An automatic parking method, device and system

Abstract

The application discloses an automatic parking method, device and system, relates to the technical field of intelligent driving, and can realize efficient and accurate automatic parking in a parallel parking scene. In the application, when automatic parking path planning is performed, an automatic parking system processes a parking process in stages, such as an entry stage and an exit stage, and fully combines entry capacity and exit capacity to improve path search efficiency and success rate of automatic parking. In the entry stage, the automatic parking system determines a target state set that is relatively easy to enter by calculation, any target state point in the target state set fully considers an easy-to-enter posture in an ultra-narrow parallel parking scene, and therefore, a relatively optimal entry target pose can be obtained, so that entry difficulty is reduced, and path search efficiency and success rate in the entry stage are improved. In the exit stage, the automatic parking system fully utilizes a drivable area by performing real-time sensing on a close-range obstacle, and adjusts the vehicle to a target pose in a parking space.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and in particular to an automatic parking method, device and system. Background Technology

[0002] With the widespread application of autonomous driving technology in advanced driver assistance systems (ADAS) or autonomous driving (AD) systems, automatic parking technology has become an important research area that can greatly improve the user experience.

[0003] Typically, autonomous driving systems can use appropriate automatic parking algorithms to complete automatic parking based on the parking scenario. Common parking scenarios include... Figure 1 The diagram shows parallel parking, perpendicular parking, and angled parking scenarios. Among these, parallel parking is the most frequently encountered. Furthermore, parallel parking is technically the most challenging because it typically requires multiple gear shifts and vehicle start-stop maneuvers, especially in scenarios with obstacles or narrow spaces. Therefore, achieving efficient and accurate automatic parking in parallel parking scenarios has become an important research direction. Summary of the Invention

[0004] This application provides an automatic parking method, apparatus, and system that can achieve efficient and accurate automatic parking in parallel parking scenarios.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] In a first aspect, an automatic parking method is provided, characterized in that it is applied to a parallel parking scenario. The method includes: acquiring an initial vehicle pose and a drivable area of ​​the vehicle, the drivable area including a parking space area; determining a first target state point based on a first constraint and a second constraint, the first target state point representing the position and attitude of the vehicle when it enters the parking space area from the initial pose; planning a first trajectory from the initial pose to the first target state point; and after the vehicle reaches the first target state point from the initial pose, determining a second trajectory of the vehicle within the parking space area based on the real-time pose of the vehicle within the parking space area; wherein the first constraint is used to ensure that the vehicle can exit the parking space area when it enters the parking space area and maintains a first curvature while moving forward from the first target state point, and the second constraint is used to ensure that the vehicle will not cross the edge of the parking space area when it enters the parking space area and maintains a second curvature while moving forward from the first target state point.

[0007] The solution provided in the first aspect above, based on the solution provided in this application, divides the parking process into stages during automatic parking trajectory planning, such as the entry stage and the caching stage. This fully combines entry and caching capabilities to improve the path search efficiency and success rate of automatic parking. Specifically, in the entry stage, the automatic parking system calculates and determines a set of target states that are relatively easy to enter the parking space. Any target state point in this set fully considers the posture that is easy to enter in ultra-narrow parallel parking scenarios, thus obtaining a better entry target posture, thereby reducing the difficulty of entering the parking space and improving the path search efficiency and success rate of the entry stage. In the caching stage, the automatic parking system, through real-time perception of nearby obstacles, fully utilizes the drivable area to adjust the vehicle's posture within the parking space to the target posture.

[0008] As one possible implementation, determining the first target state point based on the first and second constraints includes: determining a first pose based on the first constraint, whereby the first pose represents the limit pose that allows the vehicle to exit the parking space area while maintaining a first curvature; determining a second pose based on the second constraint, whereby the second pose represents the limit pose that prevents the vehicle from crossing the edge of the parking space area while maintaining a second curvature; determining a target area for the vehicle entering the parking space area based on the first and second poses, where the target area includes multiple target state points; and determining a first target state point from the target area, where the first target state point is any state point within the target area. Based on this, a set of target state points can be determined that simultaneously ensures the vehicle can smoothly exit the parking space after entering and guarantees the ability of subsequent vehicles to maneuver within the parking space. Any state point in this set of target state points can be used as a target state point for the parking entry stage.

[0009] As one possible implementation, determining the first target state point from the target area includes: determining the first target state point from the target area according to a preset strategy. Based on this, it can be ensured that the first target state point not only ensures that the vehicle can smoothly exit the parking space after entering the parking space, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0010] For example, the first target state point can be an intermediate state point during interpolation between the first pose and the second pose. Of course, the first target state point can also be other state points in the target region, and this application does not specifically limit it.

[0011] As one possible implementation, determining the target area when the vehicle enters the parking space based on the first pose and the second pose includes: determining the target area based on the trajectory traversed by the vehicle while maintaining a first curvature from the first pose and the second pose respectively. Based on this, state recursion curves corresponding to the first pose and the second pose can be obtained, wherein any state point in the area enclosed by the state recursion curves corresponding to the first pose and the second pose respectively can ensure both smooth entry and exit of the vehicle from the parking space and the vehicle's maneuverability within the parking space.

[0012] As one possible implementation, the above-mentioned planning of the first trajectory from the initial pose to the first target state point includes: determining a first path from the initial pose to the first target state point; smoothing the first path to obtain a second path; and planning the inbound speed of the second path to obtain the first trajectory. Based on this, a continuous and highly responsive inbound trajectory can be obtained.

[0013] As one possible implementation, determining the coarse path from the initial pose to the first target state point includes: determining the coarse path from the initial pose to the first target state point. Based on this, a continuous and highly responsive first trajectory can be obtained.

[0014] As one possible implementation, the path search algorithm described above considers first information. For example, the path search algorithm can determine the coarse path from the initial pose to the first target state point based on the first information, where the first information includes the shortest obstacle avoidance distance in a two-dimensional discrete space between the state points traversed by the vehicle and the first target state point. Based on this, not only obstacle avoidance can be considered during path search, but the shortest route can also be searched as much as possible.

[0015] As one possible implementation, the path search algorithm described above also considers second information. For example, the path search algorithm can determine the coarse path from the initial pose to the first target state point based on the second information, wherein the second information includes one or more of the following: the number of curvature changes of the vehicle, the number of gear shifts of the vehicle, and the length of the shortest RS curve between the initial pose and the first target state point. Based on this, the actual situation such as the number of curvature changes of the vehicle, the number of gear shifts, and the length of the RS curve can also be considered during path search to obtain a more scientific, reasonable, and accurate entry path.

[0016] As one possible implementation, the path search algorithm described above is the Hybrid A*+RS curve search algorithm. Based on this, since the RS curve between one state point and another can be a combination of many circles and straight lines, many feasible path solutions between these two state points can be searched, making the search easier and increasing the success rate, thus better meeting the needs of vehicle parking.

[0017] As one possible implementation, when the vehicle is in its initial position, the parking space area is located to the right of the vehicle; where the first curvature corresponds to the motion curvature when the vehicle makes its maximum left turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum right turn. Based on this, it can be ensured that the target state point during the parking phase not only ensures that the vehicle can smoothly exit the parking space after entering, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0018] As one possible implementation, when the vehicle is in its initial position, the parking space area is located to the left of the vehicle; where the first curvature corresponds to the motion curvature when the vehicle makes its maximum right turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum left turn. Based on this, it can be ensured that the target state point during the parking phase not only ensures that the vehicle can smoothly exit the parking space after entering, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0019] As one possible implementation, determining the vehicle's second trajectory within the parking space area based on its real-time pose includes: acquiring real-time surrounding environmental information of the vehicle within the parking space area; and performing one or more gear shifts and one or more curvature changes based on the vehicle's real-time pose and surrounding environmental information to ensure the vehicle meets a third constraint condition. Based on this, dynamically adjusting the vehicle's pose and lateral error within the parking space using the detected surrounding environmental information can significantly improve the accuracy of path planning during the parking space hopping phase.

[0020] For example, real-time surrounding environmental information of a vehicle within a parking space can be obtained using ultrasonic radar. It is understood that, in the confined space of a parking space, ultrasonic radar provides more accurate and faster detection of the parking environment compared to other positioning technologies, significantly improving the accuracy of parking space route planning.

[0021] As one possible implementation, the third constraint includes: the angle between the vehicle and the parking space is less than a preset angle; and the distance between the center point of the vehicle and the center point of the parking space is less than a preset distance. Based on this, it can be ensured that the vehicle safely parks in the parking space with a reasonable posture.

[0022] As one possible implementation, the method further includes: acquiring surrounding environmental information during the parking process, including the position, size, direction of movement, and speed of target objects within a preset range of the vehicle; and controlling the vehicle to stop when a collision with a target object is predicted. Based on this, obstacles can be accurately avoided during parking, ensuring the safety of the vehicle and its occupants.

[0023] As one possible implementation, during the process of the vehicle moving from its initial pose to the first target state point, the surrounding environment information also includes static obstacles within a preset range of the vehicle; the method further includes updating the vehicle's drivable area in real time based on the location information of the static obstacles. Based on this, obstacles can be accurately avoided during parking, ensuring the safety of the vehicle and personnel.

[0024] Secondly, an automatic parking system is provided, comprising: a data acquisition unit for acquiring an initial vehicle pose and a drivable area, the drivable area including a parking space area; a processing unit for determining a first target state point based on a first constraint and a second constraint, and planning a first trajectory from the initial pose to the first target state point; and, after the vehicle reaches the first target state point from the initial pose, determining a second trajectory of the vehicle within the parking space area based on the real-time pose of the vehicle within the parking space area; wherein the first target state point characterizes the position and attitude of the vehicle when it enters the parking space area from the initial pose; the first constraint ensures that the vehicle can exit the parking space area when it enters the parking space area and maintains a first curvature while moving forward from the first target state point; and the second constraint ensures that the vehicle will not cross the edge of the parking space area when it enters the parking space area and maintains a second curvature while moving forward from the first target state point.

