Parking path processing method, product and vehicle

Abstract

The invention provides a parking path processing method, a product and a vehicle, and belongs to the technical field of parking. The method comprises the following steps: according to a pose error corresponding to a node in a determined initial parking path, obtaining a node after smooth processing; and obtaining a target parking path after smoothing processing according to the nodes after smoothing processing. The objective of the invention is to obtain a smoother parking path.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of parking technology, and more specifically, to a parking path processing method, product, and vehicle. Background Art

[0002] With the popularization of intelligent electric vehicles, more and more vehicles are equipped with automatic parking technology. Path planning is one of the key technologies in automatic parking technology. The quality of path planning directly affects the user experience. If the path planning effect is good, the parking time will be short, the risk of collision will be reduced, and the user will have a higher sense of trust. If the path planning is not good, the parking time will be longer and the risk of collision will be greater, resulting in a poor user experience.

[0003] Current path planning methods can plan a collision-free parking path when parking, but there are mutations at the nodes in the parking path. When the vehicle parks according to the planned path, the vehicle's orientation and steering angle will change suddenly, causing the vehicle to jerk. Summary of the Invention

[0004] The embodiments of the present application provide a parking path processing method, product, and vehicle, aiming to improve the comfort during parking.

[0005] In a first aspect, an embodiment of the present application provides a parking path processing method, the method comprising: Obtain smoothed nodes based on the posture errors corresponding to the nodes in the determined initial parking path; A smoothed target parking path is obtained according to the smoothed nodes.

[0006] Optionally, the method further includes: Determining a predicted state value of the node according to the initial state value of the node by using an extended Kalman filter and a preset kinematic model; Determine a posture error corresponding to the node based on the initial state value and the predicted state value of the node.

[0007] Optionally, determining a posture error corresponding to the node according to the initial state value and the predicted state value of the node includes: Determine the position error and / or heading angle error corresponding to the node according to the initial state value and the predicted state value of the node.

[0008] Optionally, the posture error includes a position error and a heading angle error. According to the initial state value and the predicted state value of the node, a formula for determining the position error and heading angle error corresponding to the node is:

[0009] in, and is the position error; is the heading angle error; is the initial x-axis coordinate in the initial state value; is the initial y-axis coordinate in the initial state value; is the initial heading angle in the initial state value; x is the predicted x-axis coordinate in the predicted state value; y is the predicted y-axis coordinate in the predicted state value; is the predicted heading angle in the predicted state value.

[0010] Optionally, the kinematic model is:

[0011] in, for k The predicted state value at the moment; for k The predicted x-axis coordinate of the predicted state value at the moment; for k The predicted y-axis coordinate of the predicted state value at the moment; for k The predicted heading angle in the predicted state value at the moment; for k The x-axis coordinate of the updated state value; for k The y-axis coordinate of the state value after the moment is updated; for k The heading angle in the updated state value; for k Vehicle speed at the time; for k The wheel steering angle at the moment; L is the wheelbase of the vehicle.

[0012] Optionally, obtaining smoothed nodes based on the determined pose errors corresponding to the nodes in the initial parking path includes: According to the posture errors corresponding to the nodes in the initial parking path, the process noise matrix in the extended Kalman filter is adjusted to obtain smoothed nodes corresponding to the nodes.

[0013] Optionally, adjusting a process noise matrix in an extended Kalman filter according to a posture error corresponding to a node in the initial parking path to obtain a smoothed node corresponding to the node includes: According to the posture error corresponding to the node in the initial parking path, a first scaling factor corresponding to the node for adjusting the process noise matrix in the extended Kalman filter is calculated, and a smoothed node corresponding to the node is obtained at least based on the first scaling factor.

[0014] Optionally, the posture error includes a position error and a heading angle error. According to the posture error corresponding to a node in the initial parking path, a formula for calculating a first scaling factor corresponding to the node for adjusting a process noise matrix in an extended Kalman filter is:

[0015] in, is the first scaling factor corresponding to the process noise matrix; , , and m+n=1; and is the position error; is the heading angle error; is a cardinality close to 0; is the preset magnification factor.

[0016] Optionally, the method further includes: A second scaling factor for adjusting a measurement noise matrix in the extended Kalman filter is determined according to a wheel steering angle corresponding to the node, and a smoothed node corresponding to the node is obtained at least according to the second scaling factor.

[0017] Optionally, according to the wheel steering angle corresponding to the node, a formula for determining a second scaling factor for adjusting the measurement noise matrix in the extended Kalman filter is:

[0018] in, is the second scaling factor; b is the preset zoom factor; for k The wheel steering angle at time .

[0019] Optionally, obtaining a smoothed node corresponding to the node includes: Calculating an estimated value of a state covariance matrix corresponding to the node according to a process noise matrix corresponding to the node and the first scaling factor; The state value of the node is updated according to the estimated value of the state covariance matrix corresponding to the node to obtain the smoothed node corresponding to the node.

[0020] Optionally, according to the process noise matrix corresponding to the node and the first scaling factor, a formula for calculating an estimated value of the state covariance matrix corresponding to the node is:

[0021] in, for k The estimated value of the state covariance matrix corresponding to the moment; for k The state transfer matrix corresponding to the moment; for k -1 moment corresponding to the state covariance matrix; is the first scaling factor corresponding to the process noise matrix; Q is the process noise matrix.

[0022] Optionally, updating the state value of the node according to the estimated value of the state covariance matrix corresponding to the node to obtain the smoothed node corresponding to the node includes: Calculating a gain matrix corresponding to the node according to an estimated value of a state covariance matrix corresponding to the node, the second scaling factor, and the measurement noise matrix; According to the gain matrix and the predicted state value corresponding to the node, the state of the node is updated to obtain a smoothed node corresponding to the node.

[0023] Optionally, based on the estimated value of the state covariance matrix corresponding to the node, the calculation formula for updating the state value of the node is:

[0024] in, for k The gain matrix at time t, for k The updated status value at each moment; for k The estimated value of the state covariance matrix corresponding to the moment; for k The measurement matrix corresponding to the moment, R is the measurement noise matrix; is the second scaling factor; for k The predicted state value at the moment; for k The state covariance matrix after the moment update; for k The initial state value at the moment.

[0025] Optionally, obtaining a smoothed target parking path according to the smoothed nodes includes: The smoothed target parking path is obtained by combining the smoothed nodes corresponding to the nodes in the initial parking path.

