Parking position determination method and device, electronic equipment and storage medium

Abstract

The application discloses a parking position determination method and device, electronic equipment and a storage medium, and relates to the technical field of automatic parking. The parking position determination method comprises the following steps: constructing a first-precision grid map by using surrounding environment information collected by a vehicle sensor and a selected target parking space, and selecting at least one alternative parking position in the target parking space to construct a multi-dimensional cost function comprising a collision cost function, a safety distance cost function and an accessibility cost function. Then, the cost values of the alternative parking positions are calculated by using the multi-dimensional cost function, and the alternative parking position corresponding to the minimum cost value is taken as a target parking position. In this way, the problem of repeated parking caused by frequent adjustment of the target parking position due to unattainable parking paths in the automatic parking process can be avoided. Therefore, the alternative parking position with the minimum cost value is determined as the target parking position by constructing the multi-dimensional cost function, and the parking efficiency can be significantly improved.

Description

Technical Field

[0001] This application relates to the field of automatic parking technology, specifically to parking location determination methods, devices, electronic equipment, and storage media. Background Technology

[0002] With the continuous development of autonomous driving technology, automatic parking systems have become an important component of intelligent vehicles. Automatic parking systems primarily use sensors to perceive the surrounding environment, automatically plan parking paths, and control the vehicle to complete the parking operation.

[0003] Currently, some technologies rely on the Hybrid A* algorithm for parking path planning. This algorithm calculates an optimal parking path from the current location to the target parking position by inputting the target parking location and the vehicle's current position, combined with obstacle information received from parking sensors. However, the accuracy of parking sensors is significantly affected by the relative positions of the vehicle and obstacles. As the vehicle approaches the target parking location, the feasibility of the current parking path may need to be reassessed, leading to frequent adjustments to the target parking location and resulting in misalignment, thus impacting parking efficiency. Improvements are needed. Summary of the Invention

[0004] This application provides a parking location determination method, apparatus, electronic device, and storage medium to solve the problem of low parking efficiency caused by frequent adjustments to the target parking location in related technologies.

[0005] Firstly, this application provides a method for determining a parking location, including:

[0006] Acquire surrounding environmental information collected by vehicle sensors;

[0007] In response to the selection of a target parking space, a first-precision raster map is constructed based on the target parking space and the surrounding environmental information.

[0008] Based on the first-precision raster map, select at least one alternative parking location from the target parking space;

[0009] Using each alternative parking location as a benchmark, a multi-dimensional cost function is constructed, which includes a collision cost function, an accessibility cost function, and a safe distance cost function.

[0010] Based on a multi-dimensional cost function, the cost value of each alternative parking location is determined, and the alternative parking location with the minimum cost value is taken as the target parking location.

[0011] Secondly, this application provides a parking position determination device, comprising:

[0012] The environmental information acquisition module is used to acquire surrounding environmental information collected by the vehicle's sensors.

[0013] The grid map building module is used to build a first-precision grid map based on the target parking space and the surrounding environment information in response to the selection of the target parking space.

[0014] The alternative parking location determination module is used to select at least one alternative parking location from the target parking space based on a first-precision grid map;

[0015] The cost function construction module is used to construct multi-dimensional cost functions based on each alternative parking location. These multi-dimensional cost functions include collision cost function, accessibility cost function, and safe distance cost function.

[0016] The target parking location determination module is used to determine the cost value of each candidate parking location based on a multi-dimensional cost function, and to select the candidate parking location with the minimum cost value as the target parking location.

[0017] Thirdly, this application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described parking position determination methods.

[0018] Fourthly, this application also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of any of the above-described parking position determination methods.

[0019] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described parking position determination methods.

[0020] In some embodiments of this application, a first-precision grid map is constructed using surrounding environmental information collected by vehicle sensors and a selected target parking space. At least one alternative parking location is selected from the target parking space to construct a multi-dimensional cost function, including a collision cost function, a safe distance cost function, and an accessibility cost function. Then, the cost value of each alternative parking location is calculated using the multi-dimensional cost function, and the alternative parking location with the lowest cost value is selected as the target parking location. This avoids the problem of frequent adjustments to the target parking location and repeated parking maneuvers caused by unreachable parking paths during automatic parking. Therefore, by constructing a multi-dimensional cost function to determine the alternative parking location with the lowest cost value as the target parking location, parking efficiency can be significantly improved. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating the parking location determination method provided in this application embodiment;

[0023] Figure 2 A schematic diagram of the safety distance cost function provided in the embodiments of this application;

[0024] Figure 3 A schematic diagram showing the distribution of different alternative parking locations and their corresponding total costs for embodiments of this application;

[0025] Figure 4 A schematic diagram of the parking position determination device provided in the embodiments of this application;

[0026] Figure 5 A schematic diagram of the modules of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0028] It should be noted that, in the description of this application, 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. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0029] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] Automated Parking Assist (APA) is an intelligent driving technology that can assist drivers or automatically complete parking operations. The system uses onboard sensors to perceive the surrounding environment, plan parking paths, and control the vehicle to automatically complete parking operations.

