An automatic parking path planning method, an automatic parking method and related devices

Abstract

The present application discloses a path planning method for automatic parking, an automatic parking method and related devices. The path planning method includes obtaining parking space information of a target parking space, and determining a parking-in section and an adjustment section based on the parking space information; determining a first planned path corresponding to the parking-in section through a hybrid A* algorithm; determining a second planned path corresponding to the adjustment section by using a straight-in driving method; and connecting the first planned path and the second planned path to form an automatic parking path. By dividing the parking path into a parking-in section and an adjustment section, the present application uses a hybrid A* algorithm for path planning in the parking-in section and a straight-in driving method for path planning in the adjustment section, and adjusts the vehicle attitude error generated by the parking-in section through the adjustment section, reducing the requirement for the control accuracy when the vehicle parks in, and can well ensure the final attitude of the vehicle.

Description

Technical Field

[0001] The present application relates to the technical field of automatic parking, and particularly relates to a path planning method for automatic parking, an automatic parking method, and related devices. Background Art

[0002] In recent years, with the rapid development of the domestic automotive industry, automatic parking technology has become an essential function of vehicles. Automatic parking technology is to plan a feasible parking path based on the vehicle pose and the parking space pose, and then automatically control the vehicle to follow the parking path to complete parking. Among them, parking path planning is a key link in automatic parking, and its basic requirement is that the planned path is feasible and collision-free.

[0003] Existing parking path planning methods generally adopt a planning method of geometric curve splicing and a planning method of graph search (for example, the hybrid A* search method). Among them, the planning method of geometric curve splicing has very high requirements for the working accuracy of sensors and actuators, and it is difficult to compensate for the dynamic errors of system operation; for the planning method of graph search, it is difficult to guarantee the search efficiency and the optimality of the path in a narrow and tortuous search space.

[0004] Therefore, the existing technology still needs to be improved. Summary of the Invention

[0005] The technical problem to be solved by the present application is to provide a path planning method for automatic parking, an automatic parking method, and related devices in view of the deficiencies of the existing technology.

[0006] To solve the above technical problem, in the first aspect of the embodiments of the present application, a path planning method for automatic parking is provided, and the method includes:

[0007] Obtain the parking space information of the target parking space, and determine the parking-in section and the adjustment section based on the parking space information;

[0008] Determine the first planning path corresponding to the parking-in section through the hybrid A* algorithm;

[0009] Adopt a straight-line driving method to determine the second planning path corresponding to the adjustment section;

[0010] Connect the first planning path and the second planning path to form an automatic parking path.

[0011] For the path planning method for automatic parking, wherein the step of determining the first planning path corresponding to the parking-in section through the hybrid A* algorithm specifically includes:

[0012] Take the starting position of the vehicle as the initial path point;

[0013] Obtain the distance between the initial path point and the target parking space, and determine the heuristic function weight corresponding to the initial path point according to the distance;

[0014] Based on the heuristic function weight, search for the next path node of the initial path point within the search area corresponding to the vehicle through the hybrid A* algorithm;

[0015] Take the next path node as the initial path point, and continue to execute the step of obtaining the distance between the initial path point and the target parking space until the termination point of the parking section to form a first planned path.

[0016] The path planning method for automatic parking, wherein the determination process of the search area specifically includes:

[0017] Based on the parking space information, determine the initial obstacle area and the initial parking space area corresponding to the target vehicle, expand the initial obstacle area outward to obtain the target obstacle area, and expand the initial parking space area inward to obtain the target parking space area;

[0018] Based on the target obstacle area and the target parking space area, determine the search area.

[0019] The path planning method for automatic parking, wherein the specific process of determining the heuristic function weight corresponding to the initial path point according to the distance specifically includes:

[0020] Compare the distance with a first distance threshold and a second distance threshold respectively;

[0021] If the distance is greater than the first distance threshold, set the first preset weight as the heuristic function weight corresponding to the initial path point;

[0022] If the distance is less than or equal to the first distance threshold and greater than or equal to the second distance threshold, set the second preset weight as the heuristic function weight corresponding to the initial path point;

[0023] If the distance is less than the second distance threshold, set the third preset weight as the heuristic function weight corresponding to the initial path point, wherein the first preset weight is greater than the second preset weight, and the second preset weight is greater than the third preset weight.

[0024] The path planning method for automatic parking, wherein the specific process of determining the parking section and the adjustment section based on the parking space information specifically includes:

[0025] Determine the parking entry point of the target parking space based on the above-mentioned parking space information, and select a target point as the dividing point in front of the parking entry point, where the dividing point and the parking entry point are on the same straight line, and the direction from the parking entry point to the dividing point is the front direction of the vehicle;

[0026] Take the section between the vehicle starting point and the dividing point as the entry section, and take the section between the dividing point and the parking entry point as the adjustment section.

[0027] A second aspect of the embodiments of the present application provides an automatic parking method, which applies the parking path determined by the above-mentioned automatic parking path planning method; the method includes:

[0028] Take the initial path point of the parking path as the previous driving moment, and control the vehicle to drive to the current driving moment of the previous driving moment;

[0029] Obtain the candidate obstacle positions at the current driving moment and the new obstacle positions relative to the previous driving moment;

[0030] Predict the extended obstacle positions corresponding to each new obstacle position, and determine the obstacle positions corresponding to the current driving moment based on the candidate obstacle positions and the extended obstacle positions;

[0031] Adjust the parking path based on the obstacle positions, and continue to execute the step of controlling the vehicle to drive to the current driving moment of the previous driving moment until the vehicle completes automatic parking.

[0032] In the automatic parking method, where the adjusting the parking path based on the obstacle positions specifically includes:

[0033] Detect whether the parking path collides based on the obstacle positions;

[0034] If the parking path collides, re-plan the parking path to adjust the parking path;

[0035] If the parking path does not collide, keep the parking path unchanged.

