Parking trajectory prediction method and electronic equipment

Abstract

The invention discloses a parking track prediction method and electronic equipment. The method comprises the steps of obtaining local images of visual images collected by a plurality of visual sensors in response to a parking request for a target vehicle, determining a panoramic image according to the local images, segmenting the panoramic image into a plurality of areas of a first size, and determining semantic information; acquiring radar data acquired by the ultrasonic radar, determining distance information between the target vehicle and the target obstacle, and determining attribute features of the obstacle according to the semantic information and the distance information; a grid map is determined according to the target vehicle and the target parking space, a target occupation grid map is determined according to the grid map and the obstacle attribute characteristics, a target track sequence of the target obstacle is determined, a target parking track is determined, a parking control instruction is determined according to the target parking track, and the parking control instruction is sent to the parking control system. The safe and feasible parking track is intelligently planned, and the vehicle is accurately controlled to complete efficient and automatic parking.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a parking trajectory prediction method and electronic device. Background Technology

[0002] With the development of intelligent driving technology, automatic parking, as a typical application of low-speed autonomous driving, has become an important function to improve driving convenience and safety. However, parking lot environments are usually small and changeable, with a variety of obstacles and complex distribution, making it difficult for parking systems to fully perceive the environment, accurately predict obstacle behavior, and generate safe paths.

[0003] In related technologies, many rely on a single sensor (such as ultrasound or vision alone) for obstacle detection and path planning, lacking multi-source information fusion. This results in incomplete perception and inaccurate identification of obstacle attributes (such as category and motion state). Furthermore, most solutions do not fully consider the future trajectory of dynamic obstacles, nor do they impose sufficient constraints on path planning (such as safety boundaries and vehicle dynamics restrictions). This can easily lead to unsafe or unenforceable paths, or even collision risks, making it difficult to meet the requirements for high safety and intelligent parking. Summary of the Invention

[0004] This invention provides a parking trajectory prediction method and an electronic device to solve the problem in the prior art of accurately perceiving the environment, identifying obstacles, and planning a safe and feasible parking trajectory when parking.

[0005] According to one aspect of the present invention, a parking trajectory prediction method is provided, comprising:

[0006] In response to a parking request for a target vehicle, partial images of visual images collected by multiple visual sensors are acquired, a panoramic image of the target vehicle is determined based on the multiple partial images, the panoramic image is segmented into multiple regions of a first size, and semantic information of each region is determined.

[0007] The system acquires radar data collected by ultrasonic radar, determines the distance information between the target vehicle and the target obstacle based on the radar data, and determines the obstacle attribute characteristics based on the semantic information and the distance information. The target obstacle includes static obstacles and dynamic obstacles.

[0008] A grid map is determined based on the positional relationship between the target vehicle and the target parking space in the world coordinate system. A target occupancy grid map is determined based on the grid map and the obstacle attribute characteristics. A target trajectory sequence of the target obstacle is determined based on the obstacle attribute characteristics. A target parking trajectory is determined based on the target trajectory sequence, the target occupancy grid map, and preset constraints. A parking control command is determined based on the target parking trajectory and sent to the parking control system.

[0009] According to another aspect of the present invention, a parking trajectory prediction device is provided, comprising:

[0010] A semantic information determination module is used to respond to a parking request for a target vehicle by acquiring partial images of visual images collected by multiple visual sensors, determining a panoramic image of the target vehicle based on the multiple partial images, dividing the panoramic image into multiple regions of a first size, and determining the semantic information of each region.

[0011] An attribute feature determination module is used to acquire radar data collected by ultrasonic radar, determine the distance information between the target vehicle and the target obstacle based on the radar data, and determine the obstacle attribute features based on the semantic information and the distance information, wherein the target obstacle includes static obstacles and dynamic obstacles;

[0012] The parking instruction sending module is used to determine a grid map based on the positional relationship between the target vehicle and the target parking space in the world coordinate system, determine a target occupancy grid map based on the grid map and the obstacle attribute characteristics, determine a target trajectory sequence of the target obstacle based on the obstacle attribute characteristics, determine a target parking trajectory based on the target trajectory sequence, the target occupancy grid map and preset constraints, determine a parking control instruction based on the target parking trajectory, and send the parking control instruction to the parking control system.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the parking trajectory prediction method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the parking trajectory prediction method according to any embodiment of the present invention.

[0018] The technical solution of this invention, in response to a parking request for a target vehicle, acquires partial images of visual images collected by multiple visual sensors, determines a panoramic image of the target vehicle based on the multiple partial images, divides the panoramic image into multiple regions of a first size, determines the semantic information of each region, and fuses the partial images from multiple cameras to generate a panoramic semantic map, accurately identifying the vehicle's surrounding environment and obstacle categories; acquires radar data collected by ultrasonic radar, determines the distance information between the target vehicle and target obstacles based on the radar data, and determines obstacle attribute characteristics based on the semantic information and the distance information, wherein the target obstacles include static obstacles and dynamic obstacles, and fuses radar distance and visual semantic information to accurately identify the attributes of static and dynamic obstacles to ensure safety; A grid map is determined based on the positional relationship between the target vehicle and the target parking space in the world coordinate system. A target occupancy grid map is determined based on the grid map and the obstacle attribute characteristics. A target trajectory sequence of the target obstacle is determined based on the obstacle attribute characteristics. A target parking trajectory is determined based on the target trajectory sequence, the target occupancy grid map, and preset constraints. A parking control command is determined based on the target parking trajectory and sent to the parking control system. This accurately constructs an environmental model, intelligently plans a safe trajectory, and controls the vehicle to complete parking efficiently. This solves the problem in the prior art of accurately perceiving the environment, identifying obstacles, and planning a safe and feasible parking trajectory during parking. It realizes intelligent planning of a safe and feasible parking trajectory and precise control of the vehicle to complete efficient automatic parking.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a parking trajectory prediction method provided in Embodiment 1 of the present invention;

