Vehicle and parking control method and system

Abstract

This application provides a vehicle and parking control method and system. The method includes: in response to receiving a parking command, acquiring vehicle perception data and state data; determining the target parking space for the vehicle and the relative pose of the vehicle and the target parking space based on the perception data and state data; generating a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type and the relative pose of the target parking space; and controlling the vehicle to enter the target parking space based on the target path and the set of target deflection angles corresponding to the multiple path points. This application solves the technical problem of complex vehicle parking control in related technologies.

Description

Technical Field

[0001] This application relates to the fields of vehicles and parking, and more specifically, to a vehicle and parking control method and system. Background Technology

[0002] With the continuous increase in vehicle ownership, parking space is becoming increasingly scarce, especially in urban environments where narrow parking spaces place high demands on driver skills. Therefore, automatic parking has become a necessity for intelligent vehicles, and users are increasingly demanding an efficient, space-saving, and low-operation parking experience.

[0003] Currently, automatic parking methods often rely on high-precision maps, design multiple paths, and repeatedly perform forward-backward and gear-shifting adjustments, which are cumbersome and time-consuming, resulting in a relatively complex parking control of the vehicle.

[0004] There is currently no good solution to the above problems. Summary of the Invention

[0005] This application provides a vehicle and parking control method and system to at least solve the technical problem of complex vehicle parking control in related technologies.

[0006] According to one aspect of the embodiments of this application, a parking control method for a vehicle is provided, comprising: in response to receiving a parking instruction, acquiring perception data and state data of the vehicle; determining a target parking space for the vehicle and the relative pose of the vehicle and the target parking space based on the perception data and state data; generating a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type and the relative pose of the target parking space, wherein the set of target deflection angles includes the deflection angles of multiple wheels of the vehicle; and controlling the vehicle to drive into the target parking space based on the target path and the set of target deflection angles corresponding to the multiple path points.

[0007] Optionally, based on the vehicle's perception data and state data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined, including: identifying the perception data to determine the target parking space; and determining the relative pose of the vehicle and the target parking space based on the state data and the position and orientation of the target parking space.

[0008] Optionally, based on the parking space type and relative pose of the target parking space, a target path and a set of target deflection angles corresponding to multiple path points in the target path are generated, including: determining the target pose of the vehicle based on the position and orientation of the target parking space, wherein the target pose is used to characterize the pose of the vehicle parked in the target parking space; when the parking space type is a perpendicular parking space, a first arc path segment is generated based on the relative pose and the target pose, the first arc path segment is used as the target path, and a set of target deflection angles corresponding to multiple path points is determined based on the first arc path segment; when the parking space type is a parallel parking space, a second arc path segment and a third arc path segment are generated based on the relative pose, the preset position corresponding to the target parking space, and the target pose, and a target path and a set of target deflection angles corresponding to multiple path points are generated based on the second arc path segment and the third arc path segment, wherein the second arc path segment and the third arc path segment are tangent at a preset position.

[0009] Optionally, based on the first circular arc path segment, determining the target deflection angle set corresponding to multiple path points includes: determining the vehicle's pose at multiple path points based on the positions of multiple path points on the first circular arc path segment; determining the target deflection angle set corresponding to multiple path points based on the first center of the first circular arc path segment and the vehicle's pose at multiple path points; preferably, generating the first circular arc path segment based on the relative pose and the target pose includes: determining the first center and the first radius based on the relative pose and the target pose; generating the first circular arc path segment based on the first center and the first radius.

[0010] Optionally, based on the second and third circular arc path segments, a set of target deflection angles corresponding to multiple path points is generated, including: determining the vehicle's pose at multiple first path points based on the positions of multiple first path points on the second circular arc path segment, and determining the vehicle's pose at multiple second path points based on the positions of multiple second path points on the third circular arc path segment; determining the target deflection angles corresponding to multiple path points based on the second center of the second circular arc path segment, the vehicle's pose at multiple first path points, the third center of the third circular arc path segment, and the vehicle's pose at multiple second path points. The set, wherein the target deflection angle sets corresponding to multiple first path points have the same direction; preferably, based on the relative pose, the preset position corresponding to the target parking space, and the target pose, a second arc path segment and a third arc path segment are generated, including: based on the preset position, determining the preset pose of the vehicle arriving at the preset position; based on the relative pose and the preset pose, determining the second center and the second radius, and based on the preset pose and the target pose, determining the third center and the third radius; based on the second center and the second radius, generating the second arc path segment, and based on the third center and the third radius, generating the third arc path segment.

[0011] Optionally, based on the target path and the target deflection angle set corresponding to multiple path points, the vehicle is controlled to enter the target parking space, including: if the target path includes a first arc path segment, the vehicle is controlled to enter the target parking space based on the target deflection angle set corresponding to multiple path points; if the target path includes a second arc path segment and a third arc path segment, the vehicle is controlled to drive to a preset position of the target parking space based on the target deflection angle set corresponding to multiple first path points on the second arc path segment, and the vehicle is controlled to enter the target parking space based on the target deflection angle set corresponding to multiple second path points on the third arc path segment.

[0012] According to another aspect of the embodiments of this application, a vehicle parking control system is also provided, including: an environmental perception module for acquiring vehicle perception data; a state sensing module for acquiring vehicle state data; a processor for responding to receiving a parking instruction, determining the target parking space of the vehicle and the relative pose of the vehicle and the target parking space based on the perception data and the state data; generating a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type and the relative pose of the target parking space, wherein the set of target deflection angles includes the deflection angles of multiple wheels of the vehicle; and controlling the vehicle to drive into the target parking space based on the target path and the set of target deflection angles corresponding to multiple path points.

[0013] Optionally, the system may further include a chassis execution system for controlling vehicle movement and steering of multiple wheels based on control commands generated by the processor, wherein the control commands are generated based on a target path and a set of target deflection angles corresponding to multiple path points.

[0014] According to another aspect of the embodiments of this application, a vehicle is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0015] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0017] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods in various embodiments of this application.

[0018] According to another aspect of the embodiments of this application, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this application.

[0019] In this embodiment of the application, in response to receiving a parking instruction, after acquiring the vehicle's perception data and status data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined based on the perception data and status data. Then, based on the parking space type and relative pose of the target parking space, a target path and a set of target deflection angles corresponding to multiple path points in the target path are generated. Thus, based on the target path and the set of target deflection angles corresponding to multiple path points, the vehicle is controlled to drive into the target parking space. Upon responding to a parking command, this application dynamically identifies the target parking space through environmental perception and vehicle status data, and determines the precise relative pose between the vehicle and the target parking space. Based on different target parking space types and relative poses, it simultaneously constructs a target path and calculates the target deflection angle set for each wheel's coordinated deflection. Using this target deflection angle set, it can actively control the vehicle to park along the target path, achieving coordinated control of the target path and vehicle steering. This avoids the complexity of repeatedly and passively adjusting the vehicle to conform to the path during parking, thus simplifying the parking process and solving the technical problem of complex vehicle parking control in related technologies. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0021] Figure 1 This is a flowchart of a vehicle parking control method according to an embodiment of this application;

[0022] Figure 2 This is a schematic diagram of an optional parking path for a vertical parking space according to an embodiment of this application;

[0023] Figure 3 It is a schematic diagram of the parking path for related vehicles;

[0024] Figure 4 This is a schematic diagram of an optional parallel parking path according to an embodiment of this application;

[0025] Figure 5 This is a schematic diagram of a vehicle parking control system according to an embodiment of this application;

[0026] Figure 6This is a schematic diagram of an optional vehicle parking control system according to an embodiment of this application;

[0027] Figure 7 This is a schematic diagram of a vehicle parking control device according to an embodiment of this application;

[0028] Figure 8 This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] According to an embodiment of this application, a parking control method for a vehicle is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0032] This embodiment provides a vehicle parking control method. Figure 1 This is a flowchart of a vehicle parking control method according to an embodiment of this application, such as... Figure 1 As shown, the process includes the following steps:

[0033] In step S102, in response to receiving a parking instruction, the vehicle's perception data and status data are acquired.