[0036] In the automatic parking method, where the obtaining the candidate obstacle positions at the current driving moment specifically includes:

[0037] Detect the suspicious obstacle positions corresponding to the current driving moment through the sensors configured on the vehicle;

[0038] For each suspicious obstacle position, obtain the observation probability corresponding to the suspicious obstacle position and the prior probability corresponding to the suspicious obstacle position when the vehicle is at the previous driving moment;

[0039] Calculate the posterior probability corresponding to the suspicious obstacle position based on the observation probability and the prior probability;

[0040] If the posterior probability is greater than a preset probability threshold, use the suspicious obstacle position as the candidate obstacle position.

[0041] The automatic parking method, wherein predicting the extended obstacle positions corresponding to the positions of each newly added obstacle specifically includes:

[0042] Obtain the center position of the previous obstacle corresponding to the previous driving moment and the center position of the current obstacle corresponding to the current driving moment, and use the direction from the center position of the previous obstacle to the center position of the current obstacle as the extension direction;

[0043] For each position of the newly added obstacle, predict the extension estimation probability of each extension position within a preset range along the extension direction, and use the extension position with the extension estimation probability greater than the preset probability value as the extended obstacle position.

[0044] A third aspect of the embodiments of the present application provides an automatic parking path planning device, where the path planning device includes:

[0045] An acquisition module, configured to acquire the parking space information of the target parking space, and determine the parking section and the adjustment section based on the parking space information;

[0046] A first determination module, configured to determine the first planned path corresponding to the parking section through a hybrid A* algorithm;

[0047] A second determination module, configured to determine the second planned path corresponding to the adjustment section by using a straight-line driving method;

[0048] A formation module, configured to connect the first planned path and the second planned path to form an automatic parking path.

[0049] A fourth aspect of the embodiments of the present application provides an automatic parking device, using the parking path determined by the automatic parking path planning device as described above; the device includes:

[0050] A control module, configured to control the vehicle to drive according to the parking path, and acquire the candidate obstacle positions at the current driving moment and the positions of the newly added obstacles at the current driving moment relative to the previous driving moment;

[0051] A prediction module, configured to predict the extended obstacle positions corresponding to the positions of each newly added obstacle, and determine the obstacle positions corresponding to the current driving moment based on the candidate obstacle positions and the extended obstacle positions;

[0052] An adjustment module is configured to adjust the parking path based on the position of the obstacle, and continue to execute the step of controlling the vehicle to travel along the parking path until the vehicle completes automatic parking. A fifth aspect of the embodiments of the present application provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the above-mentioned path planning method for automatic parking, and / or to implement the steps in the above-mentioned automatic parking method.

[0053] A third aspect of the embodiments of the present application provides a computer-readable storage medium, the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in any one of the above-mentioned path planning methods for automatic parking and automatic parking methods.

[0054] A sixth aspect of the embodiments of the present application provides a terminal device, including: a processor, a memory, and a communication bus; a computer-readable program executable by the processor is stored on the memory;

[0055] The communication bus realizes the connection and communication between the processor and the memory;

[0056] When the processor executes the computer-readable program, it implements the steps in the path planning method for automatic parking as described in any one of claims 1-5, and / or implements the steps in the automatic parking method as described in any one of claims 6-9.

[0057] Advantageous effects: Compared with the prior art, the present application provides a path planning method for automatic parking, an automatic parking method and related devices. The path planning method includes obtaining the parking space information of a target parking space, and determining a parking-in section and an adjustment section based on the parking space information; determining a first planned path corresponding to the parking-in section through a hybrid A* algorithm; determining a second planned path corresponding to the adjustment section by using a straight-line driving method; connecting the first planned path and the second planned path to form an automatic parking path. By dividing the parking path into a parking-in section and an adjustment section, the present application uses a hybrid A* algorithm for path planning in the parking-in section and a straight-line driving method for path planning in the adjustment section, and adjusts the vehicle attitude error generated by the parking-in section through the adjustment section, reducing the requirement for the control accuracy when the vehicle parks, and can well ensure the final attitude of the vehicle. Description of the Drawings

[0058] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative labor, other drawings can also be obtained based on these drawings.

[0059] Figure 1 It is a flowchart of the path planning method for automatic parking provided by the present application.

[0060] Figure 2 It is a schematic diagram of a parking space.

[0061] Figure 3 It is a curve of the parking attitude change for a side parking space.

[0062] Figure 4 It is a curve of the parking attitude change for a backward parking space.

[0063] Figure 5 It is a schematic diagram of the parking-in section.

[0064] Figure 6 It is a schematic diagram of the parking-in path.

[0065] Figure 7 It is a schematic diagram of the initial obstacle area.

[0066] Figure 8 It is a polygon schematic diagram of the feasible region of the parking space lane.

[0067] Figure 9 It is a schematic diagram of the target obstacle area.

[0068] Figure 10 It is a schematic flow diagram of the path planning method for automatic parking.

[0069] Figure 11 It is a flowchart of the automatic parking method provided by the present application.

[0070] Figure 12 It is a schematic flow diagram of the automatic parking method provided by the present application.

[0071] Figure 13 It is a schematic diagram of the parameters of the ultrasonic sensor.

[0072] Figure 14 It is a schematic diagram of the obstacle distribution at the previous driving moment.

[0073] Figure 15 It is a schematic diagram of the candidate obstacle distribution at the current driving moment.

[0074] Figure 16 It is a schematic diagram of the obstacle distribution at the current driving moment.

[0075] Figure 17 This is the structural schematic diagram of the path planning device for automatic parking provided by this application.

[0076] Figure 18 This is the structural schematic diagram of the automatic parking device provided by this application.