[0022] Figure 2 This is a flowchart of a parking trajectory prediction method provided in Embodiment 2 of the present invention;

[0023] Figure 3 This is a schematic diagram of the structure of a parking trajectory prediction device according to Embodiment 3 of the present invention;

[0024] Figure 4 This is a schematic diagram of the structure of an electronic device that implements the parking trajectory prediction method of this invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0028] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0029] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0030] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0031] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0032] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0033] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0034] Example 1

[0035] Figure 1 The flowchart of a parking trajectory prediction method is provided in Embodiment 1 of the present invention. This embodiment is applicable to situations involving complex environment perception and precise parking trajectory planning in automatic parking systems. The method can be executed by a parking trajectory prediction device, which can be implemented in hardware and / or software. Optionally, it can be implemented through an electronic device, such as a mobile terminal, a PC, or a server.

[0036] like Figure 1 As shown, the method may specifically include:

[0037] S110. In response to a parking request for a target vehicle, acquire partial images of visual images collected by multiple visual sensors, determine a panoramic image of the target vehicle based on the multiple partial images, segment the panoramic image into multiple regions of a first size, and determine semantic information for each region.

[0038] The target vehicle can be understood as the vehicle currently performing an automatic parking operation. As the control object of the entire parking system, its position, attitude, and motion state are the basis for planning the parking trajectory and generating control commands. The parking request can be understood as an instruction initiated by the driver or the autonomous driving system, requiring the vehicle to automatically complete the parking operation, triggering the starting point of the entire automatic parking process and initiating modules such as sensor data acquisition, environmental perception, and path planning. The visual sensors can be understood as cameras installed around the vehicle (such as front, rear, left, and right) to collect images of the surrounding environment, including but not limited to fisheye lenses, to provide high-resolution environmental visual information, construct a panoramic view of the vehicle's surroundings, and identify semantic content such as lane lines, obstacle outlines, and parking space boundaries. The local images can be understood as raw images collected by a single visual sensor, covering only a certain direction area around the vehicle, serving as the basic input unit for stitching together a panoramic image. Each local image reflects the local environment on one side of the vehicle. The panoramic image can be understood as a composite image formed by merging multiple local images through image stitching and projection transformation, providing a top-down view of the vehicle and its surrounding environment. This provides a unified spatial reference for subsequent image segmentation and semantic understanding, facilitating the identification of parking spaces, road boundaries, and static obstacle layouts. The first-sized region can be understood as several sub-regions of the panoramic image divided into fixed sizes (such as pixel blocks or areas corresponding to physical dimensions), facilitating structured image processing, supporting region-by-region semantic analysis, and improving computational efficiency and perception accuracy. The semantic information can be understood as the labeling or identification results of the physical object category (such as "ground," "vehicle," "pedestrian," "curb," "parking line," etc.) represented by each region in the image. This information is used to understand the environmental content, distinguish drivable areas from obstacles, and provide a basis for obstacle attribute judgment and path planning.

[0039] S120. Acquire radar data collected by ultrasonic radar, determine the distance information between the target vehicle and the target obstacle based on the radar data, and determine the obstacle attribute characteristics based on the semantic information and the distance information, wherein the target obstacle includes static obstacles and dynamic obstacles.

[0040] The ultrasonic radar, as described above, can be understood as a sensor that uses ultrasonic ranging principles to detect nearby obstacles, providing high-precision distance information for these obstacles and serving as an important supplement to visual perception. The radar data, as understood above, refers to the raw data collected by the ultrasonic radar, generally including information such as distance and angle between the vehicle and obstacles in various directions, used to calculate the actual spatial relationship between the target vehicle and surrounding obstacles. The target obstacle can be understood as an object that affects the vehicle's movement or parking in a parking scenario. Target obstacles include, but are not limited to, static and dynamic obstacles. Static obstacles can be understood as objects with fixed positions, such as walls, parking posts, and curbs. Dynamic obstacles can be understood as objects that may move, such as pedestrians and other moving vehicles. The distance information, as understood above, refers to the actual spatial distance between the target vehicle and the target obstacle, usually derived from radar data, used to determine the existence of a collision risk and is a crucial basis for obstacle detection and obstacle avoidance strategy formulation. The obstacle attribute features can be understood as a characteristic description of the obstacle derived from semantic information (such as what type of obstacle it is) and distance information (how far away from the vehicle, relative speed, etc.). These features include category, distance, motion state (stationary / dynamic), speed, direction, etc., and are used to distinguish different obstacles (such as parked cars vs. pedestrians) and determine whether and how to avoid them.