[0026] Optionally, the method further includes: In response to the smoothing selection operation, the initial parking path is determined.

[0027] Optionally, the method further includes: The initial parking path is determined based on the starting and ending points of the parking process.

[0028] Optionally, determining an initial parking path based on the starting point and end point of the parking process includes: When the starting point and the end point satisfy the analytical extension, the initial parking path obtained by the analytical extension is used as the target parking path.

[0029] Optionally, determining an initial parking path based on the starting point and end point of the parking process includes: When the starting point and the end point do not satisfy analytical extension, the initial parking path is determined based on a Hybrid A* algorithm.

[0030] Optionally, the method further includes: When the starting point and the end point do not satisfy the parsing extension, the starting point is added to the node open list and a cyclic expansion strategy is executed until the node open list is empty or a node that satisfies the parsing extension with the end point exists, and then an initial parking path is output.

[0031] Optionally, a round-robin extension strategy is implemented, including: In the node open list, if the current node with the smallest path loss and the end point do not meet the requirements for parsing extension, the current node is added to the node closed list, and an extension node is determined based on a differential model; When the extended node passes the collision detection, the path loss corresponding to the extended node is calculated, and the extended node is added to the node open list.

[0032] Optionally, executing the cyclic extension strategy further includes: When the extended node fails the collision detection, the extended node is discarded.

[0033] Optionally, the formula for determining the extended node based on the differential model is:

[0034] in, d is the average driving distance of the vehicle, 、 are the distances traveled by the left and right wheels of the vehicle respectively, , is the position coordinate of the vehicle corresponding to the extended node in the world coordinate system, is the heading angle of the vehicle corresponding to the extended node, L is the wheelbase of the vehicle.

[0035] Optionally, before obtaining the smoothed nodes based on the determined pose errors corresponding to the nodes in the initial parking path, the method further includes: It is determined that the initial parking path satisfies a smoothing condition.

[0036] Optionally, the method further includes: When the initial parking path does not meet the smoothing condition, the initial parking path is used as the target parking path.

[0037] Optionally, when the number of nodes in the initial parking path is greater than a target number, the initial parking path meets the smoothing condition.

[0038] In a second aspect, an embodiment of the present application provides an electronic device comprising: at least one processor, and a memory, wherein the memory stores a computer program that can be run on the processor, wherein when the processor executes the computer program, the parking path processing method described in the first aspect of the embodiment is executed.

[0039] In a third aspect, an embodiment of the present application provides a non-volatile readable storage medium, wherein the non-volatile readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the parking path processing method described in the first aspect of the embodiment is executed.

[0040] In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the parking path processing method described in the first aspect of the embodiment.

[0041] In a fifth aspect, an embodiment of the present application provides a vehicle, which is used to execute the parking path processing method described in the first aspect of the embodiment.

[0042] Beneficial effects: The method provided in this embodiment obtains smoothed nodes based on the position errors corresponding to the nodes in the determined initial parking path, and obtains a smoothed target parking path based on the smoothed nodes.

[0043] By smoothing the nodes in the initial parking path according to their posture errors, the problem of uneven parking path caused by node errors in the initial parking path is reduced, thereby obtaining a smoother parking path. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0045] Figure 1 This is a flowchart of the steps of a parking path processing method proposed in one embodiment of the present application; Figure 2 This is a flowchart of the steps of a parking path processing method proposed in one embodiment of the present application; Figure 3 is an execution flow chart of a parking path processing method provided in one embodiment of the present application; Figure 4 This is a functional module diagram of a parking path processing device proposed in one embodiment of the present application; Figure 5 is a schematic diagram of an electronic device proposed in one embodiment of the present application; Figure 6 is a schematic diagram of a non-volatile readable storage medium provided in one embodiment of the present application; Figure 7 It is a schematic diagram of a computer program product proposed in one embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0047] With the popularization of intelligent electric vehicles, more and more vehicles are equipped with automatic parking technology. Path planning is one of the key technologies in automatic parking technology. The quality of path planning directly affects the user experience. If the path planning effect is good, the parking time will be short, the risk of collision will be reduced, and the user will have a higher sense of trust. If the path planning is not good, the parking time will be longer and the risk of collision will be greater, resulting in a poor user experience.

[0048] Currently, path planning for automated parking is primarily divided into two approaches: geometry-based and search-based. While the geometry-based approach offers a simple and easy-to-implement algorithm, it can only be used for simple parking spaces and is difficult to plan for more complex spaces, such as those on dead-end roads. Search-based path planning can be applied to complex parking scenarios. Among search-based approaches, A* (A*) is a direct search method with heuristics that is most effective for finding the shortest path in static road networks. However, A* does not consider the vehicle's kinematic characteristics, resulting in a planned path that does not conform to the actual vehicle trajectory. This is why Hybrid A* (Hybrid A*) was developed. Hybrid A* considers vehicle kinematics and plans a rough, collision-free trajectory. However, this trajectory may contain numerous grip points, resulting in sudden and large changes in speed, orientation, and steering angle at the path points. This can cause the vehicle to jerk, reducing parking comfort. Therefore, the planned trajectory also requires smoothing.

[0049] Currently, path smoothing algorithms mainly use graph-based curve generation methods, such as the Dubins curve. The Dubins curve mainly uses line segments plus circular arcs to calculate the shortest path between two postures in a plane, while also considering the vehicle's maximum turning rate. However, this method will cause curvature discontinuity.

[0050] In addition to graph-based curve generation methods, there are also polynomial interpolation methods, such as Lagrange polynomial interpolation. However, this method is computationally complex and prone to Runge's phenomenon. There are also optimization-based smoothing algorithms, such as the OBCA (Optimization-Based Collision Avoidance) algorithm. This algorithm establishes a model based on model predictive control and uses an optimization algorithm to solve it. However, the computational time complexity is high, the computing power requirements are high, and the calculation time is too long, making it unsuitable for deployment in actual vehicles.

[0051] Therefore, this embodiment proposes a parking path processing method, which can perform tracking and smoothing operations based on path planning, realize tracking and smoothing of path planning, obtain a smoother parking path, and thus improve the comfort during the parking process.

[0052] Reference Figure 1 , shows a flowchart of a parking path processing method in an embodiment of the present application, the method may specifically include the following steps: S101: Obtain smoothed nodes based on the determined posture errors corresponding to the nodes in the initial parking path.