[0077] Figure 19 This is the structural schematic diagram of the terminal device provided by this application. Detailed implementation manners

[0078] This application provides a path planning method for automatic parking, an automatic parking method and related devices. To make the objectives, technical solutions and effects of this application clearer and more definite, the following further describes this application in detail with reference to the accompanying drawings and by way of examples. It should be understood that the specific examples described herein are only used to explain this application and are not used to limit this application.

[0079] Those skilled in the art of this technology can understand that unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the specification of this application means the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0080] Those skilled in the art of this technology can understand that unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0081] It should be understood that the sequence numbers and magnitudes of the steps in this embodiment do not mean the order of execution is prior or posterior. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiment of this application.

[0082] It has been found through research that in recent years, with the rapid development of the domestic automotive industry, automatic parking technology has become an essential function of vehicles. Automatic parking technology is to plan a feasible parking path based on the vehicle pose and the parking space pose, and then automatically control the vehicle to follow the parking path to complete parking. Among them, parking path planning is a key link in automatic parking, and its basic requirement is that the planned path is feasible and collision-free.

[0083] Existing parking path planning methods generally adopt the planning method of geometric curve splicing and the planning method of graph search (for example, the hybrid A* search method). Among them, the planning method of geometric curve splicing is to solve the planning path by splicing multiple curves. When the pose at the entry point is not ideal, there are errors in the vehicle entry control process, and when an obstacle appears, there may be no analytical solution for the entry section route and the route is not optimal, making it difficult to ensure the final body pose when parking. The hybrid A* search method is to determine the optimal planning path based on the start point, end point, and obstacle position information. However, due to the left and right restrictions of the perpendicular parking space and the front and back restrictions of the side parking space, the parking end point is in a narrow search space that is closed in three directions, making it difficult to ensure the search efficiency and the optimality of the path. When an obstacle appears, the search algorithm may not be able to find a feasible path. And because the search domain and the detection domain are the same, in the dynamic obstacle scenario, the phenomenon of replanning will occur frequently.

[0084] In addition, during the process of using the hybrid A* search method, some path planning methods divide the route into two sections: forward driving and reverse parking. Different heuristic function weights are used for the two sections during hybrid A* search. However, using hybrid A* search for the parking section has relatively high requirements for the control accuracy when the vehicle parks, and it is difficult to ensure the final pose.

[0085] To solve the above problems, in the embodiments of the present application, the parking space information of the target parking space is obtained, and the parking section and the adjustment section are determined based on the parking space information; the first planning path corresponding to the parking section is determined by the hybrid A* algorithm; the second planning path corresponding to the adjustment section is determined by the straight driving method; the first planning path and the second planning path are connected to form an automatic parking path. In the present application, by dividing the parking path into a parking section and an adjustment section, the hybrid A* algorithm is used for path planning in the parking section, and the straight driving method is used for path planning in the adjustment section. The vehicle pose error generated by the parking section is adjusted through the adjustment section, reducing the requirements for the control accuracy when the vehicle parks, and can well ensure the final pose of the vehicle.

[0086] The following further illustrates the application content through the description of the embodiments in conjunction with the accompanying drawings.

[0087] This embodiment provides a path planning method for automatic parking, as Figure 1 shown, the method includes:

[0088] S10. Obtain the parking space information of the target parking space, and determine the parking-in section and the adjustment section based on the parking space information.

[0089] Specifically, the parking space information includes vehicle information, parking space geometric information, and obstacle information. Among them, the vehicle information may include vehicle model and vehicle position, the parking space geometric information may include parking space length, parking space width, parking space slope, safety distance, and lane width, and the obstacle information may include front and rear vehicle position information, left and right vehicle position information, and parking space internal obstacle information (such as a car stopper, guardrail, etc.). Among them, the vehicle information can be determined according to the positioning device carried by the vehicle itself, the parking space geometric information can be obtained by an ultrasonic sensor or an image acquisition device, and the obstacle information can also be obtained by an ultrasonic sensor or an image acquisition device. In addition, when the parking space internal obstacle information is not obtained, the standard information of the parking lot design specification can be adopted.

[0090] The parking-in section and the adjustment section are two components of the parking path, and the parking-in section and the adjustment section are connected. Among them, the adjustment section is used to adjust the vehicle passing through the parking-in section to maintain the final attitude of the vehicle. It can be understood that when planning the parking path, the parking path is divided into a parking-in section and an adjustment section. When the vehicle automatically parks according to the parking path, it first passes through the parking-in section and then passes through the adjustment section to complete automatic parking.

[0091] In one implementation manner, determining the parking-in section and the adjustment section based on the parking space information specifically includes:

[0092] S11. Determine the parking-in point of the target parking space based on the parking space information, and select a target point in front of the parking-in point as the dividing point;

[0093] S12. Take the section between the vehicle starting point and the dividing point as the parking-in section, and take the section between the dividing point and the parking-in point as the adjustment section.

[0094] Specifically, as Figure 2 shown, the dividing point and the parking-in point are on the same straight line, and the direction from the parking-in point to the dividing point is the vehicle's head orientation. That is to say, a dividing point is set at a preset distance forward in the head direction. The section between the vehicle starting point and the dividing point is taken as the parking-in section, and the section between the dividing point and the parking-in point is taken as the adjustment section. That is, the dividing point is the splitting point between the parking-in section and the adjustment section.