[0041] Based on the above scheme, optionally, determining the obstacle attribute features according to the semantic information and the distance information includes: clustering the radar data, determining the reference point coordinates of each target obstacle according to the clustering results, transforming the reference point coordinates to the pixel coordinates of the panoramic image according to a preset coordinate transformation matrix; and determining the obstacle attribute features according to the semantic information of each target obstacle, the pixel coordinates, the distance information between the target vehicle and the target obstacle, and a preset safety distance.

[0042] The clustering can be understood as an algorithmic process of grouping multiple points detected by radar (such as multiple distance-angle points) according to spatial proximity, in order to separate points corresponding to different objects and thus identify multiple independent obstacles. The reference point coordinates can be understood as a representative coordinate point determined for each obstacle (i.e., a clustering result) after clustering the radar data, including but not limited to the geometric center of the obstacle. After determining the reference point coordinates, a safe distance can be reserved in front of the target vehicle before the target obstacle. The coordinate transformation matrix can be understood as a matrix used to transform points in one coordinate system (such as reference point coordinates in the radar coordinate system or vehicle coordinate system) to another coordinate system (such as the pixel coordinate system used in panoramic images). The pixel coordinates can be understood as a two-dimensional coordinate representation in a panoramic image, where (x, y) represents the specific pixel position of a point in the image, used to visualize the spatial location information of obstacles or combine it with visual perception information, such as determining the approximate area of ​​an obstacle in the image, assisting semantic understanding and decision-making.

[0043] An alternative implementation involves sending a hard-wired synchronization signal or a software-triggered command based on a precision clock (such as PTP) to all ultrasonic radar and vision sensors to ensure that all sensor data frames have a unified time base (timestamp deviation less than 10ms). The ultrasonic radar data is then processed through filtering and clustering; the vision images undergo distortion correction, transformation, and stitching.

[0044] Ultrasonic radar: The left radar (ID: L2) continuously returns the distance to vehicle A, and the right radar (ID: R2) returns the distance to vehicle B. The rear radar did not detect any nearby obstacles.

[0045] A lightweight U-Net semantic segmentation model (which can run in real time on an automotive chip) was used. The model takes a stitched panoramic image as input and outputs pixel-level label maps of the same size. The label categories include: background (0), drivable area (1), parking line (2), vehicle (3), pedestrian (4), pillar (5), and curb (6). For example, the visual sensor: the left rear camera captures the space between two stationary vehicles (A and B). The semantic segmentation model identifies both A and B as VEHICLE and recognizes a blurred PARKING_LINE on the ground, and judges that this is an available horizontal parking space with a confidence level of 0.9.

[0046] The coordinates (x_radar, y_radar, z_radar) of each ultrasonic obstacle obtained from clustering are transformed to the pixel coordinates (u, v) of the bird's-eye view through the pre-calibrated transformation matrices T_radar_to_body (radar to vehicle body) and T_body_to_birdview (vehicle body to bird's-eye view coordinate system).

[0047] Iterate through the pixel coordinates (u, v) of each ultrasonic point cloud cluster after transformation. In the label map output by semantic segmentation, read the semantic label and confidence score for that coordinate (u, v). Bind this semantic information (e.g., "vehicle") to the distance value d provided by the ultrasonic waves to form obstacle attribute features {x, y, label, confidence, distance} (distance can be the actual distance minus a preset safety distance). For example, the centroid coordinates (x_rr2, y_rr2) of the ultrasonic cluster are transformed to the vehicle coordinate system and then projected onto the pixel coordinates (u, v) of the panoramic image through a pre-calibrated homography matrix H.

[0048] By using this technical solution, obstacle attribute characteristics are determined by clustering radar data and combining semantic and distance information. This enables precise identification of the location of each target obstacle and its corresponding pixel coordinates in the panoramic image. Furthermore, the actual distance between the vehicle and the obstacle and the preset safety distance are used to comprehensively judge the potential threat level of the obstacle. This allows for a more accurate and detailed distinction and safety assessment of static and dynamic obstacles, thereby improving the decision-making accuracy and safety of the automatic parking system.

[0049] S130. Determine a grid map based on the positional relationship between the target vehicle and the target parking space in the world coordinate system; determine a target occupancy grid map based on the grid map and the obstacle attribute characteristics; determine the target trajectory sequence of the target obstacle based on the obstacle attribute characteristics; determine the target parking trajectory based on the target trajectory sequence, the target occupancy grid map, and preset constraints; determine a parking control command based on the target parking trajectory; and send the parking control command to the parking control system.