[0034] The aforementioned vehicle can refer to an intelligent vehicle equipped with environmental perception modules and state sensing modules such as sensors, cameras, and radar, as well as a chassis execution system including a processor, a four-wheel active steering system, and a drive-by-wire chassis architecture. The vehicle can be a pure electric or plug-in hybrid model. Both the front and rear wheels of the vehicle have independent, programmable steering capabilities, and the steering angle can be adjusted in real time via an electronic control unit. The vehicle is equipped with a high-precision positioning system, a multi-sensor fusion perception system, and a power-brake coordinated control module, enabling centimeter-level posture control and millisecond-level steering response. The electronically controlled and intelligent chassis execution system in the vehicle is a crucial hardware component for achieving vehicle parking in this embodiment, possessing the ability to reconstruct motion with any point as the instantaneous center of rotation, thereby enabling efficient parking in confined spaces.

[0035] The aforementioned parking command is the trigger signal to initiate parking. Parking commands can originate from user operation (such as buttons on the central control screen or voice commands) or system autonomous decision-making (such as automatic recommendation after the perception module identifies a suitable parking space). The parking command can be a digital signal, which, after being received and verified by the vehicle domain controller, activates the automatic parking function module, sequentially calling the perception, planning, and control subsystems. Parking commands should not be confused with other functions such as cruise control and lane keeping assist. This parking command triggering mechanism embodies the intelligent parking concept of "human-machine collaboration" and "environmental perception-driven" operation. From a safety perspective, the system must confirm that the vehicle speed is below 5 km / h, the gear is in reverse (R / D), and there are no moving obstacles in the vicinity before the command is triggered to ensure safe execution.

[0036] The aforementioned perception data refers to the sum of environmental information collected in real time by the vehicle through external sensors. Perception data can include camera images, ultrasonic radar ranging, lidar point clouds, and millimeter-wave radar Doppler data. Cameras are used to identify lane lines, parking space boundaries, and static obstacles; ultrasonic radar detects nearby obstacles (such as wheels and curbs); lidar constructs a 3D environmental map; and millimeter-wave radar senses dynamic vehicles and pedestrians. This data is processed by multi-sensor fusion algorithms (such as Kalman filtering and deep learning semantic segmentation) to output high-precision parking space boundary coordinates (such as the upper left and lower right corners), obstacle locations, and types. Perception data is the foundation for determining target parking spaces and path planning. Perception data can also be used to identify complex scenarios (such as angled parking spaces and occluded parking spaces), improving environmental understanding capabilities.

[0037] The aforementioned state data refers to real-time measurement information of the vehicle's own operating status, such as the vehicle's center of gravity position (x, y), heading angle θ, vehicle speed v, yaw rate ω, actual front and rear wheel deflection angles, gear status, and suspension attitude. This state data can be acquired by high-precision sensors such as inertial measurement units, wheel speed sensors, and controller area network buses. The state data can be used to accurately describe the vehicle's current pose in the global coordinate system, thereby accurately determining its relative pose. Relying on a centimeter-level positioning system ensures minimal state data error, enabling path planning and tracking accuracy to reach an engineering-usable level. The synergy between state data and perception data can form a two-dimensional "vehicle-environment" information loop, providing real physical input for parking control.

[0038] In one optional embodiment, upon receiving a parking command, a high-precision multi-sensor fusion system is immediately activated, simultaneously launching cameras, LiDAR and ultrasonic arrays, an inertial measurement unit (IMU), wheel speed sensors, etc., to perform environmental scanning. The camera captures images of road markings and obstacles; the LiDAR generates point cloud data; and the ultrasonic sensors monitor near-end obstacles (such as wheels and curbs) in real time and correct point cloud drift. The system fuses the three-source data using Kalman filtering to obtain perception data. Wheel speed sensors are used to acquire vehicle speed. The IMU integrates a three-axis gyroscope and a three-axis accelerometer to measure the vehicle's three-axis angular velocity and three-axis acceleration in real time, and uses a fusion algorithm to calculate the vehicle's roll angle, pitch angle, and yaw angle, among other state data.

[0039] In another alternative embodiment, before the parking command is triggered, the processor is already in a low-power standby state, but continuously caches the perception and status data of the most recent 3 seconds to the local cache. When the user triggers the parking command, the system immediately retrieves the most recent frame of image, point cloud, and vehicle pose from the cache as initial input, and simultaneously initiates new data acquisition to achieve zero-latency response. For example, when the vehicle is slowly driving on the roadside, the system has continuously cached lane line and parking space candidate area data. Once the command is issued, the system can directly call the most recently matched perception and status data. By using pre-caching and immediate invocation, a seamless switching experience is achieved, improving the smoothness of user perception.

[0040] In another optional embodiment, when the vehicle receives a parking instruction, if it detects the presence of intelligent roadside units or other parked vehicles that support vehicle-to-the-world information exchange, the processor proactively initiates a communication request to acquire perception data. For example, a roadside camera can provide higher-resolution images of parking spaces, or adjacent parked vehicles can share their identified parking space occupancy status and geometric dimensions. The system integrates local perception data and communication data to determine the perception and status data used for subsequent parking control. This mechanism overcomes the limitations of single-vehicle perception, achieves collaborative perception, and improves parking success rates. It also expands the perception dimensions through open communication protocols.

[0041] Step S104: Based on the perception data and state data, determine the target parking space of the vehicle and the relative pose of the vehicle and the target parking space.

[0042] The target parking space mentioned above can be a parking space identified and determined as suitable for parking by the vehicle perception system. The target parking space can be defined by its geometric boundaries (start and end coordinates), size (length × width), orientation (angle with the road), type (perpendicular / parallel / diagonal), and real-time occupancy status. The target parking space must meet the minimum size requirements of the vehicle (e.g., a perpendicular parking space is ≥5.5m long and ≥2.2m wide) and be free from dynamic obstacles. Determining the target parking space is the starting point for parking path generation. The position and orientation of the target parking space directly determine the vehicle's target pose—that is, the alignment state of the vehicle relative to the target parking space when it finally comes to a complete stop. By identifying parking spaces, parking in complex urban environments (such as angled roadside parking spaces or spaces between two vehicles) is supported. As the final destination, the geometric parameters of the target parking space drive the entire path planning logic. Different parking space types may trigger different parking paths, enabling the system to have adaptive strategy selection capabilities and improving the generalization of parking scenarios.

[0043] The relative pose described above is a three-dimensional vector depicting the spatial relationship between the vehicle's current pose and the ideal parking pose of the target parking space. The relative pose can be represented as Δp = [Δx, Δy, Δθ], where Δx is the lateral offset, Δy is the longitudinal offset, and Δθ is the heading angle difference. The relative pose can be calculated from vehicle state data and the boundary coordinates of the target parking space. Based on the relative pose, parking feasibility can be assessed and strategy selection can be made. The relative pose serves as a bridge connecting perception data and path generation; high-precision acquisition of the relative pose is a prerequisite for achieving seamless parking and collision-free parking.

[0044] In one optional embodiment, after receiving the perception data and state data, the processor first performs pixel-level recognition using a deep learning semantic segmentation network to extract semantic elements such as lane lines, parking space markings, and obstacle outlines. Simultaneously, the LiDAR point cloud is used to separate candidate parking space regions using a clustering algorithm. The system performs geometric modeling on each candidate parking space, calculating its length, width, orientation, and edge continuity. It then combines this with vehicle state data to calculate a score for each candidate parking space. The scoring function incorporates weighted factors such as parking space size matching degree, distance to the vehicle body, safe distance to obstacles, and heading alignment angle. Finally, the system selects the parking space with the highest score as the target parking space, thereby determining the relative pose between the target parking space and the vehicle.

[0045] In another optional embodiment, when the vehicle is cruising at low speed or moving slowly, the system continuously caches multiple candidate parking spaces and their corresponding vehicle pose sequences detected within the past 5 seconds. The current real-time pose is matched with the "pose-parking space" mapping relationship of each candidate parking space in the historical cache, and a dynamic time warping algorithm is used to calculate the similarity between the current pose and the historical records. If a parking space has been consistently identified multiple times in the past and the current environment has not changed significantly, it is prioritized as the target parking space. This mechanism utilizes the continuity of vehicle movement and the stability of the environment to avoid repeated perception and shorten the target parking space confirmation time, making it particularly suitable for continuous parking scenarios on urban roads (such as consecutive empty parking spaces on the roadside). Through memory-based re-identification, it achieves immediate recognition upon stopping, improving user experience and system response efficiency.

