure 6 The target parking path shown; here, the second preset A* algorithm is the same as the first preset A* algorithm in the above embodiment, and will not be repeated here;

[0121] The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve provided by the above embodiment of the present invention determines the target parking point of the vehicle to be parked by using the parking information in the grid map and the minimum turning radius of the vehicle to be parked, thereby improving the accuracy and efficiency of parking path planning and reducing the consumption of computing resources; further, according to the target parking point, the curve skeleton corresponding to the grid map and the minimum turning radius, the feasible path for the vehicle to be parked is determined, and the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm. The skeletonization algorithm is used to simplify the complex environment corresponding to the grid map into a curve skeleton with a one-dimensional structure, and the key path nodes are extracted to reduce the search space. The time and computing resource consumption are reduced, and the efficiency of path planning is improved; further, in the feasible path planning stage, a preliminary intermediate path is firstly generated based on the first preset A* algorithm and the simplified target skeleton node; then a circle is drawn with the intermediate node of the intermediate path as the center and the minimum turning radius of the vehicle to be parked, and a guiding path is constructed to ensure that the subsequent path planning conforms to the vehicle kinematic characteristics, while avoiding the risk of excessive path length and collision; to ensure the safety of the path, the feasible path is tested and optimized to ensure that the planned path avoids all obstacles; if a collision between the path and the obstacle is detected, the system will re-plan the path, which improves the safety and reliability of the path and ensures the actual executability of the path;

[0122] In the special scenario of "dead-end" parking spaces, the system identifies "dead-end" parking spaces by matching the features of the parking spaces (such as the vehicle body posture angle), deletes the last section of the path, and redirects an intermediate node so that the vehicle can complete the operation of reversing into the dead-end road, effectively avoiding the risk of the vehicle being unable to enter the parking space in narrow dead-end road scenarios, thereby improving parking efficiency and safety in complex and confined spaces.

[0123] like Figure 8 As shown, an embodiment of the present invention further provides an autonomous parking path planning device 20 based on a skeleton-assisted Reeds-Shepp curve, comprising:

[0124] A map construction module 21, used to determine a grid map corresponding to a real parking environment map;

[0125] The path planning module 22 is used to determine the target parking point of the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking information includes the vehicle parking method, the parking starting point and the parking end point; according to the target parking point, the curve skeleton corresponding to the grid map and the minimum turning radius, determine the feasible path for the vehicle to be parked when parking, the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm; test and optimize the feasible path to obtain the target parking path of the vehicle to be parked.

[0126] It should be noted that this device is a device corresponding to the above-mentioned autonomous parking path planning method 10 based on skeleton-assisted Reeds-Shepp curve. All implementation methods in the above-mentioned method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0127] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0128] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0129] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0130] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention.

[0131] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0132] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code for implementing a method or device. That is to say, such a program product also constitutes the present invention, and a storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0133] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve, characterized in that: include: Determine the grid map corresponding to the real environment map of the parking lot; Determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking information includes a vehicle parking method, a parking starting point, and a parking end point; Determining a feasible path for the vehicle to be parked according to the target parking point, a curve skeleton corresponding to the grid map, and the minimum turning radius, wherein the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm; The feasible path is tested and optimized to obtain a target parking path for the vehicle to be parked.

2. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 1, characterized in that: Determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked includes: Determining a target position of the vehicle to be parked according to the vehicle parking method and the parking end point; The target parking point is determined according to the target posture, the parking starting point and the minimum turning radius.

3. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 1, characterized in that: Determining a feasible path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius includes: Determining a guidance path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the estimated minimum turning radius; A feasible path for parking the vehicle to be parked is determined according to the guidance path.

4. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 3, characterized in that: Determining a guidance path for the vehicle to be parked according to the target parking point, the curve skeleton corresponding to the grid map, and the minimum turning radius includes: Extracting target skeleton nodes on the curve skeleton; Based on a first preset A* algorithm, searching and processing the target skeleton node and the target parking point to determine an intermediate path for the vehicle to be parked; A guiding path for the vehicle to be parked is determined according to the intermediate path and the minimum turning radius.

5. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 4, characterized in that: Determining a guidance path for the vehicle to be parked according to the intermediate path and the minimum turning radius includes: Determine a plurality of intersection points between a circle having each intermediate node on the intermediate path as a circle center and the minimum turning radius as a drawing radius and the intermediate path; The guide path is determined according to the plurality of intersection points.

6. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 3, characterized in that: Determining a feasible path for the vehicle to be parked according to the guidance path includes: Each guiding node on the guiding path is used as a target node, and a preset RS curve is extended for the target node to obtain a feasible path for the vehicle to be parked.

7. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 1, characterized in that: Testing and optimizing the feasible path to obtain a target parking path for the vehicle to be parked includes: A collision test is performed on the feasible nodes in the feasible path according to a preset collision strategy. If a test result obtained by the collision test indicates that the current feasible path can avoid all obstacles, the current feasible path is determined as the target parking path; otherwise, a subsequent feasible path to which the feasible node including the collision location belongs is optimized to obtain the target parking path.

8. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 7, characterized in that: Optimizing the subsequent feasible paths including the feasible nodes where the collision occurs to obtain the target parking path, including: When the feasible node at the collision location is located at the last segment of the feasible path and a first body posture angle of the vehicle to be parked at the feasible node at the collision location is opposite to a second body posture angle at the end point of the last segment of the feasible path, a previous feasible node of the feasible node at the collision location is determined, and an extension process is performed on the path between the previous feasible node and the parking end point to obtain a first subsequent target path.

9. The autonomous parking path planning method based on skeleton-assisted Reeds-Shepp curve according to claim 7, characterized in that: Optimizing the subsequent feasible paths including the feasible nodes where the collision occurs to obtain the target parking path, including: When the feasible node at the collision location is located at the last segment of the feasible path and a first body posture angle of the vehicle to be parked at the feasible node at the collision location is not opposite to a second body posture angle at the end point of the last segment of the feasible path, a search process is performed on the feasible node at the collision location, nodes between the feasible node at the collision location and the parking end point, and the parking end point based on a second preset A* algorithm, and an expansion process is performed on the nodes in the searched path to obtain a second subsequent target path.

10. An autonomous parking path planning device based on skeleton-assisted Reeds-Shepp curve, characterized in that: include: A map construction module, used to determine a grid map corresponding to a real environment map of the parking lot; a path planning module, for determining a target parking point for the vehicle to be parked according to the parking information in the grid map and the minimum turning radius of the vehicle to be parked; the parking information includes a vehicle parking method, a parking start point, and a parking end point; and determining a feasible path for the vehicle to be parked according to the target parking point, a curve skeleton corresponding to the grid map, and the minimum turning radius, wherein the curve skeleton is obtained by skeletonizing the grid map based on a preset skeletonization algorithm; The feasible path is tested and optimized to obtain a target parking path for the vehicle to be parked.

Images