[0050] The world coordinate system can be understood as a unified coordinate system used to describe the position and orientation of target vehicles, parking spaces, obstacles, etc., in the actual physical space. It serves as the spatial reference basis for vehicle positioning, parking space identification, obstacle location calibration, and path planning, ensuring that all elements are analyzed within the same coordinate framework. The grid map can be understood as a map representation that divides the vehicle's surrounding environment (usually based on the world coordinate system) into a series of regularly arranged small squares (i.e., grids). Each grid represents a fixed-size spatial area in the real world. The grid map is a fundamental tool for environmental modeling, used to represent the state information of each small area in space, such as whether it is occupied, whether it is passable, and whether there are obstacles. The target occupancy grid map can be understood as a high-level environmental map obtained by further annotating the occupancy status (whether it is occupied by an obstacle, and the probability) of each grid, based on the standard grid map and combined with obstacle attribute characteristics (such as location, occupancy probability, timestamp, etc.). It is one of the core inputs for path planning, obstacle avoidance decisions, and trajectory generation. The target occupancy grid map clearly shows which areas around the vehicle are safe and passable, and which areas pose potential collision risks, thus providing a reliable environmental model for subsequent automatic parking trajectory planning. The target trajectory sequence can be understood as the predicted movement trajectory of the target obstacle within a future target time period (e.g., 2 seconds, 3 seconds). For example, for a dynamic obstacle, the position in the 1st second, the position in the 1.5th second, the position in the 2nd second, etc., constitute a continuous trajectory. For a static obstacle, the position remains unchanged, representing fixed trajectory points. The target parking trajectory can be understood as the optimal or feasible trajectory ultimately determined after considering obstacles, the environmental map, and constraints, used to guide the vehicle into the target parking space. This includes, but is not limited to, changes in position, speed, and direction over time. It is the core output of automatic parking, guiding the vehicle on how to move to safely and accurately park in the space. The preset constraints can be understood as physical or safety limitations that must be met during parking, such as maximum steering angle, minimum turning radius, speed limits, and minimum safe distance from obstacles, ensuring that the generated parking trajectory is feasible in vehicle dynamics and complies with safety regulations. The parking control commands can be understood as specific execution commands generated based on the target parking trajectory, such as steering wheel angle, accelerator / brake force, gear shifting, etc., transforming the abstract trajectory of the planning layer into actions that can be executed by the underlying actuators, thus achieving closed-loop control. The parking control system can be understood as the electronic control unit at the vehicle's underlying level that executes automatic parking actions, receiving control commands and driving the steering, braking, and drive systems to complete the actual parking operation.

[0051] Based on the above scheme, optionally, determining the target-occupied grid map according to the grid map and the obstacle attribute features includes: dividing the grid map into multiple grids according to a preset size; for each grid, determining the occupancy probability and timestamp of the obstacle attribute features falling into the grid; and constructing the target-occupied grid map according to the occupancy probability, the timestamp, and the grid map.

[0052] The preset size can be understood as the physical size of each grid cell manually set before constructing the grid map, determining the spatial resolution of environmental modeling. Smaller sizes result in a more detailed map but require more computation; larger sizes are more efficient but may lose detail. The occupancy probability can be understood as the likelihood that a grid cell is occupied by an obstacle, represented by a value between 0 and 1. 0 indicates the grid cell is completely empty and has no possibility of being occupied; 1 indicates the grid cell is almost certainly occupied by an obstacle. This value is used to handle perceptual uncertainty by fusing the confidence levels of obstacle presence from multiple sensor sources (such as visual semantics + ultrasonic distance). The timestamp can be understood as the data acquisition or prediction time point associated with a specific obstacle or grid state, used to identify the timeliness of the information. In dynamic environments, the position and state of obstacles change over time. Timestamps are used to determine the freshness of obstacle information, helping the system distinguish between currently valid obstacles and outdated data, which is especially crucial when dealing with dynamic obstacles (such as other vehicles and pedestrians).

[0053] By adopting this technical solution and constructing a target occupancy grid map by introducing occupancy probability and timestamp, the spatiotemporal distribution and confidence of obstacles can be dynamically reflected, thereby improving the parking system's perception accuracy and prediction capability for static and dynamic obstacles, and enhancing the safety and timeliness of trajectory planning.

[0054] Based on the above scheme, optionally, determining the target trajectory sequence of the target obstacle according to the obstacle attribute features includes: when the target obstacle is a dynamic obstacle, determining the initial trajectory sequence of the dynamic obstacle corresponding to the dynamic obstacle information in multiple target frames within the target time period, and inputting the initial trajectory sequence into a trained trajectory prediction model to obtain the target trajectory sequence of the target obstacle.

[0055] The target time period can be understood as a historical time window (e.g., the past 3 seconds) used to determine the motion behavior of dynamic obstacles within this time period. The target frame can be understood as a time point within the target time period divided at fixed time intervals (e.g., every 0.1 seconds), with each time point corresponding to a frame of historical output. The initial trajectory sequence can be understood as the historical motion trajectory of the dynamic obstacle fitted based on current and historical observation data (e.g., multiple frames of visual or radar detection results). The trajectory prediction model can be understood as a trained machine learning or deep learning model (e.g., LSTM, Transformer, etc.) used to predict the future trajectory of the obstacle based on its historical state.

[0056] In one alternative implementation, for a dynamic obstacle labeled as a "pedestrian" or "vehicle," its trajectory sequence over the past second (10 frames) is extracted and fed into a pre-trained LSTM trajectory prediction model. The network outputs the predicted positions of the target at six time points (0.5s intervals) within the next three seconds, along with a covariance matrix representing the prediction uncertainty. For example, the smoothed trajectory coordinate sequence of ID-103 over the past 10 frames (500ms) [(3.8, -1.2), (3.7, -1.2), (3.5, -1.2), ...]. LSTM Inference: The trajectory prediction module feeds the sequence into the deployed LSTM model. The model outputs a most probable predicted trajectory (80% probability) and a low-probability trajectory (15% probability, e.g., a pedestrian suddenly accelerating).