[0031] Currently, automatic parking systems typically use a Hybrid A* algorithm combined with vehicle dynamics models (such as the Ackermann steering model) for path searching when planning parking routes. The Hybrid A* algorithm requires a pre-inputted initial parking position to calculate the optimal parking path between the vehicle's current position and the target parking space. However, the detection accuracy of vehicle parking sensors is limited by the relative position between the vehicle and obstacles, as well as environmental factors. Therefore, when the distance between the vehicle and obstacles is too great or there are obstructions, the sensors cannot accurately detect the obstacle boundaries, leading to potentially inaccurate initial parking positions. Consequently, when the automatic parking system controls parking based on the planned parking path from the initial parking position, the initial parking position may become unreachable, necessitating a reset of the initial parking position and a replanning of the parking path, severely impacting parking efficiency.

[0032] Furthermore, during automatic parking, as the vehicle gradually approaches the initial parking position, the accumulated sensor detection errors cause a deviation between the actual parking position and the initial position. This necessitates repositioning the vehicle to correct the deviation. This repeated repositioning not only prolongs parking time but also diminishes the driver's parking experience.

[0033] In view of this, this application provides a parking position determination method that can solve the above problems. This parking position determination method can be applied to automatic parking systems.

[0034] See also Figure 1 , Figure 1 This is a flowchart illustrating the parking location determination method provided in an embodiment of this application, as shown below. Figure 1 As shown, the method for determining the parking location includes the following steps.

[0035] Step S101: Obtain the surrounding environment information collected by the vehicle's sensors.

[0036] Specifically, vehicle sensors include, but are not limited to, cameras and ultrasonic sensors, used to collect distance information between the vehicle and surrounding obstacles, as well as the outline features of the obstacles.

[0037] In this embodiment, when a vehicle is preparing to park, the electronic device executing the method of this application can determine the vehicle's current position, the available parking space area around the vehicle, and the distribution of obstacles based on the surrounding environment information collected by the vehicle's sensors.

[0038] Understandably, obstacles can be walls, roadblocks, parked vehicles, etc. Anything that may affect the vehicle's automatic parking operation can be called an obstacle.

[0039] In step S102, in response to the selection of a target parking space, a first-precision raster map is constructed based on the target parking space and the surrounding environment information.

[0040] Specifically, the target parking space is used to represent the vacant parking space or area where a vehicle can be parked selected by the driver; the first-precision grid map is used to show the target parking space and the distribution of obstacles around the target parking space.

[0041] In this embodiment, after acquiring the surrounding environment information collected by the vehicle sensors, the vehicle's on-board display can show at least one parking space for the driver to select. After the driver selects the target parking space, the electronic device executing the method of this application will use the target parking space as a reference and combine it with the surrounding environment information to construct a first-precision grid map.

[0042] For example, when the driver selects a target parking space on the in-vehicle display, the electronic device executing the method of this application will construct a (20m×20m) ROI (Region of Interest) area with the center point of the target parking space as the origin of the coordinates, and divide the area into a grid map with an accuracy of 0.1m. Then, the obstacle positions, contour data and drivable areas in the surrounding environment information are mapped to each grid to form a first-precision grid map.

[0043] It should be noted that the size of the ROI and the grid division accuracy can be dynamically adjusted based on the sensor's detection range and the actual parking scenario to balance the requirements of computing resources and positioning accuracy. For example, when parking in a large parking lot, where the space is relatively open and the parking spaces are well-organized, the ROI area can be appropriately expanded to 30m×30m, while the grid division accuracy can be adjusted to 0.2m, reducing the computational load while ensuring coverage. In narrow alleys or complex parking environments, the ROI area can be reduced to 15m×15m, and the grid division accuracy can be increased to 0.05m to enhance the ability to identify minor obstacles and ensure parking safety.

[0044] Step S103: Based on the first-precision grid map, select at least one alternative parking location from the target parking space.

[0045] Specifically, the alternative parking location refers to the initial parking location that needs to be input into the Hybrid A* algorithm, i.e., the ending location of the parking path, which is used by the Hybrid A* algorithm to plan the parking path.

[0046] In this embodiment, the target parking space can be divided into multiple grid areas by constructing a first-precision grid map. By analyzing the outline boundary of the target parking space and the relative positions of surrounding obstacles, multiple parking areas that meet vehicle size constraints can be identified as candidate parking locations. Thus, one or more candidate parking locations can be determined from multiple grid areas of the target parking space.