[0053] When a user of a vehicle wants to park, an initial parking path may be first determined. The initial parking path is a path that needs to be smoothed. The initial parking path includes multiple nodes. Each node in the initial parking path includes an initial state value. The initial state value includes an initial x-axis coordinate, an initial y-axis coordinate, and an initial heading angle.

[0054] In a feasible implementation manner, when acquiring the initial parking path, the initial parking path may be determined in response to a smoothing process selection operation.

[0055] Specifically, after the parking path is initially planned, the user can select a path that needs to be smoothed as the initial parking path. For example, the entire parking path obtained by the initial plan can be used as the initial parking path, or a section of the parking path obtained by the initial plan can be selected as the initial parking path.

[0056] In a feasible implementation, the initial parking path is obtained, and the initial parking path may be determined by path planning according to the starting point and the end point of the parking process.

[0057] Specifically, when parking, the starting point and end point of the parking process selected by the user can be obtained, the end point represents the parking space or position, and the starting point can represent the current position of the vehicle.

[0058] During actual implementation, the user can also set the end point of the parking process on the vehicle's central control screen or in the parking control program on a mobile terminal such as a mobile phone. This is not limited in this embodiment, and the starting point of the parking process can also be set by the user. When the user does not set the starting point, the default starting point is the current position of the vehicle.

[0059] After obtaining the starting and end points of the parking process, the automatic parking technology on the vehicle will plan a path based on the starting and end points of the parking process to determine an initial parking path. The initial parking path is a collision-free path.

[0060] In this embodiment, the Hybrid A* algorithm may be used to determine the initial parking path based on the starting and ending points of the parking process. In actual implementation, other algorithms may also be used to generate a collision-free initial parking path, which is not limited in this embodiment.

[0061] Specifically, taking the Hybrid A* algorithm as an example to determine the initial parking path, since the Hybrid A* algorithm takes into account the vehicle's kinematic characteristics, it can plan a more effective collision-free parking path. However, the path planned by the Hybrid A* algorithm may also include a large number of grip points, resulting in sudden and large changes in the speed, direction, and steering angle of the path points, which may cause the vehicle to jerk. Therefore, when determining the initial parking path based on the Hybrid A* algorithm, in order to improve the comfort of the parking process, the initial parking path can be filtered to obtain a smoothed target parking path corresponding to the initial parking path.

[0062] In actual implementation, if the number of nodes in the initial parking path determined based on the Hybrid A* algorithm is small, the jerking phenomenon is not obvious. To save processing resources, smoothing can be omitted. When the number of nodes in the initial parking path is greater than the target number, the initial parking path can be smoothed.

[0063] For example, the target number can be set to 5, which can be set according to the needs of actual applications and is not limited in this embodiment.

[0064] When filtering the initial parking path, a Kalman filtering method may be used, such as using an extended Kalman filter to smooth the initial parking path and determine the smoothed nodes corresponding to each node in the initial parking path.

[0065] Specifically, an extended Kalman filter is used to calculate the posture error corresponding to the node according to the initial state value of the node in the initial parking path. The posture error may include a position error and / or a heading angle error. Then, during the update process of the extended Kalman filter, the process noise matrix in the extended Kalman filter is dynamically adjusted according to the position error and heading angle error corresponding to the node to obtain the node after smoothing.

[0066] S102: Obtaining a smoothed target parking path according to the smoothed nodes.

[0067] After smoothing each of the multiple nodes in the initial parking path, the smoothed nodes corresponding to the nodes in the initial parking path are combined to obtain a smoothed target parking path. Specifically, the smoothed nodes corresponding to the nodes are combined to form the target parking path based on the order of the time points corresponding to the nodes.

[0068] By smoothing the nodes in the initial parking path based on their posture errors, the problem of uneven parking path caused by node errors in the initial parking path is reduced, resulting in a smoother parking path. This not only reduces parking jerks and improves parking comfort, but also allows the smoothing process to converge quickly by introducing the posture errors of the vehicle at the nodes in the path planning, such as position error and heading angle error, to smooth the initial parking path.

[0069] Reference Figure 2 , shows a flowchart of a parking path processing method in an embodiment of the present application. In a feasible implementation, the method may specifically include the following steps: S201: Determine an initial parking path according to the starting point and end point of the parking process.

[0070] In this embodiment, the initial parking path is determined according to the starting point and the end point selected by the user during the parking process.

[0071] Specifically, after the starting point and the end point of the parking process are obtained, it can be first determined whether the obtained starting point and the end point satisfy the parsing extension.

[0072] Specifically, analytic expansions (or one-shot heuristics) refer to directly solving a fixed-end optimal control problem between a certain node and the target point during the search process. If a collision-free and dynamically consistent path can be found, the planning can be terminated early.

[0073] If the start and end points set by the user can be expanded analytically, that is, there is no need to determine the initial parking path based on the Hybrid A* algorithm, and when the start and end points meet the analytical extension requirements, the path between the start and end points is smooth and no additional smoothing is required.

[0074] Therefore, when the starting point and the end point satisfy the analytical expansion, the initial parking path obtained by the analytical expansion can be used as the target parking path.

[0075] When the starting and ending points do not satisfy analytical extension, the initial parking path can be determined based on the Hybrid A* algorithm. The Hybrid A* algorithm takes into account scenarios with vehicle kinematic constraints (such as autonomous driving and parking), combines the heuristic search of the traditional A algorithm with the vehicle dynamics model, and by discretizing the continuous space and introducing analytical extension technology, it can generate a feasible path that better conforms to the vehicle's motion characteristics.

[0076] Specifically, when the starting point and the end point do not satisfy the parsing extension, the starting point can be added to the node open list, i.e., the open set. Then, a cyclic expansion strategy is executed until the node open list is empty or a node that satisfies the parsing extension with the end point exists, and the initial parking path is output.

[0077] In one feasible implementation, executing the cyclic extension strategy includes the following steps: A1: Determine the current node with the smallest path loss in the open node list.

[0078] The open set stores candidate nodes to be processed, and all nodes that may be explored can be sorted by path loss.

[0079] The path loss of a node is the sum of the path loss from the starting point to the node and the path loss from the node to the end point. In actual implementation, the Euclidean distance can be used to calculate the path loss from the starting point to the node or from the node to the end point.