[0057] This technical solution utilizes a trajectory prediction model to perform temporal modeling of dynamic obstacles, accurately predicting their future movement trends and effectively improving the obstacle avoidance capability and path planning foresight of the parking system in complex dynamic environments.

[0058] Based on the above scheme, optionally, the preset constraints include at least one of the following: the starting point of the target parking trajectory matches the first vehicle driving information of the target vehicle; the ending point of the target parking trajectory matches the second vehicle driving information of the target vehicle; the target parking trajectory must not exceed the preset safety boundary; and the third vehicle driving information of the target vehicle meets the preset driving conditions of the target vehicle.

[0059] The starting point can be understood as the initial position of the target parking trajectory, i.e., the current position of the vehicle (or the position at the start of the planning process). It is the "starting point" of the trajectory. Trajectory planning must start from the actual state of the vehicle to ensure that the starting point of the trajectory matches the current position / state of the vehicle; otherwise, the vehicle cannot "enter the planned path from its actual position." The first vehicle driving information can be understood as vehicle state information related to the starting point of the target parking trajectory, including but not limited to the vehicle's current position, current heading angle, and current speed. This information is used to constrain the starting point of the target parking trajectory to be consistent with the vehicle's current actual driving state (position, orientation, speed, etc.) to avoid situations where "the trajectory starting point deviates too much from the vehicle's actual position," leading to execution failure. The ending point can be understood as the ending position of the target parking trajectory, i.e., the final target parking space position that the vehicle must reach or the parking completion state point. The second vehicle driving information matching can be understood as the ideal state that the target vehicle should meet at the moment of parking completion, including but not limited to the desired position, desired heading, parking posture, and desired speed within the target parking space. The preset safety boundary can be understood as a predefined safe area boundary in a parking scenario. It can be determined by factors such as the parking lot structure, lane lines, obstacles, and the vehicle's dimensions—a "drivable and collision-free boundary range." All points on the vehicle's trajectory must be within this boundary and cannot exceed it; otherwise, a collision or violation may occur. This limit the spatial range of the target parking trajectory, ensuring the vehicle does not exceed the safe area during its movement. The third vehicle driving information can be understood as the dynamic state information of the target vehicle at each moment during the entire parking trajectory execution process, including but not limited to real-time speed, acceleration, and steering curvature. This information is used to verify whether the entire trajectory fully meets the vehicle's physical capabilities and safety regulations. For example, the curvature cannot exceed the vehicle's minimum turning radius, the speed cannot be too high to cause loss of control, and the distance to obstacles must always be greater than a safety threshold. The preset driving conditions can be understood as technical or safety restrictions set on the vehicle's motion state, including but not limited to maximum / minimum speed limits, maximum acceleration / deceleration, maximum steering angle rate, minimum turning radius, and vehicle stability boundaries (such as anti-skid conditions).

[0060] By adopting this technical solution, multi-dimensional constraints such as starting point matching, ending point alignment, safety boundaries, and driving conditions are introduced to ensure that the generated parking trajectory meets the actual vehicle state and environmental requirements in terms of geometry, dynamics, and safety, thus significantly improving the reliability and success rate of automatic parking.

[0061] The technical solution of this invention, in response to a parking request for a target vehicle, acquires partial images of visual images collected by multiple visual sensors, determines a panoramic image of the target vehicle based on the multiple partial images, divides the panoramic image into multiple regions of a first size, determines the semantic information of each region, and fuses the partial images from multiple cameras to generate a panoramic semantic map, accurately identifying the vehicle's surrounding environment and obstacle categories; acquires radar data collected by ultrasonic radar, determines the distance information between the target vehicle and target obstacles based on the radar data, and determines obstacle attribute characteristics based on the semantic information and the distance information, wherein the target obstacles include static obstacles and dynamic obstacles, and fuses radar distance and visual semantic information to accurately identify the attributes of static and dynamic obstacles to ensure safety; A grid map is determined based on the positional relationship between the target vehicle and the target parking space in the world coordinate system. A target occupancy grid map is determined based on the grid map and the obstacle attribute characteristics. A target trajectory sequence of the target obstacle is determined based on the obstacle attribute characteristics. A target parking trajectory is determined based on the target trajectory sequence, the target occupancy grid map, and preset constraints. A parking control command is determined based on the target parking trajectory and sent to the parking control system. This accurately constructs an environmental model, intelligently plans a safe trajectory, and controls the vehicle to complete parking efficiently. This solves the problem in the prior art of accurately perceiving the environment, identifying obstacles, and planning a safe and feasible parking trajectory during parking. It realizes intelligent planning of a safe and feasible parking trajectory and precise control of the vehicle to complete efficient automatic parking.

[0062] Example 2

[0063] Figure 2 This is a flowchart of a parking trajectory prediction method provided in Embodiment 2 of the present invention. This embodiment further refines how to determine the target parking trajectory based on the target trajectory sequence, the target occupancy grid map, and preset constraints, building upon the previous embodiments. Optionally, determining the target parking trajectory based on the target trajectory sequence, the target occupancy grid map, and preset constraints includes: determining a static driving path based on the target occupancy grid map, the position of the target vehicle, the target parking space, and the target trajectory sequence of the static obstacle; determining a target safe driving path based on the static driving path and the target trajectory sequence of the dynamic obstacle; and determining the target parking trajectory based on the target safe driving path and preset constraints. Detailed implementation can be found in the description of this embodiment. Technical features that are the same as or similar to those in the previous embodiments will not be repeated here.