[0095] As Figure 2As shown in the figure, the parking spaces include side parking spaces, perpendicular parking spaces, and inclined parking spaces. Among them, the position spaces that can be reserved in front of the vehicle heads of different parking spaces are different. Therefore, when determining the parking point of the target parking space based on the parking space information, the parking space type can be determined based on the parking space information, and the adjustment distance corresponding to the parking space type can be selected; the division point can be determined based on the adjustment distance, so that the distance between the division point and the parking point is equal to the adjustment distance. In this way, the adjustment section can be determined, and it can be avoided that when the vehicle travels to the division point through the parking section, a collision occurs or the first planned path corresponding to the parking section cannot be planned. Among them, the parking space types include side parking spaces and rear parking spaces. The side parking spaces can include side parking spaces, and the rear parking spaces can include perpendicular parking spaces and inclined parking spaces.

[0096] Furthermore, the adjustment distances corresponding to different parking space types can be different. Among them, for side parking spaces, as Figure 3 shown, the larger the adjustment distance, the better the parking attitude. However, as the adjustment distance increases, the parking space will become smaller, which will affect the planning of the first planned path and may even lead to the inability to solve the first planned path. Therefore, for side parking spaces, the adjustment distance is determined based on the first preset condition, where the first preset condition is preset. For example, the first preset condition is that the adjustment distance is less than the preset distance threshold, where the preset distance threshold is determined based on the parking space length. For example, the preset distance threshold is equal to one-fifth of the parking space length, etc. In addition, in practical applications, in order to improve the path planning speed, the adjustment distance can be a preset value. For example, the adjustment distance is 0.4m, etc.

[0097] For rear parking spaces, as Figure 4 shown, the larger the adjustment distance, the better the parking attitude. However, during the parking path planning process, the search domain is jointly composed of the parking space domain and the lane domain. As the adjustment distance increases, the requirements for the lane domain will increase, and when the lane width is insufficient, the planned path cannot be solved. Therefore, for rear parking spaces, the adjustment distance can be determined based on the second preset condition, where the second preset condition is preset. For example, the second preset condition is that the adjustment distance is less than the first distance threshold and greater than the second distance threshold, where the first distance threshold is greater than the second distance threshold, and both the first distance threshold and the second distance threshold are determined based on the parking space length. For example, the first distance is equal to the parking space length, and the second distance threshold is one-fifth of the parking space length, etc. In practical applications, in order to improve the path planning speed, the adjustment distance can be a preset value. For example, the adjustment distance is 4m, etc.

[0098] S20. Determine the first planned path corresponding to the parking section through the hybrid A* algorithm.

[0099] Specifically, the hybrid A* algorithm is used to solve the path search problem under the movement constraints of the parking space. Through the hybrid A* algorithm, the first planned path of the parking-in section can be solved. Among them, as Figure 5 shown, the first planned path is a curved path. The starting point of the vehicle is the starting point of the first planned path, and the dividing point is the end point of the first planned path. During the search process of the hybrid A* algorithm, the sum of the movement cost and the heuristic function is used as the search basis, and based on this search basis, a search is performed in the search domain to form the first planned path.

[0100] Since the heuristic function in the hybrid A* algorithm generally uses the RS curve to measure the distance from the current point to the target point, and the role of the heuristic function is the same regardless of the distance between the current point and the target point, this will affect the path planning efficiency. Based on this, in one implementation manner, determining the first planned path corresponding to the parking-in section through the hybrid A* algorithm specifically includes:

[0101] S21. Take the starting position of the vehicle as the initial path point;

[0102] S22. Obtain the distance between the initial path point and the target parking space, and determine the heuristic function weight corresponding to the initial path point according to the distance;

[0103] S23. Based on the heuristic function weight, search for the next path node of the initial path point in the search area corresponding to the vehicle through the hybrid A* algorithm;

[0104] S24. Take the next path node as the initial path point, and continue to execute the step of obtaining the distance between the initial path point and the target parking space until the end point of the parking-in section to form the first planned path.

[0105] Specifically, the distance between the initial path point and the target parking space refers to the distance between the initial path point and the parking-in point of the target parking space, which is used to reflect the relative distance between the vehicle and the target parking space. Among them, the larger the distance, the farther the vehicle is from the target parking space. On the contrary, the smaller the distance, the closer the vehicle is to the target parking space. Among them, when the distance between the vehicle and the target parking space is large, a large heuristic function weight can be set to increase the proportion of the heuristic function to improve the search speed. When the distance between the vehicle and the target parking space is small, a small heuristic function weight can be set to reduce the proportion of the heuristic function to determine the optimal path. Although this may sacrifice the optimal route of the front-end path, the front-end path of parking is generally mainly for fast passing, so the impact on the optimal path can be ignored. And in the later stage, by reducing the proportion of the heuristic function, the optimal path can be obtained. By driving along the optimal path in the later stage and adjusting the vehicle attitude in the adjustment section, it can be ensured that the vehicle has a good attitude when parked in the target parking space, so that both the automatic parking speed can be improved and the vehicle can maintain a good final attitude.

[0106] In one implementation, the specific process of determining the heuristic function weight corresponding to the initial path point according to the distance includes:

[0107] Compare the distance with a first distance threshold and a second distance threshold respectively;

[0108] If the distance is greater than the first distance threshold, set a first preset weight as the heuristic function weight corresponding to the initial path point;

[0109] If the distance is less than or equal to the first distance threshold and greater than or equal to the second distance threshold, set a second preset weight as the heuristic function weight corresponding to the initial path point;

[0110] If the distance is less than the second distance threshold, set a third preset weight as the heuristic function weight corresponding to the initial path point.

[0111] Specifically, both the first distance threshold and the second distance threshold are pre-set and serve as the basis for determining the heuristic function weight, and the first distance threshold is less than the second distance threshold. The first preset weight, the second preset weight, and the third preset weight are all pre-set. The first preset weight is greater than the second preset weight, and the second preset weight is greater than the third preset weight, such that when the distance is greater than the first distance threshold, the maximum heuristic function weight is used to achieve the purpose of fast search; when the distance is less than or equal to the first distance threshold and greater than the second distance threshold, the intermediate heuristic function weight is used to seek the optimal path while ensuring the search speed; when the distance is less than or equal to the second distance threshold, the minimum heuristic function weight is used to increase the attention to the movement cost and ensure that the end of the parking section can search for the final parking path, thereby ensuring the vehicle posture after completing the parking section.