[0046] In another optional embodiment, upon detecting the presence of a smart roadside unit supporting vehicle-to-the-outside information exchange or a parked vehicle in the vicinity, the processor proactively initiates a communication request to obtain high-precision parking space occupancy status and geometric parameters (such as the four corner coordinates, length, width, orientation, and whether the parking space is occupied). The system performs spatiotemporal alignment and fusion of local sensing data, status data, and communication data, and uses a weighted consistency rate algorithm to determine the credibility of the target parking space. For example, if the communication data reports a parking space as "vacant," but the vehicle's sensing data and status data fail to identify it due to occlusion, the system can still adopt the communication data as the basis for identifying the target parking space.

[0047] In another optional embodiment, the system constructs a multimodal feature vector of the parking space in real time based on perceived and state data, such as the straight-line fitting residual of the parking space boundary, corner distribution density, longitudinal / lateral distance ratio to the vehicle body, heading angle distribution, and symmetry of surrounding obstacles. This feature vector is input into a lightweight classification model to automatically determine the parking space type and target parking space. Subsequently, the system uses the target parking space as a reference to calculate the relative pose between the vehicle's current pose and the target parking space. By realizing recognition, classification, and localization, path planning can directly call the parking model that matches the type, improving the overall system response efficiency and parking success rate.

[0048] Step S106: Based on the parking space type and relative pose of the target parking space, generate a target path and a set of target deflection angles corresponding to multiple path points in the target path.

[0049] The target deflection angle set includes the deflection angles of multiple wheels of the vehicle.

[0050] The target path described above is the continuous spatial trajectory planned by the system for the vehicle to reach the target parking space. The target path can be composed of simplified arcs. Specifically, the target path can consist of a series of discrete path points, representing the movement route of the vehicle's center of mass in a two-dimensional plane. The target path can be classified into single-arc, double-arc, and triple-arc paths based on the number of arcs, to distinguish it from broken lines and S-shaped paths, thereby simplifying the control logic. The trajectory curvature of the target path is continuous and without abrupt changes, avoiding tire sideslip and vibration. The target path does not rely on a preset trajectory library but is dynamically calculated based on the vehicle's real-time status and the target parking space, possessing strong adaptability. Compressing complex library manipulation into one or two arcs reduces computational load, improves response speed, and shortens parking time.

[0051] The aforementioned target deflection angle set refers to the steering angles of the four wheels corresponding to each path point on the target path. This target deflection angle set is not a fixed value but is dynamically calculated based on the vehicle's real-time pose and the target path. By analyzing the target path, asymmetric, continuously changing deflection angles are derived, enabling the vehicle to maintain pure rolling during circular motion. This embodiment can simultaneously control the front and rear wheels, achieving "crab-like" movement or extremely small-radius turns. The system refreshes this set at a frequency of 10ms, ensuring a trajectory tracking error of <3cm.

[0052] In one optional embodiment, before generating the target path, the system calls a pre-trained deep reinforcement learning policy network. The input is the parking space type and relative pose of the target parking space, and the output is the target path and a set of target deflection angles. This network learns the path morphology and trajectory smoothness under different scenarios (such as narrow parking spaces, slopes, and obstacle interference) through multiple simulations. Its reward function comprehensively considers parking time, path length, deflection angle changes, and collision risk. After generating the path, the system still uses the target deflection angle set predicted by this network. By encoding empirical knowledge into neural network weights, scene-adaptive path generation is achieved, improving the parking success rate under complex conditions and demonstrating high intelligence and generalization capabilities.

[0053] In another alternative embodiment, the system can pre-construct a high-precision trajectory-deflection angle database based on typical parking space types (such as standard perpendicular, angled, and narrow parallel parking), with each template containing a sequence of path points and a corresponding set of four-wheel deflection angles. During parking, the system performs nearest neighbor matching in the database based on the relative pose extracted from the perception data (such as lateral offset, longitudinal distance, and angular deviation), selects the best-matching template, and then fine-tunes the path and deflection angle through linear interpolation or affine transformation to obtain the target path and the set of target deflection angles corresponding to multiple path points in the target path.

[0054] In another alternative embodiment, after generating a high-order continuous target path using cubic spline curves based on the parking space type and relative pose of the target parking space, ensuring the path curvature is continuously differentiable, a physical feasibility check is performed on the theoretical deflection angle set for each path point. This involves calculating, based on a vehicle dynamics model (such as a two-wheel steering kinematics model), whether each deflection angle exceeds the actuator limits (front wheel ±30°, rear wheel ±15°), whether there is a wheel speed difference causing excessive sideslip, and whether there is a conflict between front and rear deflection angles (such as left front +15° and left rear -12° causing reverse torque). If any path point does not meet the constraints, the system adaptively adjusts the path radius and iteratively generates a new path until all points meet the physical feasibility boundaries. Control execution capability is used as a hard constraint for path generation, rather than post-processing compensation, ensuring that each path is executable, trackable, and jitter-free. By co-designing planning and execution, tracking errors and system jitter are reduced.

[0055] Step S108: Based on the target path and the set of target deflection angles corresponding to multiple path points, control the vehicle to drive into the target parking space.

[0056] In one optional embodiment, a model predictive control algorithm based on the vehicle's nonlinear kinematics model constructs a rolling window using the target path and the corresponding set of target yaw angles as reference inputs. Within each 10ms control cycle, the target yaw angles of all four wheels are directly input as control variables, executed by the steer-by-wire system, and state corrections are performed. This achieves joint control of the path and steering.

[0057] In another optional embodiment, when the vehicle is traveling at low speed along the target path, the tires are prone to unexpected sideslip due to the accumulation of lateral forces, causing the center of gravity to deviate from the target trajectory. While executing the target yaw angle set, the system monitors the wheel speed difference and lateral acceleration of each wheel in real time and calculates the estimated tire sideslip angle. If the sideslip angle of a wheel exceeds a threshold, the system immediately activates the dynamic yaw angle compensation module, which calculates the required yaw angle correction Δδ based on the tire model and adds it to the original target yaw angle set to form the compensated yaw angle. This compensation is calculated independently for each wheel. For example, if the left rear wheel needs to increase its camber angle due to sideslip, the system simultaneously fine-tunes the right front wheel to maintain instantaneous rotation center stability, enabling the vehicle to maintain trajectory accuracy on slippery surfaces or in narrow parking spaces, thus improving parking robustness.

[0058] In another optional embodiment, during vehicle control, if minor jitter occurs in path tracking due to sensor noise or actuator delay, the system does not directly correct the wheel slip angle. Instead, it dynamically triggers an adaptive resampling mechanism for path points. For example, every 200ms, based on the current measured vehicle pose, the system resamples the next five path points, generates a smooth transition trajectory using cubic spline interpolation, and simultaneously recalculates the corresponding wheel slip angle set. This allows for fine-tuning of the path during execution without altering the overall geometry (e.g., the center and radius remain unchanged), ensuring consistency in the parking target. For instance, when the vehicle slightly deviates due to uneven road surfaces, the system shifts the original path point inward by 0.1m, generating a new continuous curve, avoiding sudden steering changes caused by hard correction. By fine-tuning the path during execution, comfort and stability are improved.

[0059] In this embodiment of the application, in response to receiving a parking instruction, after acquiring the vehicle's perception data and status data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined based on the perception data and status data. Then, based on the parking space type and relative pose of the target parking space, a target path and a set of target deflection angles corresponding to multiple path points in the target path are generated. Thus, based on the target path and the set of target deflection angles corresponding to multiple path points, the vehicle is controlled to drive into the target parking space. Upon responding to a parking command, this application dynamically identifies the target parking space through environmental perception and vehicle status data, determines the precise relative pose between the vehicle and the target parking space, and then constructs a simplified target path based on different target parking space types and relative poses. It also inversely calculates the set of target deflection angles for coordinated wheel deflection. By using the dynamically coordinated set of target deflection angles corresponding to the target path, the vehicle avoids path redundancy and cumbersome operation caused by kinematic rigidity, ultimately driving the vehicle to complete parking. This simplifies the parking process and solves the complex technical problems associated with vehicle parking control in related technologies.