[0064] like Figure 2 As shown, the method may specifically include:

[0065] S210. In response to a parking request for a target vehicle, acquire partial images of visual images collected by multiple visual sensors, determine a panoramic image of the target vehicle based on the multiple partial images, segment the panoramic image into multiple regions of a first size, and determine semantic information for each region.

[0066] S220. Acquire radar data collected by ultrasonic radar, determine the distance information between the target vehicle and the target obstacle based on the radar data, and determine the obstacle attribute characteristics based on the semantic information and the distance information, wherein the target obstacle includes static obstacles and dynamic obstacles.

[0067] S230. Determine a grid map based on the positional relationship between the target vehicle and the target parking space in the world coordinate system; determine a target-occupied grid map based on the grid map and the obstacle attribute characteristics; and determine the target trajectory sequence of the target obstacle based on the obstacle attribute characteristics.

[0068] S240. Based on the target occupancy grid map, determine the static driving path according to the position of the target vehicle, the target parking space, and the target trajectory sequence of the static obstacle.

[0069] The static driving path can be understood as a collision-free geometric path planned in a static environment, considering only static obstacles, the current position of the target vehicle, and the target parking space. This path avoids all static obstacle areas marked as "occupied" in the target occupancy grid map. As the initial path of the parking trajectory, it provides a basic passage route. It assumes that there is no dynamic interference in the environment and is a prerequisite for subsequent dynamic adjustments.

[0070] Optionally, based on the above scheme, determining the static driving path according to the target vehicle's position, the target parking space, and the target trajectory sequence of the static obstacle based on the target occupancy grid map includes: determining a first path based on the target vehicle's position and the target parking space based on the target occupancy grid map, and determining the static driving path based on the first path and the target trajectory sequence of the static obstacle, wherein the width of the static driving path along the vehicle body direction is greater than the vehicle width.

[0071] The first path can be understood as a path initially planned from the current location of the vehicle to the area near the target parking space, based only on the current location of the target vehicle, the target parking space, and the target occupancy grid map (i.e., obstacle distribution). It is the first-stage result of path planning and provides a basic passable path from the "starting point" to the "end point".

[0072] By combining the target occupancy grid map with static obstacle information, a static driving path with a width greater than the vehicle width is generated, effectively reserving lateral safety margin and avoiding scratches caused by positioning errors or vehicle size deviations, thus significantly improving the robustness and safety of the parking path.

[0073] S250. Determine the target safe driving path based on the static driving path and the target trajectory sequence of the dynamic obstacles.

[0074] The target safe driving path can be understood as a driving path that avoids both static obstacles and potential future dynamic obstacles, based on the static driving path and further combined with the target trajectory sequence of dynamic obstacles (i.e. how they will move in the future). It is the second stage result of path planning, and the goal is to ensure that the vehicle will not collide with any obstacle (whether static or dynamic) during driving.

[0075] Based on the above scheme, optionally, determining the target safe driving path according to the static driving path and the target trajectory sequence of the dynamic obstacle includes: removing the dynamic obstacle from the static driving path according to the target trajectory sequence of the dynamic obstacle and its timestamp, and determining the target safe driving path according to the removal result.

[0076] For example, after a static driving path that can avoid static obstacles has been planned, it is determined how the surrounding moving objects (such as other vehicles and pedestrians) will move in the future (target trajectory sequence) and when they will go to where (time stamp). Dynamic obstacles that may be hit in the future are excluded from the path, and finally a safe driving path that avoids both static and dynamic obstacles is planned.

[0077] An alternative implementation, static obstacles: On the occupied grid map, from the vehicle's current position to the target parking space, a jump point search (JPS) algorithm is used to quickly search for an initial path that avoids all static obstacle grids, and a static driving path wider than the vehicle body is generated along this path.

[0078] Dynamic obstacles: The predicted future trajectory of dynamic obstacles is depicted as a "spacetime cuboid" that occupies a specific space within a specific time period, and these cuboids are "cut out" from the static driving path.

[0079] By adopting this technical solution, safe and feasible driving paths are dynamically generated by eliminating areas in the static path that conflict with dynamic obstacles in time and space, effectively avoiding collision risks and improving the adaptability and safety of the parking process in complex dynamic environments.

[0080] S260. Determine the target parking trajectory based on the target safe driving path and preset constraints, determine the parking control command based on the target parking trajectory, and send the parking control command to the parking control system.

[0081] The technical solution of this invention integrates static environment modeling, dynamic obstacle prediction and multi-dimensional constraints in stages. It can first construct a safe static path that takes into account the width of the vehicle body, and then combine the spatiotemporal trajectory of dynamic obstacles to eliminate conflict areas, generating a highly reliable parking trajectory that meets constraints such as vehicle dynamics, start / end point matching and safety boundaries. This significantly improves the environmental adaptability, obstacle avoidance foresight and trajectory execution success rate of the automatic parking system in complex and dynamic scenarios, taking into account safety, feasibility and comfort.