[0112] Based on this, the search objective function of the hybrid A* algorithm can be expressed as:

[0113]

[0114] where f(N) represents the search objective function, g(N) represents the movement cost, h(N) represents the heuristic function, w1, w2, and w3 all represent the heuristic function weights, N represents the search position, M represents the parking position, and D1, D2, and D3 all represent the distance thresholds.

[0115] Furthermore, the specific process of determining the search area includes:

[0116] Determine the initial obstacle area and the initial parking space area corresponding to the target vehicle based on the parking space information, expand the initial obstacle area outward to obtain the target obstacle area, and expand the initial parking space area inward to obtain the target parking space area;

[0117] Determine the search area based on the target obstacle area and the target parking space area.

[0118] Specifically, the initial obstacle area is the area formed by the obstacles corresponding to the target parking space, which can be obtained when collecting parking space information. For example, it can be collected by the ultrasonic sensor or the image acquisition device of the vehicle device. The target obstacle area is obtained by expanding the initial obstacle area outward, and the target obstacle area includes the initial obstacle area. The expansion refers to expanding the initial obstacle area. For example, Figure 6 the initial obstacle area after expansion is as Figure 7 the target obstacle area. Among them, the expansion can be achieved by expanding the initial obstacle area outward by a preset distance, or by expanding the area range of the initial obstacle area by a preset multiple, etc.

[0119] After obtaining the target obstacle area, convert the parking space geometric information included in the parking space information into a polygon area of the parking space lane feasible region, where the polygon area includes the initial parking space area and the lane area. For example, Figure 8 the parking space lane feasible region. After the initial parking space area, the parking space area is expanded inward to shrink the parking space area to obtain the target parking space area, and then the overlapping part of the target polygon area composed of the target parking space area and the lane area and the target obstacle area is removed to obtain the search area. This implementation method reserves a moving area for the obstacle by expanding the obstacle area outward and expanding the parking space area inward, which can avoid the problem of frequent replanning of the path caused by obstacle movement, detection blind spots, and obstacle position errors, and improves the parking efficiency of automatic parking.

[0120] S30. Determine the second planned path corresponding to the adjustment section by using the straight-line driving method.

[0121] Specifically, the starting point of the adjustment section is directly in front of the end point of the adjustment section. As Figure 9 shown, the second planned path can be directly planned by using the straight-line driving method. By approaching the end straight line in a way of decreasing speed gradient in the adjustment section, the attitude error can be eliminated, and the requirement for the attitude accuracy of the parking point is reduced.

[0122] S40. Connect the first planned path and the second planned path to form an automatic parking path.

[0123] Specifically, after obtaining the first planned path and the second planned path, the first planned path and the second planned path are connected, that is, the end point of the first planned path is merged with the start point of the second planned path to obtain an automatic parking path. In addition, as Figure 10 shown, after obtaining the parking path, it is also possible to detect whether the parking path will collide with an obstacle. If a collision occurs, the path is replanned. If no collision occurs, the speed gradient of the parking path is smoothed and the parking path is published.

[0124] In summary, in this embodiment, a path planning method for automatic parking is provided. The path planning method includes obtaining the parking space information of the target parking space and determining the parking-in section and the adjustment section based on the parking space information; determining the first planned path corresponding to the parking-in section through the hybrid A* algorithm; using a straight-line driving method to determine the second planned path corresponding to the adjustment section; connecting the first planned path and the second planned path to form an automatic parking path. In this application, the parking path is divided into a parking-in section and an adjustment section. The hybrid A* algorithm is used for path planning in the parking-in section, and the straight-line driving method is used for path planning in the adjustment section. The adjustment section adjusts the vehicle attitude error generated by the parking-in section, reducing the requirement for the control accuracy when the vehicle parks and ensuring the final attitude of the vehicle well.

[0125] Based on the above path planning method for automatic parking, this embodiment provides an automatic parking method, which applies the parking path determined by the above path planning method for automatic parking; as Figure 11 and 12 shown, the method includes:

[0126] H10. Control the vehicle to drive according to the parking path, and obtain the candidate obstacle position at the current driving moment and the new obstacle position relative to the previous driving moment at the current driving moment.

[0127] Specifically, the candidate obstacle position is detected when the vehicle travels to the current driving moment, and the new obstacle position is the new one detected by the vehicle at the current driving moment relative to the previous obstacle position detected by the vehicle at the previous driving moment. That is, the new obstacle position is included in the candidate obstacle position but not included in the previous obstacle position.

[0128] In one implementation, the obtaining of the candidate obstacle position at the current driving moment specifically includes:

[0129] H21. Detect the suspicious obstacle position corresponding to the current driving moment through the sensors configured on the vehicle;

[0130] H22. For each suspicious obstacle position, obtain the observation probability corresponding to the suspicious obstacle position and the prior probability corresponding to the suspicious obstacle position when the vehicle was at the previous driving moment;

[0131] H23. Calculate the posterior probability corresponding to the suspicious obstacle position based on the observation probability and the prior probability;

[0132] H24. If the posterior probability is greater than a preset probability threshold, then use the suspicious obstacle position as a candidate obstacle position.

[0133] Specifically, the vehicle is equipped with a number of ultrasonic sensors. Each ultrasonic sensor is independent of each other and can detect suspicious obstacle positions according to the beam angle and the obstacle distance. That is to say, the transmitter of each ultrasonic sensor emits ultrasonic waves and records the emission time of the emitted wave. Thus, when the ultrasonic wave meets an obstacle, it will be reflected. In this way, the receiver of the ultrasonic sensor receives the reflected wave. Using the reception time of the reflected wave, the emission time of the emitted wave, and the transmission rate of the light wave as ultrasonic distance data, the vehicle terminal can obtain the position of the obstacle relative to the vehicle according to the ultrasonic distance data, the driving speed of the vehicle, etc.

