Auxiliary parking method and device, storage medium, electronic equipment and vehicle

Abstract

The invention provides an auxiliary parking method and device, a storage medium, electronic equipment and a vehicle, and relates to the technical field of auxiliary parking.The method comprises the steps that in the process of conducting auxiliary parking according to a first parking track, in response to an adjusting instruction for the first parking track, an adjusted second parking track is obtained; and then auxiliary parking is carried out according to the second parking track. According to the technical scheme, the adjusting instruction of the first parking track is responded in real time, the second parking track is generated, a driver is allowed to directly correct the parking track in the parking process, the parking process does not need to be interrupted, flexible interaction of parking and adjusting at the same time can be achieved, operation is easy, efficiency is high, the user operation cost can be reduced, and the user experience is improved. And personalized parking preferences of different drivers are met.

Description

Technical Field

[0001] This disclosure relates to the field of assisted parking technology, and more particularly to an assisted parking method, device, storage medium, electronic device, and vehicle. Background Technology

[0002] With the continuous development of the automotive industry and the ongoing advancements in artificial intelligence, the Internet of Things, and sensor technologies, automobiles are gradually evolving from traditional means of transportation into mobile intelligent terminals that integrate safety, intelligence, and comfort. Assisted parking functions, as a frequently used and readily perceived intelligent scenario, have received widespread attention. Summary of the Invention

[0003] This disclosure provides an assisted parking method, device, storage medium, electronic device, and vehicle. The main purpose is to solve the problem that in related technologies, assisted parking can only execute according to the initially planned fixed parking trajectory. If the user needs to adjust the parking trajectory during the assisted parking process, the parking process must be interrupted, the system restarted, and replanned, resulting in cumbersome operation and low efficiency.

[0004] According to a first aspect of the present disclosure, an assisted parking method is provided, comprising:

[0005] During the assisted parking process according to the first parking trajectory, in response to the adjustment command of the first parking trajectory, the adjusted second parking trajectory is obtained;

[0006] Assisted parking is performed according to the second parking trajectory.

[0007] By responding to the adjustment command of the first parking trajectory in real time and generating a second parking trajectory, the driver can directly correct the parking trajectory during the parking process without interrupting the parking process. This enables flexible interaction of adjusting while parking, is simple to operate and highly efficient, and can reduce the user's operating costs.

[0008] Optionally, the method further includes:

[0009] An adjustment instruction for the first parking trajectory is determined based on at least one of the user's voice information, user gesture information, and operation information on the vehicle control screen.

[0010] In this way, diverse human-computer interaction interfaces are provided, such as voice control and gesture control, allowing drivers to initiate trajectory adjustments in the most natural way, covering different usage habits.

[0011] Optionally, the method further includes:

[0012] The vehicle control screen displays functions that allow adjustment of the first parking trajectory.

[0013] In response to the trigger command of the function item, operation information for the vehicle control screen is obtained.

[0014] By using a visual interface, the user operation threshold is lowered, the adjustable content is clearly explained, and the ineffectiveness of user operations is avoided, making the operation simple and convenient for users.

[0015] Optionally, the method further includes:

[0016] If a parking risk is determined based on the images around the vehicle and the second parking trajectory, a warning message is output.

[0017] By comparing the second parking trajectory with real-time environmental images, potential risks can be identified and warned in advance to avoid collisions or operational errors.

[0018] Optionally, when a parking risk is determined based on the image around the vehicle and the second parking trajectory, the output of a prompt message includes:

[0019] Obstacles in the image around the vehicle are identified, and the position coordinates of the obstacles in the coordinate system of the second parking trajectory are obtained;

[0020] Calculate the distance between the path point coordinates of the second parking trajectory and the position coordinates of the obstacle;

[0021] If at least one waypoint is less than a preset safety threshold from the obstacle, a collision risk warning message will be output.

[0022] This approach solves the problem of accurately identifying and effectively alerting users to static collision risks in assisted parking scenarios, ensuring parking safety, avoiding collisions, and also taking into account user experience.

[0023] Optionally, when a parking risk is determined based on the image around the vehicle and the second parking trajectory, the output of a prompt message includes:

[0024] The effective parking area of ​​the parking space is determined by identifying the boundary line of the parking space through the image around the vehicle;

[0025] If the target parking location of the second parking trajectory exceeds the effective parking area, a warning message indicating a risk of exceeding the boundary will be output.

[0026] By accurately identifying parking space boundaries, quantifying effective areas, and identifying the risk of the final target crossing the boundary, the effectiveness of assisted parking is ensured. At the same time, user experience is optimized through advance warnings and precise guidance, while avoiding subsequent safety hazards caused by crossing the boundary.

[0027] Optionally, when a parking risk is determined based on the image around the vehicle and the second parking trajectory, the output of a prompt message includes:

[0028] Analyze the motion trends of obstacles using multiple consecutive frames of images surrounding the vehicle;

[0029] If the motion trend predicts that the obstacle will intersect with the second parking trajectory within a preset time period, a warning message indicating a dynamic collision risk will be output.

[0030] By analyzing the movement trends and predicting the intersections of dynamic obstacles, sudden and uncontrollable collision risks can be avoided in advance. This not only fills the gap in traditional static risk assessment, but also balances safety and efficiency through accurate prediction, ultimately providing users with a safer and more reliable assisted parking experience.

[0031] Optionally, the method further includes:

[0032] During the assisted parking process following the second parking trajectory, the vehicle's expected parking path is displayed based on the second parking trajectory.

[0033] In this way, users can clearly understand the system's planning logic, enhance their trust in assisted parking, and give them the control to make advance judgments and actively correct errors, thus avoiding the limitations of the system's one-way control.

[0034] Optionally, the step of obtaining the adjusted second parking trajectory in response to the adjustment command for the first parking trajectory includes:

[0035] In response to an adjustment command for the target parking position in the first parking trajectory, the updated vehicle target pose is obtained according to the adjustment command for the target parking position;

[0036] The second parking trajectory is generated based on the current vehicle status and the updated vehicle target pose.

[0037] In this way, the user's adjustment requirements for parking results can be accurately transformed into a trajectory plan that the system can execute. The adjustment command clearly defines the updated vehicle target pose. Starting from the current state and combining it with the updated vehicle target pose, a new parking trajectory is generated. This avoids planning a completely new trajectory from scratch, does not interrupt the current parking, and can connect the current situation state to the updated vehicle target pose, ensuring a smooth transition.

[0038] Optionally, obtaining the updated vehicle target pose based on the adjustment command of the target parking position includes:

[0039] The offset parameters are determined according to the adjustment command of the target parking position, wherein the offset parameters include at least one of lateral offset, longitudinal offset and yaw angle adjustment.

[0040] The updated vehicle target pose is determined based on the offset parameters and the vehicle target pose before the update.

[0041] In this way, the user's vague adjustment needs are transformed into precise quantitative parameters for the system. The vague needs are then precisely broken down using offset parameters. The calculation is accurate based on the original target pose, providing a precise and feasible endpoint benchmark for the subsequent generation of the second parking trajectory. This avoids trajectory planning deviations caused by fuzzy target poses, ensuring the reliability of the assisted parking adjustment function and user satisfaction from the source.

[0042] Optionally, generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes:

[0043] Obtain a reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle.

[0044] Based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and combined with the control input change rate corresponding to the vehicle in at least one control cycle, a second parking trajectory is generated, wherein the control input change rate includes the steering angle change rate and / or acceleration change rate.

[0045] In this way, a second parking trajectory is generated based on the error information between the reference trajectory and the predicted future trajectory, combined with the rate of change of the control input. This approach balances trajectory accuracy and driving smoothness, resulting in a second parking trajectory that is both accurate and stable, feasible and reliable. For example, minimizing the error ensures the vehicle accurately reaches the new target pose, while minimizing the rate of change of the control input avoids vehicle jerking.

[0046] Optionally, generating the second parking trajectory based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and in conjunction with the rate of change of the control input corresponding to the vehicle in at least one control cycle, includes:

[0047] The error information, control input change rate, and end state error corresponding to the first control cycle are substituted into the target cost function to calculate the cost information. The end state error is the deviation between the vehicle state in the last step of the prediction time domain corresponding to the first control cycle and the updated vehicle target pose.

[0048] Under the conditions of satisfying vehicle physical constraints and environmental safety constraints, with the goal of reducing the cost information, the target control sequence in the prediction time domain is analyzed by a quadratic programming algorithm.

[0049] In the first control cycle, the control instructions of the target control sequence are executed to obtain the vehicle state updated in the first control cycle;

[0050] Based on the vehicle state updated in the first control cycle, the cost information calculation, target control sequence analysis, and vehicle state updated in the next control cycle are performed iteratively until the finally updated vehicle state reaches the updated vehicle target pose.

[0051] The second parking trajectory is generated based on the vehicle status corresponding to each control cycle.

[0052] This approach provides a concrete method for generating a second parking trajectory based on the error information between the reference trajectory and the predicted future trajectory, combined with the rate of change of the control input. This results in a precise new parking trajectory, with multi-dimensional error constraints ensuring accurate arrival at the target pose. Multiple layers of protection through physical and environmental constraints avoid risky operations, thus guaranteeing driving safety. Constraints on the rate of change of the control input prevent jerking, resulting in smoother control.

[0053] Optionally, predicting the future trajectory of the vehicle in at least one control cycle based on the current state of the vehicle includes:

[0054] Based on the current state of the vehicle and in conjunction with the dynamic model, the future trajectory of the vehicle in at least one control cycle is predicted, wherein the dynamic model is used to determine the dynamic relationship between the vehicle target parameters and the steering angle and acceleration, and the vehicle target parameters include at least one of vehicle position, yaw angle and speed.