[0082] Example 3

[0083] Figure 3 This is a schematic diagram of a parking trajectory prediction device provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes: a semantic information determination module 310, an attribute feature determination module 320, and a parking instruction sending module 330. Among them,

[0084] The semantic information determination module 310 is used to, in response to a parking request for a target vehicle, acquire partial images of visual images collected by multiple visual sensors, determine a panoramic image of the target vehicle based on the multiple partial images, segment the panoramic image into multiple regions of a first size, and determine the semantic information of each region; the attribute feature determination module 320 is used to acquire radar data collected by ultrasonic radar, determine the distance information between the target vehicle and the target obstacle based on the radar data, and determine the obstacle attribute features based on the semantic information and the distance information, wherein the target obstacle includes static obstacles and dynamic obstacles; the parking instruction sending module 330 is used to determine a grid map based on the positional relationship between the target vehicle and the target parking space in the world coordinate system, determine a target occupancy grid map based on the grid map and the obstacle attribute features, determine a target trajectory sequence of the target obstacle based on the obstacle attribute features, determine a target parking trajectory based on the target trajectory sequence, the target occupancy grid map and preset constraints, determine a parking control instruction based on the target parking trajectory, and send the parking control instruction to the parking control system.

[0085] The technical solution of this invention, in response to a parking request for a target vehicle, involves a semantic information determination module acquiring partial images from multiple visual sensors, determining a panoramic image of the target vehicle based on these partial images, segmenting the panoramic image into multiple regions of a first size, determining the semantic information of each region, and fusing the partial images from multiple cameras to generate a panoramic semantic map, thereby accurately identifying the vehicle's surrounding environment and obstacle categories. Furthermore, an attribute feature determination module acquires radar data from ultrasonic radar, determines the distance information between the target vehicle and target obstacles based on the radar data, and determines obstacle attribute features based on the semantic information and the distance information. The target obstacles include static and dynamic obstacles. By fusing radar distance and visual semantic information, the attributes of static and dynamic obstacles are accurately identified. To ensure parking safety, the parking command sending module determines a grid map based on the positional relationship between the target vehicle and the target parking space in the world coordinate system. It then determines a target occupancy grid map based on the grid map and obstacle attribute characteristics, determines the target obstacle's trajectory sequence based on the obstacle attribute characteristics, determines the target parking trajectory based on the target trajectory sequence, the target occupancy grid map, and preset constraints, and determines a parking control command based on the target parking trajectory. This command is then sent to the parking control system, accurately constructing an environmental model, intelligently planning a safe trajectory, and controlling the vehicle to efficiently complete parking. This solves the problem in existing technologies where it is difficult to accurately perceive the environment, identify obstacles, and plan a safe and feasible parking trajectory. It achieves intelligent planning of a safe and feasible parking trajectory and precise control of the vehicle to complete efficient automatic parking.

[0086] Optionally, based on the above scheme, the attribute feature determination module includes an attribute feature determination submodule. This submodule is used to cluster the radar data, determine the reference point coordinates of each target obstacle based on the clustering results, transform the reference point coordinates to pixel coordinates of the panoramic image according to a preset coordinate transformation matrix, and determine obstacle attribute features based on the semantic information of each target obstacle, the pixel coordinates, the distance information between the target vehicle and the target obstacle, and a preset safety distance.

[0087] Optionally, based on the above solution, the parking instruction sending module includes: an occupancy grid map determination submodule. The occupancy grid map determination submodule is used to divide the grid map into multiple grids according to a preset size; for each grid, determine the occupancy probability and timestamp of the obstacle attribute features falling into the grid; and construct a target occupancy grid map based on the occupancy probability, the timestamp, and the grid map.

[0088] Optionally, based on the above scheme, the parking instruction sending module includes: a first trajectory sequence determination submodule. The first trajectory sequence determination submodule is used to, when the target obstacle is a static obstacle, determine the predicted position of the target obstacle for multiple frames based on the obstacle attribute features and a preset prediction model, and determine the target trajectory sequence of the target obstacle based on the predicted position, obstacle attribute features, and the corresponding static obstacle according to a preset algorithm.

[0089] Optionally, based on the above scheme, the parking instruction sending module includes a second trajectory sequence determination submodule. The second trajectory sequence determination submodule is used to determine, when the target obstacle is a dynamic obstacle, the initial trajectory sequence of the dynamic obstacle corresponding to the dynamic obstacle information within multiple target frames during a target time period, and input the initial trajectory sequence into a trained trajectory prediction model to obtain the target trajectory sequence of the target obstacle.

[0090] Based on the above solution, optionally, the parking instruction sending module includes: a static driving path determination submodule, a safe driving path determination submodule, and a parking trajectory determination submodule. Wherein,

[0091] The static driving path determination submodule is used to determine a static driving path based on the target occupancy grid map, according to the position of the target vehicle, the target parking space, and the target trajectory sequence of the static obstacles; the safe driving path determination submodule is used to determine a target safe driving path based on the static driving path and the target trajectory sequence of the dynamic obstacles; the parking trajectory determination submodule is used to determine a target parking trajectory based on the target safe driving path and preset constraints.

[0092] Based on the above scheme, optionally, the static driving path determination submodule is specifically used to determine a first path based on the target occupied grid map, according to the position of the target vehicle and the target parking space, and to determine a static driving path based on the first path and the target trajectory sequence of the static obstacles, wherein the width of the static driving path along the vehicle body direction is greater than the vehicle width.