[0025] The solution provided in the second aspect above, based on the solution provided in this application, divides the parking process into stages during automatic parking trajectory planning, such as the entry stage and the caching stage. This fully combines entry and caching capabilities to improve the path search efficiency and success rate of automatic parking. Specifically, in the entry stage, the automatic parking system calculates and determines a set of target states that are relatively easy to enter the parking space. Any target state point in this set fully considers the posture that is easy to enter in ultra-narrow parallel parking scenarios, thus obtaining a better entry target pose, thereby reducing the difficulty of entering the parking space and improving the path search efficiency and success rate of the entry stage. In the caching stage, the automatic parking system, through real-time perception of nearby obstacles, fully utilizes the drivable area to adjust the vehicle's pose within the parking space to the target pose.

[0026] As one possible implementation, the data acquisition unit includes a perception unit for collecting surrounding environmental data to determine the vehicle's drivable area. Based on this, obstacles can be accurately avoided during parking, ensuring the safety of the vehicle and its occupants.

[0027] As one possible implementation, the processing unit determines the first target state point based on the first and second constraints, including: the processing unit determines a first pose based on the first constraint, the first pose representing the limit pose that allows the vehicle to exit the parking space area while maintaining a first curvature; the processing unit determines a second pose based on the second constraint, the second pose representing the limit pose that prevents the vehicle from crossing the edge of the parking space area while maintaining a second curvature; the processing unit determines a target area when the vehicle enters the parking space area based on the first and second poses, the target area including multiple target state points; and the processing unit determines a first target state point from the target area, the first target state point being any state point within the target area. Based on this, a set of target state points can be determined that simultaneously ensures the vehicle can smoothly exit the parking space after entering and ensures the subsequent vehicle's maneuverability within the parking space. Any state point in this set of target state points can be used as a target state point for the parking entry stage.

[0028] As one possible implementation, the processing unit determines a first target state point from the target area, including: the processing unit determines the first target state point from the target area according to a preset strategy. Based on this, it can be ensured that the first target state point not only ensures that the vehicle can smoothly exit the parking space after entering the parking space, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0029] For example, the first target state point can be an intermediate state point during interpolation between the first pose and the second pose. Of course, the first target state point can also be other state points in the target region, and this application does not specifically limit it.

[0030] As one possible implementation, the processing unit determines the target area when the vehicle enters the parking space based on the first pose and the second pose, including: the processing unit determines the target area based on the trajectory traversed by the vehicle while maintaining a first curvature from the first pose and the second pose respectively. Based on this, state recursion curves corresponding to the first pose and the second pose can be obtained, wherein any state point in the area enclosed by the state recursion curves corresponding to the first pose and the second pose respectively can ensure both smooth entry and exit of the vehicle from the parking space and the vehicle's maneuverability within the parking space.

[0031] As one possible implementation, the processing unit plans a first trajectory from the initial pose to the first target state point, including: the processing unit determining a first path from the initial pose to the first target state point; the processing unit smoothing the first path to obtain a second path; and the processing unit planning the inbound speed of the second path to obtain the first trajectory. Based on this, a continuous and highly responsive inbound trajectory can be obtained.

[0032] As one possible implementation, the processing unit determines a coarse path from the initial pose to the first target state point, including: the processing unit uses a path search algorithm to determine the coarse path from the initial pose to the first target state point. Based on this, a continuous and highly responsive inbound trajectory can be obtained.

[0033] As one possible implementation, the path search algorithm described above considers first information. For example, the processing unit can use the path search algorithm to determine the coarse path from the initial pose to the first target state point based on the first information, wherein the first information includes the shortest obstacle avoidance distance in a two-dimensional discrete space between the state points traversed by the vehicle and the first target state point. Based on this, not only obstacle avoidance can be considered during path search, but the shortest route can also be searched as much as possible.

[0034] As one possible implementation, the path search algorithm described above also considers second information. For example, the processing unit can use the path search algorithm to determine the coarse path from the initial pose to the first target state point based on the second information, wherein the second information includes one or more of the following: the number of curvature changes of the vehicle, the number of gear shifts of the vehicle, and the length of the shortest RS curve between the initial pose and the first target state point. Based on this, the actual situation such as the number of curvature changes of the vehicle, the number of gear shifts, and the length of the RS curve can also be considered during path search to obtain a more scientific, reasonable, and accurate entry path.

[0035] As one possible implementation, the path search algorithm described above is the Hybrid A*+RS curve search algorithm. Based on this, since the RS curve between one state point and another can be a combination of many circles and straight lines, many feasible path solutions between these two state points can be searched, making the search easier and increasing the success rate, thus better meeting the needs of vehicle parking.

[0036] As one possible implementation, when the vehicle is in its initial position, the parking space area is located to the right of the vehicle; where the first curvature corresponds to the motion curvature when the vehicle makes its maximum left turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum right turn. Based on this, it can be ensured that the target state point during the parking phase not only ensures that the vehicle can smoothly exit the parking space after entering, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0037] As one possible implementation, when the vehicle is in its initial position, the parking space area is located to the left of the vehicle; where the first curvature corresponds to the motion curvature when the vehicle makes its maximum right turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum left turn. Based on this, it can be ensured that the target state point during the parking phase not only ensures that the vehicle can smoothly exit the parking space after entering, but also ensures the ability of subsequent vehicles to maneuver within the parking space.

[0038] As one possible implementation, the processing unit determines the vehicle's second trajectory within the parking space area based on the vehicle's real-time pose within the parking space area. This includes: the processing unit acquiring real-time surrounding environmental information of the vehicle within the parking space area; and the processing unit performing one or more gear shifts and one or more curvature changes based on the vehicle's real-time pose and surrounding environmental information within the parking space area, so that the vehicle satisfies a third constraint condition. Based on this, dynamic adjustment of the vehicle's pose and lateral error within the parking space, based on the detected surrounding environmental information, can significantly improve the accuracy of path planning during the parking space hopping phase.

[0039] For example, real-time surrounding environmental information of a vehicle within a parking space can be obtained using ultrasonic radar. It is understood that, in the confined space of a parking space, ultrasonic radar provides more accurate and faster detection of the parking environment compared to other positioning technologies, significantly improving the accuracy of parking space route planning.

[0040] As one possible implementation, the third constraint includes: the angle between the vehicle and the parking space is less than a preset angle; and the distance between the center point of the vehicle and the center point of the parking space is less than a preset distance. Based on this, it can be ensured that the vehicle safely parks in the parking space with a reasonable posture.

[0041] As one possible implementation, the processing unit is also used to: acquire surrounding environmental information during the parking process; and control the vehicle to stop when a collision with a target object is predicted; wherein the surrounding environmental information includes the position, size, direction of movement, and speed of the target object within a preset range of the vehicle. Based on this, obstacles can be accurately avoided during parking, ensuring the safety of the vehicle and its occupants.

[0042] As one possible implementation, during the process of the vehicle moving from its initial pose to the first target state point, the aforementioned surrounding environment information also includes static obstacles within a preset range of the vehicle; the processing unit is further used to update the vehicle's drivable area in real time based on the location information of the static obstacles. Based on this, obstacles can be accurately avoided during the parking process, ensuring the safety of the vehicle and personnel.

[0043] Thirdly, an automatic parking system is provided, the automatic parking system including a communication interface for communication between devices; a memory for storing computer program instructions; and a processor for executing the computer program instructions to support the automatic parking system in implementing the method as in any possible implementation of the first aspect.

[0044] Fourthly, a vehicle is provided that includes an automated parking system as described in any of the second or third aspects. The vehicle is used to implement the method as described in any possible implementation of the first aspect.

[0045] Fifthly, a computer-readable storage medium is provided that stores computer program instructions that, when executed by a processor, implement the method as described in any possible implementation of the first aspect.

[0046] In a sixth aspect, a computer program product containing instructions is provided, which, when run on a computer, causes the computer to implement the method as in any possible implementation of the first aspect.

[0047] In a seventh aspect, a chip system is provided, comprising processing circuitry and a storage medium storing computer program instructions; when executed by the processor, the computer program instructions implement the method as described in any possible implementation of the first aspect. The chip system may be composed of chips or may include chips and other discrete devices. Attached Figure Description

[0048] Figure 1 These are schematic diagrams of several common parking scenarios;

[0049] Figure 2 These are schematic diagrams illustrating two parallel parking scenarios provided in the embodiments of this application;

[0050] Figure 3 Schematic diagrams of parking results under two parallel parking scenarios provided in the embodiments of this application;

[0051] Figure 4 A structural block diagram of an automatic parking system provided in an embodiment of this application;

[0052] Figure 5 A schematic diagram of an automatic parking process provided in an embodiment of this application;

[0053] Figure 6 This application provides a schematic diagram of a gridded area for vehicle driving.

[0054] Figure 7 A schematic diagram illustrating the conditions satisfied by the first pose in the warehousing stage, as provided in an embodiment of this application.

[0055] Figure 8 A schematic diagram illustrating the conditions satisfied by the second pose during the warehousing stage, as provided in an embodiment of this application.

[0056] Figure 9 This is a schematic diagram illustrating a method for determining the first pose during the warehousing stage, as provided in an embodiment of this application.

[0057] Figure 10 This is a schematic diagram illustrating a method for determining the second pose during the warehousing stage, provided in an embodiment of this application.