[0047] The following example further illustrates that after the grid map is constructed through step S102 above, the vehicle size can be compared with the outline of the target parking space to determine multiple candidate areas that meet the parking conditions within the target parking space. Then, the grid area located at the center of the candidate areas is determined, and the center point of the grid area is a candidate parking location.

[0048] Step S104: Using each alternative parking location as a benchmark, construct a multi-dimensional cost function, which includes a collision cost function, an accessibility cost function, and a safe distance cost function.

[0049] Specifically, a multi-dimensional cost function is used to calculate the cost of the aforementioned alternative parking locations. Specifically, the collision cost function calculates the probability of a collision between the vehicle and an obstacle; the accessibility cost function assesses whether the vehicle can reach the aforementioned alternative parking locations; and the safe distance cost function measures whether the distance between the vehicle and the obstacle meets the preset safe distance requirements.

[0050] Step S105: Based on the multi-dimensional cost function, determine the cost value of each candidate parking location, and take the candidate parking location with the minimum cost value as the target parking location.

[0051] Specifically, the cost value of each candidate parking location is calculated using the constructed multi-dimensional cost function. A smaller cost value indicates higher parking efficiency when the parking location corresponding to that cost value is used as the target parking location for path planning. Therefore, the candidate parking location with the lowest cost value is selected as the target parking location for subsequent automatic parking control system planning and parking operations.

[0052] The parking location determination method provided in this embodiment constructs a first-precision grid map using surrounding environmental information collected by vehicle sensors and the selected target parking space. It then selects at least one alternative parking location from the target parking space to construct a multi-dimensional cost function, including a collision cost function, a safe distance cost function, and an accessibility cost function. Next, the cost value of each alternative parking location is calculated using the multi-dimensional cost function, and the alternative parking location with the lowest cost value is selected as the target parking location. This avoids the problem of frequent adjustments to the target parking location and repeated parking maneuvers caused by unreachable parking paths during automatic parking. Therefore, by constructing a multi-dimensional cost function to determine the alternative parking location with the lowest cost value as the target parking location, parking efficiency can be significantly improved.

[0053] In some embodiments, the surrounding environment information includes obstacle distribution information; using each candidate parking location as a benchmark, a multi-dimensional cost function is constructed respectively, and the parking location determination method includes:

[0054] Based on obstacle distribution information and vehicle collision volume model, the probability of a collision between the vehicle and an obstacle is determined, and a collision cost function is constructed based on the probability.

[0055] And / or, determine whether the alternative parking location is within the reachable range of the vehicle's driving path, and construct an reachability function based on whether the alternative parking location is reachable;

[0056] And / or, obtain the difference between the distance between the obstacle and the vehicle and the preset safe distance, and construct a safe distance cost function based on the difference.

[0057] Based on the above description, it is understandable that constructing the cost function is crucial for determining the target parking location, and the rationality of the selected target parking location directly affects the planning of the parking path and parking efficiency. Therefore, constructing a reasonable multi-dimensional cost function is of paramount importance.

[0058] The three cost functions mentioned above will be explained in detail below.

[0059] 1) Collision Cost Function: In the first-precision grid map mentioned above, a collision volume model of the vehicle is established. Based on the distribution of obstacles, it is determined whether the vehicle's volume model and the obstacles occupy the same grid area in different alternative parking positions. If they occupy the same grid area, the alternative parking position is determined to have a collision risk.

[0060] Furthermore, collisions during vehicle parking are unacceptable safety events. Therefore, when constructing the collision cost function, the cost of alternative parking positions with collision risk can be directly set to 1, while that of alternative parking positions without collision risk is set to 0. In other words, a step function is used to characterize the collision cost function.

[0061] 2) Accessibility Cost Function: Since the accessibility cost function is used to evaluate whether a vehicle can reach the above alternative parking locations, it is understandable that alternative parking locations with collision risks are inaccessible.

[0062] Furthermore, a vehicle's motion is influenced by its own dynamic constraints. For example, when a vehicle is turning, an Ackermann steering model is typically used to constrain all wheels to rotate around the same instantaneous steering center. In actual parking, whether reversing into a parking space or parallel parking, the vehicle needs to perform a steering operation. Therefore, some alternative parking positions may not be accessible via a continuous path due to the constraints of the Ackermann steering model. In such cases, it is necessary to exit the parking space, adjust the vehicle's posture, and then perform the parking operation again.

[0063] Therefore, in order to avoid repeated parking adjustments that reduce parking efficiency and lead to a poor experience for drivers when using automatic parking, the accessibility cost function should be constructed by considering collision risks and vehicle dynamics constraints in order to determine whether alternative parking locations are accessible.