[0080] In the actual implementation process, the path loss of a node in the node open list can also be calculated based on other cost functions, which is not limited in this embodiment; lower path loss means faster path planning convergence speed and better path selection. By reasonably setting the cost function, it can better adapt to different environments and needs, such as considering obstacles and terrain factors.

[0081] A2: Determine whether the current node and the end point satisfy the parsing extension.

[0082] After determining the current node with the smallest path loss in the node open list, determine whether the current node with the smallest path loss and the end point satisfy the analytical extension. In other words, determine whether a fixed-end optimal control problem can be directly solved between the current node and the end point. If the current node and the end point satisfy the analytical extension, the path planning process of the Hybrid A* algorithm can be terminated early to determine the initial parking path.

[0083] A3: When the current node and the end point do not satisfy the parsing extension, the current node is added to the node closing list, and an extension node is determined based on the differential model, and a collision detection is performed on the extension node.

[0084] If the current node and the end point do not meet the requirements for parsing and expansion, path planning needs to continue. First, the current node is added to the node close list, i.e., the close set. The close set is used to record processed nodes to prevent repeated traversal of a node and repeated expansion of the same state, thereby reducing invalid calculations.

[0085] Then, the expansion node is determined based on the differential model. For example, the formula for determining the expansion node based on the differential model is:

[0086] in, d is the average driving distance of the vehicle, 、 are the distances traveled by the left and right wheels of the vehicle respectively, , is the position coordinate of the vehicle corresponding to the extended node in the world coordinate system, is the heading angle of the vehicle corresponding to the extended node, L is the wheelbase of the vehicle.

[0087] The distance traveled by the left and right wheels of the vehicle and It is related to the parameter setting in path planning. It is a preset value that can be set according to the actual application needs. The distance traveled by the left and right wheels and After determination, the average vehicle driving distance can be further calculated d .

[0088] After the expansion node is determined, collision detection is also required for the expansion node, that is, detecting whether the expansion node collides with obstacles in the map.

[0089] When the extended node fails the collision detection, the extended node is discarded and the cyclic extension strategy is continued.

[0090] A4: When the extended node passes the collision detection, the path loss corresponding to the extended node is calculated, and the extended node is added to the node open list.

[0091] When the extended node passes the collision detection, the path loss between the starting point to the extended node and the path loss between the extended node and the end point can be calculated as the path loss corresponding to the extended node by calculating the Euclidean distance, and the extended node is added to the open set to continue executing the cyclic extension strategy.

[0092] The initial parking path is output until the node open set is empty or a node that satisfies the parsing extension of the end point exists. The initial state value of each node in the initial parking path obtained based on the Hybrid A* algorithm includes the initial x-axis coordinate, initial y-axis coordinate, and initial heading angle.

[0093] S202: When the initial parking path satisfies a smoothing condition, a predicted state value of the node is determined according to the initial state value of the node by using an extended Kalman filter and a preset kinematic model.

[0094] When the initial parking path satisfies the smoothing condition, smoothed nodes may be obtained based on position errors and heading angle errors corresponding to the nodes in the initial parking path.

[0095] For example, after the path planning process based on the Hybrid A* algorithm outputs an initial parking path, it can be first determined whether the initial parking path meets the smoothing processing condition.

[0096] When the number of nodes in the initial parking path is greater than the target number, the initial parking path satisfies the smoothing condition, and the initial parking path may be smoothed.

[0097] If the number of nodes in the initial parking path is less than or equal to the target number, it indicates that the initial parking path has a small number of nodes, and the vehicle will not experience significant jerking when parking according to the initial parking path. Therefore, to conserve processing resources, if the initial parking path does not meet the smoothing conditions, that is, if the number of nodes in the initial parking path is less than or equal to the target number, the initial parking path may be used as the target parking path without smoothing.

[0098] The target number can be set according to the actual application requirements, for example, the target number can be set to 5.

[0099] Through the extended Kalman filter based on the kinematic model, the vehicle kinematic characteristics can be fully considered, so that the smoothed path will not have sudden changes in curvature.

[0100] Extended Kalman Filter (EKF) is a state estimation method for nonlinear systems. Its core steps are divided into two processes: prediction and update.

[0101] In the prediction process of the extended Kalman filter, the predicted state value of the node in the initial parking path is first determined based on the initial state value of the node and the kinematic model. The predicted state value includes the predicted x-axis coordinate, the predicted y-axis coordinate and the predicted heading angle.

[0102] Specifically, the kinematic model is:

[0103] in, for k The predicted state value at the moment; for kThe predicted x-axis coordinate of the predicted state value at the moment; for k The predicted y-axis coordinate of the predicted state value at the moment; for k The predicted heading angle in the predicted state value at the moment; for k The x-axis coordinate of the updated state value; for k The y-axis coordinate of the state value after the moment is updated; for k The heading angle in the updated state value; for k Vehicle speed at the time; for k The wheel steering angle at the moment; L is the wheelbase of the vehicle.

[0104] S203: Determine the position error and heading angle error corresponding to the node according to the initial state value and the predicted state value of the node.

[0105] Specifically, in this embodiment, the posture error includes a position error and a heading angle error, that is, the position error and heading angle error corresponding to the node are determined based on the initial state value and the predicted state value of the node.

[0106] For example, according to the initial state value and the predicted state value of the node, the formula for determining the position error and heading angle error corresponding to the node is:

[0107] in, and is the position error; is the heading angle error; is the initial x-axis coordinate in the initial state value; is the initial y-axis coordinate in the initial state value; is the initial heading angle in the initial state value; x is the predicted x-axis coordinate in the predicted state value; y is the predicted y-axis coordinate in the predicted state value; is the predicted heading angle in the predicted state value.

[0108] After determining the position error and heading angle error corresponding to the node in the initial parking path, the process noise matrix in the extended Kalman filter can be adjusted according to the position error and heading angle error corresponding to the node in the initial parking path, and the node corresponds to the smoothed node.

[0109] S204: Calculating a first scaling factor corresponding to a node in the initial parking path for adjusting a process noise matrix in an extended Kalman filter according to the position error and the heading angle error corresponding to the node.

[0110] Specifically, based on the posture error corresponding to the node in the initial parking path, a first scaling factor corresponding to the node for adjusting the process noise matrix in the extended Kalman filter can be calculated, and the smoothed node corresponding to the node can be obtained based on at least the first scaling factor.