[0058] Figure 11 This is a schematic diagram of an attitude angle interpolation method provided in an embodiment of this application;

[0059] Figure 12 This is a schematic diagram illustrating a method for obtaining a target region provided in an embodiment of this application;

[0060] Figure 13 A schematic diagram illustrating a path search process using Hybrid A*+RS (Reeds and Shepp) provided in this application embodiment;

[0061] Figure 14 A schematic diagram of the shortest obstacle avoidance distance from the initial pose to the target state point in a two-dimensional discrete space, provided for an embodiment of this application;

[0062] Figure 15 A schematic diagram of the shortest RS curve from the initial pose to the target state point is provided for an embodiment of this application;

[0063] Figure 16 This is a schematic diagram of a kneading process provided in an embodiment of this application. Detailed Implementation

[0064] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. In the description of the embodiments of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B; "and / or" in this text is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the description of the embodiments of this application, "multiple" refers to two or more than two.

[0065] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this embodiment, unless otherwise stated, "a plurality of" means two or more.

[0066] In this embodiment, parallel parking refers to a parking scenario where the parking space and the vehicle are nearly parallel. For example, please refer to... Figure 2 , Figure 2 (a) and Figure 2 (b) shows a schematic diagram of two parallel parking scenarios provided in the embodiments of this application.

[0067] Typically, in parallel parking scenarios, if there is sufficient space outside the parking space, the vehicle can be parked after simple adjustments. For example, in... Figure 2 In the parallel parking scenario shown in (a), vehicles can usually be parked in one go.

[0068] However, in ultra-narrow parallel parking scenarios, vehicles cannot easily maneuver into the parking space. In this embodiment, the ultra-narrow parallel parking scenario satisfies, but is not limited to, one or more of the following conditions: close proximity to an obstacle, a narrow drivable area outside the parking space (e.g., a narrow passageway outside the parking space), and a small parking space. For example, compared to... Figure 2 The scene shown in (a) is as follows. Figure 2 In scenario (b) shown, vehicles are parked both in front of and behind the parking space. When a vehicle attempts to park, these parked vehicles act as obstacles, effectively reducing the vehicle's drivable area and significantly increasing the difficulty of parking. Furthermore, compared to... Figure 2 The scene shown in (a) is as follows. Figure 2 In scenario (b) shown, the aisle outside the parking space is narrow (e.g., aisle width < 3.5m), therefore... Figure 2 As shown in (b), the vehicle needs to repeatedly adjust its gears (such as switching between forward and reverse gears multiple times) and direction, repeatedly "entering" and "exiting" the parking space to fully utilize the space inside and outside the parking space to adjust the vehicle's position and posture (including position and attitude) in order to successfully park. This process makes parking difficult, results in excessively long path search times, and leads to a poor user experience. Furthermore, since each gear shift involves starting and stopping the vehicle, repeated starts and stops introduce greater control errors, leading to a significant positional error between the actual path and the planned path. Therefore, in... Figure 2 In scenario (b) shown, the success rate and reliability of automatic parking are relatively low. Or, in Figure 2 In the scenario shown in (b), the following will occur Figure 3 As shown in (a), when parking is completed, part of the vehicle body is outside the parking space or Figure 3 As shown in (b), the vehicle's posture has a large angle with the center line of the parking space, which affects the user experience.

[0069] To achieve efficient and accurate automatic parking in parallel parking scenarios, especially in ultra-narrow parallel parking scenarios, this application provides an automatic parking method. This method divides the parking process into stages, such as an entry stage and a parking adjustment stage. In the entry stage, the automatic parking system calculates and determines a set of target states that are relatively easy to enter the parking space. Any target state point in this set fully considers the posture that is easy to enter the parking space in ultra-narrow parallel parking scenarios. Therefore, the automatic parking method provided by this application has high entry path search efficiency and high success rate. In the parking adjustment stage, the automatic parking system uses real-time perception of nearby obstacles and makes full use of the drivable area to adjust the vehicle's posture inside the parking space to the target posture.

[0070] The method provided in this application can be implemented based on an automatic parking system on a vehicle. For example, the automatic parking system can be an intelligent control device on the vehicle, or it can be an intelligent control system composed of multiple cooperating modules on the vehicle; this application does not impose any limitations.

[0071] For example, please refer to Figure 4 , Figure 4 A structural block diagram of an automatic parking system provided in an embodiment of this application is shown. Figure 4 As shown, the automatic parking system may include Figure 4 The diagram shows a perception module 410, a localization module 420, a planning module 430, and a control module 440.

[0072] The perception module 410 is used to perceive information about the vehicle's surrounding environment, such as dynamic and static obstacles. Dynamic obstacles include moving vehicles, pedestrians, and animals, while static obstacles include parking lines, stationary vehicles, trees, curbs, railings, streetlights, steps, and potholes. This embodiment does not specifically limit the types of obstacles. The perception module 410 may include one or more sensors, such as a time-of-flight (TOF) sensor, a camera (e.g., an infrared camera), or radar (e.g., millimeter-wave radar, lidar, ultrasonic radar). This embodiment does not specifically limit the types of sensors used.

[0073] The positioning module 420 is used to provide the vehicle's location information.

[0074] In some embodiments, the positioning module 420 may provide the vehicle's absolute position information based on the Global Navigation Satellite System (GNSS).

[0075] For example, GNSS may include, but is not limited to, the Global Positioning System (GPS), the Global Navigation Satellite System (GLONASS), the BeiDou Navigation Satellite System (BDS), the Quasi-Zenith Satellite System (QZSS), and / or satellite-based augmentation systems (SBAS).

[0076] In some embodiments, such as in scenarios without GNSS coverage (e.g., underground parking lots), the positioning module 420 can provide the vehicle's relative position information based on the vehicle's surrounding environment information perceived by the sensing module 410 and the environmental map in the positioning module 420. The environmental map includes a map of the area where the vehicle is currently located.

[0077] For example, the environmental map in the positioning module 420 may be pre-set in the positioning module 420, or it may be obtained by the positioning module 420 from the cloud in real time via the Internet, or it may be drawn and saved in the positioning module 420 by automatic mapping during historical use. This application embodiment does not make specific limitations.

[0078] In other embodiments, the positioning module 420 can provide fused position information of the vehicle based on the vehicle's absolute position information and the vehicle's relative position information.

[0079] The planning module 430 is used to plan parking paths based on data such as the vehicle's surrounding environment information perceived by the perception module 410 and the vehicle's location information provided by the positioning module 420, including determining the target state set in the parking phase, planning the parking trajectory, and adjusting the vehicle's posture in the parking phase.

[0080] In this embodiment, the planning module 430 is further configured to perform speed planning on the parking path based on the planned parking path, the vehicle's initial posture, and the vehicle's motion information, including parking speed planning during the parking entry phase and parking speed planning during the parking maneuver phase.

[0081] The control module 440 controls the vehicle to move to the target position according to the parking path and parking speed planned by the planning module 430. As an example, the automatic parking system may also include... Figure 4The serial bus module 450 is shown. The serial bus module 450 serves as a medium for information and command communication between the control module 440 and the vehicle chassis. Exemplarily, the serial bus module 450 is a controller area network bus (CANBUS).

[0082] In this embodiment, the control module 440 is also used to send vehicle chassis information such as the front wheel steering angle and wheel speed to the planning module 430, so that the planning module 430 can determine the vehicle's initial posture and / or vehicle motion information.

[0083] It should be noted that, Figure 4 The structures illustrated in this application are merely examples of possible structures for an automated parking system and do not constitute a specific limitation on the system. In other embodiments of this application, the automated parking system may include more or fewer components than illustrated, or combine or separate certain components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of both.

[0084] For example, the automatic parking system may also include a prediction module, used to predict the future trajectory of dynamic obstacles in real time based on information about objects (such as dynamic obstacles) in the vehicle's surrounding environment perceived by the perception module 410 during vehicle movement. This allows the vehicle to stop when a collision with the target is predicted, enabling emergency obstacle avoidance and ensuring driving safety. For example, the information about the target may include, but is not limited to, the position, size, direction of movement, and speed of the target within a preset range of the vehicle.

[0085] Alternatively, for example, the automatic parking system may not include a planning module. The function of the planning module can be implemented by other devices with path planning capabilities. The automatic parking system can send vehicle information and information about the vehicle's surrounding environment to other devices with path planning capabilities so that the parking trajectory can be planned by those devices.

[0086] The following will describe in detail, with reference to the accompanying drawings, an automatic parking method provided by an embodiment of this application.

[0087] Please refer to Figure 5 , Figure 5 A schematic diagram of an automatic parking process provided in an embodiment of this application is shown. Figure 5 As shown, the automatic parking process provided in this application embodiment mainly includes S501-S510:

[0088] S501: The automatic parking system acquires vehicle information and information about the vehicle's surrounding environment.

[0089] Vehicle information may include, but is not limited to, vehicle motion information and vehicle initial position.

[0090] Vehicle motion information may include, but is not limited to, information such as vehicle speed and acceleration. The initial vehicle pose is used to characterize the vehicle's initial position and initial attitude. For example, vehicle motion information can be determined based on vehicle chassis information such as vehicle position information and / or wheel speed over a period of time; the vehicle's initial position can be determined based on the vehicle's position information, which can be represented by the vehicle's coordinates in a local coordinate system; the vehicle's initial attitude can be determined based on vehicle chassis information such as the front wheel steering angle or the steering wheel rotation state.

[0091] The information includes information about the vehicle's surrounding environment, such as dynamic and static obstacles. Dynamic obstacles include moving vehicles, pedestrians, and animals, while static obstacles include parking lines, stationary vehicles, trees, curbs, railings, streetlights, steps, and potholes. This application does not impose specific limitations on these aspects.

[0092] by Figure 4 Taking the illustrated automatic parking system as an example, as one possible implementation, the automatic parking system can acquire information about the vehicle's surrounding environment through the perception module 410. For example, the perception module 410 may include one or more sensors, such as a TOF sensor, a camera (e.g., an infrared camera), or radar (e.g., millimeter-wave radar, lidar, ultrasonic radar, etc.). The perception module 410 can detect and acquire information about the vehicle's surrounding environment through at least one sensor.