[0055] In this way, the future trajectory of a vehicle in at least one control cycle can be accurately predicted based on the laws of physical motion, providing a predictive benchmark that conforms to the actual motion characteristics of the vehicle for subsequent trajectory planning and control.

[0056] Optionally, generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes:

[0057] Based on the current state of the vehicle and the updated vehicle target pose, the trajectory adjustment parameters are analyzed, and the trajectory adjustment parameters include at least one of lateral adjustment, longitudinal adjustment and angle adjustment.

[0058] The coordinates of the path points of the first parking trajectory are transformed according to the trajectory adjustment parameters to generate a temporary transition trajectory;

[0059] The smoothness of the temporary transition trajectory is checked. If the rate of change of trajectory curvature exceeds a preset threshold, the transition segment is optimized by spline curve interpolation.

[0060] The feasibility of the optimized transition trajectory is verified based on vehicle dynamics constraints and environmental obstacle information;

[0061] The trajectory segment that has passed the feasibility verification is connected with the updated vehicle target pose to generate the second parking trajectory.

[0062] In this way, precise adjustments and a safe, smooth transition are achieved by reusing the first parking trajectory to obtain the adjusted second parking trajectory. This eliminates the need for full trajectory replanning, reducing computational load and response latency, thus improving the efficiency of parking trajectory generation and avoiding the safety hazards associated with simple adjustments. Through smoothness verification and optimization, inflection points in the adjusted trajectory are eliminated, ensuring precise alignment with the new target pose and preventing trajectory decoupling from the target.

[0063] Optionally, generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes:

[0064] Using the current state of the vehicle as the root node, the tree branch is iteratively expanded by randomly sampling points in the state space, and valid nodes are filtered out by collision detection until a search tree connecting the current state of the vehicle and the updated vehicle target pose is generated.

[0065] The second parking trajectory is selected from the search tree.

[0066] This method generates a search tree with the vehicle's current state as the root node, and then filters the trajectory from the tree to obtain the second parking trajectory. It enables trajectory generation in complex environments with no preset path dependency and strong obstacle avoidance capabilities. Random sampling covers the entire state space, independent of preset paths. Collision detection is used to filter valid nodes in real time, dynamically avoiding obstacles.

[0067] Optionally, the step of iteratively expanding the tree by randomly sampling points in the state space, using the current state of the vehicle as the root node, includes:

[0068] Using the current state of the vehicle as the root node, the tree is iteratively expanded by randomly sampling points in the state space according to the direction of the updated vehicle target pose in the state space.

[0069] In this way, the method of generating the second parking trajectory is further optimized. By sampling in the target direction, the efficiency and quality of trajectory planning can be greatly improved, while retaining the advantages of random sampling. It can still flexibly avoid obstacles in complex obstacle environments without losing its anti-interference ability.

[0070] Optionally, the assisted parking according to the second parking trajectory includes:

[0071] Determine the angle between the tangent direction of the current point of the first parking trajectory and the tangent direction of the starting point of the second parking trajectory;

[0072] If the included angle is less than a preset angle threshold, then assisted parking is performed according to the second parking trajectory; or, if the included angle is greater than or equal to the preset angle threshold, then a transition trajectory segment is generated, which is used for the vehicle to smoothly transition from the first parking trajectory to the second parking trajectory.

[0073] This approach addresses the smoothness and safety issues when switching from the first parking trajectory to the second, preventing vehicle jerking or loss of control due to sudden trajectory changes. It strikes an optimal balance between the necessity of trajectory adjustment and the smoothness and safety of driving. For example, small-angle switching balances efficiency and smoothness, while large-angle transitions avoid risks and abruptness. Ultimately, it ensures that regardless of the adjustment magnitude, the vehicle can smoothly and safely transition from the first to the second trajectory. This avoids the jerking and risks caused by hard switching and enhances user trust in the system through dynamic adaptation to different scenarios.

[0074] According to a second aspect of the present disclosure, an auxiliary parking device is provided, comprising:

[0075] The acquisition module is configured to, during the process of assisted parking according to the first parking trajectory, in response to an adjustment command for the first parking trajectory, acquire the adjusted second parking trajectory;

[0076] The control module is configured to perform assisted parking according to the second parking trajectory.

[0077] Optionally, the acquisition module is specifically configured to, in response to an adjustment instruction for the target parking position in the first parking trajectory, acquire an updated vehicle target pose based on the adjustment instruction for the target parking position; and generate a second parking trajectory based on the current vehicle state and the updated vehicle target pose.

[0078] Optionally, the acquisition module is specifically configured to determine offset parameters based on the adjustment instruction of the target parking position, wherein the offset parameters include at least one of lateral offset, longitudinal offset, and yaw angle adjustment; and determine the updated vehicle target pose based on the offset parameters and the vehicle target pose before the update.

[0079] Optionally, the acquisition module is specifically configured to acquire a reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and to predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle; and to generate a second parking trajectory based on the error information between the future state trajectory of the vehicle in the at least one control cycle and the reference trajectory, and in combination with the control input change rate corresponding to the vehicle in the at least one control cycle, wherein the control input change rate includes the steering angle change rate and / or acceleration change rate.

[0080] Optionally, the acquisition module is specifically configured to substitute the error information, control input change rate, and end-state error corresponding to the first control cycle into the target cost function to calculate the cost information, wherein the end-state error is the deviation between the vehicle state at the last step in the prediction time domain corresponding to the first control cycle and the updated vehicle target pose; under the condition of satisfying vehicle physical constraints and environmental safety constraints, with the goal of reducing the cost information, the target control sequence in the prediction time domain is analyzed by a quadratic programming algorithm; the control instructions of the target control sequence are executed in the first control cycle to obtain the vehicle state updated in the first control cycle; based on the vehicle state updated in the first control cycle, the cost information calculation, target control sequence analysis, and vehicle state updated in the next control cycle are iteratively executed until the finally updated vehicle state reaches the updated vehicle target pose; and the second parking trajectory is generated based on the vehicle state corresponding to each control cycle.

[0081] Optionally, the acquisition module is specifically configured to predict the future trajectory of the vehicle in at least one control cycle based on the current state of the vehicle and in conjunction with a dynamic model, wherein the dynamic model is used to determine the dynamic relationship between the vehicle target parameters and the steering angle and acceleration, and the vehicle target parameters include at least one of vehicle position, yaw angle and speed.

[0082] According to a third aspect of the present disclosure, an electronic device is provided, comprising:

[0083] processor;

[0084] A memory connected to the processor, the memory storing a computer program that, when executed by the processor, implements the assisted parking method described in the first aspect.

[0085] According to a fourth aspect of the present disclosure, a vehicle is provided, comprising:

[0086] processor;

[0087] A memory connected to the processor, the memory storing a computer program that, when executed by the processor, implements the assisted parking method described in the first aspect.

[0088] According to a fifth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the assisted parking method described in the first aspect.

[0089] According to a sixth aspect of the present disclosure, a computer program product is provided, comprising a computer program, characterized in that the computer program, when executed by a processor, implements the assisted parking method described in the first aspect.

[0090] According to a seventh aspect of the present disclosure, a chip is provided, including one or more interface circuits and one or more processors; the interface circuits are configured to receive signals from a memory of an electronic device and send the signals to the processors, the signals including computer instructions stored in the memory; when the processor executes the computer instructions, the electronic device performs the assisted parking method described in the first aspect.

[0091] By employing the above technical solution, this disclosure provides an assisted parking method, device, storage medium, electronic device, and vehicle. During assisted parking according to a first parking trajectory, in response to an adjustment command for the first parking trajectory, an adjusted second parking trajectory is first acquired; then, assisted parking is performed according to the second parking trajectory. This technical solution, by responding in real-time to the adjustment command for the first parking trajectory and generating a second parking trajectory, allows the driver to directly correct the parking trajectory during the parking process without interrupting the parking procedure. It enables flexible interaction of adjusting while parking, is simple to operate, highly efficient, reduces user operating costs, and meets the personalized parking preferences of different drivers.

[0092] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0093] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0094] Figure 1 A schematic flowchart of an assisted parking method provided in an embodiment of this disclosure is shown;

[0095] Figure 2A flowchart illustrating another assisted parking method provided in an embodiment of this disclosure is shown;

[0096] Figure 3 A flowchart illustrating yet another assisted parking method provided in an embodiment of this disclosure is shown;

[0097] Figure 4 A schematic flowchart of another assisted parking method provided in an embodiment of this disclosure is shown;

[0098] Figure 5 A schematic flowchart of another assisted parking method provided in an embodiment of this disclosure is shown;

[0099] Figure 6 A schematic diagram of the structure of an auxiliary parking device provided in an embodiment of this disclosure is shown;

[0100] Figure 7 A schematic diagram of the structure of a vehicle provided in an embodiment of this disclosure is shown. Detailed Implementation

[0101] Some embodiments of this disclosure will be described in detail herein, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. Various changes, modifications, and equivalents of the methods, apparatus, and / or systems described herein will become apparent upon understanding this disclosure. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but can be changed as will become apparent upon understanding this disclosure, except for operations that must be performed in a particular order. Furthermore, for clarity and brevity, descriptions of features known in the art may be omitted. It should be noted that, without conflict, the embodiments and features in the embodiments of this disclosure can be combined with each other.