[0064] Based on the above description, it can also be understood that the accessibility cost function can also be constructed using a step function. The cost of unreachable alternative parking locations, i.e., those with collision risks and those that do not meet vehicle kinematic constraints, is set to 1, while the cost of reachable alternative parking locations, i.e., those without collision risks and those that meet vehicle dynamic constraints, is set to 0.

[0065] 3) Safety Distance Cost Function: The collision cost function described above only considers whether a direct collision occurs between the vehicle and an obstacle, without considering the safety distance between the vehicle and the obstacle. Therefore, when the alternative parking positions pose no collision risk and are accessible, the distance between the vehicle and surrounding obstacles can be calculated by comparing it to the preset safety distance when the vehicle is parked at each alternative parking position. If the difference is positive, it means that the actual distance between the vehicle and the obstacle is greater than the preset safety distance, and the cost of the corresponding alternative parking position is 0. If the difference is negative, the cost can be set based on the actual distance between the vehicle and the obstacle; the smaller the actual distance, the higher the cost.

[0066] Optionally, refer to Figure 2 , Figure 2The safety distance cost function provided for the embodiments of this application is, for example... Figure 2 As shown, the safety distance cost function can be constructed using a piecewise function. When the difference is greater than or equal to zero, the cost is always zero. When the difference is less than zero, the cost can be represented by a quadratic function. The smaller the difference, the higher the cost.

[0067] In the above embodiments, by constructing a collision cost function, an accessibility cost function, and a safe distance cost function, a multi-dimensional evaluation of each candidate parking location can be performed. This can effectively filter out target parking locations that meet the requirements of safety, accessibility, and safe distance, thereby ensuring the safety and reliability of the automatic parking process, while improving parking efficiency and the driver's parking experience.

[0068] In some embodiments, the multi-dimensional cost function further includes a centering cost function, which is used to characterize the degree of deviation between the candidate parking location and the center point of the target parking space; the parking location determination method further includes the following steps a1 to a2:

[0069] Step a1: Based on the first-precision raster map, determine the coordinates of the center point of the candidate parking position and the target parking space, and obtain the lateral offset and longitudinal offset of the center point of the candidate parking position and the target parking space.

[0070] Step a2: Construct a centering cost function based on the horizontal and vertical offsets.

[0071] Specifically, during the parking process, when the vehicle is parked in the center of the target parking space, it can usually ensure that there is enough safety space on both sides and in front and behind the vehicle to the greatest extent. Therefore, when designing a multi-dimensional cost function, a centralization cost function can be constructed to characterize the degree of deviation between the alternative parking position and the center point of the target parking space.

[0072] It should be noted that, since a first-precision raster map has been constructed, the coordinate information of the center point of the candidate parking position and the target parking space can be accurately obtained based on this map, and then the degree of deviation between the two can be calculated.

[0073] Optionally, the offset between the center point of the alternative parking location and the center point of the target parking space can be measured by Euclidean distance, which is calculated by taking the square root of the sum of the squares of the differences between their coordinates, thereby quantifying the centering cost.

[0074] However, since the vehicle's trajectory is constrained by the Ackerman steering model, the Euclidean distance, while reflecting the straight-line distance between the alternative parking position and the center point of the target parking space, cannot reflect the accessibility of the actual parking path. The actual trajectory of the vehicle can be approximated as a step-by-step process of first adjusting the lateral offset and then adjusting the longitudinal offset. Therefore, using the Manhattan distance, which is the sum of the absolute values ​​of the lateral offset and the longitudinal offset, can more reasonably reflect the actual adjustment cost between the alternative parking position and the center point of the target parking space, thus better reflecting the vehicle's motion characteristics.

[0075] Therefore, when constructing the centering cost function, the Manhattan distance can be used as a way to evaluate the centering cost. The expression of the centering cost function can be as follows:

[0076]

[0077] in, Represents the centering cost function; The x-coordinate represents the alternative parking positions; The vertical coordinate represents the alternative parking positions; The x-coordinate of the center point of the target parking space; This represents the ordinate of the center point of the target parking space; This indicates taking the absolute value.

[0078] In the above embodiments, by introducing Manhattan distance to construct a centering cost function, the actual adjustment difficulty between the alternative parking position and the center point of the target parking space can be assessed more accurately, so that the automatic parking system can plan the parking path more reasonably, thereby improving the parking accuracy of the parking process, while reducing repeated parking due to position deviation, and further optimizing the driver's parking experience.

[0079] In some embodiments, the multi-dimensional cost function further includes a bias cost function, which characterizes the degree of lateral deviation of the vehicle's parking posture from its center position in the target parking space; the parking position determination method further includes the following steps b1 to b2:

[0080] Step b1: If there are obstacles on the left and right sides of the target parking space, obtain the first distance between the vehicle and the obstacle on the left, and the second distance between the vehicle and the obstacle on the right, and determine the first difference between the first distance and the second distance.