[0111] When the posture error includes position error and heading angle error, the formula for calculating the first scaling factor corresponding to the process noise matrix corresponding to the node is as follows based on the position error and heading angle error of the node:

[0112] in, is the first scaling factor corresponding to the process noise matrix; , , and m+n=1; and is the position error; is the heading angle error; is a cardinality close to 0; is the preset amplification factor, and a can be set according to the needs of actual application.

[0113] It can be seen from the calculation formula of the first scaling factor that if the position error and heading angle error between the initial state value and the predicted state value are larger, then If the position error and heading angle error between the initial state value and the predicted state value are smaller, then The bigger it will be.

[0114] S205: Calculate an estimated value of a state covariance matrix corresponding to the node according to the process noise matrix corresponding to the node and the first scaling factor.

[0115] After determining the first scaling factor according to the posture error of the node, the process noise matrix in the extended Kalman filter can be scaled by the first scaling factor, that is, the process noise of the smoothing process is dynamically adjusted according to the posture error of the node.

[0116] Specifically, according to the process noise matrix corresponding to the node and the first scaling factor, the formula for calculating the estimated value of the state covariance matrix corresponding to the node is:

[0117] in, fork The estimated value of the state covariance matrix corresponding to the moment; for k The state transfer matrix corresponding to the moment; for k -1 moment corresponding to the state covariance matrix; is the first scaling factor corresponding to the process noise matrix; Q is the process noise matrix.

[0118] State transfer matrix Solving the Jacobian matrix yields:

[0119] Process noise matrix Q For a 3x3 matrix:

[0120] State covariance matrix For a 3x3 matrix:

[0121] Wherein, subscript 1 represents position x; subscript 2 represents position y; and subscript 3 represents heading angle.

[0122] If the position error and heading angle error between the initial state value and the predicted state value are larger, then The smaller the process noise matrix Q The smaller it is, the updated state value of the node on the parking path is closer to the initial state value.

[0123] If the position error and heading angle error between the initial state value and the predicted state value are smaller, then The larger the process noise matrix Q The larger the value, the updated state value of the node on the parking path tends to be the predicted state value obtained according to the kinematic model.

[0124] S206: Determine a second scaling factor for adjusting a measurement noise matrix in an extended Kalman filter according to the wheel steering angle corresponding to the node.

[0125] During the smoothing process, in addition to determining a first scaling factor to adjust the process noise matrix based on the position error and heading angle error of the node, a second scaling factor for adjusting the measurement noise matrix in the extended Kalman filter can also be determined based on the wheel steering angle corresponding to the node. At least based on the second scaling factor, the smoothed node corresponding to the node, for example, can be obtained by the combined action of the first scaling factor and the second scaling factor to obtain the target parking path after smoothing.

[0126] Specifically, according to the wheel steering angle corresponding to the node, a formula for determining the second scaling factor for adjusting the measurement noise matrix in the extended Kalman filter is:

[0127] in, is the second scaling factor; b is the preset zoom factor; for k The wheel steering angle at time .

[0128] Second scaling factor Inversely proportional to the node's wheel steering angle, the preset scaling factor b It is related to actual vehicle debugging in actual application. Generally b The value range can be 0.1-1.

[0129] The Hybrid A* algorithm supports multi-directional expansion, so the Hybrid A* algorithm can support forward, reverse and different steering angle combinations, such as left turn (+30°), right turn (-30°), and straight ahead (0°). As a result, the wheel steering angles corresponding to the nodes in the initial parking path may be different. If the wheel steering angle corresponding to a node is larger, it indicates that the vehicle is turning at the node. In this case, the second scaling factor The smaller it is.

[0130] S207: updating the state value of the node according to the estimated value of the state covariance matrix corresponding to the node, and obtaining a smoothed node corresponding to the node.

[0131] Specifically, the process of updating the state value of the node according to the estimated value of the state covariance matrix corresponding to the node to obtain the smoothed node corresponding to the node includes the following process: Calculating a gain matrix corresponding to the node according to an estimated value of a state covariance matrix corresponding to the node, the second scaling factor, and the measurement noise matrix; According to the gain matrix and the predicted state value corresponding to the node, the state of the node is updated to obtain a smoothed node corresponding to the node.

[0132] For example, based on the estimated value of the state covariance matrix corresponding to the node, the calculation formula for updating the state value of the node is:

[0133] in, for k The gain matrix at time t, fork The updated status value at each moment; for k The estimated value of the state covariance matrix corresponding to the moment; for k The measurement matrix corresponding to the moment, R is the measurement noise matrix; is the second scaling factor; for k The predicted state value at the moment; for k The state covariance matrix after the moment update; for k The initial state value at the moment.

[0134] The larger the value of the wheel steering angle corresponding to a node is, the more it indicates that the vehicle is turning at the node. In this case, the second scaling factor The smaller it is, the smaller the measurement noise matrix R The smaller it is, the more the updated state value of the node tends to be closer to the predicted state value obtained according to the kinematic model.

[0135] S208: Obtaining a smoothed target parking path according to the smoothed nodes.

[0136] Specifically, after smoothing multiple nodes in the initial parking path, the smoothed nodes corresponding to the nodes are combined into the target parking path according to the sequence of time points corresponding to the nodes.

[0137] Reference Figure 3 , shows an execution flow chart of the parking path processing method provided by this embodiment, the method includes the following processes: First, after obtaining the starting point and end point of the parking process, the starting point is added to the open list of nodes, namely the open set.

[0138] Then, determine whether the starting point and the end point meet the parsing extension.

[0139] If the starting point and the end point satisfy the analytical extension, the initial parking path obtained by the analytical extension is used as the target parking path; If the starting point and the end point do not satisfy the parsing expansion, the execution process of the loop expansion strategy is entered.

[0140] In the execution process of the cyclic expansion strategy: First, traverse the open set and select the node with the smallest path loss as the current node; Next, determine whether the current node and the end point meet the parsing extension.

[0141] If the current node and the end point do not satisfy the parsing extension, the current node is added to the node closing list, i.e., closeset; Then, the expansion nodes are determined based on the differential model; Next, collision detection is performed on the extended nodes; When the extended node fails the collision detection, the extended node is discarded and the cyclic expansion strategy continues; When the extended node passes the collision detection, the path loss corresponding to the extended node is calculated, the extended node is added to the open set, and the cyclic extension strategy is continued; Until the current node and the end point meet the analytical extension, the initial parking path is output and the execution process of the loop expansion strategy ends.