[0093] Based on the above scheme, optionally, the safe driving path determination submodule is specifically used to remove the dynamic obstacle from the static driving path according to the target trajectory sequence of the dynamic obstacle and its timestamp, and determine the target safe driving path according to the removal result.

[0094] Based on the above scheme, optionally, the preset constraints include at least one of the following: the starting point of the target parking trajectory matches the first vehicle driving information of the target vehicle; the ending point of the target parking trajectory matches the second vehicle driving information of the target vehicle; the target parking trajectory must not exceed the preset safety boundary; and the third vehicle driving information of the target vehicle meets the preset driving conditions of the target vehicle.

[0095] The parking trajectory prediction device provided in this embodiment of the invention can execute the parking trajectory prediction method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0096] Example 4

[0097] Figure 4 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0098] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0099] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0100] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a parking trajectory prediction method.

[0101] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0102] In some embodiments, a parking trajectory prediction method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the parking trajectory prediction method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform a parking trajectory prediction method by any other suitable means (e.g., by means of firmware).

[0103] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0104] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0105] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0106] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0107] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0108] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0109] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0110] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A parking trajectory prediction method, characterized in that, include: In response to a parking request for a target vehicle, partial images of visual images collected by multiple visual sensors are acquired, a panoramic image of the target vehicle is determined based on the multiple partial images, the panoramic image is segmented into multiple regions of a first size, and semantic information of each region is determined. The system acquires radar data collected by ultrasonic radar, determines the distance information between the target vehicle and the target obstacle based on the radar data, and determines the obstacle attribute characteristics based on the semantic information and the distance information. The target obstacle includes static obstacles and dynamic obstacles. A grid map is determined based on the positional relationship between the target vehicle and the target parking space in the world coordinate system. A target occupancy grid map is determined based on the grid map and the obstacle attribute characteristics. A target trajectory sequence of the target obstacle is determined based on the obstacle attribute characteristics. A target parking trajectory is determined based on the target trajectory sequence, the target occupancy grid map, and preset constraints. A parking control command is determined based on the target parking trajectory and sent to the parking control system.

2. The method according to claim 1, characterized in that, Determining obstacle attribute features based on the semantic information and the distance information includes: The radar data is clustered, and the reference point coordinates of each target obstacle are determined according to the clustering results. The reference point coordinates are then transformed to the pixel coordinates of the panoramic image according to a preset coordinate transformation matrix. The obstacle attribute features are determined based on the semantic information of each target obstacle, the pixel coordinates, the distance information between the target vehicle and the target obstacle, and a preset safe distance.

3. The method according to claim 1, wherein determining the target-occupied grid map based on the grid map and the obstacle attribute features comprises: The raster map is divided into multiple grids according to a preset size; For each grid cell, determine the occupancy probability and timestamp of the obstacle attribute feature falling into the grid cell, and construct a target occupancy grid map based on the occupancy probability, the timestamp, and the grid map.

4. The method according to claim 1, characterized in that, Determining the target trajectory sequence of the target obstacle based on the obstacle attribute characteristics includes: When the target obstacle is a static obstacle, the predicted position of the target obstacle in multiple frames is determined according to the obstacle attribute features and a preset prediction model. The predicted position, obstacle attribute features and the corresponding static obstacle are used to determine the target trajectory sequence of the target obstacle according to a preset algorithm.

5. The method according to claim 1, characterized in that, Determining the target trajectory sequence of the target obstacle based on the obstacle attribute characteristics includes: When the target obstacle is a dynamic obstacle, the initial trajectory sequence of the dynamic obstacle corresponding to the dynamic obstacle information in multiple target frames within the target time period is determined, and the initial trajectory sequence is input into the trained trajectory prediction model to obtain the target trajectory sequence of the target obstacle.

6. The method according to claim 1, characterized in that, Determining the target parking trajectory based on the target trajectory sequence, the target occupancy grid map, and preset constraints includes: Based on the target occupancy grid map, a static driving path is determined according to the position of the target vehicle, the target parking space, and the target trajectory sequence of the static obstacles. Determine the target safe driving path based on the target trajectory sequence of the static driving path and the dynamic obstacles; The target parking trajectory is determined based on the target safe driving path and preset constraints.

7. The method according to claim 6, characterized in that, The step of determining the static driving path based on the target occupancy grid map, according to the position of the target vehicle, the target parking space, and the target trajectory sequence of the static obstacles, includes: Based on the target occupancy grid map, a first path is determined according to the position of the target vehicle and the target parking space. A static driving path is determined according to the first path and the target trajectory sequence of the static obstacles, wherein the width of the static driving path along the vehicle body direction is greater than the vehicle width.

8. The method according to claim 6, characterized in that, Determining the target safe driving path based on the target trajectory sequence of the static driving path and the dynamic obstacles includes: The dynamic obstacles are removed from the static driving path based on their target trajectory sequence and timestamps, and the target safe driving path is determined based on the removal results.

9. The method according to claim 1, wherein the preset constraint condition includes at least one of the following: The starting point of the target parking trajectory is matched with the first vehicle driving information of the target vehicle; The endpoint of the target parking trajectory matches the second vehicle driving information of the target vehicle; The target parking trajectory must not exceed the preset safety boundary; The third vehicle driving information of the target vehicle meets the preset driving conditions of the target vehicle.

10. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the parking trajectory prediction method according to any one of claims 1-9.

Images