[0093] by Figure 4 Taking the automatic parking system with the structure shown as an example, as a possible implementation, the automatic parking system can obtain the vehicle's starting position through the positioning module 420 and / or the sensing module 410.

[0094] For example, in scenarios with GNSS coverage (such as outdoor parking lots), an automated parking system can provide the vehicle's starting position based on GNSS, which is the vehicle's absolute position information. Examples of GNSS include GPS, GLONASS, BDS, QZSS, and SBAS.

[0095] For example, in scenarios without GNSS coverage (such as underground parking lots), the automatic parking system can provide the vehicle's starting position based on the vehicle's surrounding environment information perceived by the perception module 410 and the environmental map in the positioning module 420. This starting position is the vehicle's relative position information. The environmental map includes a map of the area where the vehicle is currently located.

[0096] For example, to ensure the accuracy of determining the vehicle's starting position, the automatic parking system can provide fused position information of the vehicle based on its absolute and relative position information. The specific methods and processes for obtaining the vehicle's starting position are not limited in the embodiments of this application, and conventional techniques can be referenced.

[0097] As one possible implementation, an automatic parking system can be implemented through a specific interface (such as...). Figure 4 The serial bus module 450 shown communicates with the vehicle chassis to obtain the vehicle's initial posture, such as the front wheel steering angle and wheel speed, and other vehicle chassis information.

[0098] S502: The automatic parking system determines the parking space boundary information and the drivable area around the parking space based on the surrounding environment information of the vehicle.

[0099] The parking space boundary information is used to characterize the position and size of each side of the parking space (including the inner long side line, outer long side line, front short side line, and rear short side line). For example, the length of the parking space is L (i.e., the size of the inner and outer long side lines is L), and the width is W (i.e., the size of the front and rear short side lines is W).

[0100] As one possible implementation, an automated parking system can grid the parking space and the surrounding drivable area as a drivable area for the vehicle, thereby preparing for subsequent parking path planning.

[0101] As an example, an automated parking system can divide the drivable area of ​​a vehicle in ultra-narrow parallel parking scenarios into units of △L. Figure 6 The discrete space is shown. Where, Figure 6 The area L shown is the drivable area for vehicles, area L1 is the parking space area, and area L2 is the drivable area surrounding the parking space. For example, △L can be 25cm. Figure 6 In the discrete space shown, a1 is the inner long side line, a2 is the outer long side line, b1 is the front short side line, and b2 is the rear short side line.

[0102] It is understandable that when meshing the drivable area of ​​a vehicle, if the basic unit ΔL is too large, the search may fail due to the low discretization precision of the drivable area; however, if the basic unit ΔL is too small, the computational load may be too large due to the high discretization precision of the drivable area, affecting the real-time performance of the algorithm. In the embodiments of this application, the automatic parking system can comprehensively consider one or more of the following factors to select an appropriate ΔL for meshing the drivable area: the computing power of the automatic parking system, the size of the drivable area of ​​the vehicle, the surrounding environment of the vehicle, and user needs.

[0103] It should be noted that the embodiments of this application only use the example of an automatic parking system automatically determining parking space boundary information based on the vehicle's surrounding environment information, but these embodiments do not limit the scope of the application. In other embodiments, the automatic parking system may also determine parking space boundary information based on the area selected by the user on the user interface (UI). In some embodiments, after the user selects an area on the UI, the automatic parking system may also perform a feasibility analysis on the selected area to avoid safety accidents caused by user misoperation.

[0104] S503: The automatic parking system determines the target area when the vehicle enters the parking space. Any target state point of the vehicle within the target area satisfies both the first and second constraints.

[0105] As one possible implementation, an automated parking system can calculate and obtain target state points that satisfy the first and second constraints, respectively. Then, based on the first and second constraints and the state recursion curve during maximum forward steering, it determines the target area for the vehicle to enter the parking space. Therefore, the target area is a set of multiple target state points, and any target state point included therein fully considers the posture conducive to parking in ultra-narrow parallel parking scenarios; that is, any target state point in the target area satisfies both the first and second constraints.

[0106] Taking a parking space located to the right of the vehicle and the vehicle driving on the right side of the road as an example, the automatic parking system can calculate and obtain the state points that satisfy the first constraint and the second constraint respectively. Then, based on the first constraint and the second constraint and the state recursion curve when the vehicle moves forward with the maximum left turn (i.e., the steering wheel is turned to the left all the way), the system can determine the target area when the vehicle enters the parking space.

[0107] Taking a parking space located to the left of the vehicle and the vehicle driving on the left side of the road as an example, the automatic parking system can calculate and obtain the state points that satisfy the first constraint and the second constraint respectively. Then, based on the first constraint and the second constraint and the state recursion curve when the vehicle moves forward with the maximum right turn (i.e., the steering wheel is turned to the right all the way), the system can determine the target area when the vehicle enters the parking space.

[0108] As an example, the first constraint condition ensures that the vehicle can exit the parking space when it enters the parking area and maintains a first curvature while moving forward from the first target state point. The second constraint condition ensures that the vehicle will not cross the edge of the parking space when it enters the parking area and maintains a second curvature while moving forward from the first target state point. In the parking phase of this embodiment, "entering the parking area" means that the vehicle enters the parking area to its maximum extent, such as when the rear of the vehicle does not exceed the critical state of the edge of the parking area.

[0109] As an example, an automatic parking system can determine a first pose based on a first constraint and a second pose based on a second constraint. The first pose represents the limit pose at which the vehicle can exit the parking space while entering the parking area and maintaining a first curvature, while the second pose represents the limit pose at which the vehicle will not cross the edge of the parking space while entering the parking area and maintaining a second curvature.

[0110] For example, the attitude angle of the first pose is the minimum attitude angle. The attitude angle of the second pose is the maximum attitude angle.

[0111] Taking a parking space located on the right side of the vehicle, and the vehicle driving on the right side of the road as an example, such as Figure 7 As shown, the vehicle's attitude angle within the parking space is equal to the minimum attitude angle. At that time, the right rear boundary of the vehicle is tangent to the inner long side line a1 of the parking space, and when the vehicle is at its minimum attitude angle When making a maximum left turn, the vehicle should be able to exit the parking space without crossing its edge (specifically, it should not cross the short side line b1 in front of the parking space). Figure 8 As shown, the vehicle's attitude angle within the parking space is equal to the maximum attitude angle. When turning right at its maximum turning speed, the car will be able to enter the parking space without crossing the short side line b1 in front of the parking space.

[0112] Taking a parking space located to the left of the vehicle and the vehicle traveling on the left side of the road as an example, the vehicle's attitude angle within the parking space area is equal to the minimum attitude angle. At that time, the left rear boundary of the vehicle is tangent to the inner long side line a1 of the parking space, and the vehicle is at its minimum attitude angle. When turning right at its maximum turning angle, the vehicle can just exit the parking space without crossing the short side line b1 in front of the parking space. The vehicle's attitude angle within the parking space area is equal to its maximum attitude angle. When turning left at its maximum turning speed, the car will be able to enter the parking space without crossing the short side line b1 in front of the parking space.

[0113] For ease of understanding, Figure 9 The parking space shown is located to the right of the vehicle. Taking the vehicle driving on the right side of the road as an example, this section introduces how an automatic parking system determines the minimum attitude angle. The specific methods and processes. (Based on...) Figure 10 The parking space shown is located to the right of the vehicle. Taking the vehicle driving on the right side of the road as an example, this section explains how the automatic parking system determines the maximum attitude angle. The specific methods and processes.

[0114] like Figure 9 As shown, L is the length of the parking space, W is the width of the parking space, (x k y k ( ) represents the vehicle at its minimum attitude angle The coordinates of the rear axle center are given, (x0, y0) is the coordinates of the center of the vehicle's turning circle, (x1, y1) is the coordinates of the vertex of the outer edge of the parking space, slightly off-center from the front of the vehicle, and (x2, y2) is the coordinates of the vertex of the front edge of the parking space, slightly off-center from the inside of the parking space. R min D1 is the minimum turning radius of the vehicle, D2 is the distance from the center of the turning circle to the outer edge of the parking space near the front of the vehicle (i.e., coordinates (x1, y1)), and D2 is the distance from the center of the turning circle to the inner edge of the parking space in front of the vehicle (i.e., coordinates (x2, y2)). v d is the width of the vehicle. v For the vehicle length, t v It is the length from the center of the rear axle of the vehicle to the rear edge of the vehicle body.

[0115] based on Figure 9 As shown in the schematic diagram, the automatic parking system can calculate the minimum attitude angle that meets the conditions according to the following formulas 1-11.

[0116]

[0117]

[0118]

[0119]

[0120] x1 = L (Formula 5)

[0121] y1=W(Formula 6)

[0122]

[0123]

[0124]

[0125]

[0126] D2 = D1 (Formula 11).

[0127] Understandable, such as Figure 9 As shown, when the vehicle is at its minimum attitude angle When making the maximum left turn, the vehicle can safely exit the parking space, and the innermost vertex of the front of the vehicle (i.e., coordinates (x2, y2)) will not cross the outer edge of the parking space, near the front vertex of the vehicle (i.e., coordinates (x1, y1)). Therefore, the minimum attitude angle... It can be used to represent the extreme posture of a vehicle that can successfully exit the parking space during the parking phase.