[0081] Step b2: Construct a bias cost function based on the first difference and the second difference of the preset bias distance.

[0082] Specifically, the first distance and the second distance refer to the distances between the vehicle and obstacles on the left and right, respectively, which can be obtained through vehicle sensors.

[0083] In this embodiment, an offset cost function is constructed to determine the deviation distance between the vehicle's left and right sides and obstacles, thereby evaluating the rationality of alternative parking positions under different needs. For example, when the passenger seat is unoccupied, the driver can choose an offset parking position to provide more space on the driver's side for easy entry and exit. Similarly, when the vehicle needs to be parked in a parallel parking space on the side of the road, an offset parking position can also be chosen to move away from the roadside and avoid causing safety accidents.

[0084] Optionally, by presetting the deviation distance, if the vehicle needs to be parked with a left-side offset, the expression for the offset cost function can be as follows:

[0085]

[0086] in, Represents the bias cost function; This indicates the first distance, which is the distance between the vehicle and the obstacle on the left. This indicates the second distance, namely the distance between the vehicle and the obstacle on the right. This indicates the preset deviation distance.

[0087] In the above embodiments, an offset cost function is constructed by calculating the deviation distance, enabling the driver to choose a suitable parking posture based on their own vehicle conditions, thereby improving the driver's parking experience.

[0088] In addition to the cost function described above, the electronic device implementing this method can further construct a cost function based on the driver's parking needs when evaluating each alternative parking location. For example, due to the inconsistent size of vehicles, and the fact that some parking spaces may be too large in design, smaller vehicles may be difficult to see if all vehicles are parked in the center of the space. Therefore, the position of the vehicle's front end can be considered when parking, aligning the front end of the vehicle with the front ends of surrounding parked vehicles to avoid the aforementioned problem.

[0089] In some embodiments, the parking location determination method further includes the following steps c1 to c4:

[0090] Step c1: Within the preset area of ​​the target parking space, each grid point is identified as a candidate parking location.

[0091] Step c2, and the fact that different cost functions have different weights, based on the multi-dimensional cost function, yields the target parking position corresponding to the minimum cost, including:

[0092] Step c3: Based on the weights, perform weighted fusion of the various cost functions in the multi-dimensional cost function to obtain the total cost function;

[0093] Step c4: Within the preset area, calculate the total value corresponding to each candidate parking location, and select the candidate parking location with the lowest total value as the target parking location.

[0094] Specifically, since a first-precision grid map has been constructed based on the target parking space, and the grid map can intuitively show the distance relationship between the vehicle and the obstacle, in order to facilitate the calculation of the cost of each alternative parking location, the grid points can be determined as alternative parking locations within the preset area of ​​the target parking space.

[0095] Furthermore, since this method constructs multiple cost functions, the total cost value can be calculated for each constructed cost function during cost value calculation to evaluate the rationality of each alternative parking location. However, if both the collision cost function and the accessibility cost function are constructed using step functions, since the step function can only substitute values ​​between 0 and 1, although the cost value of each alternative parking location can be determined, the step function is not differentiable, which may lead to abrupt changes in cost value due to errors in data collected by some vehicle sensors. Therefore, the step function can be replaced with a differentiable sigmoid function to construct the collision cost function and the accessibility cost function.

[0096] In this embodiment, different cost functions can have different weights. By weighting, the cost functions are weighted and fused to calculate the total cost value of each candidate parking location. The candidate parking location with the lowest total cost value is then selected as the target parking location.

[0097] For example, when parking, the most basic requirement is that the parking spot is accessible and collision-free. Therefore, the weights of the collision cost function and the accessibility cost function can be set as high as possible as these basic constraints. The remaining cost functions can be dynamically adjusted based on the actual parking scenario and the driver's parking needs.

[0098] See also Figure 3 , Figure 3 This is a schematic diagram illustrating the distribution of different alternative parking locations and their corresponding total costs, as provided in the embodiments of this application. Figure 3 As shown in (a), the coordinate axes X and Y represent the first-precision grid map constructed in the vertical and horizontal directions of the target parking space, respectively, while the coordinate axis Z represents the cost of the current alternative parking location. The points in the figure are the aforementioned alternative parking locations.

[0099] Understandably, the darker the color of the alternative parking space, the lower its corresponding value, and the lighter the color of the alternative parking space, the higher its corresponding value.

[0100] For ease of understanding Figure 3(a), combined Figure 3 (b) Provide an explanation, such as Figure 3 As shown in (b), a grid map is constructed with the center point of the target parking space as the origin. The X-axis represents the vertical direction of the target parking space, and the Y-axis represents the horizontal direction. Figure 3 Projecting some points in (a) to Figure 3 In (b), it can be understood that the darker the color of the alternative parking location located in the center area of ​​the target parking space, the lower its corresponding cost value; the farther away from the center area of ​​the target parking space, the lighter the color, indicating the higher its corresponding cost value. Furthermore, when alternative parking locations are determined along the X-axis (vertical direction of the target parking space), the cost value fluctuation is not significant; however, when alternative parking locations are determined along the Y-axis (horizontal direction of the target parking space), the cost value fluctuation is significant.