[0142] After determining the initial parking path, first determine whether the number of nodes in the initial parking path is greater than the target number; When the number of nodes in the initial parking path is less than or equal to the target number, the initial parking path is used as the target parking path.

[0143] When the number of nodes in the initial parking path is greater than the target number, the initial parking path is smoothed based on an extended Kalman filter to output a target parking path.

[0144] This method has at least the following beneficial effects: 1. This method not only smoothes the nodes in the original rough initial parking path based on their pose errors, reducing the unevenness of the initial parking path caused by node errors and obtaining a smoother target parking path, thereby reducing parking jerks and improving parking comfort, but also ensures rapid convergence without distortion when there is an error between the initial state value and the predicted state value during the filter update process, thus ensuring a smooth final parking path. 2. During the filtering update process of the extended Kalman filter, pose errors such as position error and heading angle error are introduced to scale the process noise in the filtering process, that is, the scaled process noise matrix is calculated based on the pose error Q A first scaling factor is set such that when the error between the initial state value of the node in the initial parking path and the predicted state value obtained by the kinematic model increases, the state value of the node updated by the extended Kalman filter is more biased towards the initial state value; and when the error decreases, the state value of the node updated by the extended Kalman filter is more biased towards the predicted state value; 3. In the filter update process of the extended Kalman filter, the vehicle steering angle is also introduced to determine the scaling measurement noise, that is, the scaling measurement noise matrix RThe second scaling factor of the wheel steering angle corresponding to a node is larger, and the second scaling factor is smaller. At this time, the measurement noise matrix R The smaller it is, the more the updated state value of the node tends to be closer to the predicted state value obtained according to the kinematic model.

[0145] 4. By increasing the first scaling factor of process noise and the second scaling factor of measurement noise in the dynamic scaling filtering process, the extended Kalman filter can adaptively adjust the process noise matrix when a rough initial parking path is given. Q and the measurement noise matrix R The matrix is used to smooth the initial parking path to obtain a smoother target parking path.

[0146] 5. The extended Kalman filter makes predictions based on a kinematic model, fully considering the vehicle's kinematic characteristics, ensuring that the smoothed target parking path does not experience sudden changes in curvature. 6. This method has a small delay and low computational cost.

[0147] Reference Figure 4 , shows a functional module diagram of a parking path processing device provided in an embodiment of the present application, the device comprising: The node smoothing processing module 100 is used to obtain smoothed nodes according to the posture errors corresponding to the nodes in the determined initial parking path; The target parking path determining module 200 is configured to obtain a smoothed target parking path according to the smoothed nodes.

[0148] Optionally, the device further comprises: A prediction module, configured to determine a predicted state value of the node according to the initial state value of the node by using an extended Kalman filter and a preset kinematic model; The error calculation module is used to determine the posture error corresponding to the node based on the initial state value and the predicted state value of the node.

[0149] Optionally, the error calculation module is further configured to: Determine the position error and / or heading angle error corresponding to the node according to the initial state value and the predicted state value of the node.

[0150] Optionally, the posture error includes a position error and a heading angle error, and the formula for determining the position error and heading angle error corresponding to the node is:

[0151] in, and is the position error; is the heading angle error; is the initial x-axis coordinate in the initial state value; is the initial y-axis coordinate in the initial state value; is the initial heading angle in the initial state value; x is the predicted x-axis coordinate in the predicted state value; y is the predicted y-axis coordinate in the predicted state value; is the predicted heading angle in the predicted state value.

[0152] Optionally, the kinematic model is:

[0153] in, for k The predicted state value at the moment; for k The predicted x-axis coordinate of the predicted state value at the moment; for k The predicted y-axis coordinate of the predicted state value at the moment; for k The predicted heading angle in the predicted state value at the moment; for k The x-axis coordinate of the updated state value; for k The y-axis coordinate of the state value after the moment is updated; for k The heading angle in the updated state value; for k Vehicle speed at the time; for k The wheel steering angle at the moment; L is the wheelbase of the vehicle.

[0154] Optionally, the node smoothing processing module includes: The node smoothing processing unit is used to adjust the process noise matrix in the extended Kalman filter according to the posture error corresponding to the node in the initial parking path to obtain a smoothed node corresponding to the node.

[0155] Optionally, the node smoothing processing unit includes: a first scaling factor calculation unit, configured to calculate, based on a posture error corresponding to a node in the initial parking path, a first scaling factor corresponding to the node for adjusting a process noise matrix in an extended Kalman filter; The node smoothing subunit is configured to obtain a smoothed node corresponding to the node at least according to the first scaling factor.

[0156] Optionally, the posture error includes a position error and a heading angle error. According to the posture error corresponding to a node in the initial parking path, a formula for calculating a first scaling factor corresponding to the node for adjusting a process noise matrix in an extended Kalman filter is:

[0157] in, is the first scaling factor corresponding to the process noise matrix; , , and m+n=1; and is the position error; is the heading angle error; is a cardinality close to 0; is the preset magnification factor.

[0158] Optionally, the device further comprises: a second scaling factor calculation unit, configured to determine, according to the wheel steering angle corresponding to the node, a second scaling factor for adjusting a measurement noise matrix in the extended Kalman filter; The node smoothing subunit is configured to obtain a smoothed node corresponding to the node at least according to the second scaling factor.

[0159] Optionally, according to the wheel steering angle corresponding to the node, a formula for determining a second scaling factor for adjusting the measurement noise matrix in the extended Kalman filter is:

[0160] in, is the second scaling factor; b is the preset zoom factor; for k The wheel steering angle at time .

[0161] Optionally, the node smoothing processing subunit is used to: a calculation subunit, configured to calculate an estimated value of a state covariance matrix corresponding to the node based on a process noise matrix corresponding to the node and the first scaling factor; The updating subunit is used to update the state value of the node according to the estimated value of the state covariance matrix corresponding to the node, and obtain the smoothed node corresponding to the node.