[0128] like Figure 10 As shown, L is the length of the parking space, W is the width of the parking space, (xu y u (This refers to the vehicle being at its maximum attitude angle) The coordinates of the rear axle center are given by: (x0, y0) being the center of the vehicle's turning circle; (x1, y1) being the coordinates of the vertex of the outer edge of the parking space, slightly off-center from the front of the vehicle; and (x2, y2) being the coordinates of the vertex of the front edge of the parking space, slightly off-center from the front of the vehicle. R... min D1 is the minimum turning radius of the vehicle, D2 is the distance from the center of the turning circle to the outer edge of the parking space near the front of the vehicle (i.e., coordinates (x1, y1)), and D2 is the distance from the center of the turning circle to the inner edge of the parking space in front of the vehicle (i.e., coordinates (x2, y2)). v d is the width of the vehicle. v For the vehicle length, t v It is the length from the center of the rear axle of the vehicle to the rear edge of the vehicle body.

[0129] based on Figure 10 As shown in the schematic diagram, the automatic parking system can calculate the maximum attitude angle that meets the conditions according to the following formulas 12-22.

[0130]

[0131]

[0132]

[0133]

[0134] x1 = L (Formula 16)

[0135] y1=W (Formula 17)

[0136]

[0137]

[0138]

[0139]

[0140] D2 = D1 (Formula 22)

[0141] Understandable, such as Figure 10 As shown, when the vehicle is at its maximum attitude angle When turning right at its maximum turning angle, the vehicle can safely enter the parking space, and the innermost vertex of the front of the vehicle (i.e., coordinates (x2, y2)) will not cross the outer edge of the parking space, near the front vertex of the vehicle (i.e., coordinates (x1, y1)). Therefore, the maximum attitude angle... It can be used to ensure the ability of subsequent vehicles to maneuver within the parking space.

[0142] It should be noted that, Figure 9 and Figure 10 Taking the lower left corner of the parking space as the origin of the local coordinate system, with the longer inner side of the parking space as the x-axis and the shorter rear side of the parking space as the y-axis, the minimum attitude angles are shown below. and maximum attitude angle The origin of the local coordinate system can also be set in other ways, and this application embodiment does not limit it.

[0143] Furthermore, after determining the minimum attitude angle that ensures the vehicle can smoothly exit the parking space after entering it, The first posture and the maximum posture angle that can ensure the ability of subsequent vehicles to maneuver within the parking space. After the second pose, it can be understood that the target pose for the vehicle to enter the parking space is to satisfy... (include and For any value of ), the system can ensure both smooth entry and exit of vehicles from and exit from parking spaces, as well as the vehicle's maneuverability within the parking space. Based on this, in this embodiment, the automatic parking system can determine the parking space that meets the following steps 1 and 2. Target area:

[0144] Step 1: Attitude angle interpolation

[0145] Step 1 is used to obtain the parking space that meets the requirements by interpolating between the first pose and the second pose. Multiple state points. Among them, satisfying Any state point can ensure that vehicles can smoothly enter and exit parking spaces while also ensuring the vehicle's maneuverability within the parking space.

[0146] As one possible implementation, such as Figure 11 As shown, the automatic parking system can perform equidistant interpolation between the first and second parking positions to obtain a parking space that meets the requirements. Multiple state points. The interpolation difference scale value can be set by the automatic parking system based on one or more factors such as the system's computing power, the size of the vehicle's drivable area, the vehicle's surrounding environment, and user needs; this embodiment does not impose such limitations.

[0147] Step 2: State Recursion

[0148] Step 2 is used to pass through the state The maximum steering angle of the vehicle at multiple state points is used for state recursion to obtain the state recursion curve when the vehicle is moving forward with maximum steering angle. Any state point on the state recursion curve can ensure both smooth entry and exit from parking spaces and the vehicle's maneuverability within the parking space.

[0149] Taking a parking space located on the right side of the vehicle, with the vehicle driving on the right side of the road, as an example, one possible implementation method is... Figure 12 As shown, the automatic parking system can... The maximum left-turn steering of the vehicle at multiple state points is recursively analyzed to obtain multiple recursive state curves for the maximum left-turn steering forward. Among them, Figure 12 The regions corresponding to the multiple state recursion curves shown are the target regions, and each grid in the target region is a target state point.

[0150] As an example, the target state points in the target region can be used Let (x) represent, where (x) i y i () represents the position of the i-th target state point. This represents the maximum attitude angle (upperbound) of the i-th target state point. It represents the minimum attitude angle (lower bound) of the i-th target state point.

[0151] As one possible implementation, the automatic parking system can uniformly sample the curves within each grid, acquiring the attitude angles of multiple sampling points, and recording the maximum and minimum attitude angles among these sampling points, which are the maximum and minimum attitude angles corresponding to that grid. Taking the i-th target state point as an example, the maximum and minimum attitude angles corresponding to the i-th target state point are respectively... and During the loading phase, when the vehicle reaches the location of the i-th target state point, if the vehicle attitude angle satisfies... This ensures that vehicles can smoothly enter and exit parking spaces while also maintaining their maneuverability within those spaces.

[0152] S504: The automatic parking system uses a path search algorithm to determine the path of the vehicle from its initial pose to the first target state point in the target area, wherein the initial state is determined by the automatic parking system based on the acquired vehicle information.

[0153] For example, the first target state point can be used The first target state point serves as the termination point of the warehousing stage and the starting point of the warehousing kneading stage.

[0154] As an example, the first target state point can be any state point in the target region.

[0155] As an example, an automatic parking system can determine a first target state point from a target area based on a preset strategy. Exemplarily, the preset strategy could be that the first target state point is an intermediate state point between a first pose and a second pose. For example, the first target state point could be an intermediate state point interpolated between the first pose and the second pose. Of course, the preset strategy could be other strategies, and the first target state point could be other state points within the target area; this application does not impose specific limitations.

[0156] As an example, pathfinding algorithms such as the Hybrid A*+RS curve algorithm.

[0157] The automatic parking system employs the Hybrid A*+RS curve algorithm for path planning from the initial pose to the first target state point. This can be achieved by: in discrete space, given the coordinates and attitude angles of the starting point (i.e., the initial pose) and the ending point (i.e., the first target state point); calculating the path cost and heuristic cost for each state point; obtaining the priority of the state point in the path search process based on the path cost and heuristic cost; finding the highest priority state point and calculating the cost of the neighboring state points around it; adding the highest priority state point to the queue for further expansion until the end point (i.e., the first target state point); and then backtracking through the traversed state points to generate a coarse path.

[0158] The path cost is related to the distance the vehicle travels from its initial pose to the current state, the cost of the change in front wheel angle (also known as steering wheel cost), and the cost of gear shifting (also known as gear change cost). For example, the path cost can be a weighted sum of the distance the vehicle travels from its initial pose to the current grid, the steering wheel cost, and the gear shifting cost. Heuristic cost is used to measure the cost incurred from the current grid to the first target state. The values ​​of each component in the path cost can be adjusted according to specific needs, and this application does not impose limitations. For example, if the desired path requires minimal gear shifts, the gear shifting cost needs to be increased; if the desired path is a straight line, the steering wheel rotation cost needs to be increased.

[0159] As one possible implementation, please refer to Figure 13 , Figure 13 This illustration shows a schematic diagram of a path search process using Hybrid A*+RS provided in an embodiment of this application. For example... Figure 13 As shown, the path search process using Hybrid A*+RS can specifically include S1-S5:

[0160] S1: Detect whether the initial pose and the first target state point are within the vehicle's drivable area.

[0161] It is understandable that if the initial pose and / or the first target state point (i.e., the state point the vehicle is about to reach) are not within the vehicle's drivable area, such as on an obstacle, then the Hybrid A*+RS algorithm cannot find a path from the initial pose to the first target state point. Only when both the initial pose and the first target state point (i.e., the state point the vehicle is about to reach) are within the vehicle's drivable area can the vehicle be guaranteed to reach the first target state point. Therefore, S1 is a fundamental prerequisite for using Hybrid A*+RS for path search.

[0162] If the initial pose and / or the first target state point are not within the vehicle's drivable area, the algorithm terminates; if both the initial pose and the first target state point are within the vehicle's drivable area, proceed with steps S2 and subsequent steps.

[0163] S2: Calculate the first type of heuristic value.

[0164] It can be understood that the heuristic value of a state point represents the preference of the path search process for that state point. The smaller the heuristic value, the stronger the preference for that state point (i.e., the higher the priority of that state point), and thus the path exploration for that state point will be prioritized.

[0165] As one possible implementation, the first type of heuristic value can be the shortest distance from each state point in the discrete space to the first target state point in the two-dimensional discrete space, considering obstacle avoidance (or simply the shortest obstacle avoidance distance in the two-dimensional discrete space). For example, please refer to... Figure 14 , Figure 14 This application illustrates an embodiment of the shortest obstacle avoidance distance from an initial pose to a first target state point in a two-dimensional discrete space.

[0166] For example, in the embodiments of this application, the automatic parking system can determine the shortest obstacle avoidance distance between two state points in a two-dimensional discrete space based on algorithms such as Dijkstra's algorithm and R* algorithm, etc. For specific details, conventional techniques can be referred to, and the embodiments of this application do not make specific limitations.

[0167] As another possible implementation, the first type of heuristic value can be a heuristic value calculated from the navigation path, considering obstacle avoidance from each state point in the discrete space to the first target state point. For example, if the vehicle has a pre-stored parking map, the automatic parking system can use the stored parking map to calculate the heuristic value from the navigation path, considering obstacle avoidance from each state point in the discrete space to the first target state point.

[0168] S3: Construct a priority queue, which includes the initial pose.

[0169] For example, a priority queue is an open set.

[0170] The priority of each state point in the priority queue is related to two factors: one is the path cost, which considers the distance the vehicle travels from its initial pose to the current state point, steering wheel costs, gear shifting costs, etc. The other factor is the heuristic cost, or heuristic value, which measures the cost to be incurred from the current state point to the first target state point. The priority of the current state point is determined by weighted summing of its path cost and heuristic cost. For example, a smaller weighted sum means a higher priority for subsequent path searches based on that state point; conversely, a larger weighted sum means a lower priority for subsequent path searches based on that state point.