[0060] Optionally, based on the vehicle's perception data and state data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined, including: identifying the perception data to determine the target parking space; and determining the relative pose of the vehicle and the target parking space based on the state data and the position and orientation of the target parking space.

[0061] In one optional embodiment, the perceived data, such as data collected by sensors like cameras, LiDAR, and ultrasonic sensors, is identified. Algorithms such as image semantic segmentation, point cloud clustering, and edge detection are then used to automatically identify available spaces that meet parking standards and determine the type (perpendicular, parallel, or diagonal, etc.) and geometric boundaries of the parking spaces. This allows for the selection of feasible parking targets from the perceived data, avoiding misidentification of curbs, billboards, or temporary obstacles as parking spaces, and ensuring the physical feasibility of parking actions. It provides accurate target anchor points for subsequent path planning, enabling the parking system to have autonomous perception capabilities. This eliminates reliance on high-precision maps or manual markings, allowing the system to operate reliably in real-world scenarios such as unmarked, non-standard parking spaces, and dynamic interference, thus improving environmental adaptability and market universality.

[0062] Then, the location of the target parking space can be represented by coordinates or latitude and longitude. Based on state data (such as positioning and heading angle), and the location and orientation of the target parking space, the relative pose of the vehicle's current center of mass relative to the target parking space is calculated. This is achieved by representing the relative pose through lateral offset, longitudinal offset, and heading deviation, quantifying the spatial relationship between the vehicle and the parking space, and providing precise geometric input for path generation. By establishing a mathematical bridge between perception and planning, it is ensured that the path is not generated based on experience, but rather on the geometric inverse solution of real spatial relationships, enabling parking control to achieve precise alignment with centimeter-level positioning accuracy.

[0063] Optionally, based on the parking space type and relative pose of the target parking space, a target path and a set of target deflection angles corresponding to multiple path points in the target path are generated, including: determining the target pose of the vehicle based on the position and orientation of the target parking space, wherein the target pose is used to characterize the pose of the vehicle parked in the target parking space; when the parking space type is a perpendicular parking space, a first arc path segment is generated based on the relative pose and the target pose, the first arc path segment is used as the target path, and a set of target deflection angles corresponding to multiple path points is determined based on the first arc path segment; when the parking space type is a parallel parking space, a second arc path segment and a third arc path segment are generated based on the relative pose, the preset position corresponding to the target parking space, and the target pose, and a target path and a set of target deflection angles corresponding to multiple path points are generated based on the second arc path segment and the third arc path segment, wherein the second arc path segment and the third arc path segment are tangent at a preset position.

[0064] The first circular path segment mentioned above is the target path generated in a perpendicular parking scenario. It is a continuous circular arc trajectory followed by the vehicle to reach the target pose. The first circular path segment can be determined by the center and radius. The center can be located at the right rear corner (or left rear corner) of the parking space, and the radius is the distance from the vehicle's center of mass to that point. Determining the first circular path segment allows the vehicle to glide directly into the parking space with a single continuous action, without reversing or shifting gears, achieving a one-shot parking. The generation of this first circular path segment does not rely on empirical rules, but is based on the geometric inverse solution of the relative pose and the target pose. It breaks through the limitation of the turning radius being greater than or equal to the vehicle length in the traditional Ackerman model, achieving a minimum radius turning where the radius is smaller than the vehicle length. This allows the vehicle to park in narrow spaces, improving urban parking efficiency.

[0065] The aforementioned preset position is the geometric connection point between the second and third circular path segments in parallel parking, i.e., the switching point where the vehicle transitions from the lateral approach phase to the parking phase. The preset position is the system's intermediate target for parallel parking. The preset position can be determined by the target pose, vehicle dimensions, and rear axle spacing. Generally, the preset position can be the midpoint of the side of the parallel parking space on the same side as the curb.

[0066] At this preset position, the second and third circular arc path segments are tangent, meaning the tangent directions of the two arcs are consistent, thus achieving continuous and abrupt path curvature and preventing vehicle vibration or tire slippage. By pre-determining the target deflection angle set, switching can be completed without interrupting power, improving smoothness and safety.

[0067] The aforementioned second circular arc path segment is the first stage of the parallel parking process, consisting of a large-radius arc (e.g., 5-8m) along which the vehicle moves laterally towards the preset position. This second circular arc path segment allows the vehicle to smoothly approach the parking space entrance from the side of the road, avoiding collisions with other vehicles or the curb. Its center is located in the external space to the side and front of the vehicle, and can be calculated using the relative posture and preset position to ensure a smooth trajectory. This path segment has a small curvature, allowing for a slightly higher speed to quickly complete lateral displacement. The large arc of the second circular arc path segment enables continuous steering, improving comfort. The second and third circular arc path segments can be smoothly switched.

[0068] The aforementioned third circular arc path segment is the second stage of the parallel parking path. It is a small-radius arc (e.g., radius 1.5~2.5m) along which the vehicle slides from the preset position into the target parking space, with its center located at the rear corner of the target parking space. This path segment achieves the final precise positioning of the vehicle, ensuring that the vehicle body is parallel to the parking space and aligned within the wheelbase. The radius of the third circular arc path segment can be determined by the geometric relationship between the preset position and the target pose, ensuring that the vehicle completes the sliding action within a minimal space. In this embodiment, the rear wheels are deflected in the opposite direction (e.g., the right rear wheel swings outward), causing the vehicle to rotate around the rear corner with a very small radius, completing the parking maneuver in one step.

[0069] In one optional embodiment, the final posture the vehicle should achieve when parking in the target parking space can be determined based on the location (e.g., corner coordinates) and orientation of the target parking space. The target posture can be ensured by using centroid coordinates and heading angles, guaranteeing that the vehicle body is parallel to the parking space line, the four wheels are centered, and there is a safety margin in front and behind. By providing a clear "end point" for the parking process, path planning becomes deterministic rather than tentative. The fuzzy "parking in the space" requirement is transformed into precise mathematical posture constraints, serving as the benchmark and convergence basis for subsequent path generation. This upgrades parking to precise alignment, ensuring the neatness, safety, and user satisfaction of the vehicle after parking.

[0070] When the parking space type is perpendicular, a circular arc trajectory is derived from the target pose as the endpoint and the relative pose as the starting point through inverse kinematics. A set of target deflection angles corresponding to multiple path points is determined. This trajectory uses the corner point of the parking space as the instantaneous rotation center, allowing the vehicle to slide directly from its current position into the target pose along a single-segment circular arc. This utilizes four-wheel active steering capability, overcoming the turning radius limitation of the traditional Ackerman model, achieving a one-shot parking maneuver along a single arc. This compresses the complex multi-segment parking maneuver into a smooth, continuous curve, simplifying path complexity and control difficulty.

[0071] In another optional embodiment, the path is planned in two stages in parallel parking spaces. The first stage is a second circular arc path segment that approaches the parking space entrance laterally. The second stage is a third circular arc path segment that turns at a preset position and slides into the final target position with a very small radius. The tangent directions of the two circular arcs are consistent at the preset position, ensuring path continuity. By achieving a smooth transition from lateral movement to vertical sliding in within a limited space, both path feasibility and turning feasibility are considered. The double-arc tangent structure breaks through the spatial limitations of traditional straight-ahead and reverse modes, achieving efficient parking. This transforms parallel parking from a "three-step operation" to a "one-step completion," improving parking efficiency and user experience. At the same time, the preset position tangential constraint ensures that the trajectory is physically feasible and the control is vibration-free.

[0072] Optionally, based on the first circular arc path segment, determine the set of target deflection angles corresponding to multiple path points, including: determining the pose of the vehicle arriving at multiple path points based on the positions of multiple path points on the first circular arc path segment; and determining the set of target deflection angles corresponding to multiple path points based on the first center of the first circular arc path segment and the pose of the vehicle arriving at multiple path points.