[0162] Optionally, according to the process noise matrix corresponding to the node and the first scaling factor, a formula for calculating an estimated value of the state covariance matrix corresponding to the node is:

[0163] in, for k The estimated value of the state covariance matrix corresponding to the moment; for k The state transfer matrix corresponding to the moment; for k -1 moment corresponding to the state covariance matrix; is the first scaling factor corresponding to the process noise matrix; Q is the process noise matrix.

[0164] Optionally, the updating subunit includes: a gain matrix calculation subunit, configured to calculate a gain matrix corresponding to the node based on an estimated value of a state covariance matrix corresponding to the node, the second scaling factor, and the measurement noise matrix; The state updating subunit is used to update the state of the node according to the gain matrix and the predicted state value corresponding to the node, and obtain the smoothed node corresponding to the node.

[0165] Optionally, based on the estimated value of the state covariance matrix corresponding to the node, the calculation formula for updating the state value of the node is:

[0166] in, for k The gain matrix at time t, for k The updated status value at each moment; for k The estimated value of the state covariance matrix corresponding to the moment; for k The measurement matrix corresponding to the moment, R is the measurement noise matrix; is the second scaling factor; for k The predicted state value at the moment; for k The state covariance matrix after the moment update; for k The initial state value at the moment.

[0167] Optionally, the target parking path determination module includes: The target parking path determining unit is configured to obtain a smoothed target parking path by combining the smoothed nodes corresponding to the nodes in the initial parking path.

[0168] Optionally, the device further comprises: The initial parking path selection module is configured to determine the initial parking path in response to the smoothing selection operation.

[0169] Optionally, the device further comprises: The initial parking path determination module is used to determine the initial parking path according to the starting point and end point of the parking process.

[0170] Optionally, the initial parking path determination module includes: The first initial parking path determining unit is configured to use the initial parking path obtained by parsing and expanding as the target parking path when the starting point and the end point satisfy parsing and expanding conditions.

[0171] Optionally, the initial parking path determination module includes: The second initial parking path determining unit is configured to determine the initial parking path based on a Hybrid A* algorithm when the starting point and the end point do not satisfy analytical extension.

[0172] Optionally, the initial parking path determination module includes: The third initial parking path determining unit is configured to, when the start point and the end point do not satisfy the parsing extension, add the start point to an open node list and execute a cyclic expansion strategy until the open node list is empty or a node that satisfies the parsing extension with the end point exists, and then output an initial parking path.

[0173] Optionally, the third initial parking path determining unit is configured to: In the node open list, if the current node with the smallest path loss and the end point do not meet the requirements for parsing extension, the current node is added to the node closed list, and an extension node is determined based on a differential model; When the extended node passes the collision detection, the path loss corresponding to the extended node is calculated, and the extended node is added to the node open list.

[0174] Optionally, the third initial parking path determination unit is further configured to: When the extended node fails the collision detection, the extended node is discarded.

[0175] Optionally, the formula for determining the extended node based on the differential model is:

[0176] in, d is the average driving distance of the vehicle, 、 are the distances traveled by the left and right wheels of the vehicle respectively, , is the position coordinate of the vehicle corresponding to the extended node in the world coordinate system, is the heading angle of the vehicle corresponding to the extended node, L is the wheelbase of the vehicle.

[0177] Optionally, the device further comprises: The smoothing condition judgment unit is configured to determine whether the initial parking path satisfies a smoothing condition.

[0178] Optionally, the device further comprises: The path determining unit is configured to use the initial parking path as a target parking path when the initial parking path does not meet a smoothing condition.

[0179] Optionally, when the number of nodes in the initial parking path is greater than a target number, the initial parking path meets the smoothing condition.

[0180] Reference Figure 5 , shows a schematic diagram of an electronic device provided in an embodiment of the present application, comprising: at least one processor, and a memory, wherein the memory stores a computer program that can be run on the processor, wherein the processor executes the parking path processing method described in the embodiment when executing the computer program.

[0181] Reference Figure 6 , shows a schematic diagram of a non-volatile readable storage medium provided in an embodiment of the present application, wherein the non-volatile readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the parking path processing method described in the embodiment is executed.

[0182] Reference Figure 7 , shows a schematic diagram of a computer program product provided in an embodiment of the present application, including a computer program / instruction, which, when executed by a processor, implements the parking path processing method described in the embodiment.

[0183] An embodiment of the present application further provides a vehicle, which is used to execute the parking path processing method described in the embodiment.

[0184] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0185] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, apparatuses, or computer program products. Therefore, the embodiments of the present application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware. Furthermore, the embodiments of the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0186] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0187] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0188] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0189] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0190] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0191] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A parking path processing method, characterized in that: The method comprises: Obtain smoothed nodes based on the posture errors corresponding to the nodes in the determined initial parking path; A smoothed target parking path is obtained according to the smoothed nodes.

2. The method according to claim 1, characterized in that The method further comprises: Determining a predicted state value of the node according to the initial state value of the node by using an extended Kalman filter and a preset kinematic model; Determine a posture error corresponding to the node based on the initial state value and the predicted state value of the node.

3. The method according to claim 2, characterized in that Determining a posture error corresponding to the node according to the initial state value and the predicted state value of the node includes: Determine the position error and / or heading angle error corresponding to the node according to the initial state value and the predicted state value of the node.

4. The method according to claim 3, characterized in that The posture error includes position error and heading angle error. According to the initial state value and the predicted state value of the node, the formula for determining the position error and heading angle error corresponding to the node is: in, and is the position error; is the heading angle error; is the initial x-axis coordinate in the initial state value; is the initial y-axis coordinate in the initial state value; is the initial heading angle in the initial state value; x is the predicted x-axis coordinate in the predicted state value; y is the predicted y-axis coordinate in the predicted state value; is the predicted heading angle in the predicted state value.

5. The method according to claim 2, characterized in that The kinematic model is: in, for k The predicted state value at the moment; for k The predicted x-axis coordinate of the predicted state value at the moment; for k The predicted y-axis coordinate of the predicted state value at the moment; for k The predicted heading angle in the predicted state value at the moment; for k The x-axis coordinate of the updated state value; for k The y-axis coordinate of the state value after the moment is updated; for k The heading angle in the updated state value; for k Vehicle speed at the time; for k The wheel steering angle at the moment; L is the wheelbase of the vehicle.

6. The method according to claim 1, characterized in that According to the determined pose errors corresponding to the nodes in the initial parking path, the smoothed nodes are obtained, including: According to the posture errors corresponding to the nodes in the initial parking path, the process noise matrix in the extended Kalman filter is adjusted to obtain smoothed nodes corresponding to the nodes.