[0171] S4: Take the highest priority state point (denoted as the i-th state point) from the priority queue.

[0172] Among them, the state point with the highest priority in the priority queue is the state point with the minimum cost after weighted sum of path cost and heuristic cost.

[0173] S5: Determine whether the i-th state point is located in the target area of ​​the parking space.

[0174] The target area is the area defined by the automatic parking system in S503.

[0175] If the i-th state point is located in the target area of ​​the parking space, then execute S13 directly; if the i-th state point is not located in the target area of ​​the parking space, then continue to execute S6.

[0176] S6: Recursively deduce the neighboring state points of the i-th state point according to the action sampling strategy.

[0177] For example, the automatic parking system can use the front wheel turning angle and gear position as motion sampling points, and recursively deduce through the kinematic model to obtain a series of neighboring state points of the i-th state point.

[0178] S7: Calculate the path cost of neighboring state points of state point i.

[0179] The path cost of a nearby state point is related to the cost of the distance traveled from the initial pose to the nearby state point, the cost of the change in front wheel angle (also known as steering wheel cost), and the cost of gear shifting (also known as shifting cost). For example, the path cost of a nearby state point can be obtained by weighted summation of the cost of the distance traveled from the initial pose to the nearby state point, the cost of the change in front wheel angle (also known as steering wheel cost), and the cost of gear shifting (also known as shifting cost).

[0180] S8: Calculate the RS curve between the i-th state point and the first target state point.

[0181] The RS curve is essentially a circle plus a straight line.

[0182] As an example, the RS curve between the i-th state point and the first target state point can be any of the following nine motion mode combinations: C|C|C, CC|C, C|CC, CSC, CC β |C β C、C|C β C β |C、C|C π / 2 SC, CSC π / 2 |C、C|C π / 2SC π / 2 |C. In the above combination of motion modes, C represents a circle, S represents a straight line, "|" indicates that the vehicle's direction of motion changes from forward to reverse or from reverse to forward (i.e., the vehicle shifts gears), the subscript β represents the radian of C's rotation as β, and similarly, the subscript π / 2 represents the radian of C's rotation as π / 2.

[0183] It is understandable that since a circle has two possibilities, such as L for a left turn and R for a right turn; and since both C and S allow forward and reverse movement, L, S, and R can be positive or negative. For example, the above nine combinations of movement modes can be further subdivided into 48 RS curves as shown in Table 1 below:

[0184] Table 1

[0185]

[0186] It is understandable that since the RS curve from the i-th state point to the first target state point can be a combination of many circles and straight lines, many feasible path solutions can be found from the i-th state point to the first target state point. This makes the search easier and has a higher success rate, thus better meeting the needs of vehicle parking. Of course, this application does not specifically limit the specific path search algorithm used.

[0187] S9: Determine whether there is a collision-free RS curve between the i-th state point and the first target state point.

[0188] If there is a collision-free RS curve between the i-th state point and the first target state point, execute S13 directly; if there is no collision-free RS curve between the i-th state point and the first target state point, continue executing S10.

[0189] S10: Calculate the second type of heuristic value for the i-th state point.

[0190] The second type of heuristic value (H) consists of two parts: H euc and H RS Among them, H eucH is a heuristic value based on the shortest obstacle avoidance distance in a two-dimensional discrete space. RS It is a heuristic value based on the RS curve.

[0191] As an example, the second type of heuristic value H = W euc H euc +H RS .

[0192] Among them, H euc H is the shortest obstacle avoidance distance from the i-th state point in the discrete space to the first target state point in the two-dimensional discrete space. RS It can be calculated based on the following formula 23:

[0193] H RS = (1+W) kappa N kappa +W gear N gear +W other N other )L RS (Formula 23).

[0194] In formula 23 above, L RS This represents the length of the shortest RS curve from the i-th state point to the first target state point, i.e. Figure 15 The length of curve 1; N kappa This represents the number of curvature changes from the i-th state point to the first target state point (e.g., Figure 15 In the example shown, the curvature changes 2 times; N gear This represents the number of gear shifts from the i-th state point to the first target state point (e.g., Figure 15 In the example shown, the number of gear changes is 1; N other Representing other factors that can be considered (the automatic parking system can be expanded according to the actual situation, formula 23), W kappa W gear and W other The weights are respectively for curvature change factor, shift factor and other factors.

[0195] It is understandable that while the first type of heuristic value roughly measures which direction will lead to the first target state point from the initial pose more quickly, this method treats the vehicle as a point mass and does not consider the vehicle's motion model, thus only providing a coarse directional guide. To provide a path search adapted to the vehicle's motion model, this application introduces a heuristic function design based on RS curves. This fully considers the impact of vehicle gear shifting and curvature changes on path search in ultra-narrow parallel parking scenarios, making the calculation of heuristic values ​​more reasonable and efficient, and greatly improving the efficiency and success rate of path search.

[0196] S11: Determine whether the i-th state point is in the priority queue.

[0197] If the i-th state point is in the priority queue, increment i by 1 and continue searching for the next state point according to the methods shown in S5-S11; if the i-th state point is not in the priority queue, continue executing S12.

[0198] S12: Add the i-th state point to the priority queue.

[0199] S13: Determine whether the i-th state point is the first target state point.

[0200] If the i-th state point is the first target state point, the automatic parking system continues to execute S14.

[0201] If the i-th state point is not the first target state point, the automatic parking system increments i by 1 and continues to search for the next state point according to the methods shown in S3-S11 until the first target state point is found.

[0202] It should be noted that, as a possible scenario, if the automatic parking system has completed the search of all state points in the priority queue but still has not found the first target state point, then the path search for the first target state point can be considered a failure. In this case, the automatic parking system can retry the path search, or it can change the target state point and then re-search the path, or it can perform other preset processing.

[0203] S14: Draw the first path.

[0204] As one possible implementation, the automatic parking system can first trace back from the first target state point to the searched state points to obtain the tracing path, and then divide the tracing path into segments according to the gear position of each state point on the tracing path to obtain the first path.

[0205] It is understandable that the curvature of the first path searched using Hybrid A*+RS is usually discontinuous, and therefore can only be considered a coarse path. To ensure the path curvature is within the vehicle's following capability and improve its tracing ability, in this embodiment, the automatic parking system can smooth the path from the vehicle's initial pose to the first target state point, ensuring the parking path curvature is within the vehicle's following capability, thereby improving path tracing ability. Specifically, the automatic parking system can execute the following S505.

[0206] S505: The automatic parking system smooths the path of the vehicle from its initial pose to the first target state point to obtain a second path.

[0207] It is understandable that since the state points in the coarse path searched by Hybrid A*+RS contain different gears, the subsequent path smoothing needs to ensure that the gears in the same path are consistent. Therefore, before smoothing, the automatic parking system can first perform path segmentation on the coarse path, dividing the state points of the same gear into one path.

[0208] As one possible implementation, the automatic parking system can employ nonlinear optimization methods to smooth out coarse path points along the vehicle's path from its initial pose to the first target state point. For example, the automatic parking system can combine curvature constraints and / or obstacle avoidance conditions to smooth the path from the initial pose to the first target state point.

[0209] As an example, a curvature constraint condition is provided, such as the curvature of any state point on the second path cannot exceed the vehicle's steering capability. It is understood that the vehicle's minimum radius of curvature determines its steering capability. Based on this, in this embodiment, the automatic parking system can use curvature constraints as constraints to smooth the path from the initial pose to the first target state point.

[0210] As an example, obstacle avoidance conditions can be determined based on an accurate obstacle avoidance optimization model, such as a convex optimization model (e.g., sequential convex optimization). The basic idea of ​​convex optimization is to use the results of a coarse path search as initial values ​​to determine local optima near the coarse path. For a detailed introduction to accurate obstacle avoidance optimization models, such as convex optimization models, please refer to conventional techniques; this application's embodiments will not elaborate further.

[0211] S506: The automatic parking system determines the parking speed strategy corresponding to the second path and obtains the first trajectory (the first trajectory is the parking trajectory).

[0212] Among them, the automatic parking system determines the entry speed strategy corresponding to the second path to ensure the effectiveness and comfort of path following.

[0213] As an example, the inbound speed strategy for the second path has the following characteristics: the starting speed and ending speed of the second path are both 0. This is to ensure the effectiveness of path following.

[0214] As an example, the entry speed strategy for the second path has the following characteristics: the acceleration (jerk) value of the movement speed is less than a preset value. This is to ensure the comfort of path following.

[0215] For the specific planning methods and processes of speed strategy in the path planning process, conventional techniques can also be referenced, and the embodiments of this application do not impose specific limitations.

[0216] S507: The automatic parking system instructs the vehicle to travel along the planned first trajectory to the first target state point.

[0217] As one possible implementation, the planning module of an automated parking system (such as...) Figure 4 The planning module 430 shown can send data to the control module (such as...). Figure 4 The control module 440 shown sends an entry start command to instruct the control module to control the vehicle to travel to the first target state point according to the planned first trajectory.

[0218] Furthermore, as one possible implementation, the control module of the automatic parking system (such as...) Figure 4 The control module 440 shown can be accessed through a specific interface (such as...). Figure 4 The serial bus module 450 shown sends control commands to the vehicle chassis, causing the vehicle to travel to the first target state point according to the planned first trajectory.

[0219] S508: The automatic parking system determines the third path of the vehicle within the parking space area based on the vehicle's real-time position and orientation (where the third path is the swiping path).

[0220] During the parking maneuvering phase, the automatic parking system's main task is to adjust the vehicle's posture and lateral error within the parking space.