[0102] The embodiments described in the following examples of this disclosure are not representative of all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0103] In some examples, assisted parking can be achieved using Parking Assist (APA) or Intelligent Parking Assist (IPA) systems, whose operation typically follows a fixed and linear pattern. This process may involve the driver manually or automatically activating the parking system while driving at low speeds (e.g., below 30 km / h). As the vehicle moves along the lane, lateral ultrasonic sensors or cameras continuously scan to detect and identify appropriately sized parallel or perpendicular parking spaces. The system then presents one or more identified available parking spaces to the driver via the vehicle's central display screen and requests confirmation. Once the driver confirms the target parking space, the system calculates a single, complete, and fixed parking trajectory. The system then takes over vehicle steering control. In semi-automatic systems, the driver needs to control the accelerator, brake, and gear shifting as prompted; in fully automatic systems, the system controls all longitudinal and lateral movements. Throughout the parking process, the driver's primary task is to monitor the surrounding environment. At this point, the user's interactive options are extremely limited, typically limited to completely aborting the assisted parking process by pressing the brake or turning the steering wheel.

[0104] The limitation of the above example lies in the rigidity of its operation: once parking begins, the system's trajectory planning enters an open-loop state, no longer accepting any fine-tuning instructions for the parking trajectory (such as adjusting the target parking position). If, during the parking process, the driver discovers that neighboring vehicles are parked improperly, encroaching on part of the parking space; or that there are obstacles in the parking space that were not previously fully detected by the sensors, such as shopping carts or low bollards; or that although the intended parking position is within the parking lines, it would cause the driver's side door to be pressed against a pillar or wall, making it difficult to get out of the car, then the driver's only option is to abort the entire assisted parking process, manually move the vehicle, and then find a new parking space or try to restart the assisted parking function. This "interrupt-restart" operation mode is inefficient, severely impacting the user experience and weakening the user's trust and reliance on the assisted parking system. In other words, this example treats assisted parking as a deterministic geometric problem: once the initial conditions (i.e., the parking space position) are set, the system solves for an optimal, unchanging solution. This "one-time instruction, never to be modified" logic assumes that the machine can always make the best decision after receiving the initial parameters. However, real-world parking scenarios are full of uncertainties, and user needs may change during the process.

[0105] To address the problems described in the examples above, embodiments of this disclosure provide an assisted parking method that views assisted parking as an interactive, human-centered control problem. This changes the driver's role from a "one-time commander" issuing a single instruction before the process begins to a "continuous supervisor" providing ongoing, nuanced guidance throughout the process. This shift from rigid automation to human-machine collaboration can meet the personalized parking preferences of different drivers, thereby enhancing the driving experience. Figure 1 As shown, this method can be applied to end-side execution of control devices or equipment for assisted parking, and the method includes the following steps.

[0106] Step 101: During the assisted parking process according to the first parking trajectory, in response to the adjustment command of the first parking trajectory, obtain the adjusted second parking trajectory.

[0107] The first parking trajectory can be a parking trajectory planned before the vehicle performs assisted parking. Before the vehicle performs assisted parking, the user can select and confirm a target parking location (such as a target parking space), and then plan an initial parking trajectory from the vehicle's current position to the target parking location. The user can adjust this initial parking trajectory, such as adjusting the shape of the parking trajectory or adjusting the target parking location corresponding to the parking trajectory. The adjustment methods can include various methods, such as adjusting to the left, right, up, down, forward, and backward, etc. Specifically, parameters such as distance and / or yaw angle can be adjusted. For example, parking a certain distance to the left or right, etc. If there are other large vehicles behind the target parking location, to avoid safety hazards or situations where the trunk cannot be opened, it is necessary to adjust the target parking location forward, etc.

[0108] When the user confirms the execution of assisted parking, the system enters assisted parking mode and performs assisted parking according to the first parking trajectory. For example, the vehicle has activated the assisted parking function and is driving according to the initially planned first parking trajectory, which is a parking trajectory that cuts straight in from the current position and then reverses into the parking space.

[0109] In some examples, the process of assisted parking according to the first parking trajectory can be counted from the moment the assisted parking begins, until the moment the vehicle completes parking at the target location and exits the assisted parking state. In some examples, if the vehicle completes parking at the target location but has not exited the assisted parking state, this can also be considered as being in the process of assisted parking according to the first parking trajectory.

[0110] The second parking trajectory can be a new, feasible, safe, and smooth trajectory recalculated based on adjustment commands and current vehicle status and environmental information. For example, during assisted parking according to the first parking trajectory, the user can adjust the first parking trajectory, such as adjusting its shape or the target parking position. Adjustments can be made in various ways, such as adjusting to the left, right, up, down, forward, or backward, and can specifically adjust parameters such as distance and / or yaw angle. For example, parking a certain distance to the left or to the right. After the user completes the trajectory adjustment, the adjusted second parking trajectory can be obtained in this embodiment.

[0111] Step 102: Perform assisted parking according to the second parking trajectory.

[0112] The technical solution of this disclosure responds in real time to the adjustment command of the first parking trajectory and generates a second parking trajectory, and continues to perform assisted parking according to the newly generated second parking trajectory. This allows the driver to directly correct the parking trajectory during the parking process without interrupting the parking process, enabling flexible interaction of parking and adjustment. The operation is simple and efficient, reducing user operating costs and meeting the personalized parking preferences of different drivers.

[0113] In some embodiments, the method provided by this disclosure allows the driver to dynamically and in real-time adjust the vehicle's parking trajectory while assisted parking maneuvers are in progress. To achieve this, the method may include the following key functional modules:

[0114] Human-Machine Interface (HMI) module: such as a user interface used to receive user target correction commands during parking. This interface can be integrated into the vehicle's infotainment screen or central infotainment display, receiving user input in an intuitive and low-interference manner (such as touch and drag), while also supporting user voice commands and gesture commands.

[0115] Dynamic Target Status Module: This module is responsible for converting user input commands from the HMI (e.g., "translate to the right") into an updated vehicle target pose in real time, which is a new set of target world coordinates x, y (used to determine the vehicle position and yaw angle).

[0116] Real-time motion planning and control module: This module receives the target pose updated by the dynamic target state module, and immediately starts from the vehicle's current state (current position, attitude, speed, etc.) to recalculate an optimal trajectory that is collision-free, satisfies the vehicle's kinematic constraints, and smoothly transitions to the new target state, and seamlessly controls the vehicle to execute this new trajectory.

[0117] In some embodiments, the adjustment command for the first parking trajectory can be determined based on at least one of user voice information, user gesture information, and operation information on the vehicle control screen. Diverse human-machine interaction interfaces are provided, allowing drivers to initiate trajectory adjustments in the most natural way, covering different usage habits.

[0118] For example, voice information: when the driver says "move 10 centimeters to the right" or "lift the front of the car a little more," the system converts the voice into adjustment commands through voice recognition. Gesture information: when the driver makes gestures such as "wave to the left" or "push the palm forward" in front of the in-car camera, the system interprets them as "move left" or "move forward" commands through image recognition. Control screen operation: the driver can directly click the directional keys on the screen, drag the virtual vehicle icon, or input a specific offset value (such as lateral +5cm).

[0119] In some embodiments, a function item that allows adjustment of the first parking trajectory can be displayed on the vehicle control screen; in response to the trigger command of the function item, operation information for the vehicle control screen is obtained. A visual interface lowers the user's operational threshold, clearly informing them of the adjustable content and avoiding ineffective operations.

[0120] For example, a trajectory adjustment panel pops up on the control screen, containing buttons such as "Horizontal Offset," "Vertical Offset," and "Angle Fine Adjustment," or directly displays the "Left / Right / Forward / Backward" directional keys, intuitively indicating the adjustable dimensions to the user. When the user clicks the "Move Right" button or drags the slider, the system records the operation amplitude in real time (such as moving 5cm to the right with one click or other values), and converts it into specific adjustment parameters (such as horizontal offset + 5cm or other values).

[0121] In some embodiments, a warning message may be output if a parking risk is determined based on images of the vehicle's surroundings and a second parking trajectory. This disclosure embodiment identifies potential risks in advance and issues warnings by comparing the second parking trajectory with real-time environmental images, thus avoiding collisions or operational errors.

[0122] In some examples, when a parking risk is determined based on the image around the vehicle and the second parking trajectory, a warning message is output. Specifically, this may include: first, identifying obstacles in the image around the vehicle and obtaining the position coordinates of the obstacles in the coordinate system of the second parking trajectory, i.e., the obstacles and the second parking trajectory are in the same coordinate system, such as in the world coordinate system or in the vehicle coordinate system; then, calculating the distance between the path point coordinates of the second parking trajectory and the position coordinates of the obstacles; if at least one path point is less than the distance to the obstacle by a preset safety threshold, a warning message indicating a collision risk is output.

[0123] For example, the system obtains the location of obstacles through image recognition (such as identifying pillars, bollards, and other vehicles next to the parking space), and then checks whether parking according to the second parking trajectory would allow the vehicle to brush against the obstacle. For instance, if a path point in the second parking trajectory is only 0.2m away from a nearby vehicle (the safety threshold can be set to 0.3m, etc.), the system immediately prompts the message "Too close to the right, please be careful."

[0124] In some examples, when a parking risk is determined based on images around the vehicle and a second parking trajectory, a warning message is output. Specifically, this may include: identifying the parking space boundary line through images around the vehicle to determine the effective parking area of ​​the parking space; if the target parking position of the second parking trajectory exceeds the effective parking area, a warning message indicating a risk of exceeding the boundary is output.

[0125] An effective parking area refers to a dedicated space defined by the parking space boundary line (or the physical outline of the parking space) where vehicles can be parked compliantly (e.g., without occupying public areas or violating parking rules) and safely (e.g., ensuring normal entry and exit without affecting the surrounding area). This area not only meets the requirement of not affecting the surrounding environment (e.g., neighboring vehicles, road traffic) but also the requirement of vehicles being able to park completely and enter and exit normally. For example, vehicles parked within an effective parking area must ensure that no part of the vehicle extends beyond the parking space boundary line, does not obstruct fire lanes or no-parking zones, that car doors can be fully opened, and that sufficient space is left in front of and behind the vehicle.