[0073] The first center of the circle is the geometric center of the first circular arc path segment, that is, the instantaneous rotation center around which the vehicle rotates during perpendicular parking. Its position can be determined jointly by the corner point of the target parking space (such as the right rear corner) and the vehicle's initial pose, rather than by the vehicle structure. The first center of the circle is actively set, such as by sensing the coordinates of the parking space corner point, so that the vehicle's center of mass moves precisely along the arc centered at that point. This setting makes the parking path geometrically strongly coupled with the target parking space, realizing motion reconstruction "driven by the target". The determination of the first center of the circle depends on high-precision coordinate calculation, and its position selection directly affects parking safety and efficiency.

[0074] In one optional embodiment, a number of discrete path points can be generated on the first circular arc path by sampling at fixed intervals (e.g., every 0.1 meters) or at fixed time points. For each path point, the coordinates of the vehicle's center of mass at that point are calculated based on the circular arc geometry, and the vehicle's heading angle at that position, i.e., the angle between the vehicle's forward direction and the global coordinate axis, is derived simultaneously. This discretizes the continuous first circular arc path into a set of spatiotemporal pose sequences that can be executed point-by-point by the controller, making the abstract path concrete as an executable input. By establishing a mapping relationship between path geometry and vehicle motion state, a precise target basis is provided for subsequent deflection angle calculation. This is a prerequisite for trajectory tracking control, ensuring that the vehicle's position is clear at every moment.

[0075] Furthermore, using the first center of the first circular arc path segment as the instantaneous rotation center, and combining the vehicle's pose at each path point, the required deflection angles of the front and rear wheels at that point are calculated in reverse using a four-wheel steering kinematic model. For example, based on the angle between the vector direction from the vehicle's center of mass to the center of the circle and the vehicle's longitudinal axis, and combined with vehicle parameters such as wheelbase and track width, the target deflection angles of each wheel are calculated using geometric relationships, ensuring that the vehicle's instantaneous motion trajectory at that point strictly matches the circular arc path. This converts the geometric path into steering parameters that the actuator can execute, achieving a precise mapping from path to control, ensuring that the vehicle moves along the designed trajectory at each path point, and avoiding trajectory deviation or tire slippage due to incorrect deflection angles. Thus, in perpendicular parking, all-wheel coordinated deflection control based on a single circular arc is achieved, enabling the vehicle to complete single-arc parking within a very small space.

[0076] Preferably, generating a first circular arc path segment based on the relative pose and the target pose includes: determining a first center and a first radius based on the relative pose and the target pose; and generating the first circular arc path segment based on the first center and the first radius.

[0077] The first radius is the radius of the first circular arc path segment, approximately equal to the Euclidean distance from the vehicle's center of mass to the center of the first circle. This first radius directly determines the curvature of the vehicle's turn and the required yaw angle. The first radius can be smaller than the vehicle's wheelbase, and through the coordinated counter-yawing of the rear wheels, the vehicle rotates around the corner point, achieving parking in minimal space. The determination of the first radius is the mathematical basis for path generation, and its value can be calculated in real time from the parking space dimensions and the vehicle's position and posture.

[0078] In one alternative embodiment, the first center and radius of a circle can be derived from the relative pose and the target pose through geometric constraints, allowing the vehicle to smoothly transition from its current state to the target pose along a single circular arc. Under four-wheel active steering, the vehicle trajectory can be considered as a circular arc rotating around a certain spatial point. The system determines the first center and radius of the continuously rotating circle by solving for the perpendicular intersection of the trajectories of the vehicle's front and rear axle centers, or by using an inverse kinematics algorithm, ensuring that the tangent at the starting point of the path is consistent with the current heading and the tangent at the ending point is consistent with the target heading. Constructing a unique, continuous, and inflection-free single circular arc trajectory provides the geometric basis for parking. Transforming parking requirements into precise circular arc parameters ensures that path generation possesses both mathematical uniqueness and physical feasibility.

[0079] Then, using the determined first circle center as a reference and the first radius as a distance, a first circular arc path segment is generated between the vehicle's current pose and the target pose. This transforms the abstract combination of the circle center and radius into an executable first circular arc path segment for the controller, realizing the conversion from geometric parameters to control. This provides a clear motion reference for subsequent parking, ensuring the vehicle runs stably along the designed trajectory during execution. Because this first circular arc path segment has characteristics such as low curvature change, no sudden steering changes, and no gear shifting requirements, combined with four-wheel cooperative steering, it can achieve perpendicular parking in a single forward motion, improving parking efficiency and user experience.

[0080] like Figure 2 As shown, the bolded segment in arc 2 illustrates an optional first arc path segment in this embodiment. For perpendicular parking spaces, arc 2 has a smaller radius compared to the conventional arc 1. Figure 2 The midpoint O is the center of the first circle, and R is the first radius. Specifically, determine the centerline 1 of the target parking space and the centerline 2 before the vehicle parks, and then obtain circle PQ that is tangent to both centerline 1 and centerline 2. PQ is also the intersection point of the line from the center O to the circle and the circle itself. Figure 2 As shown, the target deflection angle corresponding to one of the path points of the first circular path segment is also displayed. For example, the angle between the perpendicular line from the center O to the central axis 2 and the line connecting the wheel to the center of the circle is the deflection angle. For example, angles 1, 2, 3, and 4 are the deflection angles of the right rear wheel, left rear wheel, left front wheel, and right front wheel of the vehicle, respectively.

[0081] Optionally, based on the second and third circular arc path segments, a set of target deflection angles corresponding to multiple path points is generated, including: determining the vehicle's pose at multiple first path points based on the positions of multiple first path points on the second circular arc path segment, and determining the vehicle's pose at multiple second path points based on the positions of multiple second path points on the third circular arc path segment; determining the set of target deflection angles corresponding to multiple path points based on the second center of the second circular arc path segment, the poses of the vehicle at multiple first path points, the third center of the third circular arc path segment, and the poses of the vehicle at multiple second path points, wherein the directions of the set of target deflection angles corresponding to multiple first path points are the same.

[0082] The second center is the geometric center of the second circular path segment, located in the external space to the side and front of the vehicle's initial pose. The coordinates of the second center can be obtained by inversely solving the geometric relationship between the relative pose and the target pose. The second radius is the distance from the vehicle's center of mass to the second center. Together, the second center and the second radius define the vehicle's lateral approach trajectory to the parking space, ensuring the path does not collide with roadside obstacles. The second center can be dynamically generated based on the parking space width, vehicle length, and initial pose, allowing the path to adapt to different scenarios. The second radius is typically greater than 5 meters, with a small curvature, allowing for higher vehicle speeds and improving parking efficiency.

[0083] The third center is the rotation center of the third arc path segment, located at the rear corner of the target parking space. The third radius is the distance from the vehicle's center of mass to this corner when it is in the preset position. This center and radius together determine the precise trajectory of the vehicle as it finally slides into the parking space. The third radius is typically less than 2.5 meters, much smaller than the vehicle length, enabling extremely small-radius turns. This parameter can be calculated by the system in real time without manual intervention, ensuring accurate parking every time. Furthermore, through rear wheel counter-rotation coordination, the vehicle adaptively rotates around the corner, achieving millimeter-level alignment. This design allows vehicles to park in smaller parking spaces, improving urban parking utilization.

[0084] In one optional embodiment, the parallel parking space double-arc path can be divided into two independent arc segments. For each first path point on the second arc, the system calculates the vehicle's centroid coordinates and heading angle at that point based on its geometric position under the second arc's center, forming the vehicle's pose for multiple first path points. Similarly, for each second path point on the third arc, the corresponding pose is deduced from the third arc's center. This discretizes the composite path into an executable set of spatiotemporal poses, ensuring that the vehicle has a clear motion target in each path segment.