[0221] For example, during the parking maneuvering phase, the automatic parking system can perform one or more gear shifts and one or more curvature changes based on the vehicle's real-time position and orientation within the parking space and real-time surrounding environment information, so that the vehicle meets the third constraint condition.

[0222] For example, the third constraint includes: the vehicle's posture is approximately parallel to the parking space (e.g., the angle between the long side of the vehicle and the long side of the parking space is less than a first preset angle); and the vehicle is approximately located at the center of the parking space (e.g., the distance between the center point of the vehicle and the center point of the parking space is less than a first preset distance). As an example, the parking space manipulation stage mainly includes the following two stages:

[0223] Forward Phase: The forward phase is mainly used to adjust the vehicle's posture. For example, an automatic parking system can set the front wheel steering angle to its maximum value and move forward until the vehicle boundary reaches the safe driving area boundary within the parking space or the vehicle posture is approximately parallel to the parking space.

[0224] Reverse segment: The reverse segment is mainly used to adjust for lateral errors. For example, an automatic parking system can calculate the vehicle's reverse path by using tangent circular arcs, so that the vehicle reverses to the boundary of the drivable area.

[0225] In some embodiments, during the parking maneuvering phase, the automatic parking system is subject to a preset number of gear shifts when planning the parking maneuvering path for the forward and reverse segments. For example, the number of gear shifts cannot be too many, so as to ensure that the operation during the parking maneuvering phase is not too complicated.

[0226] As an example, please refer to Figure 16 , Figure 16 A schematic diagram of a kneading process provided in an embodiment of this application is shown. Figure 16 As shown, during the vehicle's forward movement, the automatic parking system can plan a gliding path to allow the vehicle to move forward. If, before reaching the safe driving zone boundary, the system detects that the vehicle's posture is parallel to the long side of the parking space, the system can continue planning the gliding path based on the current posture, causing the vehicle to approximately adjust to the center of the parking space. If, before reaching the safe driving zone boundary, the system detects that the vehicle's posture is not approximately parallel to the long side of the parking space, the system can continue planning the gliding path, allowing the vehicle to maintain its current posture and continue moving forward. Subsequently, if, before reaching the safe driving zone boundary, the system detects that the vehicle's posture is not approximately parallel to the long side of the parking space, the system continues planning the gliding path to allow the vehicle to reverse; if, before reaching the safe driving zone boundary, the system detects that the vehicle's posture is parallel to the long side of the parking space, the system continues planning the gliding path, causing the vehicle to approximately adjust to the center of the parking space. Furthermore, during the reversing phase, if the vehicle's posture is detected to be approximately parallel to the long side of the parking space before reaching the boundary of the safe driving area, the automatic parking system can continue to plan the gliding path based on the current posture, so that the vehicle is approximately adjusted to the center of the parking space. If the vehicle's posture is not detected to be approximately parallel to the long side of the parking space before reaching the boundary of the safe driving area, the automatic parking system can determine whether the number of gear shifts exceeds the preset number of gear shifts. If the number of gear shifts exceeds the preset number of gear shifts, the algorithm ends; if the number of gear shifts does not exceed the preset number of gear shifts, the automatic parking system can continue to plan the gliding path, repeating the gliding path planning of the forward and reverse phases.

[0227] It should be noted that, Figure 16 This application serves only as an illustration of the parking space shunting process and does not specify the concrete implementation details of the shunting path planning process. For example, in other examples, if the number of gear shifts exceeds a preset number of gear shifts, but the vehicle's posture is not approximately parallel to the long side of the parking space and / or not approximately located at the center of the parking space, the automatic parking system can relax the criteria for approximately parallel and / or approximately located at the center of the parking space. For example, the angle between the long side of the vehicle and the long side of the parking space can be less than a second preset angle, and the distance between the center point of the vehicle and the center point of the parking space can be less than a second preset distance. Wherein, the second preset angle is greater than the first preset angle, and the second preset distance is greater than the first preset distance.

[0228] As an example, during the parking space selection phase, the automatic parking system can adjust the vehicle's posture and lateral error within the parking space based on information about the surrounding environment detected by ultrasonic radar. It's understandable that, since ultrasonic radar has high accuracy within a detection range of 0.3-5 meters, dynamically adjusting the vehicle's posture and lateral error based on this information during the selection phase can significantly improve the accuracy of the parking space selection path planning.

[0229] For other related information regarding the kneading stage, please refer to conventional techniques; the embodiments in this application do not impose specific limitations.

[0230] S509: The automatic parking system determines the speed strategy corresponding to the third path and obtains the second trajectory.

[0231] Among them, the automatic parking system determines the speed strategy corresponding to the third path to ensure the effectiveness and comfort of path following.

[0232] As an example, the speed strategy corresponding to the third path has the following characteristics: the starting speed and ending speed of the forward segment are both 0, and the starting speed and ending speed of the backward segment are both 0. This is to ensure the path following effect.

[0233] As an example, the speed strategy corresponding to the third path has the following characteristics: the acceleration (jerk) value of the movement speed is less than a preset value. This is to ensure the comfort of path following.

[0234] For the specific planning methods and processes of speed strategy in the path planning process, conventional techniques can also be referenced, and the embodiments of this application do not impose specific limitations.

[0235] S510: The automatic parking system instructs the vehicle to park in a parking space according to a planned second trajectory.

[0236] As one possible implementation, the planning module of an automated parking system (such as...) Figure 4 The planning module 430 shown can send data to the control module (such as...). Figure 4 The control module 440 shown sends a parking start command to instruct the control module to control the vehicle to park in the parking space according to the planned second trajectory.

[0237] Furthermore, as one possible implementation, the control module of the automatic parking system (such as...) Figure 4 The control module 440 shown can be accessed through a specific interface (such as...). Figure 4 The serial bus module 450 shown sends control commands to the vehicle chassis, causing the vehicle to park in the parking space according to the planned second trajectory.

[0238] In some examples, after the vehicle has successfully parked in the parking space, the automatic parking system can also send a parking end command to the control module to instruct the control module to end control of the vehicle's movement. For example, the automatic parking system can detect the relative positional relationship between the vehicle and the parking space in real time during the vehicle's movement, and send a parking end command to the control module when it determines that the relative positional relationship between the vehicle and the parking space satisfies the relative positional relationship at the termination of the third path.

[0239] It should be noted that, Figure 5 The flowchart shown is merely an example, and the embodiments of this application do not limit the timing between specific trajectory planning steps during automatic parking. For example, the automatic parking system may first plan the entry path and the rubbing path, then determine the entry speed strategy corresponding to the entry path and the rubbing speed strategy corresponding to the rubbing path, and finally instruct the vehicle to park in the parking space according to the planned entry path and rubbing path, based on the corresponding speed strategy.

[0240] It is understood that, based on the automatic parking method provided in this application embodiment, the automatic parking system processes the parking process in stages during automatic parking trajectory planning, such as the entry stage and the caching stage, fully combining entry and caching capabilities to improve the path search efficiency and success rate of automatic parking. Specifically, in the entry stage, the automatic parking system calculates and determines a set of target states that are relatively easy to enter the parking space. Any target state point in this set fully considers the posture that is easy to enter in ultra-narrow parallel parking scenarios, thus obtaining a better entry target pose, thereby reducing the difficulty of entering the parking space and improving the path search efficiency and success rate of the entry stage. In the caching stage, the automatic parking system, through real-time perception of nearby obstacles, fully utilizes the drivable area to adjust the vehicle's pose within the parking space to the target pose.

[0241] It should be understood that the various solutions in the embodiments of this application can be used in a reasonable combination, and the explanations or descriptions of the various terms appearing in the embodiments can be referenced or explained to each other in the various embodiments, without limitation.

[0242] It should also be understood that, in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0243] It should also be understood that, in various embodiments of this application, the vehicle may integrate a device or unit with path planning functionality. Alternatively, the automatic parking system described in the embodiments of this application may be a device or unit with path planning functionality. Alternatively, the automatic parking system described in the embodiments of this application may establish a communication connection with a device or unit with path planning functionality, sending vehicle information and surrounding environment information to the device or unit with path planning functionality, so as to rely on the device or unit with path planning functionality to complete parking path planning.

[0244] It is understood that, in order to achieve the functions of any of the above embodiments, the automatic parking system includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0245] This application embodiment can divide the automatic parking system into functional modules. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0246] It should also be understood that the various modules in an automatic parking system can be implemented in software and / or hardware, without specific limitations. In other words, electronic devices are presented in the form of functional modules. Here, "module" can refer to application-specific integrated circuits (ASICs), circuits, processors and memories that execute one or more software or firmware programs, integrated logic circuits, and / or other devices that can provide the above functions.

[0247] In an alternative approach, when data transmission is implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are implemented. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disk (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0248] The steps of the methods or algorithms described in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in RAM, flash memory, ROM, EPROM, EEPROM, registers, hard disk, portable hard disk, CD-ROM, or any other form of storage medium known in the art. One exemplary embodiment couples a storage medium to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in an electronic device. Of course, the processor and storage medium can also exist as discrete components in an automatic parking system.

[0249] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

Claims

1. An automatic parking method, characterized in that, The method includes: Obtain the initial vehicle pose and the vehicle's drivable area, which includes the parking space area; The first pose is determined according to the first constraint condition. The first pose is used to characterize the limit pose that the vehicle can drive out of the parking space area when it enters the parking space area and maintains the first curvature. A second pose is determined based on a second constraint condition. The second pose is used to characterize the limit pose of the vehicle when it enters the parking space area and maintains the second curvature while moving forward without crossing the edge of the parking space area. The target area when the vehicle enters the parking space area is determined based on the first pose and the second pose, and the target area includes multiple target state points; A first target state point is determined from the target area. The first target state point is used to characterize the position and posture of the vehicle when it enters the parking space area from the initial pose. The first target state point is any state point in the target area. Plan the first trajectory from the initial pose to the first target state point; After the vehicle reaches the first target state point from the initial pose, the second trajectory of the vehicle in the parking space area is determined based on the real-time pose of the vehicle in the parking space area. The first constraint condition is used to ensure that the vehicle can drive out of the parking space area when it enters the parking space area and moves forward with a first curvature from the first target state point, and the second constraint condition is used to ensure that the vehicle will not cross the edge of the parking space area when it enters the parking space area and moves forward with a second curvature from the first target state point.