[0126] For example, the system first identifies the white boundary line of the parking space using image recognition to determine the compliant parking area (such as a rectangular area within the boundary line), and then checks whether the final parking position is on or outside the boundary line. For instance, if the target parking position of the second parking trajectory exceeds the boundary line of the parking space by 0.3m, the system prompts the message "Target position exceeds the parking space, correct?".

[0127] In some examples, when a parking risk is determined based on images of the vehicle's surroundings and a second parking trajectory, a warning message is output. Specifically, this may include: analyzing the movement trend of obstacles through multiple consecutive frames of images of the vehicle's surroundings; if the movement trend predicts that the obstacle will intersect with the second parking trajectory within a preset time period in the future, a warning message indicating a dynamic collision risk is output.

[0128] For example, when dealing with moving obstacles (such as a pedestrian walking across or a neighboring car suddenly opening its door), the system calculates the direction and speed of movement using consecutive frame images to predict whether a collision with the second parking trajectory is imminent. For instance, if a neighboring car door is detected opening at a speed of 0.5 m / s, and it is predicted that the car will intersect with the second parking trajectory in 2 seconds (distance < 0.1 m), the system immediately displays the message "Neighboring car door open, collision imminent!"

[0129] In some examples, during assisted parking following a second parking trajectory, the vehicle's expected parking path can be displayed based on that trajectory. This visualization allows the driver to intuitively understand the vehicle's upcoming route, enhancing their sense of control and trust in the assisted parking process.

[0130] For example, the complete path of the second parking trajectory can be displayed in real time on the vehicle's central control screen or dashboard using dynamic lines (such as green dashed lines), including the current position → future path → target position, and updated synchronously as the vehicle moves, allowing the driver to intuitively judge whether the trajectory meets expectations.

[0131] The assisted parking method provided in this disclosure supports pre-operation selection of the target parking space and provides dynamic adjustments during the process. For offset adjustment, dynamic adjustments can be made at any time during the parking process. For trajectory planning, real-time, iterative replanning can be performed based on user input. For user demand response, continuous and smooth adjustments are allowed without interrupting the automated process. A user workflow of "select parking space → confirm → execute → adjust target → complete" can be implemented. The entire process is a closed-loop control relative to user preferences.

[0132] Furthermore, as Figure 1 An alternative approach to the method shown is provided in this disclosure embodiment as follows: Figure 2 The specific method shown includes:

[0133] Step 201: During the assisted parking process according to the first parking trajectory, in response to the adjustment command of the target parking position in the first parking trajectory, the updated vehicle target pose is obtained according to the adjustment command of the target parking position.

[0134] For example, if a vehicle is driving along its first parking trajectory (e.g., it has started reversing into a parking space and is executing the initially planned trajectory), and the driver finds that the initial target parking position is not ideal (e.g., it is parked too far to the left of the parking space, or the front of the vehicle is angled), the driver can actively input a command to adjust the target parking position (e.g., issuing a voice control message "move 10cm to the right", or dragging a virtual parking space icon on the touch screen). The system will then initiate a target pose update, that is, based on the target parking position adjustment command, it will obtain the updated vehicle target pose, which may include the vehicle position (e.g., lateral position, longitudinal position) and yaw angle.

[0135] In some examples, the updated vehicle target pose is obtained based on the adjustment command for the target parking position. Specifically, this may include: determining offset parameters based on the adjustment command for the target parking position, wherein the offset parameters include at least one of the following: lateral offset (the amount of adjustment of the vehicle in the direction perpendicular to the length of the parking space (e.g., in a longitudinal parking space, lateral means left and right direction, +5cm means moving 5cm to the right, -3cm means moving 3cm to the left)); longitudinal offset (the amount of adjustment of the vehicle in the direction parallel to the length of the parking space (e.g., in a longitudinal parking space, longitudinal means front and back direction, +8cm means moving forward 8cm more, -6cm means moving backward 6cm more)); and yaw angle adjustment (the adjustment angle of the vehicle's orientation, e.g., if the original target yaw angle is parallel to the edge of the parking space, +3° means the front of the vehicle yaws 3° to the right, -2° means the front of the vehicle yaws 2° to the left, ensuring the vehicle body is aligned with the edge of the parking space); and then determining the updated vehicle target pose based on the offset parameters and the vehicle target pose before the update.

[0136] For example, the parameters of the vehicle target pose are position (x, y coordinates) + yaw angle (θ). The update logic is based on the original target pose + offset parameters: assuming the vehicle target pose before the update is (x0, y0, θ0), and the offset parameters are lateral Δx = +10cm, longitudinal Δy = 0, and yaw angle Δθ = -2°, then the updated target pose is: new position (x0 + 10cm, y0 + 0), new yaw angle (θ0 - 2°). In this way, the driver's fuzzy adjustment requirements are transformed into precise target parameters that the system can recognize, providing a clear endpoint target for the subsequent generation of the second parking trajectory.

[0137] Step 202: Generate a second parking trajectory based on the current vehicle status and the updated vehicle target pose.

[0138] The vehicle's current state can be the vehicle's real-time dynamic parameters, such as the current position x1, y1, current yaw angle θ1, and current speed v1, which can be collected in real time by onboard sensors (such as IMU, wheel speed sensors, surround view cameras, etc.).

[0139] In some embodiments, the current state of the vehicle and the updated target pose of the vehicle can be input into the model through model prediction to obtain a second parking trajectory.

[0140] Step 203: Perform assisted parking according to the second parking trajectory.

[0141] In some embodiments, assisted parking according to a second parking trajectory may specifically include: determining the angle between the tangent direction of the current point of the first parking trajectory and the tangent direction of the starting point of the second parking trajectory; if the angle is less than a preset angle threshold, then assisted parking is performed according to the second parking trajectory. If the angle is greater than or equal to the preset angle threshold, then a transition trajectory segment is generated, which can be used for the vehicle to smoothly transition from the first parking trajectory to the second parking trajectory.

[0142] For example, directly switching from the first parking trajectory to the second parking trajectory might cause the vehicle to turn sharply due to the significant difference in the initial directions of the two trajectories. For instance, if the tangent direction at the current point of the first parking trajectory is "45° backward," while the tangent direction at the starting point of the second parking trajectory is "15° backward," the angle between them is 30°. A direct switch would cause the steering wheel to turn suddenly, affecting smoothness. Therefore, this embodiment determines the switching method by judging the angle, i.e., selecting whether a transition is needed based on the directional difference.

[0143] When the initial direction difference between the two tracks is small, the vehicle can smoothly switch without additional adjustments. For example, if the vehicle is reversing along the first track (the tangent direction at the current point is "5° backward," meaning the front of the car is slightly turned 5° to the right while reversing), and the tangent direction at the starting point of the second track is "7° backward," with an angle of only 2° (<5°), then only a slight adjustment to the steering wheel is needed (0.5° from the original steering angle). The vehicle can then naturally transition from the first parking track to the second parking track without the need to generate an additional transition segment, ensuring smooth driving.

[0144] When the initial directions of two parking trajectories differ significantly, a transitional trajectory segment is needed to connect them and avoid sharp turns. The main function of the transitional trajectory segment is to serve as a buffer path from the current point of the first parking trajectory to the starting point of the second parking trajectory. It must satisfy the requirements of continuous curvature and smooth changes in steering angle (e.g., generated using B-spline curves, ensuring a gradual transition from the tangent direction of the first parking trajectory to the tangent direction of the second trajectory without abrupt angle changes). For example, if the tangent direction of the current point of the first trajectory is "10° backward" and the tangent direction of the starting point of the second trajectory is "25° backward", with an included angle of 15° (≥10°), the system generates a transitional trajectory segment approximately 1m in length, gradually transitioning from "10° backward" to "25° backward". The vehicle first executes the transitional trajectory segment (the steering wheel is slowly turned, and the steering angle is gradually adjusted from the original angle to the angle required by the second trajectory). After reaching the end of the transitional trajectory segment (i.e., the starting point of the second parking trajectory), the vehicle continues to execute the second parking trajectory, without any sharp turns or jerks throughout the entire process.

[0145] The methods disclosed herein relate to advanced driver assistance systems (ADAS) and intelligent parking assistance. They can be used for autonomous vehicle motion planning and trajectory control in confined environments such as parking lots. They not only encompass traditional technologies for assisted parking systems, such as perception, planning, and control, but also innovatively integrate human-machine interface (HMI) design with real-time motion planning algorithms. This disclosure establishes a continuously effective, user-guided closed-loop control system during the parking process. This disclosure enables a real-time, user-guided dynamic trajectory correction system in semi-autonomous vehicles. Furthermore, compared to a single parking scenario, it can be extended to other low-speed, precise control tasks requiring human-machine collaboration.

[0146] Furthermore, as an optional approach to the method of step 202, embodiments of this disclosure provide, as follows: Figure 3 The specific method shown includes:

[0147] Step 301: Obtain the reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle.

[0148] The reference trajectory can be an ideal path from the vehicle's current state to the updated target pose, or the optimal route the system expects the vehicle to travel. It can be generated based on the updated target pose (such as endpoint coordinates and yaw angle) through polynomial fitting (such as a fifth-order polynomial) or curve generation algorithms (such as B-splines), ensuring a continuous and smooth path (without abrupt changes in position, velocity, or acceleration), and theoretically, it can accurately reach the target pose.