[0134] The vehicle is equipped with a number of ultrasonic sensors. Each ultrasonic sensor will detect a number of suspicious obstacle positions. Thus, the suspicious obstacle positions corresponding to the current driving moment include the suspicious obstacle positions collected by each sensor equipped on the vehicle. For example, if the vehicle includes ultrasonic sensor 1, ultrasonic sensor 2,..., ultrasonic sensor N, then the suspicious obstacle positions include the suspicious obstacle positions collected by ultrasonic sensor 1, ultrasonic sensor 2,..., ultrasonic sensor N.

[0135] Furthermore, after obtaining the suspicious obstacle positions corresponding to the current driving moment, the suspicious obstacle positions can be merged and then converted to the map coordinate system to form an obstacle probability map. Among them, the obstacle probability map contains all the suspicious obstacle positions and probabilities in the map coordinate system at the current driving moment. After merging and converting the suspicious obstacle positions to the map coordinate system, under the ultrasonic sensor detection parameters as shown in Figure 13 The calculation formula of the observation probability can be:

[0136]

[0137] where a represents the distance between the suspicious obstacle position and the ultrasonic sensor, b represents the deviation angle between the suspicious obstacle position and the central axis of the ultrasonic sensor, s represents the distance measurement value returned by the ultrasonic sensor, w represents the beam angle of the ultrasonic sensor, ds = 0.2×s represents the model credible half-width of the ultrasonic sensor. When the point k(x, y) is not within the model credible half-width range, the probability of the presence of an obstacle is regarded as 0.

[0138] Based on the Bayesian principle, the posterior probability of the position of a suspicious obstacle can be determined based on the observation probability of the position of the suspicious obstacle and the prior probability at the previous driving moment. That is, when a position of a suspicious obstacle appears in the possible area of an obstacle detected by an ultrasonic wave, it will not be immediately determined as an obstacle. Instead, when this position of the suspicious obstacle is continuously determined as an obstacle for a period of time, it will be regarded as the position of an obstacle. Using probability to represent the event of the appearance of an obstacle in this way can improve the accuracy of obstacle detection and avoid the measurement error of traditional ultrasonic sensors. For example, it can avoid the problem that the vehicle has no way to go because all possible areas within the beam angle are regarded as the positions of obstacles. Or, it can avoid the problem that when using the ultrasonic data of two adjacent positions to judge the appearance position of an obstacle, and the judgment logic is that when both ultrasonic sensors detect an obstacle, it is considered that the obstacle appears in the overlapping area of the detection ranges of the two ultrasonic sensors. Because there are two smaller obstacles, obstacle 1 appears in the left half of the detection range of the left ultrasonic sensor, and obstacle 2 appears in the right half of the detection range of the right ultrasonic sensor, resulting in the misjudgment that only one obstacle appears in the overlapping area of the left and right ultrasonic sensors.

[0139] In one implementation, the calculation process of the posterior probability of the position of a suspicious obstacle can be as follows:

[0140] Let k(x,y) be the event A that there is a real obstacle, and the event B that k(x,y) is detected as having an obstacle at this moment. Then, according to the conditional probability, P(B|A) = P(A|B)P(B) / P(A). From the total probability formula, we can get Based on this, the posterior probability T i (t,k) that there is an obstacle at point k(x,y) under the measurement of ultrasonic sensor i at time t and is correctly detected, where i the calculation formula of T

[0141]

[0142] P(O) = P(t, k) × T(t - 1, k)

[0143]

[0144] where i T(t,k) represents the posterior probability, T(t - 1,k) represents the prior probability, and P(t,k) represents the observation probability.

[0145] Furthermore, after obtaining the posterior probability T i (t,k) of each ultrasonic sensor, each posterior probability T iTake the maximum value of (t,k) as the posterior probability of the suspicious obstacle position. The calculation formula for the posterior probability can be:

[0146]

[0147] where N represents the number of ultrasonic sensors.

[0148] H20, predict the extended obstacle positions corresponding to each newly added obstacle position, and determine the obstacle position corresponding to the current driving moment based on the candidate obstacle position and the extended obstacle position.

[0149] Specifically, the extended obstacle position is the position determined by extending the newly added obstacle position in a preset direction. The preset direction can be pre-set or determined according to the obstacle information corresponding to the previous driving moment and the obstacle information corresponding to the current driving moment. In this embodiment, by obtaining the extended obstacle positions of each newly added obstacle position, the phenomenon that a large obstacle appears, the sensor cannot see the whole picture at a certain moment, and the obstacle will continuously appear as the vehicle continuously moves can be solved.

[0150] In one implementation manner, the predicting the extended obstacle positions corresponding to each newly added obstacle position specifically includes:

[0151] Obtain the previous obstacle center position corresponding to the previous driving moment and the current obstacle center position corresponding to the current driving moment, and use the direction from the previous obstacle center position to the current obstacle center position as the extension direction;

[0152] For each newly added obstacle position, predict the extended estimation probabilities of each extended position within a preset range along the extension direction, and use the extended position with the extended estimation probability greater than the preset probability value as the extended obstacle position.

[0153] Specifically, the obstacle distribution map obtained by Bayesian estimation at the previous driving moment can be as Figure 14 shown, where the darker the color, the higher the probability, and the cross arrow represents the obstacle center position at time t-1. The obstacle distribution map obtained by Bayesian estimation at the current driving moment can be as Figure 15 shown, where the obstacle center position represented by the circular cross moves to the right. Thus, the extension direction is the connection direction from the cross arrow at the previous driving moment to the circular cross at the current driving moment. In other words, the extension direction is the direction from the previous obstacle center position to the current obstacle center position.