2. The method according to claim 1, characterized in that, Determining the first target state point from the target region includes: The first target state point is determined from the target region according to a preset strategy.

3. The method according to claim 2, characterized in that, Determining the target area when the vehicle enters the parking space area based on the first pose and the second pose includes: The target area is determined based on the trajectories traversed by the vehicle as it moves forward while maintaining a first curvature from the first pose and the second pose, respectively.

4. The method according to any one of claims 1-3, characterized in that, The planning of the first trajectory from the initial pose to the first target state point includes: Determine a first path from the initial pose to the first target state point; The first path is smoothed to obtain the second path; The first trajectory is obtained by planning the inbound speed of the second path.

5. The method according to claim 4, characterized in that, Determining a coarse path from the initial pose to the first target state point includes: A path search algorithm is used to determine the coarse path from the initial pose to the first target state point.

6. The method according to claim 5, characterized in that, The step of using a path search algorithm to determine the coarse path from the initial pose to the first target state point includes: using a path search algorithm to determine the coarse path from the initial pose to the first target state point based on first information, wherein the first information includes the shortest distance for obstacle avoidance in a two-dimensional discrete space between the state points traversed by the vehicle and the first target state point.

7. The method according to claim 6, characterized in that, The step of using a path search algorithm to determine the coarse path from the initial pose to the first target state point further includes: using a path search algorithm to determine the coarse path from the initial pose to the first target state point based on second information, wherein the second information includes one or more of the following: the number of curvature changes of the vehicle, the number of gear shifts of the vehicle, and the length of the shortest RS curve between the initial pose and the first target state point.

8. The method according to any one of claims 1-3 and 5-7, characterized in that, When the vehicle is in the initial position, the parking space area is located to the right of the vehicle; wherein, the first curvature corresponds to the motion curvature when the vehicle makes its maximum left turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum right turn; or, When the vehicle is in the initial pose, the parking space area is located to the left of the vehicle; wherein, the first curvature corresponds to the motion curvature when the vehicle makes its maximum right turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum left turn.

9. The method according to any one of claims 1-3 and 5-7, characterized in that, The step of determining the second trajectory of the vehicle within the parking space area based on the real-time pose of the vehicle within the parking space area includes: Obtain real-time surrounding environmental information of the vehicle within the parking space area; Based on the real-time pose of the vehicle within the parking space area and the real-time surrounding environment information, the vehicle performs one or more gear shifts and one or more curvature changes to satisfy the third constraint condition.

10. The method according to claim 9, characterized in that, The third constraint includes: The angle between the vehicle and the parking space is less than a preset angle; and... The distance between the center point of the vehicle and the center point of the parking space is less than a preset distance.

11. The method according to claim 9, characterized in that, The step of obtaining real-time surrounding environment information of the vehicle within the parking space area includes: The vehicle acquires real-time surrounding environmental information within the parking space area using ultrasonic radar.

12. The method according to any one of claims 1-3, 5-7, and 10-11, characterized in that, The method further includes: During the process of parking the vehicle into the parking space, the surrounding environment information is acquired, including the position, size, direction of movement and speed of the target object within the preset range of the vehicle; When it is predicted that the vehicle is about to collide with the target object, the vehicle is controlled to stop.

13. The method according to claim 12, characterized in that, During the process of the vehicle moving from the initial pose to the first target state point, the surrounding environment information also includes static obstacles within a preset range of the vehicle; The method further includes: The vehicle's drivable area is updated in real time based on the location information of the static obstacles.

14. An automatic parking system, characterized in that, The automatic parking system includes: The data acquisition unit is used to acquire the initial position and orientation of the vehicle and the drivable area of ​​the vehicle, wherein the drivable area of ​​the vehicle includes the parking space area. The processing unit is configured to determine a first pose based on a first constraint condition, determine a second pose based on a second constraint condition, determine a target area when the vehicle enters the parking space area based on the first pose and the second pose, determine a first target state point from the target area, and plan a first trajectory from the initial pose to the first target state point; and, after the vehicle reaches the first target state point from the initial pose, determine a second trajectory of the vehicle within the parking space area based on the real-time pose of the vehicle within the parking space area; wherein, the first pose is used to characterize the limit pose at which the vehicle can drive out of the parking space area while entering the parking space area and maintaining a first curvature, and the second pose is used to characterize the limit pose at which the vehicle will not cross the edge of the parking space area while entering the parking space area and maintaining a second curvature, and the target area includes multiple target state points, wherein the first target state point is any state point within the target area. Wherein, the first target state point is used to characterize the position and posture of the vehicle when it enters the parking space area from the initial pose; the first constraint condition is used to ensure that the vehicle can drive out of the parking space area when it enters the parking space area and moves forward with a first curvature from the first target state point; the second constraint condition is used to ensure that the vehicle will not cross the edge of the parking space area when it enters the parking space area and moves forward with a second curvature from the first target state point.

15. The system according to claim 14, characterized in that, The data acquisition unit includes: The sensing unit is used to collect surrounding environmental data to obtain the drivable area of ​​the vehicle.

16. The system according to claim 15, characterized in that, The processing unit determines the first target state point from the target region, including: The processing unit determines the first target state point from the target region according to a preset strategy.

17. The system according to claim 16, characterized in that, The processing unit determines the target area when the vehicle enters the parking space area based on the first pose and the second pose, including: The processing unit determines the target area based on the trajectories traversed by the vehicle as it moves forward while maintaining a first curvature from the first pose and the second pose, respectively.

18. The system according to any one of claims 14-17, characterized in that, The processing unit plans a first trajectory from the initial pose to the first target state point, including: The processing unit determines a first path from the initial pose to the first target state point; The processing unit performs smoothing processing on the first path to obtain the second path; The processing unit performs inbound speed planning on the second path to obtain the first trajectory.

19. The system according to claim 18, characterized in that, The processing unit determines a coarse path from the initial pose to the first target state point, including: The processing unit uses a path search algorithm to determine the coarse path from the initial pose to the first target state point.

20. The system according to claim 19, characterized in that, The processing unit uses a path search algorithm to determine the coarse path from the initial pose to the first target state point, including: the processing unit uses the path search algorithm to determine the coarse path from the initial pose to the first target state point based on first information, the first information including the shortest obstacle avoidance distance in a two-dimensional discrete space between the state point traversed by the vehicle and the first target state point.

21. The system according to claim 20, characterized in that, The processing unit uses a path search algorithm to determine the coarse path from the initial pose to the first target state point, including: the processing unit uses the path search algorithm to determine the coarse path from the initial pose to the first target state point based on second information, the second information including one or more of the following: the number of curvature changes of the vehicle, the number of gear shifts of the vehicle, and the length of the shortest RS curve between the initial pose and the first target state point.

22. The system according to any one of claims 14-17 and 19-21, characterized in that, When the vehicle is in the initial position, the parking space area is located to the right of the vehicle; wherein, the first curvature corresponds to the motion curvature when the vehicle makes its maximum left turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum right turn; or, When the vehicle is in the initial pose, the parking space area is located to the left of the vehicle; wherein, the first curvature corresponds to the motion curvature when the vehicle makes its maximum right turn, and the second curvature corresponds to the motion curvature when the vehicle makes its maximum left turn.

23. The system according to any one of claims 14-17 and 19-21, characterized in that, The processing unit determines the second trajectory of the vehicle within the parking space area based on the real-time pose of the vehicle within the parking space, including: The processing unit acquires real-time surrounding environmental information of the vehicle within the parking space area; The processing unit performs one or more gear shifts and one or more curvature changes based on the real-time pose of the vehicle within the parking space area and the real-time surrounding environment information, so that the vehicle satisfies the third constraint condition.

24. The system according to claim 23, characterized in that, The third constraint includes: The angle between the vehicle and the parking space is less than a preset angle; and... The distance between the center point of the vehicle and the center point of the parking space is less than a preset distance.

25. The system according to claim 23 or 24, characterized in that, The processing unit acquires real-time surrounding environment information of the vehicle within the parking space area, including: The processing unit acquires real-time surrounding environment information of the vehicle within the parking space area based on ultrasonic radar.

26. The system according to any one of claims 14-17, 19-21, and 24, characterized in that, The processing unit is also used for: During the process of parking the vehicle into the parking space, information about the surrounding environment is acquired; and when it is predicted that the vehicle is about to collide with the target object, the vehicle is controlled to stop. The surrounding environment information includes the position, size, direction of movement, and speed of the target object within the preset range of the vehicle.

27. The system according to claim 26, characterized in that, During the process of the vehicle moving from the initial pose to the first target state point, the surrounding environment information also includes static obstacles within a preset range of the vehicle; the processing unit is further configured to: The vehicle's drivable area is updated in real time based on the location information of the static obstacles.

28. A vehicle, characterized in that, The vehicle includes an automatic parking system as described in any one of claims 14-27.

29. A readable storage medium, characterized in that, The readable storage medium stores program instructions that, when executed by the device, implement the method as described in any one of claims 1-13.

30. A program product containing instructions, characterized in that, When the program product is run on the device, the device performs the method as described in any one of claims 1-13.

31. A chip system, characterized in that, The chip system includes a processing circuit and a storage medium, wherein the storage medium stores program instructions; when the program instructions are executed by the processing circuit, they implement the method as described in any one of claims 1-13.

Images