[0085] Then, based on the second center of the second circular path segment, the vehicle's pose at multiple first path points, the third center of the third circular path segment, and the vehicle's pose at multiple second path points, the target deflection angle set corresponding to multiple path points is determined. The fact that the target deflection angle sets corresponding to multiple first path points have the same direction means that the system independently applies the four-wheel steering inverse kinematics model to the path points of the second and third circular arcs, combining the center positions of each segment with the corresponding vehicle pose, to solve for the target deflection angle required for each path point, thus obtaining the target deflection angle set. In the second circular path segment, the target deflection angle combination for all first path points remains consistent or only slightly adjusted to maintain the stable lateral sliding motion of the vehicle. In the third circular path segment, the deflection angle combination is switched to another set (e.g., front wheels fully turned to the left, rear wheels deflected to the right) to achieve small-radius rotation. This allows for differentiated control of two motion modes—large-radius stable coasting and small-radius precise positioning—while ensuring path continuity. This allows the vehicle to smoothly switch steering modes without stopping or shifting gears, avoiding track vibration or tire slippage caused by sudden changes in yaw angle.

[0086] Preferably, the generation of a second and a third circular arc path segment based on the relative pose, the preset position corresponding to the target parking space, and the target pose includes: determining a preset pose for the vehicle to reach the preset position based on the preset position; determining a second center and a second radius based on the relative pose and the preset pose, and determining a third center and a third radius based on the preset pose and the target pose; generating a second circular arc path segment based on the second center and the second radius, and generating a third circular arc path segment based on the third center and the third radius.

[0087] In one optional embodiment, a preset pose (heading angle and attitude, etc.) for the vehicle to reach the preset position can be determined based on the preset position, such that this pose is exactly the common tangent point state of the second and third circular arc path segments. By ensuring that the tangent directions of the two arcs are consistent at the preset position, a smooth transition with continuous path, no abrupt changes, and no steering jitter is achieved. Transforming the preset position into a dynamic geometric node uniquely determined by kinematic constraints, making path generation both mathematically rigorous and physically feasible, is a prerequisite for realizing dual-arc parking.

[0088] Then, the geometric parameters of the two circular path segments can be inversely calculated in two steps. For example, based on the relative pose and the preset pose, the second center and the second radius that allow the vehicle to smoothly approach the parking space laterally along the arc can be inversely solved. Furthermore, based on the preset pose and the target pose, the third center and the third radius that allow the vehicle to complete a perpendicular parking maneuver around the smaller radius can be inversely solved. By determining the curvature characteristics of the two circular arc segments while ensuring path continuity, such as a larger radius and gentler steering for safe approach of the second arc, and a smaller radius and sharper steering for precise parking, the complex parking task is decomposed into two independently solvable kinematic subproblems, achieving segmented analysis and overall coordination.

[0089] Then, based on the second center and the second radius, two discrete path point sequences are generated by sampling along the circumferential trajectory at fixed intervals, forming a complete second circular arc path segment. A third circular arc path segment is then generated based on the third center and the third radius. By transforming the abstract center-radius parameters into a spatiotemporal reference trajectory executable by the controller, input is provided for subsequent deflection angle calculations and path tracking, ensuring that the vehicle accurately follows the designed path during execution.

[0090] Optionally, based on the target path and the target deflection angle set corresponding to multiple path points, the vehicle is controlled to enter the target parking space, including: if the target path includes a first arc path segment, the vehicle is controlled to enter the target parking space based on the target deflection angle set corresponding to multiple path points; if the target path includes a second arc path segment and a third arc path segment, the vehicle is controlled to drive to a preset position of the target parking space based on the target deflection angle set corresponding to multiple first path points on the second arc path segment, and the vehicle is controlled to enter the target parking space based on the target deflection angle set corresponding to multiple second path points on the third arc path segment.

[0091] In one optional embodiment, when the target path includes a first circular arc path segment, executing a single circular arc path utilizes a pre-calculated set of target deflection angles corresponding to each path point, enabling the vehicle to move continuously and stably along the arc until it precisely stops in the target parking space. This achieves fully continuous control for single-arc parking, reducing the gear shifting and pauses required by traditional multi-segment parking maneuvers. The planned geometric path is directly converted into precise steering parameters at the execution layer, ensuring minimal trajectory tracking error and enabling one-time steering completion in perpendicular parking, improving parking efficiency and user experience.

[0092] In another optional embodiment, when the target path includes a second and a third circular arc path segment, the vehicle can be driven to move laterally with a large radius based on the target deflection angle set corresponding to multiple first path points on the second circular arc path segment, stably approaching the preset position of the parking space entrance. Upon reaching the preset position, the vehicle switches to the target deflection angle set corresponding to multiple second path points on the third circular arc path segment (e.g., front wheels fully turned to the left, rear wheels deflected in the opposite direction), allowing the vehicle to complete a vertical sliding motion with a very small radius. This achieves seamless switching between lateral approach and precise parking movement modes. Furthermore, through segmented deflection angle control, both path stability and parking flexibility are considered, avoiding insufficient space or trajectory drift caused by a single steering mode.

[0093] like Figure 3 As shown, Figure 3 Arc 3 in the diagram illustrates a parking path for parallel parking spaces in related technologies, where there may be a risk of collision during the parking process.

[0094] like Figure 4 As shown, an optional second circular arc path segment (such as...) is illustrated in this embodiment. Figure 4 (The bolded AB segment) and the third circular path segment (such as...) Figure 4 The diagram shows the bolded section BC, where point B is the preset position. Generally, point B can be the midpoint of the parking space on the curb side. Point B is also the midpoint of the rear axle of the vehicle in the preset position; that is, at preset position B, the line tangent to the second and third circular arc path segments is perpendicular to the rear axle axis. In the target pose, the midpoint of the rear axle is point C. Based on points B and C, the center and radius of the third circle can be determined. Then, based on the same radius, the circle corresponding to the second circular arc path segment tangent at point B can be obtained. The starting point A of the second circular arc path segment is the center point of the rear axle of the vehicle. Furthermore, in... Figure 4 The text shows the deflection angles corresponding to the path points at the preset positions. For example, the angle between the line connecting the third center and the wheel and the line connecting the third center and point B is the deflection angle. In other words, at the preset position, the deflection angle of the rear wheel is 0, and the deflection angles of the front axle are angle 1 and angle 2, respectively.

[0095] According to an embodiment of this application, an embodiment of a vehicle parking control system is provided. Figure 5 This is a schematic diagram of a vehicle parking control system according to an embodiment of this application, as shown below. Figure 5 As shown, the system includes the following:

[0096] The environmental perception module 40 is used to acquire the vehicle's perception data.

[0097] The status sensing module 42 is used to acquire vehicle status data.

[0098] The processor 44 is configured to, in response to receiving a parking instruction, determine the target parking space for the vehicle and the relative pose of the vehicle and the target parking space based on perception data and state data; generate a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type and the relative pose of the target parking space; and control the vehicle to drive into the target parking space based on the target path and the set of target deflection angles corresponding to multiple path points.

[0099] The aforementioned environmental perception module is a visual and spatial perception system, which may involve wide-angle cameras, ultrasonic radar, lidar, millimeter-wave radar, and their deep fusion algorithms. This module acquires geometric and semantic information about the vehicle's surroundings in real time. Specifically, the camera identifies lane lines, parking space boundaries, and obstacle outlines. Ultrasonic sensors detect nearby objects (such as wheels and curbs). LiDAR constructs a 360° point cloud map, accurately extracting the coordinates of the four corners of the parking space. Millimeter-wave radar detects dynamic obstacles. The environmental perception module can output structured data, such as parking space dimensions, location, orientation, occupancy status, obstacle distance, and speed.

[0100] The aforementioned state sensing module is part of the vehicle's body perception system, and may consist of wheel speed sensors, steering wheel angle encoders, and bus interfaces. This module can acquire real-time state data such as the vehicle's precise pose, speed, acceleration, and measured yaw angle. It can output state data including the vehicle's center of gravity coordinates, heading angle, vehicle speed, and yaw rate. The module relies on centimeter-level positioning to ensure minimal path tracking error. Its high-frequency (e.g., 100Hz) sampling capability provides a real-time basis for dynamic yaw angle calculation.

[0101] The aforementioned processor can be an in-vehicle domain controller, running algorithms such as path planning, yaw angle calculation, and trajectory tracking. The processor can receive perception data and status data, calculate the target parking space and its relative pose, generate a target path based on the parking space type, and calculate the set of target yaw angles for all four wheels in real time, outputting control commands. The processor features high computing power, low latency, and real-time operating system support.