[0154] After obtaining the extension direction, predict the extended estimation probabilities of each extended position within a preset range along the extension direction, where the extended estimation probability can be expressed as:

[0155]

[0156] Among them, g represents the probability calculation formula for the prediction point j, the posterior probability T(t,k) of the position k of the newly added obstacle, the distance d between the position k of the newly added obstacle and the prediction point j, and ds represents the preset distance threshold.

[0157] After obtaining the candidate obstacle position and the extended obstacle position, the obstacle area formed by the candidate obstacle position and the extended obstacle position is used as the obstacle position corresponding to the current driving moment, as Figure 16 shown.

[0158] H30. Adjust the parking path based on the obstacle position, and continue to execute the step of controlling the vehicle to drive along the parking path until the vehicle completes automatic parking.

[0159] Specifically, after obtaining the obstacle position, it is possible to detect whether the parking path collides based on the obstacle position, which can avoid collisions during automatic parking. Among them, when detecting whether the parking path collides based on the obstacle position, if the parking path collides, the parking path is re-planned to adjust the parking path; if the parking path does not collide, the parking path remains unchanged. In addition, when re-planning the parking path, the path planning method for automatic parking provided in the above embodiments can be used for re-path planning, or other existing methods can be used for path planning. For example, the hybrid A* algorithm can be directly used for path planning, etc.

[0160] In this embodiment, when using Bayesian estimation to predict obstacles, the ultrasonic raw data and the prior experience of the previous moment are used to estimate the obstacles, which can improve the accuracy of obstacle estimation. At the same time, when estimating the obstacles, the change from the obstacle position at the previous moment is used as the prediction basis to predict the possible position of the obstacles in their growth direction. On the one hand, as the vehicle gradually approaches the obstacle during movement, the true obstacle position is continuously estimated and iterated, and finally is confirmed as an obstacle. For non-obstacle positions, during the iteration process, the estimation threshold is not met, and although they are within the detection range of the ultrasonic wave at some moments, they will not be regarded as obstacles. This improves the estimation accuracy of the obstacle position, and thus gives the vehicle a larger driving space during autonomous parking. On the other hand, using the change of the obstacle at adjacent moments as the prediction basis, compared with the existing method of predicting obstacles based on the entire detection area of the ultrasonic wave, predicting in the growth direction of the obstacle, and the prediction basis is the more real obstacle distribution probability obtained from the previous Bayesian calculation, making the prediction result more reasonable and capable of reducing the estimation blind area caused by the obstacle size exceeding the detection range.

[0161] Based on the above path planning method for automatic parking, in this embodiment, a path planning device for automatic parking is provided. As Figure 17 shown, the path planning device includes:

[0162] An acquisition module 101, configured to acquire the parking space information of the target parking space, and determine the parking-in section and the adjustment section based on the parking space information;

[0163] A first determination module 102, configured to determine a first planned path corresponding to the parking-in section through a hybrid A* algorithm;

[0164] A second determination module 103, configured to determine a second planned path corresponding to the adjustment section by using a straight-line driving-in method;

[0165] A formation module 104, configured to connect the first planned path and the second planned path to form an automatic parking path.

[0166] Based on the above automatic parking method, in this embodiment, an automatic parking device is provided. As Figure 18 shown, the device includes:

[0167] A control module 201, configured to control the vehicle to travel along the parking path, and acquire the candidate obstacle positions at the current driving moment and the new obstacle positions at the current driving moment relative to the previous driving moment;

[0168] A prediction module 202, configured to predict the extended obstacle positions corresponding to the new obstacle positions, and determine the obstacle positions corresponding to the current driving moment based on the candidate obstacle positions and the extended obstacle positions;

[0169] An adjustment module 203, configured to adjust the parking path based on the obstacle positions, and continue to execute the step of controlling the vehicle to travel along the parking path until the vehicle completes automatic parking.

[0170] This embodiment provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the path planning method for automatic parking and / or the automatic parking method as described in the above embodiments.

[0171] This application also provides a terminal device. As Figure 19As shown, it includes at least one processor 20; a display screen 21; and a memory 22, and may also include a communications interface 23 and a bus 24. Among them, the processor 20, the display screen 21, the memory 22, and the communications interface 23 can communicate with each other through the bus 24. The display screen 21 is set to display a user guidance interface preset in the initial setting mode. The communications interface 23 can transmit information. The processor 20 can call the logical instructions in the memory 22 to execute the method in the above embodiments.

[0172] In addition, when the logical instructions in the above-mentioned memory 22 are implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a computer-readable storage medium.

[0173] The memory 22, as a computer-readable storage medium, can be set to store software programs and computer-executable programs, such as the program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 executes functional applications and data processing by running the software programs, instructions, or modules stored in the memory 22, that is, implements the methods in the above embodiments.

[0174] The memory 22 may include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created according to the use of the terminal device, etc. In addition, the memory 22 may include a high-speed random access memory and may also include a non-volatile memory. For example, various media such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs that can store program codes can also be transient storage media.

[0175] In addition, the specific processes of loading and executing multiple instructions by the above-mentioned storage medium and the instruction processor in the terminal device have been described in detail in the above method and will not be repeated here one by one.