[0149] For example, if the vehicle is currently at (x=2, y=3, θ=0°), and the target pose after the update is (x=5, y=3, θ=0°), the reference trajectory can be a straight line from (2,3) to (5,3) (simple scenario), or a smooth curve containing steering (complex scenario), etc.

[0150] In some embodiments, the possible driving trajectory of the vehicle in the future (e.g., the next 10 control cycles, each 10ms) can be predicted based on the vehicle's current actual state (position, speed, yaw angle, etc.) and combined with the vehicle motion model.

[0151] In some embodiments, predicting the future trajectory of a vehicle in at least one control cycle based on its current state may specifically include: predicting the future trajectory of a vehicle in at least one control cycle based on its current state and a dynamic model, wherein the dynamic model can be used to determine the dynamic relationship between vehicle target parameters and steering angle and acceleration. The vehicle target parameters include at least one of vehicle position, yaw angle, and velocity. The dynamic model can be a kinematic bicycle model, a kinematic single-track model, a differential drive model, etc. For example, the dynamic model can be used to describe how the vehicle target parameters (position (x,y), yaw angle θ, velocity v) dynamically change with control inputs (steering angle δ, acceleration a).

[0152] For example, the steering angle δ determines the vehicle's turning direction and curvature, while the acceleration a determines the vehicle's speed increase or decrease. Given the vehicle's current state (x0, y0, θ0, v0), and assuming the steering angle and acceleration for the next 10 control cycles (a set of initial values ​​can be preset), the model calculates the state for each cycle:

[0153] First cycle: Based on δ1 and a1, calculate the new position (x1, y1), new yaw angle θ1, and new velocity v1;

[0154] Second cycle: Based on x1, y1, θ1, v1 and δ2, a2, calculate (x2, y2, θ2, v2); and so on, to obtain the future state trajectory (composed of 10 state points) for the next 10 cycles. Through this model, the system can predict the vehicle's driving path under different control inputs, providing a trial-and-error benchmark for subsequent optimization of control commands.

[0155] Step 302: Generate a second parking trajectory based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and in combination with the rate of change of the control input corresponding to the vehicle in at least one control cycle.

[0156] Among them, the error information represents the deviation between the future state trajectory and the reference trajectory (such as position deviation Δx, Δy, yaw angle deviation Δθ), reflecting the gap between the actual path and the ideal path.

[0157] The rate of change of control input includes the rate of change of steering angle and / or the rate of change of acceleration. The difference in control quantities between adjacent control cycles (e.g., the rate of change of steering angle Δδ = δ2 - δ1, the rate of change of acceleration Δa = a2 - a1) reflects the smoothness of the control action; that is, the smaller the rate of change, the smoother the vehicle travels, avoiding sharp turns and sudden braking. This embodiment of the present disclosure finds a set of control commands by comprehensively evaluating the magnitude of the error and the rate of change of control input, enabling the vehicle to both closely follow the reference trajectory (small error) and travel smoothly (small rate of change of control input), ultimately forming a second parking trajectory.

[0158] In some embodiments, step 302 may specifically include: substituting the error information, control input change rate, and end-state error corresponding to the first control cycle into the target cost function to calculate the cost information, wherein the end-state error is the deviation between the vehicle state at the last step in the prediction time domain corresponding to the first control cycle and the updated vehicle target pose; under the condition of satisfying vehicle physical constraints and environmental safety constraints, with the goal of reducing cost information, analyzing the target control sequence in the prediction time domain through a quadratic programming algorithm; executing the control instructions of the target control sequence in the first control cycle to obtain the vehicle state updated in the first control cycle; based on the vehicle state updated in the first control cycle, iteratively executing the cost information calculation, analysis of the target control sequence, and obtaining the vehicle state updated in the next control cycle for the next control cycle, until the finally updated vehicle state reaches the updated vehicle target pose; generating a second parking trajectory based on the vehicle state corresponding to each control cycle.

[0159] For example, by substituting error information, control input change rate, and terminal state error into the target cost function, quantified cost information is obtained (the smaller the value, the better the trajectory). Here, error information represents the deviation between the future state trajectory and the reference trajectory within the current prediction period (e.g., Δx, Δy, Δθ at each step); control input change rate represents the steering angle change rate Δδ and acceleration change rate Δa (penalizing excessive changes to ensure smoothness); terminal state error represents the deviation between the vehicle state and the target pose at the last step in the prediction time domain (e.g., step 10) (heavily penalized to ensure accurate target arrival). As an optional approach, the cost function can be the sum of squared errors + the sum of squared control input change rates + the sum of squared terminal errors (weighted), adjusting the importance of each factor through weights (e.g., higher weight for terminal error).

[0160] Under the condition of meeting the vehicle's physical constraints (such as steering angle δ≤±45°, acceleration a≤±2m / s²), 2 Under the premise of environmental safety constraints (such as a distance of ≥0.3m from obstacles), a quadratic programming algorithm is used to find the target control sequence that minimizes the cost information, such as the steering angle and acceleration for the next 10 cycles.

[0161] Then, only the first control command of the target control sequence (such as δ1 and a1 in the first cycle) can be executed to drive the vehicle and obtain the updated vehicle state (i.e., the new x, y, θ, v) at the end of that cycle. For example, the first control command could be "steering angle 5°, acceleration 0.5 m / s²". 2 After execution, the vehicle moves to a new location, and the sensors provide real-time feedback on the new status.

[0162] Upon entering the next control cycle, the above process is repeated based on the updated vehicle state: re-predict the future trajectory; recalculate the error and cost information; re-solve the new target control sequence; execute the first instruction of the new sequence... This process iterates until the vehicle state reaches the updated target pose. Finally, the vehicle states from all cycles are concatenated to form the complete second parking trajectory.

[0163] As an example, rolling optimization can be performed based on the current state of the system in each control cycle, and various complex constraints can be explicitly handled, which is crucial for ensuring the safety and smoothness of the parking process. Vehicle Model: Since parking is performed at low speeds, a kinematic bicycle model can be used to accurately describe the vehicle's motion behavior. The system state vector x of this model can be the position of the rear axle center in the world coordinate system, ψ is the vehicle's yaw angle, and v is the vehicle speed. The control input vector u is defined as u = [a, δ]T, where a is the longitudinal acceleration and δ is the front wheel steering angle.

[0164] To achieve optimal control in a finite time domain at each discrete time step k, an objective function is designed: a cost function J is designed and minimized over a prediction time domain N. This function aims to penalize the deviation between the predicted trajectory and the reference trajectory, the magnitude of the control input, and the rate of change of the control input (to ensure smoothness and comfort).

[0165] Constraints: The above optimization problem must be solved under a series of constraints to ensure that the generated trajectory is safe and physically feasible:

[0166] Vehicle dynamics constraints: that is, the motion of a vehicle must follow a predefined kinematic bicycle model.

[0167] Actuator physical limitations: The front and rear wheel steering angles, acceleration, and steering angular velocity must all be within the vehicle's physical limits.

[0168] Collision avoidance constraint: The vehicle's geometry must maintain a safe distance from static obstacles (such as other vehicles, walls, and pillars) detected by the perception system throughout the entire prediction time domain.

[0169] This approach is not only technically feasible, but more importantly, it provides an inherent safety guarantee mechanism. While searching for the optimal path, all preset constraints must be strictly adhered to. This means that if the user issues an unsafe command through the HMI (e.g., setting the target location inside a wall or on a path that would lead to a collision), collision avoidance constraints will be prioritized. While a solution that precisely reaches the user-specified target may not be found during the process, a solution that is closest to the target will be found while all safety constraints are satisfied. This fail-safe characteristic allows the system to safely delegate a portion of control (i.e., the right to set the target) to non-expert users, a key technical prerequisite for realizing this advanced human-computer interaction function.

[0170] Furthermore, as another alternative approach to the method of step 202, embodiments of this disclosure provide, as follows: Figure 4 The specific method shown includes:

[0171] Step 401: Adjust the first parking trajectory based on the current vehicle status and the updated vehicle target pose to obtain the second parking trajectory.

[0172] Without interrupting the parking process, the original first parking trajectory is adjusted to a second parking trajectory that can accurately reach the new target.

[0173] In some embodiments, step 401 may specifically include: analyzing trajectory adjustment parameters based on the current vehicle state and the updated vehicle target pose. The trajectory adjustment parameters include lateral adjustment (the vehicle's positional offset perpendicular to the driving direction (e.g., the parking space width direction) (e.g., "adjusting 0.1m to the right"), calculated using the lateral coordinate difference between the current state and the new target pose), longitudinal adjustment (the vehicle's positional offset parallel to the driving direction (e.g., the parking space length direction) (e.g., "adjusting 0.05m forward"), calculated using the longitudinal coordinate difference), and angle adjustment (the correction amount of the vehicle's yaw angle (e.g., "from the original trajectory..."). The trajectory adjustment parameters are used to perform coordinate transformation on the path points of the first parking trajectory to generate a temporary transition trajectory. The smoothness of the temporary transition trajectory is verified; if the trajectory curvature change rate exceeds a preset threshold, the transition segment is optimized by spline curve interpolation. Based on vehicle dynamics constraints and environmental obstacle information, the feasibility of the optimized transition trajectory is verified. The trajectory segment that passes the feasibility verification is connected to the updated vehicle target pose to generate the second parking trajectory.

[0174] For example, if the vehicle is currently at (x = 2.0, y = 5.0, θ = 3°), and the new target pose is (x = 2.1, y = 5.05, θ = 0°), then the adjustment parameters are "lateral +0.1m, longitudinal +0.05m, angle -3°". In this way, the original path points of the first parking trajectory are offset or rotated as a whole, generating a preliminary transition trajectory that fits the new target, avoiding the need to plan the trajectory from scratch (saving computational resources).