[0101] In the above embodiments, by weighted fusion of multi-dimensional cost functions, the total cost value of each candidate parking location is obtained. The candidate parking location with the minimum total cost value is selected as the target parking location. In this way, the target parking location can be dynamically adjusted according to the actual parking scenario. Due to the constraints of multiple cost functions, the determined target parking location not only meets the basic requirements of collision-free and reachable parking, but also takes into account the driver's parking needs, effectively reducing repeated parking maneuvers caused by improper selection of the target parking location, thus reducing parking efficiency.

[0102] In some embodiments, within a preset area, the total value corresponding to each candidate parking location is calculated, and the candidate parking location with the lowest total value is selected as the target parking location. Step c4 above also includes the following:

[0103] The center point of the target parking space is taken as the first alternative parking location, and the total agency value corresponding to the first alternative parking location is calculated.

[0104] Find the grid point adjacent to the first alternative parking location as the second alternative parking location, and calculate the total cost of the second alternative parking location;

[0105] If the total value of the second alternative parking location is not less than the total value of the first alternative parking location, then the first alternative parking location will be used as the target parking location.

[0106] If the total value of the second alternative parking location is less than the total value of the first alternative parking location, then the second alternative parking location is used as the new first alternative parking location, and the search continues to find grid points adjacent to the new first alternative parking location as the new second alternative parking location.

[0107] If the total value of the new second alternative parking location is not less than the total value of the new first alternative parking location, then the new first alternative parking location will be used as the target parking location.

[0108] If the total value of the new second alternative parking location is less than the total value of the new first alternative parking location, then the new second alternative parking location will be used as the benchmark to continue searching for the target parking location.

[0109] Specifically, when determining the target parking location, calculating the total cost of each candidate parking location sequentially and then comparing them to find the candidate parking location with the lowest total cost is not only inefficient but also results in a large amount of redundant computation. Therefore, the center point of the target parking space can be used as the first candidate parking location, and the grid points adjacent to the first candidate parking location can be used as the second candidate parking location. The total cost of the first and second candidate parking locations can then be calculated. If the total cost of the first candidate parking location is not the minimum, the direction of the fastest gradient descent can be determined, and the second candidate parking location with the lowest current total cost can be used as the new first candidate parking location. The search continues to find the grid points adjacent to the new first candidate parking location as the new second candidate parking location.

[0110] Based on the above description, it can be understood that by taking the center point of the target parking space as the first alternative parking location and taking the direction with the fastest gradient descent speed as the main direction for determining the target parking location, and by continuously changing the search reference, i.e. the new first alternative parking location, the target parking location can be quickly determined, which can significantly improve parking efficiency.

[0111] In some embodiments, the parking location determination method further includes the following steps d1 to d2:

[0112] Step d1: Based on the first-precision raster map, construct a second-precision raster map within a preset area of ​​the target parking space, wherein the second precision is less than the first precision;

[0113] Step d2: Based on the second-precision raster map, select at least one alternative parking location from the target parking space.

[0114] Specifically, in order to more accurately determine the target parking location, after the first-precision raster map is constructed, a second-precision rasterization process can be performed again within the preset area of ​​the target parking space or the target parking space and its adjacent areas to obtain a second-precision raster map. Then, based on the second-precision raster map, at least one alternative parking location is selected from the target parking space.

[0115] In the above embodiments, by performing secondary rasterization processing, the accuracy requirements of the parking position can be improved, so that a more accurate target parking position can be obtained when planning the parking path.

[0116] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0117] Embodiments of this application also provide a parking position determination device. (See also...) Figure 4 , Figure 4 This is a schematic diagram of the parking position determination device provided in the embodiments of this application, such as... Figure 4 As shown, the parking location determination device includes the following modules.

[0118] The environmental information acquisition module 401 is used to acquire surrounding environmental information collected by vehicle sensors.

[0119] The grid map building module 402 is used to construct a first-precision grid map based on the target parking space and surrounding environmental information in response to the selection of the target parking space.

[0120] The alternative parking location determination module 403 is used to select at least one alternative parking location from the target parking space based on a first-precision grid map.

[0121] The cost function construction module 404 is used to construct multi-dimensional cost functions based on each alternative parking location. The multi-dimensional cost functions include collision cost function, accessibility cost function, and safe distance cost function.

[0122] The target parking location determination module 405 is used to determine the cost value of each candidate parking location based on a multi-dimensional cost function, and to take the candidate parking location with the minimum cost value as the target parking location.