[0176] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automatic parking method, characterized in that, The method includes: Controlling the vehicle to travel along a parking path, and obtaining the candidate obstacle positions at the current travel moment and the newly added obstacle positions at the current travel moment relative to the previous travel moment; Predicting the extended obstacle positions corresponding to the newly added obstacle positions, and determining the obstacle positions corresponding to the current travel moment based on the candidate obstacle positions and the extended obstacle positions; Adjusting the parking path based on the obstacle positions, and continuing to execute the step of controlling the vehicle to travel along the parking path until the vehicle completes automatic parking; Wherein, the predicting the extended obstacle positions corresponding to the newly added obstacle positions specifically includes: Obtaining the previous obstacle center position corresponding to the previous travel moment and the current obstacle center position corresponding to the current travel moment, and taking the direction from the previous obstacle center position to the current obstacle center position as the extension direction; For each newly added obstacle position, predicting the extension estimation probabilities of the extension positions within a preset range along the extension direction, and taking the extension positions with the extension estimation probabilities greater than a preset probability value as the extended obstacle positions; The process of obtaining the parking path specifically includes: Obtaining the parking space information of the target parking space, and determining the parking-in section and the adjustment section based on the parking space information; Determining the first planned path corresponding to the parking-in section through the hybrid A* algorithm; Determining the second planned path corresponding to the adjustment section by using the straight-line driving-in method; Connecting the first planned path and the second planned path to form an automatic parking path.

2. The automatic parking method according to claim 1, wherein, The determining the first planned path corresponding to the parking-in section through the hybrid A* algorithm specifically includes: Taking the starting position of the vehicle as the initial path point; Obtaining the distance between the initial path point and the target parking space, and determining the heuristic function weight corresponding to the initial path point according to the distance; Based on the heuristic function weight, searching for the next path node of the initial path point within the search area corresponding to the vehicle through the hybrid A* algorithm; Taking the next path node as the initial path point, and continuing to execute the step of obtaining the distance between the initial path point and the target parking space until the termination point of the parking-in section to form the first planned path.

3. The automatic parking method according to claim 2, wherein The process of determining the search area specifically includes: Determining the initial obstacle area and the initial parking space area corresponding to the target vehicle based on the parking space information, expanding the initial obstacle area outward to obtain the target obstacle area, and expanding the initial parking space area inward to obtain the target parking space area; Determining the search area based on the target obstacle area and the target parking space area.

4. The automatic parking method according to claim 2, wherein The determining the heuristic function weight corresponding to the initial path point according to the distance specifically includes: Comparing the distance with a first distance threshold and a second distance threshold respectively; If the distance is greater than the first distance threshold, setting a first preset weight as the heuristic function weight corresponding to the initial path point; If the distance is less than or equal to the first distance threshold and greater than or equal to the second distance threshold, setting a second preset weight as the heuristic function weight corresponding to the initial path point; If the distance is less than the second distance threshold, set the third preset weight to the heuristic function weight corresponding to the initial path point, where the first preset weight is greater than the second preset weight, and the second preset weight is greater than the third preset weight.

5. The automatic parking method according to claim 1, wherein The determining the parking-in section and the adjustment section based on the parking space information specifically includes: Determine the parking-in point of the target parking space based on the parking space information, and select a target point in front of the parking-in point as the dividing point, where the dividing point and the parking-in point are on the same straight line, and the direction from the parking-in point to the dividing point is the front direction of the vehicle; Take the section between the starting position of the vehicle and the dividing point as the parking-in section, and take the section between the dividing point and the parking-in point as the adjustment section.

6. The automatic parking method according to claim 1, wherein The adjusting the parking path based on the obstacle position specifically includes: Detect whether the parking path collides based on the obstacle position; If the parking path collides, re-plan the parking path to adjust the parking path; If the parking path does not collide, keep the parking path unchanged.

7. The automatic parking method according to claim 1, wherein The obtaining the candidate obstacle positions at the current driving moment specifically includes: Detect the suspicious obstacle positions corresponding to the current driving moment through the sensors configured on the vehicle; For each suspicious obstacle position, obtain the observation probability corresponding to the suspicious obstacle position and the prior probability corresponding to the suspicious obstacle position when the vehicle was at the previous driving moment; Calculate the posterior probability corresponding to the suspicious obstacle position based on the observation probability and the prior probability; If the posterior probability is greater than the preset probability threshold, use the suspicious obstacle position as the candidate obstacle position.

8. An automatic parking device, characterized in that, The device includes: A control module, which controls the vehicle to drive along the parking path, and obtains the candidate obstacle positions at the current driving moment and the newly added obstacle positions at the current driving moment relative to the previous driving moment; A prediction module, which is used to predict the extended obstacle positions corresponding to the newly added obstacle positions, and determine the obstacle positions corresponding to the current driving moment based on the candidate obstacle positions and the extended obstacle positions; An adjustment module, which is used to adjust the parking path based on the obstacle positions, and continue to execute the step of controlling the vehicle to drive along the parking path until the vehicle completes automatic parking; Among them, the predicting the extended obstacle positions corresponding to the newly added obstacle positions specifically includes: Obtain the previous obstacle center position corresponding to the previous driving moment and the current obstacle center position corresponding to the current driving moment, and use the direction from the previous obstacle center position to the current obstacle center position as the extension direction; For each newly added obstacle position, predict the extension estimation probabilities of the extension positions within a preset range along the extension direction, and use the extension positions with the extension estimation probabilities greater than the preset probability value as the extended obstacle positions; The obtaining process of the parking path specifically includes: Obtain the parking space information of the target parking space, and determine the parking-in section and the adjustment section based on the parking space information; Determine the first planned path corresponding to the parking-in section through the hybrid A* algorithm; Determine a second planned path corresponding to the adjustment section by using a straight-in driving method; Connect the first planned path and the second planned path to form an automatic parking path.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the automatic parking method according to any one of claims 1-7.

10. A terminal device, characterized in that, Comprising: A processor, a memory and a communication bus; A computer-readable program executable by the processor is stored on the memory; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the steps in the automatic parking method according to any one of claims 1-7 are implemented.

Images