[0102] In this embodiment, the environmental perception module acquires the vehicle's perception data to construct a digital model of the vehicle's external environment. This provides reliable spatial semantic information for parking decisions, enabling the system to autonomously discover and determine parking space availability, eliminating reliance on high-precision maps or manual markings. This allows automatic parking to generalize to real-world road conditions, a prerequisite for achieving unassisted parking in all scenarios. Furthermore, the state sensing module acquires the vehicle's state data to accurately grasp its current motion state, providing real-time feedback for pose calculation and path tracking. After receiving a parking command, the processor determines the target parking space and the relative pose between the vehicle and the target parking space based on the perception and state data. This allows for the identification of feasible parking targets from multi-source data and the quantification of spatial relationships, preventing mis-parking and improving parking safety and user trust.

[0103] Optionally, the system may further include a chassis execution system for controlling vehicle movement and steering of multiple wheels based on control commands generated by the processor, wherein the control commands are generated based on a target path and a set of target deflection angles corresponding to multiple path points.

[0104] The chassis actuator system is the physical execution terminal, which may involve four-wheel active steering actuators, electronic braking system, electric drive system (motor controller), and bus communication network to receive processor instructions and precisely control wheel steering, power, and braking. Both the front and rear wheels can be independently, quickly, and with high precision, enabling the target deflection angle set to be accurately executed.

[0105] This embodiment utilizes a chassis execution system to control vehicle movement and the steering of multiple wheels based on control commands generated by the processor, achieving high-precision tracking of the planned trajectory. This transforms abstract path and steering parameters into actual motion responses in the physical world, ensuring the vehicle can stably, smoothly, and without delay execute the required steering angle and speed at every moment, even under complex operating conditions, especially achieving millisecond-level steering angle switching at dual-arc switching points. The chassis execution system enables the vehicle to possess the motion capabilities of a crab, such as lateral movement or small-radius rotation in place, breaking through the kinematic limitations of traditional vehicles that rely solely on front-wheel steering.

[0106] like Figure 6 As shown, an optional parking control system is illustrated. This system includes an environmental perception module, a state sensing module, a central processing unit (CPU), and a chassis execution system. Specifically, the environmental perception module in the perception layer transmits environmental data to the CPU. The state sensing module transmits its own data to the CPU. The CPU in the planning and control layer performs path planning and yaw angle calculations, and outputs control commands to the chassis execution system. The chassis execution system in the execution layer enables steer-by-wire, power steering, and braking.

[0107] The technical solution proposed in this application will be described below with reference to an optional embodiment. This application proposes a path planning and control method for automatic parking or assisted parking.

[0108] In the field of vehicle automatic control technology, vehicles with traditional Ackerman steering mechanisms have a turning radius limited by wheelbase and maximum deflection angle. When entering narrow parking spaces, they often need to perform multiple "forward-backward" maneuvers, a cumbersome, time-consuming, and labor-intensive process. Most existing automatic parking systems are based on the traditional Ackerman steering model for path planning, and their paths are usually complex curves composed of multiple arcs and straight lines. This does not fundamentally overcome the limitations of vehicle kinematics and still requires sufficient space to complete multiple turns.

[0109] This embodiment provides a method to simplify parking paths by dynamically changing tire deflection angles, enabling parking with minimal operations and in a smaller space. By giving the wheels (especially the rear wheels) additional deflection freedom, the vehicle can move laterally like a "crab" or achieve a smaller turning radius.

[0110] Specifically, it involves perception and positioning, such as using sensors like cameras, ultrasonic radar, and lidar to identify available parking spaces and accurately obtain the relative position and attitude of the vehicle and the parking space.

[0111] Path planning is performed, such as the central processing unit planning an ideal simplified path based on perception information. The simplified paths in this embodiment include: single-arc path parking and double-arc tangential path parking.

[0112] Single-arc path parking refers to parking in a perpendicular space by simultaneously turning the front and rear wheels on one side (for example, turning both the left front wheel and the left rear wheel to the left). This allows the vehicle to rotate around a virtual point on the right side (the instantaneous turning center) and reverse directly into the parking space in a smooth arc without any reversing adjustments.

[0113] Double-arc tangential path parking refers to a method for parallel parking where all wheels are first controlled to deflect in the same direction at a certain angle, allowing the vehicle to approach the parking space laterally in an arc with a large radius. Upon reaching the predetermined position, the deflection angle is quickly switched to another combination, allowing the vehicle to slide into the parking space in an arc with a small radius. The two arcs are tangent at the switching point, ensuring a smooth path.

[0114] Then, the yaw angle is calculated and controlled. For example, based on the planned target path, the target yaw angle required for each wheel at each moment (or each path segment) is calculated in reverse. The front wheel yaw is controlled by the steer-by-wire system, and if the vehicle is equipped with a rear-wheel active steering system, the rear wheel yaw is controlled simultaneously.

[0115] Finally, path tracking and execution are achieved. For example, the vehicle control system controls the vehicle's power and braking systems according to the planned path and the calculated deflection angle sequence, so that the vehicle accurately tracks the target path until parking is completed.

[0116] Therefore, this embodiment can greatly simplify the path, reducing the complexity of complex multi-segment parking paths to one or two simple arcs, thus lowering the complexity of path planning. It reduces the required space; due to the smaller turning radius and more flexible movement, vehicles can park in tighter spaces. It improves parking efficiency; by reducing the number of forward and backward maneuvers, it can even achieve one-shot parking, shortening parking time. It enhances the user experience; for automatic parking systems, the success rate is higher, and for assisted parking systems, it provides drivers with more intuitive and simple guidance.

[0117] The parking control method is explained below using a specific parking space as an example. For instance, in a single-arc path parking maneuver for a perpendicular parking space, the vehicle intends to enter a right-hand perpendicular space. The perception system confirms that the parking space is available and the space is narrow. The central processing unit plans a single circular arc path with point O as the center and radius R. To achieve this path, the system calculates a specific set of deflection angles: the left front wheel deflection angle is δfl, the left rear wheel deflection angle is δrl, the right front wheel deflection angle is δfr (usually small or 0), and the right rear wheel deflection angle is δrr (usually negative, i.e., reverse deflection). This combination ensures that the vehicle's instantaneous turning center is precisely at point O. The chassis actuator drives each wheel to deflect to the specified angle, and the vehicle control system controls the vehicle speed, ensuring the vehicle travels strictly along the circular arc PQ and smoothly reverses into the parking space.

[0118] For example, in a parallel parking space with a double-arc tangential path, the vehicle intends to enter the parallel parking space on the right side of the road. The perception system confirms the parking space information. The central processing unit plans a path consisting of arcs AB and BC, which are tangent at point B. In the first stage (driving along arc AB), the system controls both the front and rear wheel sets to deflect to the left at a moderate angle, causing the vehicle to move forward and to the left with a larger radius, approaching the parking space. When the vehicle reaches the switching point B, the system quickly adjusts the vehicle's yaw angle to a new combination (e.g., the front wheels fully turned to the left, and the rear wheels turned to the right), drastically reducing the vehicle's instantaneous turning radius, thus allowing it to smoothly "glide" into the parking space along arc BC. Throughout the entire process, the driver only needs to trigger the "start parking" command or simply maintain the gear; all yaw angle adjustments and path tracking are automatically completed by the system.

[0119] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0120] According to an embodiment of this application, a parking control device for a vehicle is provided. It should be noted that the device can be used to execute the above-described parking control method for a vehicle. Figure 7 This is a schematic diagram of a vehicle parking control device according to an embodiment of this application, such as... Figure 7 As shown, the device includes the following:

[0121] The acquisition module 60 is used to acquire the vehicle's perception data and status data in response to receiving a parking command.

[0122] The determination module 64 is used to determine the target parking space of the vehicle and the relative pose of the vehicle and the target parking space based on the perception data and state data.

[0123] The generation module 66 is used to generate a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type and relative pose of the target parking space. The set of target deflection angles includes the deflection angles of multiple wheels of the vehicle.

[0124] The control module 68 is used to control the vehicle to drive into the target parking space based on the target path and the set of target deflection angles corresponding to multiple path points.

[0125] Optionally, the determining module is also used to: identify the perceived data to determine the target parking space; and determine the relative pose of the vehicle and the target parking space based on the state data, as well as the position and orientation of the target parking space.