[0175] The first parking trajectory consists of a series of discrete path points (such as P1(x1,y1,θ1), P2(x2,y2,θ2)...P n (x n ,y n ,θ n The conversion process involves adjusting each point according to the adjustment parameters. Specifically, the x and y coordinates of each path point are adjusted by adding lateral / longitudinal adjustments (e.g., x becomes x1+0.1m and y becomes y1+0.05m for P1); the yaw angle θ of each path point is adjusted by adding angle adjustments (e.g., θ becomes θ1-3° for P1). At the same time, the position coordinates are corrected using a coordinate rotation formula (to ensure that the position is consistent with the heading after the angle adjustment).

[0176] The smoothness of the temporary transition trajectory is checked. If the rate of change of trajectory curvature exceeds a preset threshold, spline curve interpolation is used to optimize the transition segment to resolve potential trajectory unsmoothness issues (such as sharp turns and inflection points) after coordinate transformation, ensuring a smooth driving experience. The smoothness check indicators may include: rate of change of curvature (the difference in curvature between adjacent points on the trajectory; curvature reflects the steepness of a turn, and a large rate of change means the steering wheel needs to be turned suddenly). If the rate of change of curvature of a certain trajectory segment exceeds the limit (e.g., a sharp angle appears between two path points due to coordinate transformation), spline curve interpolation (such as B-splines or cubic splines) is used: new intermediate points are inserted between the path points of that trajectory segment, transforming the trajectory from a broken line into a smooth curve, ensuring continuous curvature change (smooth steering wheel rotation).

[0177] Based on vehicle dynamics constraints and environmental obstacle information, the feasibility verification of the optimized transition trajectory serves to ensure that the optimized trajectory can be actually executed by the vehicle without colliding with obstacles, filtering out trajectories that are theoretically feasible but practically infeasible. Vehicle dynamics constraints include: whether the trajectory conforms to the vehicle's physical limits, for example, minimum turning radius (the turning radius corresponding to the trajectory curvature ≥ the vehicle's minimum turning radius, avoiding situations where the steering wheel cannot be turned); maximum steering angle (the steering angle required by the trajectory ≤ the vehicle's maximum steering angle, such as ≤45°); and environmental safety constraints: whether the trajectory avoids obstacles, by comparing the trajectory path points with obstacle positions to ensure that the distance between all points and obstacles is ≥ a safety threshold (such as 0.3m). Finally, the feasibility-verified trajectory segment is connected with the updated vehicle target pose to generate a second parking trajectory, ensuring that the adjusted trajectory accurately terminates at the new target pose, forming a complete and continuous executable path.

[0178] Furthermore, as another alternative to the method shown in step 202, embodiments of this disclosure provide, as follows: Figure 5 The specific method shown includes:

[0179] Step 501: Using the current state of the vehicle as the root node, the tree branch is iteratively expanded by randomly sampling points in the state space, and valid nodes are filtered out through collision detection until a search tree connecting the current state of the vehicle and the updated vehicle target pose is generated.

[0180] In the parking scenario, the state space typically contains the vehicle's pose parameters, such as position (x, y) and yaw angle (θ). Each state can be represented as (x, y, θ), covering the vehicle's "position + orientation" information in the environment.

[0181] The root node represents the starting point of the search tree and corresponds to the current state of the vehicle (e.g., current position x0, y0, current orientation θ0).

[0182] A tree branch is a path segment connecting two states (such as a feasible trajectory from node A to node B), and a node is a discrete point in the state space (each node corresponds to a vehicle state).

[0183] For example, first, random sampling is performed, generating a random sampling point in the state space as the target direction for branch expansion. For instance, a point can be randomly selected within a parking area (e.g., a 10m x 5m parking space and its surroundings), and a random orientation (e.g., -30° to 30°) can be assigned. Then, the nearest node is found in the generated search tree, identifying the node (x_near, y_near, θ_near) closest to the sampling point, which serves as the expansion starting point. The distance can be calculated using pose distance (e.g., a weighted sum of Euclidean distance and yaw angle differences), ensuring the expansion direction is close to the sampling point. Then, the branch is expanded and collisions are detected. Starting from the nearest node, a new node (x_new, y_new, θ_new) is generated along the direction from the nearest node to the sampling point. The distance between the new node and the nearest node can be a preset step size (e.g., 0.5m, to avoid excessively long branches). Simultaneously, a collision detection algorithm (e.g., comparing the path from the new node to the nearest node with the obstacle outline) is used to determine if the branch is collision-free: if the path and obstacle do not intersect, the new node and branch are valid and added to the search tree; otherwise, they are discarded. Finally, iterative expansion continues until the target is connected, that is, repeating the process of "sampling → finding the nearest node → expanding branches → collision detection," continuously adding nodes and branches to the tree until the endpoint of a branch (the new node) is close enough to the updated vehicle target pose (e.g., distance < 0.3m, yaw angle difference < 5°). At this point, the search tree has connected the current state with the new target pose.

[0184] In some embodiments, the current state of the vehicle is used as the root node, and the tree is iteratively expanded by randomly sampling points in the state space. Specifically, this may include: taking the current state of the vehicle as the root node, and according to the direction of the updated vehicle target pose in the state space, iteratively expanding the tree by randomly sampling points in the state space.

[0185] For example, compared to random sampling throughout space, which may waste computational resources in areas far from the target, embodiments of this disclosure can use target-biased sampling: increasing the sampling probability in areas near the new target pose, causing branches to grow preferentially towards the target, and generating paths connecting the target faster.

[0186] Step 502: Select the second parking trajectory from the search tree.

[0187] For example, starting from a node near the target, trace back to its parent node (each node records its own parent node) until the root node is reached, forming a complete candidate path; calculate the cost of the candidate path, and if there are multiple candidate paths, select the one with the lowest cost as the second parking trajectory; check whether the newly generated node can form a lower-cost path with other nodes in the tree, replace the original branches, and further improve the path optimality.

[0188] Based on the content of the above embodiments, as an optional implementation, the system architecture of this disclosure mainly consists of four layers:

[0189] The perception layer consists of a standard set of onboard sensors, including but not limited to multiple ultrasonic sensors, at least four fisheye cameras for generating 360-degree surround-view images, and optional short-range millimeter-wave radar. These sensors provide raw data for parking space recognition, surrounding obstacle mapping, and vehicle localization. The system can leverage an advanced bird's-eye view perception framework to fuse multi-sensor data, generating a unified and accurate semantic understanding of the surrounding environment.

[0190] Processing Unit (ECU): Typically a high-performance automotive-grade central computing platform, such as a domain controller employing a high-performance System-on-Chip (SoC). This unit is the "brain" of the system, responsible for running core software algorithm modules, including: multi-sensor data fusion algorithms; rendering and logic processing engines for the human-machine interface (HMI); dynamic motion planning modules; and low-level vehicle control modules that send commands to lower-level actuators.

[0191] Human-Machine Interface (HMI): Typically, this is the vehicle's central infotainment touchscreen. It presents the driver with a 360-degree surround-view image rendered in real-time from camera data, overlaying an interactive layer on top of this image. This layer provides intuitive UI controls, such as directional arrows, draggable vehicle silhouettes, or adjustable target boxes, allowing users to easily issue adjustment commands. The design principle of the HMI is to minimize driver distraction and ensure intuitive and safe interaction.

[0192] The execution layer (actuators) consists of the vehicle's drive-by-wire system, mainly including the electronic power steering (EPS) system, brake-by-wire system, and powertrain control system (used to control vehicle speed). These actuators receive precise commands from the ECU (such as target steering angle and target acceleration) and translate them into the vehicle's actual physical motion.

[0193] As an example, the operation process and user interaction are as follows:

[0194] Step 1: System Activation and Parking Space Detection: The driver activates the system while driving at low speed. The system uses sensors to automatically search for and identify available parking spaces.

[0195] Step 2: Initial trajectory calculation: After the driver selects and confirms a target parking space on the central control screen, the dynamic motion planning module calculates an initial optimal trajectory from the current position to the target parking space.

[0196] Step 3: Operation Execution and Monitoring: The system begins to control the vehicle and park it according to the initial trajectory. Simultaneously, the HMI display shows a real-time 360-degree surround view image, overlaid with the vehicle's predicted driving path.

[0197] Step 4: User-Initiated Adjustment: During parking, the driver observes the surround-view camera or the external environment to determine if the final parking position needs adjustment. The driver interacts with the interactive controls on the HMI touchscreen. For example, incremental adjustment: tapping the "left / right" arrow on the screen moves the preset target parking position by a fixed increment (e.g., 10 centimeters) in the corresponding direction with each tap. This interaction is parsed by the HMI system and generates a target adjustment vector.

[0198] Step 5: Dynamic replanning: The HMI sends the target adjustment vector to the dynamic motion planning module in the ECU.

[0199] The planning module updates the final target pose based on this vector. Immediately starting with the vehicle's current real-time state (position, attitude, velocity, etc.) and ending with the updated target pose, the planning module solves a new optimal control problem. Its output is a completely new trajectory that smoothly transitions from the current state to the new target state.

[0200] Step 6: Seamlessly execute the new trajectory: After receiving the new trajectory, the underlying control module immediately begins tracking and executing it, and the vehicle's movement smoothly changes accordingly, heading towards the new target. The driver can repeat steps 4 and 5 as needed for multiple fine-tuning steps.

[0201] Step 7: Parking Completed: The vehicle precisely reaches the final adjusted target position, the system automatically brakes the vehicle to a stop, shifts into P gear, and notifies the driver that parking is complete via HMI.