[0123] In some embodiments, the raster map building module 402 is further configured to:

[0124] Based on the first-precision raster map, a second-precision raster map is constructed within a preset area of ​​the target parking space, wherein the second precision is less than the first precision;

[0125] Based on the second-precision raster map, select at least one alternative parking location from the target parking space.

[0126] In some embodiments, the surrounding environment information includes obstacle distribution information; based on the first-precision raster map, a multi-dimensional cost function is constructed, and the cost function construction module 404 is further configured to:

[0127] Based on obstacle distribution information and vehicle collision volume model, the probability of a collision between the vehicle and an obstacle is determined, and a collision cost function is constructed based on the probability.

[0128] And / or, determine whether the alternative parking location is within the reachable range of the vehicle's driving path, and construct an accessibility cost function based on whether the alternative parking location is reachable;

[0129] And / or, obtain the difference between the distance between the obstacle and the vehicle and the preset safe distance, and construct a safe distance cost function based on the difference.

[0130] In some embodiments, the multi-dimensional cost function further includes a centering cost function, which characterizes the degree of deviation between the candidate parking location and the center point of the target parking space; the cost function construction module 404 is also used for:

[0131] Based on the first-precision raster map, the coordinates of the center point of the candidate parking position and the target parking space are determined, and the lateral offset and longitudinal offset of the center point of the candidate parking position and the target parking space are obtained.

[0132] A centering cost function is constructed based on the horizontal and vertical offsets.

[0133] In some embodiments, the multi-dimensional cost function further includes a bias cost function, which characterizes the degree of lateral deviation of the vehicle's parking posture from its center position in the target parking space; the cost function construction module 404 is also used for:

[0134] If there are obstacles on the left and right sides of the target parking space, obtain the first distance between the vehicle and the obstacle on the left, and the second distance between the vehicle and the obstacle on the right, and determine the first difference between the first distance and the second distance.

[0135] A bias cost function is constructed based on the first difference and the second difference of the preset bias distance.

[0136] In some embodiments, the target parking location determination module 405 is further configured to:

[0137] Within the preset area of ​​the target parking space, each grid point is designated as a candidate parking location;

[0138] Furthermore, different cost functions have different weights. Based on the multi-dimensional cost function, the target parking position corresponding to the minimum cost value is obtained, including:

[0139] Based on the weights, the individual cost functions in the multi-dimensional cost function are weighted and fused to obtain the total cost function;

[0140] Within the preset area, calculate the total value of each alternative parking location, and select the alternative parking location with the lowest total value as the target parking location.

[0141] In some embodiments, within a preset area, the total value corresponding to each candidate parking location is calculated, and the candidate parking location with the lowest total value is selected as the target parking location. The target parking location determination module 405 is further configured to:

[0142] The center point of the target parking space is taken as the first alternative parking location, and the total agency value corresponding to the first alternative parking location is calculated.

[0143] Find the grid point adjacent to the first alternative parking location as the second alternative parking location, and calculate the total cost of the second alternative parking location;

[0144] If the total value of the second alternative parking location is not less than the total value of the first alternative parking location, then the first alternative parking location will be used as the target parking location.

[0145] If the total value of the second alternative parking location is less than the total value of the first alternative parking location, then the second alternative parking location is used as the new first alternative parking location, and the search continues to find grid points adjacent to the new first alternative parking location as the new second alternative parking location.

[0146] If the total value of the new second alternative parking location is not less than the total value of the new first alternative parking location, then the new first alternative parking location will be used as the target parking location.

[0147] If the total value of the new second alternative parking location is less than the total value of the new first alternative parking location, then the new second alternative parking location will be used as the benchmark to continue searching for the target parking location.

[0148] The parking position determination device provided in this application can execute the parking position determination method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the various modules and units described above are the same as in the corresponding embodiments described above, and will not be repeated here.

[0149] Embodiments of this application also provide an electronic device, such as... Figure 5 As shown, it includes a memory 10 and a processor 20. The memory 10 stores a computer program, and the processor 20 is configured to run the computer program to perform the steps in any of the above-described embodiments of the parking location determination method.

[0150] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above-described parking location determination method embodiments when it is run.

[0151] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0152] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above-described parking position determination method embodiments.

[0153] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above-described parking position determination method embodiments.