[0126] Optionally, the generation module is further configured to: determine the target pose of the vehicle based on the position and orientation of the target parking space, wherein the target pose is used to characterize the pose of the vehicle parked in the target parking space; when the parking space type is a perpendicular parking space, generate a first arc path segment based on the relative pose and the target pose, use the first arc path segment as the target path, and determine a set of target deflection angles corresponding to multiple path points based on the first arc path segment; when the parking space type is a parallel parking space, generate a second arc path segment and a third arc path segment based on the relative pose, the preset position corresponding to the target parking space, and the target pose, generate a target path and a set of target deflection angles corresponding to multiple path points based on the second arc path segment and the third arc path segment, wherein the second arc path segment and the third arc path segment are tangent at a preset position.

[0127] Optionally, the generation module is further configured to: determine the pose of the vehicle arriving at multiple path points based on the positions of multiple path points on the first circular arc path segment; determine the set of target deflection angles corresponding to multiple path points based on the first center of the first circular arc path segment and the pose of the vehicle arriving at multiple path points; preferably, the generation module is further configured to: determine the first center and the first radius based on the relative pose and the target pose; and generate the first circular arc path segment based on the first center and the first radius.

[0128] Optionally, the generation module is further configured to: determine the vehicle's pose at multiple first path points based on the positions of multiple first path points on the second arc path segment, and determine the vehicle's pose at multiple second path points based on the positions of multiple second path points on the third arc path segment; determine the target deflection angle set corresponding to multiple path points based on the second center of the second arc path segment, the vehicle's pose at multiple first path points, the third center of the third arc path segment, and the vehicle's pose at multiple second path points, wherein the direction of the target deflection angle set corresponding to multiple first path points is the same; preferably, the generation module is further configured to: determine the preset pose of the vehicle at the preset position based on the preset position; determine the second center and second radius based on the relative pose and the preset pose, and determine the third center and third radius based on the preset pose and the target pose; generate the second arc path segment based on the second center and second radius, and generate the third arc path segment based on the third center and third radius.

[0129] Optionally, the control module is further configured to: control the vehicle to enter the target parking space based on the target deflection angle set corresponding to multiple path points when the target path includes a first arc path segment; control the vehicle to drive to the preset position of the target parking space based on the target deflection angle set corresponding to multiple first path points on the second arc path segment when the target path includes a second arc path segment and a third arc path segment; and control the vehicle to enter the target parking space based on the target deflection angle set corresponding to multiple second path points on the third arc path segment.

[0130] Embodiments of this application also provide a vehicle, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods described in various embodiments of this application when it runs.

[0131] This application also provides an electronic device 90, please refer to... Figure 8 It includes a memory 910 and a processor 920, wherein the memory 910 is used to store computer programs; and the processor 920 is used to execute the programs stored in the memory 910 to implement the methods in the various embodiments of this application.

[0132] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0133] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0134] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium for storing a computer program that, when executed by a processor, implements the methods in various embodiments of this application.

[0135] Embodiments of this application also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of this application.

[0136] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0137] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0138] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0139] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0140] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0141] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A parking control method for a vehicle, characterized in that, include: In response to receiving a parking instruction, it acquires the vehicle's perception data and status data; Based on the perceived data and the state data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined. Based on the parking space type of the target parking space and the relative pose, a target path is generated, as well as a set of target deflection angles corresponding to multiple path points in the target path, wherein the set of target deflection angles includes the deflection angles of multiple wheels of the vehicle. Based on the target path and the set of target deflection angles corresponding to the multiple path points, the vehicle is controlled to drive into the target parking space.

2. The method according to claim 1, characterized in that, Based on the vehicle's perception data and state data, the target parking space of the vehicle and the relative pose of the vehicle and the target parking space are determined, including: The perceived data is identified to determine the target parking space; Based on the state data, as well as the position and orientation of the target parking space, the relative pose of the vehicle and the target parking space is determined.

3. The method according to claim 1, characterized in that, Based on the parking space type and the relative pose of the target parking space, a target path is generated, along with a set of target deflection angles corresponding to multiple path points within the target path, including: Based on the location and orientation of the target parking space, the target pose of the vehicle is determined, wherein the target pose is used to characterize the pose of the vehicle parked in the target parking space; When the parking space type is a perpendicular parking space, a first arc path segment is generated based on the relative pose and the target pose. The first arc path segment is used as the target path, and the target deflection angle set corresponding to the plurality of path points is determined based on the first arc path segment. When the parking space type is a parallel parking space type, based on the relative pose, the preset position corresponding to the target parking space and the target pose, a second arc path segment and a third arc path segment are generated. Based on the second arc path segment and the third arc path segment, the target path and the target deflection angle set corresponding to the multiple path points are generated, wherein the second arc path segment and the third arc path segment are tangent at the preset position.

4. The method according to claim 3, characterized in that, Based on the first circular arc path segment, determine the set of target deflection angles corresponding to the plurality of path points, including: Based on the positions of the multiple path points on the first circular path segment, the pose of the vehicle upon reaching the multiple path points is determined. Based on the first center of the first arc path segment and the pose of the vehicle when it arrives at the multiple path points, determine the set of target deflection angles corresponding to the multiple path points. Preferably, generating a first circular arc path segment based on the relative pose and the target pose includes: Based on the relative pose and the target pose, determine the first center and the first radius; The first circular arc path segment is generated based on the first center and the first radius.

5. The method according to claim 3, characterized in that, Based on the second and third circular arc path segments, a set of target deflection angles corresponding to the plurality of path points is generated, including: Based on the positions of multiple first path points on the second arc path segment, the pose of the vehicle arriving at the multiple first path points is determined, and based on the positions of multiple second path points on the third arc path segment, the pose of the vehicle arriving at the multiple second path points is determined. Based on the second center of the second arc path segment, the pose of the vehicle when it arrives at the plurality of first path points, the third center of the third arc path segment, and the pose of the vehicle when it arrives at the plurality of second path points, the target deflection angle set corresponding to the plurality of path points is determined, wherein the direction of the target deflection angle set corresponding to the plurality of first path points is the same. Preferably, based on the relative pose, the preset position corresponding to the target parking space, and the target pose, a second arc path segment and a third arc path segment are generated, including: Based on the preset position, determine the preset posture of the vehicle when it arrives at the preset position; Based on the relative pose and the preset pose, the second center and the second radius are determined, and based on the preset pose and the target pose, the third center and the third radius are determined. Based on the second center and the second radius, the second arc path segment is generated, and based on the third center and the third radius, the third arc path segment is generated.

6. The method according to any one of claims 1 to 5, characterized in that, Based on the target path and the set of target deflection angles corresponding to the plurality of path points, controlling the vehicle to enter the target parking space includes: If the target path includes a first circular arc path segment, the vehicle is controlled to enter the target parking space based on the target deflection angle set corresponding to the plurality of path points; When the target path includes a second arc path segment and a third arc path segment, the vehicle is controlled to drive to the preset position of the target parking space based on the target deflection angle set corresponding to multiple first path points on the second arc path segment, and the vehicle is controlled to drive into the target parking space based on the target deflection angle set corresponding to multiple second path points on the third arc path segment.

7. A parking control system for a vehicle, characterized in that, include: The environmental perception module is used to acquire the vehicle's perception data; A status sensing module is used to acquire the status data of the vehicle; The processor is configured to, in response to receiving a parking instruction, determine the target parking space of the vehicle and the relative pose of the vehicle and the target parking space based on the perception data and the state data; generate a target path and a set of target deflection angles corresponding to multiple path points in the target path based on the parking space type of the target parking space and the relative pose, wherein the set of target deflection angles includes the deflection angles of multiple wheels of the vehicle. Based on the target path and the set of target deflection angles corresponding to the multiple path points, the vehicle is controlled to drive into the target parking space.

8. The system according to claim 7, characterized in that, The system also includes: A chassis execution system is used to control the vehicle's movement and the steering of the plurality of wheels based on control instructions generated by the processor, wherein the control instructions are generated based on the target path and a set of target deflection angles corresponding to the plurality of path points.

9. A vehicle, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 6.

Images