7. The method according to claim 6, characterized in that According to the posture error corresponding to the node in the initial parking path, the process noise matrix in the extended Kalman filter is adjusted to obtain a smoothed node corresponding to the node, including: According to the posture error corresponding to the node in the initial parking path, a first scaling factor corresponding to the node for adjusting the process noise matrix in the extended Kalman filter is calculated, and a smoothed node corresponding to the node is obtained at least based on the first scaling factor.

8. The method according to claim 7, characterized in that The pose error includes position error and heading angle error. According to the pose error corresponding to the node in the initial parking path, the formula for calculating the first scaling factor corresponding to the node for adjusting the process noise matrix in the extended Kalman filter is: in, is the first scaling factor corresponding to the process noise matrix; , , and m+n=1; and is the position error; is the heading angle error; is a cardinality close to 0; is the preset magnification factor.

9. The method according to claim 7, characterized in that The method further comprises: A second scaling factor for adjusting a measurement noise matrix in the extended Kalman filter is determined according to a wheel steering angle corresponding to the node, and a smoothed node corresponding to the node is obtained at least according to the second scaling factor.

10. The method according to claim 9, characterized in that According to the wheel steering angle corresponding to the node, a formula for determining a second scaling factor for adjusting the measurement noise matrix in the extended Kalman filter is: in, is the second scaling factor; b is the preset zoom factor; for k The wheel steering angle at time .

11. The method according to claim 9, characterized in that Obtaining a smoothed node corresponding to the node includes: Calculating an estimated value of a state covariance matrix corresponding to the node according to a process noise matrix corresponding to the node and the first scaling factor; The state value of the node is updated according to the estimated value of the state covariance matrix corresponding to the node to obtain the smoothed node corresponding to the node.

12. The method according to claim 11, characterized in that According to the process noise matrix corresponding to the node and the first scaling factor, the formula for calculating the estimated value of the state covariance matrix corresponding to the node is: in, for k The estimated value of the state covariance matrix corresponding to the moment; for k The state transfer matrix corresponding to the moment; for k -1 moment corresponding to the state covariance matrix; is the first scaling factor corresponding to the process noise matrix; Q is the process noise matrix.

13. The method according to claim 11, characterized in that Updating the state value of the node according to the estimated value of the state covariance matrix corresponding to the node to obtain a smoothed node corresponding to the node, including: Calculating a gain matrix corresponding to the node according to an estimated value of a state covariance matrix corresponding to the node, the second scaling factor, and the measurement noise matrix; According to the gain matrix and the predicted state value corresponding to the node, the state of the node is updated to obtain a smoothed node corresponding to the node.

14. The method according to claim 11, characterized in that According to the estimated value of the state covariance matrix corresponding to the node, the calculation formula for updating the state value of the node is: in, for k The gain matrix at time t, for k The updated status value at each moment; for k The estimated value of the state covariance matrix corresponding to the moment; for k The measurement matrix corresponding to the moment, R is the measurement noise matrix; is the second scaling factor; for k The predicted state value at the moment; for k The state covariance matrix after the moment update; for k The initial state value at the moment.

15. The method according to any one of claims 1 to 14, characterized in that Obtaining a smoothed target parking path based on the smoothed nodes includes: The smoothed target parking path is obtained by combining the smoothed nodes corresponding to the nodes in the initial parking path.

16. The method according to any one of claims 1 to 14, characterized in that The method further comprises: In response to the smoothing selection operation, the initial parking path is determined.

17. The method according to any one of claims 1 to 14, characterized in that The method further comprises: The initial parking path is determined based on the starting and ending points of the parking process.

18. The method according to claim 17, characterized in that Based on the starting and ending points of the parking process, determine the initial parking path, including: When the starting point and the end point satisfy the analytical extension, the initial parking path obtained by the analytical extension is used as the target parking path.

19. The method according to claim 17, wherein Based on the starting and ending points of the parking process, determine the initial parking path, including: When the starting point and the end point do not satisfy analytical extension, the initial parking path is determined based on a Hybrid A* algorithm.

20. The method according to claim 17, wherein The method further comprises: When the starting point and the end point do not satisfy the parsing extension, the starting point is added to the node open list and a cyclic expansion strategy is executed until the node open list is empty or a node that satisfies the parsing extension with the end point exists, and then an initial parking path is output.

21. The method according to claim 20, characterized in that Execute a cyclic expansion strategy, including: In the node open list, if the current node with the smallest path loss and the end point do not meet the requirements for parsing extension, the current node is added to the node closed list, and an extension node is determined based on a differential model; When the extended node passes the collision detection, the path loss corresponding to the extended node is calculated, and the extended node is added to the node open list.

22. The method according to claim 21, characterized in that Execute the cyclic expansion strategy, which also includes: When the extended node fails the collision detection, the extended node is discarded.

23. The method according to claim 21, characterized in that The formula for determining the expansion node based on the differential model is: in, d is the average driving distance of the vehicle, 、 are the distances traveled by the left and right wheels of the vehicle respectively, , is the position coordinate of the vehicle corresponding to the extended node in the world coordinate system, is the heading angle of the vehicle corresponding to the extended node, L is the wheelbase of the vehicle.

24. The method according to claim 1, wherein Before obtaining smoothed nodes based on the determined posture errors corresponding to the nodes in the initial parking path, the method further includes: It is determined that the initial parking path satisfies a smoothing condition.

25. The method according to claim 24, characterized in that The method further comprises: When the initial parking path does not meet the smoothing condition, the initial parking path is used as the target parking path.

26. The method according to claim 25, characterized in that When the number of nodes in the initial parking path is greater than a target number, the initial parking path satisfies the smoothing condition.

27. An electronic device, characterized in that: include: At least one processor and a memory, wherein the memory stores a computer program that can be run on the processor, wherein the processor executes the parking path processing method according to any one of claims 1 to 26 when executing the computer program.

28. A non-volatile readable storage medium, characterized in that The non-volatile readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the parking path processing method according to any one of claims 1 to 26 is executed.

29. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the parking path processing method according to any one of claims 1 to 26 is implemented.

30. A vehicle, characterized in that: The vehicle is used to execute the parking path processing method according to any one of claims 1 to 26.

Images