[0202] The following aspects can be achieved by applying the embodiments of this disclosure:

[0203] Improved convenience and efficiency: Drivers no longer need to interrupt and restart the entire parking process due to minor adjustments, significantly saving time and reducing operational complexity.

[0204] Enhanced safety and adaptability: Allows drivers to make evasive adjustments to unforeseen static obstacles or tight spaces during the parking process, improving parking success rate and safety.

[0205] Improving user experience and trust: Giving drivers greater control over the automation process makes them feel more confident and at ease, thereby enhancing their overall trust and satisfaction with the assisted parking system.

[0206] Figure 6This is a block diagram of an auxiliary parking device according to some embodiments of the present disclosure, which can be configured to perform... Figures 1 to 5 The method shown. Refer to... Figure 6 The device includes an acquisition module 61 and a control module 62.

[0207] The acquisition module 61 is configured to acquire the adjusted second parking trajectory in response to an adjustment command for the first parking trajectory during the assisted parking process according to the first parking trajectory.

[0208] The control module 62 is configured to perform assisted parking according to the second parking trajectory.

[0209] In some embodiments, the control module 62 is further configured to determine an adjustment instruction for the first parking trajectory based on at least one of user voice information, user gesture information, and operation information on the vehicle control screen.

[0210] In some embodiments, the control module 62 is further configured to display a function item in the vehicle control screen that allows adjustment of the first parking trajectory; and to obtain operation information of the vehicle control screen in response to a trigger command of the function item.

[0211] In some embodiments, the control module 62 is further configured to output a prompt message if a parking risk is determined based on the image around the vehicle and the second parking trajectory.

[0212] In some embodiments, the control module 62 is further configured to identify obstacles in the image around the vehicle, obtain the position coordinates of the obstacles in the coordinate system of the second parking trajectory; calculate the distance between the path point coordinates of the second parking trajectory and the position coordinates of the obstacles; and output a collision risk warning message if at least one path point is less than a preset safety threshold.

[0213] In some embodiments, the control module 62 is further configured to identify the parking space boundary line through the image around the vehicle to determine the effective parking area of ​​the parking space; if the target parking position of the second parking trajectory exceeds the effective parking area, a warning message indicating a risk of exceeding the boundary is output.

[0214] In some embodiments, the control module 62 is further configured to analyze the motion trend of the obstacle through multiple consecutive frames of images around the vehicle; if the motion trend predicts that the obstacle will intersect with the second parking trajectory within a preset time period in the future, a warning message indicating a dynamic collision risk is output.

[0215] In some embodiments, the control module 62 is further configured to display the vehicle's expected parking path based on the second parking trajectory during the assisted parking process.

[0216] In some embodiments, the acquisition module 61 is specifically configured to, in response to an adjustment instruction for the target parking position in the first parking trajectory, acquire an updated vehicle target pose according to the adjustment instruction for the target parking position; and generate a second parking trajectory according to the current vehicle state and the updated vehicle target pose.

[0217] In some embodiments, the acquisition module 61 is specifically configured to determine offset parameters according to the adjustment instruction of the target parking position, wherein the offset parameters include at least one of lateral offset, longitudinal offset and yaw angle adjustment; and determine the updated vehicle target pose based on the offset parameters and the vehicle target pose before the update.

[0218] In some embodiments, the acquisition module 61 is specifically configured to acquire a reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and to predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle; and to generate a second parking trajectory based on the error information between the future state trajectory of the vehicle in the at least one control cycle and the reference trajectory, and in combination with the control input change rate corresponding to the vehicle in the at least one control cycle, wherein the control input change rate includes the steering angle change rate and / or acceleration change rate.

[0219] In some embodiments, the acquisition module 61 is specifically configured to substitute the error information, control input change rate, and end-state error corresponding to the first control cycle into the target cost function to calculate the cost information, wherein the end-state error is the deviation between the vehicle state at the last step in the prediction time domain corresponding to the first control cycle and the updated vehicle target pose; under the condition of satisfying vehicle physical constraints and environmental safety constraints, with the goal of reducing the cost information, the target control sequence in the prediction time domain is analyzed by a quadratic programming algorithm; the control instructions of the target control sequence are executed in the first control cycle to obtain the vehicle state updated in the first control cycle; based on the vehicle state updated in the first control cycle, the cost information calculation, target control sequence analysis, and vehicle state updated in the next control cycle are iteratively executed until the finally updated vehicle state reaches the updated vehicle target pose; and the second parking trajectory is generated based on the vehicle state corresponding to each control cycle.

[0220] In some embodiments, the acquisition module 61 is specifically configured to predict the future trajectory of the vehicle in at least one control cycle based on the current state of the vehicle and in conjunction with a dynamic model, wherein the dynamic model is used to determine the dynamic relationship between vehicle target parameters and steering angle and acceleration, and the vehicle target parameters include at least one of vehicle position, yaw angle and speed.

[0221] In some embodiments, the acquisition module 61 is specifically configured to adjust the first parking trajectory based on the current state of the vehicle and the updated vehicle target pose to obtain the second parking trajectory.

[0222] In some embodiments, the acquisition module 61 is specifically configured to: analyze trajectory adjustment parameters based on the current state of the vehicle and the updated vehicle target pose, wherein the trajectory adjustment parameters include at least one of lateral adjustment, longitudinal adjustment, and angular adjustment; perform coordinate transformation on the path points of the first parking trajectory according to the trajectory adjustment parameters to generate a temporary transition trajectory; perform smoothness verification on the temporary transition trajectory, and if the trajectory curvature change rate exceeds a preset threshold, optimize the transition segment through spline curve interpolation; perform feasibility verification on the optimized transition trajectory based on vehicle dynamics constraints and environmental obstacle information; and connect the trajectory segment that has passed the feasibility verification with the updated vehicle target pose to generate the second parking trajectory.

[0223] In some embodiments, the acquisition module 61 is specifically configured to take the current state of the vehicle as the root node, iteratively expand the tree by randomly sampling points in the state space, and filter out valid nodes by collision detection until a search tree connecting the current state of the vehicle and the updated vehicle target pose is generated; and select the second parking trajectory from the search tree.

[0224] In some embodiments, the acquisition module 61 is specifically configured to take the current state of the vehicle as the root node and expand the tree branch iteratively by randomly sampling points in the state space according to the direction of the updated vehicle target pose in the state space.

[0225] In some embodiments, the control module 62 is specifically configured to determine the angle between the tangent direction of the current point of the first parking trajectory and the tangent direction of the starting point of the second parking trajectory; if the angle is less than a preset angle threshold, then assisted parking is performed according to the second parking trajectory.

[0226] In some embodiments, the control module 62 is specifically configured to generate a transition trajectory segment if the included angle is greater than or equal to the preset angle threshold. The transition trajectory segment is used for the vehicle to smoothly transition from the first parking trajectory to the second parking trajectory.

[0227] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0228] It should be noted that other corresponding descriptions of the functional units involved in the assisted parking device provided in this disclosure embodiment can be found by referring to... Figures 1 to 5 The corresponding descriptions in [the document] will not be repeated here.

[0229] Based on the above, Figures 1 to 5 Accordingly, this disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figures 1 to 5 The method shown.

[0230] Based on this understanding, the technical solution of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, or portable hard drive) and includes several instructions to cause a computer device (such as a personal computer, server, or network device) to execute the methods of various implementation scenarios of this disclosure.

[0231] Based on the above, Figures 1 to 5 The method shown, and Figure 6 To achieve the above objectives, this disclosure also provides an electronic device, comprising a storage medium and a processor; the storage medium for storing a computer program; and the processor for executing the computer program to implement the above-described virtual device embodiments. Figures 1 to 5 The method shown.

[0232] Optionally, the aforementioned electronic device may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0233] Those skilled in the art will understand that the physical device structure provided in the embodiments of this disclosure does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0234] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0235] Figure 7 This is a block diagram illustrating a vehicle 600 according to an exemplary embodiment. For example, vehicle 600 can be a hybrid vehicle, a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicle. Vehicle 600 can be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0236] Reference Figure 7 The vehicle 600 may include various subsystems, such as an infotainment system 610, a perception system 620, a decision control system 630, a drive system 640, and a computing platform 650. The vehicle 600 may also include more or fewer subsystems, and each subsystem may include multiple components. Furthermore, each subsystem and each component of the vehicle 600 can be interconnected via wired or wireless means.

[0237] In some embodiments, the infotainment system 610 may include a communication system, an entertainment system, and a navigation system, etc.

[0238] The perception system 620 may include several sensors for sensing information about the environment surrounding the vehicle 600. For example, the perception system 620 may include a global positioning system (which may be GPS, BeiDou, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0239] The decision control system 630 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.

[0240] The drive system 640 may include components that provide powered motion to the vehicle 600. In one embodiment, the drive system 640 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of internal combustion engines, electric motors, and compressed air engines. The engine is capable of converting energy provided by the energy source into mechanical energy.

[0241] Some or all of the functions of vehicle 600 are controlled by computing platform 650. Computing platform 650 may include at least one processor 651 and memory 652, processor 651 can execute instructions 653 stored in memory 652.

[0242] Processor 651 can be any conventional processor, such as a commercially available CPU. The processor may also include, for example, a Graphics Processing Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), or a combination thereof.

[0243] The memory 652 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0244] In addition to instruction 653, memory 652 can also store data, such as road maps, route information, vehicle position, direction, speed, and other data. The data stored in memory 652 can be used by computing platform 650.

[0245] In this embodiment of the disclosure, the processor 651 may execute instructions 653 to complete all or part of the steps of the audio processing method described above.