[0154] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0155] The foregoing has provided a detailed description of a parking location determination method, apparatus, electronic device, and storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to aid in understanding the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A method for determining a parking location, characterized in that, The method includes: Acquire surrounding environmental information collected by vehicle sensors; In response to the selection of a target parking space, a first-precision grid map is constructed based on the target parking space and the surrounding environmental information. Based on the first precision grid map, at least one alternative parking location is selected from the target parking space; Using each of the candidate parking locations as a benchmark, a multi-dimensional cost function is constructed, which includes a collision cost function, an accessibility cost function, and a safety distance cost function. The collision cost function is constructed based on the obstacle distribution information in the surrounding environment and the vehicle collision volume model, by determining the probability of a collision between the vehicle and an obstacle. The accessibility cost function is constructed based on whether the candidate parking location is within the reachable range of the vehicle's driving path, and the determination of the reachable range is constrained by vehicle dynamics. The safety distance cost function is constructed based on the difference between the distance between the obstacle and the vehicle and a preset safety distance, using a piecewise function. Based on the multi-dimensional cost function, the cost value of each of the candidate parking locations is determined, and the candidate parking location corresponding to the minimum cost value is taken as the target parking location.

2. The method according to claim 1, characterized in that, The multi-dimensional cost function further includes a centering cost function, which is used to characterize the degree of deviation between the candidate parking position and the center point of the target parking space; the method further includes: Based on the first precision grid map, the coordinates of the candidate parking position and the center point of the target parking space are determined, and the lateral offset and longitudinal offset of the candidate parking position and the center point of the target parking space are obtained. Based on the horizontal offset and the vertical offset, the centering cost function is constructed.

3. The method according to claim 1, characterized in that, The multi-dimensional cost function further includes a bias cost function, which characterizes the degree of lateral deviation of the vehicle's parking posture from its center position in the target parking space; the method further includes: If there are obstacles on the left and right sides of the target parking space, obtain the first distance between the vehicle and the obstacle on the left, and the second distance between the vehicle and the obstacle on the right, and determine the first difference between the first distance and the second distance; The bias cost function is constructed based on the first difference and the second difference of the preset deviation distance.

4. The method according to claim 2 or 3, characterized in that, The method further includes: Within the preset area of ​​the target parking space, each grid point is designated as a candidate parking location; Furthermore, different cost functions have different weights, and obtaining the target parking position corresponding to the minimum cost value based on the multi-dimensional cost function includes: Based on the weights, the cost functions in the multi-dimensional cost function are weighted and fused to obtain the total cost function; Within the preset area, the total value corresponding to each of the candidate parking locations is calculated, and the candidate parking location with the lowest total value is selected as the target parking location.

5. The method according to claim 4, characterized in that, The step of calculating the total value corresponding to each of the candidate parking locations within the preset area, and selecting the candidate parking location with the lowest total value as the target parking location, includes: The center point of the target parking space is taken as the first alternative parking location, and the total value corresponding to the first alternative parking location is calculated. Find the grid point adjacent to the first alternative parking location as the second alternative parking location, and calculate the total cost of the second alternative parking location; If the total value of the second alternative parking location is not less than the total value of the first alternative parking location, then the first alternative parking location is taken as the target parking location. If the total value of the second alternative parking location is less than the total value of the first alternative parking location, then the second alternative parking location is used as the new first alternative parking location, and the grid point adjacent to the new first alternative parking location is searched as the new second alternative parking location. If the total value of the new second alternative parking location is not less than the total value of the new first alternative parking location, then the new first alternative parking location will be used as the target parking location. If the total value of the new second alternative parking location is less than the total value of the new first alternative parking location, then the new second alternative parking location will be used as the benchmark to continue searching for the target parking location.

6. The method according to claim 1, characterized in that, The method further includes: Based on the first precision raster map, a second precision raster map is constructed within a preset area of ​​the target parking space, wherein the second precision is less than the first precision; Based on the second precision raster map, at least one alternative parking location is selected from the target parking space.

7. A parking position determination device, characterized in that, The device includes: The environmental information acquisition module is used to acquire surrounding environmental information collected by the vehicle's sensors. A grid map building module is used to construct a first-precision grid map based on the target parking space and the surrounding environment information in response to the selection of a target parking space. The alternative parking location determination module is used to select at least one alternative parking location from the target parking space based on the first precision grid map. The cost function construction module is used to construct multi-dimensional cost functions based on each of the candidate parking locations. These multi-dimensional cost functions include a collision cost function, an accessibility cost function, and a safety distance cost function. The collision cost function is constructed by determining the probability of a collision between the vehicle and an obstacle based on obstacle distribution information and a vehicle collision volume model in the surrounding environment. The accessibility cost function is constructed based on whether the candidate parking location is within the reachable range of the vehicle's driving path, and the determination of the reachable range is constrained by vehicle dynamics. The safety distance cost function is constructed using a piecewise function based on the difference between the distance between the obstacle and the vehicle and a preset safety distance. The target parking location determination module is used to determine the cost value of each of the candidate parking locations based on the multi-dimensional cost function, and to take the candidate parking location with the minimum cost value as the target parking location.

8. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory stores computer instructions, and the processor executes the parking location determination method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the parking location determination method according to any one of claims 1 to 6.

Images