[0246] Based on the above, Figures 1 to 5 The method shown, and Figure 6 The virtual device embodiment shown in this disclosure also provides a chip, including one or more interface circuits and one or more processors; the interface circuits are used to receive signals from the memory of an electronic device and send the signals to the processors, the signals including computer instructions stored in the memory; when the processor executes the computer instructions, it causes the electronic device to perform the above-described... Figures 1 to 5 The method shown.

[0247] Through the above description of the embodiments, those skilled in the art can clearly understand that this disclosure can be implemented using software plus necessary general-purpose hardware platforms, or it can be implemented using hardware. By applying the technical solutions of the embodiments of this disclosure, by responding in real time to the adjustment command of the first parking trajectory and generating the second parking trajectory, the driver is allowed to directly correct the parking trajectory during the parking process without interrupting the parking process. This enables flexible interaction of adjusting while parking, is simple to operate and highly efficient, reduces user operating costs, and meets the personalized parking preferences of different drivers.

[0248] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0249] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A parking assistance method, characterized in that, include: During the assisted parking process according to the first parking trajectory, in response to the adjustment command of the first parking trajectory, the adjusted second parking trajectory is obtained; Assisted parking is performed according to the second parking trajectory.

2. The method according to claim 1, characterized in that, The method further includes: An adjustment instruction for the first parking trajectory is determined based on at least one of the user's voice information, user gesture information, and operation information on the vehicle control screen.

3. The method according to claim 2, characterized in that, The method further includes: The vehicle control screen displays functions that allow adjustment of the first parking trajectory. In response to the trigger command of the function item, operation information for the vehicle control screen is obtained.

4. The method according to claim 1, characterized in that, The method further includes: If a parking risk is determined based on the images around the vehicle and the second parking trajectory, a warning message is output.

5. The method according to claim 4, characterized in that, When a parking risk is determined based on the image around the vehicle and the second parking trajectory, a prompt message is output, including: Obstacles in the image around the vehicle are identified, and the position coordinates of the obstacles in the coordinate system of the second parking trajectory are obtained; Calculate the distance between the path point coordinates of the second parking trajectory and the position coordinates of the obstacle; If at least one waypoint is less than a preset safety threshold from the obstacle, a collision risk warning message will be output.

6. The method according to claim 4, characterized in that, When a parking risk is determined based on the image around the vehicle and the second parking trajectory, a prompt message is output, including: The effective parking area of ​​the parking space is determined by identifying the boundary line of the parking space through the image around the vehicle; If the target parking location of the second parking trajectory exceeds the effective parking area, a warning message indicating a risk of exceeding the boundary will be output.

7. The method according to claim 4, characterized in that, When a parking risk is determined based on the image around the vehicle and the second parking trajectory, a prompt message is output, including: Analyze the motion trends of obstacles using multiple consecutive frames of images surrounding the vehicle; If the motion trend predicts that the obstacle will intersect with the second parking trajectory within a preset time period, a warning message indicating a dynamic collision risk will be output.

8. The method according to claim 1, characterized in that, The method further includes: During the assisted parking process following the second parking trajectory, the vehicle's expected parking path is displayed based on the second parking trajectory.

9. The method according to any one of claims 1 to 8, characterized in that, The step of obtaining the adjusted second parking trajectory in response to the adjustment command for the first parking trajectory includes: In response to an adjustment command for the target parking position in the first parking trajectory, the updated vehicle target pose is obtained according to the adjustment command for the target parking position; The second parking trajectory is generated based on the current vehicle status and the updated vehicle target pose.

10. The method according to claim 9, characterized in that, The step of obtaining the updated vehicle target pose based on the adjustment instruction of the target parking position includes: The offset parameters are determined according to the adjustment command of the target parking position, wherein the offset parameters include at least one of lateral offset, longitudinal offset and yaw angle adjustment. The updated vehicle target pose is determined based on the offset parameters and the vehicle target pose before the update.

11. The method according to claim 9, characterized in that, The step of generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes: Obtain a reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle. Based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and combined with the control input change rate corresponding to the vehicle in at least one control cycle, a second parking trajectory is generated, wherein the control input change rate includes the steering angle change rate and / or acceleration change rate.

12. The method according to claim 11, characterized in that, The step of generating the second parking trajectory based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and in combination with the rate of change of the control input corresponding to the vehicle in at least one control cycle, includes: The error information, control input change rate, and end state error corresponding to the first control cycle are substituted into the target cost function to calculate the cost information. The end state error is the deviation between the vehicle state in the last step of the prediction time domain corresponding to the first control cycle and the updated vehicle target pose. Under the conditions of satisfying vehicle physical constraints and environmental safety constraints, with the goal of reducing the cost information, the target control sequence in the prediction time domain is analyzed by a quadratic programming algorithm. In the first control cycle, the control instructions of the target control sequence are executed to obtain the vehicle state updated in the first control cycle; Based on the vehicle state updated in the first control cycle, the cost information calculation, target control sequence analysis, and vehicle state updated in the next control cycle are performed iteratively until the finally updated vehicle state reaches the updated vehicle target pose. The second parking trajectory is generated based on the vehicle status corresponding to each control cycle.

13. The method according to claim 11, characterized in that, The prediction of the vehicle's future trajectory in at least one control cycle based on the vehicle's current state includes: Based on the current state of the vehicle and in conjunction with the dynamic model, the future trajectory of the vehicle in at least one control cycle is predicted, wherein the dynamic model is used to determine the dynamic relationship between the vehicle target parameters and the steering angle and acceleration, and the vehicle target parameters include at least one of vehicle position, yaw angle and speed.

14. The method according to claim 9, characterized in that, The step of generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes: Based on the current state of the vehicle and the updated vehicle target pose, the trajectory adjustment parameters are analyzed, and the trajectory adjustment parameters include at least one of lateral adjustment, longitudinal adjustment and angle adjustment. The coordinates of the path points of the first parking trajectory are transformed according to the trajectory adjustment parameters to generate a temporary transition trajectory; The smoothness of the temporary transition trajectory is checked. If the rate of change of trajectory curvature exceeds a preset threshold, the transition segment is optimized by spline curve interpolation. The feasibility of the optimized transition trajectory is verified based on vehicle dynamics constraints and environmental obstacle information; The trajectory segment that has passed the feasibility verification is connected with the updated vehicle target pose to generate the second parking trajectory.

15. The method according to claim 9, characterized in that, The step of generating the second parking trajectory based on the current vehicle state and the updated vehicle target pose includes: Using the current state of the vehicle as the root node, the tree branch is iteratively expanded by randomly sampling points in the state space, and valid nodes are filtered out by collision detection until a search tree connecting the current state of the vehicle and the updated vehicle target pose is generated. The second parking trajectory is selected from the search tree.

16. The method according to claim 15, characterized in that, The step of iteratively expanding the tree branch by randomly sampling points in the state space, with the current state of the vehicle as the root node, includes: Using the current state of the vehicle as the root node, the tree is iteratively expanded by randomly sampling points in the state space according to the direction of the updated vehicle target pose in the state space.

17. The method according to any one of claims 1 to 8, characterized in that, The assisted parking according to the second parking trajectory includes: Determine the angle between the tangent direction of the current point of the first parking trajectory and the tangent direction of the starting point of the second parking trajectory; If the included angle is less than a preset angle threshold, then assisted parking is performed according to the second parking trajectory; or, if the included angle is greater than or equal to the preset angle threshold, then a transition trajectory segment is generated, which is used for the vehicle to smoothly transition from the first parking trajectory to the second parking trajectory.

18. A parking assistance device, characterized in that, include: The acquisition module is configured to, during the process of assisted parking according to the first parking trajectory, in response to an adjustment command for the first parking trajectory, acquire the adjusted second parking trajectory; The control module is configured to perform assisted parking according to the second parking trajectory.

19. The apparatus according to claim 18, characterized in that, The acquisition module is specifically configured to respond to an adjustment instruction for the target parking position in the first parking trajectory, and acquire the updated vehicle target pose according to the adjustment instruction for the target parking position; The second parking trajectory is generated based on the current vehicle status and the updated vehicle target pose.

20. The apparatus according to claim 19, characterized in that, The acquisition module is specifically configured to determine offset parameters based on the adjustment command of the target parking position, wherein the offset parameters include at least one of lateral offset, longitudinal offset, and yaw angle adjustment; and to determine the updated vehicle target pose based on the offset parameters and the vehicle target pose before the update.

21. The apparatus according to claim 19, characterized in that, The acquisition module is specifically configured to acquire a reference trajectory from the current state of the vehicle to the updated target pose of the vehicle, and to predict the future state trajectory of the vehicle in at least one control cycle based on the current state of the vehicle. Based on the error information between the future state trajectory of the vehicle in at least one control cycle and the reference trajectory, and combined with the control input change rate corresponding to the vehicle in at least one control cycle, a second parking trajectory is generated, wherein the control input change rate includes the steering angle change rate and / or acceleration change rate.

22. An electronic device, characterized in that, include: processor; A memory connected to the processor, the memory storing a computer program that, when executed by the processor, implements the method of any one of claims 1 to 17.

23. A vehicle, characterized in that, include: processor; A memory connected to the processor, the memory storing a computer program that, when executed by the processor, implements the method of any one of claims 1 to 17.

24. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 17.

25. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 17.

26. A chip, characterized in that, The device includes one or more interface circuits and one or more processors; the interface circuits are configured to receive signals from the memory of an electronic device and send the signals to the processors, the signals including computer instructions stored in the memory; when the processor executes the computer instructions, the electronic device performs the method of any one of claims 1 to 17.

Images