Parking optimization method based on mixed A* algorithm, vehicle and electronic equipment

Abstract

The invention provides a parking optimization method based on a hybrid A algorithm, a vehicle and electronic equipment. The method comprises the steps that a plurality of off-line parking scenes are constructed, each off-line parking scene comprises a terminal point, a plurality of reachable points and an obstacle, the terminal point represents a parking position point, and the reachable points represent passing points for a vehicle to reach the terminal point; determining a plurality of costs from the terminal point to each reachable point in the off-line parking scene, and generating a three-dimensional heuristic cost table of the vehicle in each off-line parking scene based on the costs; obtaining a current parking scene of the vehicle, determining the matching degree between the current parking scene and each off-line parking scene, and determining a target off-line parking scene according to the matching degree; and obtaining a target cost table, determining a current cost table of the current parking scene according to the target cost table by adopting a hybrid A algorithm, and determining a parking path of the vehicle based on the current cost table. The problem that in an existing scheme, the search efficiency is low in the process of determining the parking path through a mixed A algorithm is solved.

Description

Technical Field

[0001] This application relates to the field of parking optimization technology, specifically, to a hybrid A... Parking optimization methods based on algorithms, vehicles, computer-readable storage media, and electronic devices. Background Technology

[0002] With the continuous development of robotics, autonomous vehicles, and smart mobile devices, efficient path planning algorithms have become a core requirement in these technological fields. In these applications, A... The A-Star heuristic search algorithm is widely used because it can find the shortest path from the starting point to the destination in complex environments. The algorithm combines the completeness of breadth-first search with the efficiency of heuristic search. It guides the search priority by calculating the sum of g-cost (the actual cost from the starting point to the current point) and h-cost (the heuristically estimated cost from the current point to the destination), f-cost, thereby minimizing the search space and improving computational efficiency while ensuring that the optimal path is found.

[0003] Traditional Hybrid A The heuristic function is max(h_1, h_2), where h_1 is the Euclidean distance considering obstacles, and h_2 is the RS curve distance without considering obstacles. This calculation method has a large deviation in cost estimation when the obstacles are complex.

[0004] In addition, A When dealing with a search space with high dimensionality, the depth and width of the search tree will increase significantly. This not only consumes more computing resources, but may also lead to getting stuck in a local optimum due to the rapid expansion of the search tree, resulting in a decrease in the search efficiency of the global optimum path. Summary of the Invention

[0005] The main objective of this application is to provide a hybrid A-based The algorithm's parking optimization method, vehicle, computer-readable storage medium, and electronic devices aim to at least address the shortcomings of existing solutions employing hybrid A The algorithm suffers from low search efficiency in determining parking routes.

[0006] To achieve the above objectives, according to one aspect of this application, a hybrid A-based... The parking optimization method of the algorithm includes: constructing multiple offline parking scenarios for the vehicle, wherein each offline parking scenario includes a destination, multiple reachable points, and obstacles, the destination represents a parking position point, and the reachable points represent the points the vehicle passes through to reach the destination; determining multiple costs from the destination to each reachable point in each offline parking scenario, and generating a three-dimensional heuristic cost table for the vehicle in each offline parking scenario based on the costs; obtaining the current parking scenario of the vehicle, determining the matching degree between the current parking scenario and each offline parking scenario, and determining a target offline parking scenario based on the matching degree, wherein the target offline parking scenario represents the offline parking scenario with the highest matching degree to the current parking scenario; obtaining the target cost table, and using a hybrid A / B algorithm... The algorithm determines the current cost table of the current parking scenario based on the target cost table, and determines the parking path of the vehicle in the current parking scenario based on the current cost table, wherein the target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario.

[0007] Optionally, after determining the target offline parking scenario based on the matching degree, the method further includes: determining whether the matching degree between the current parking scenario and the target offline parking scenario is greater than a preset matching degree value; if the matching degree between the current parking scenario and the target offline parking scenario is less than or equal to the preset matching degree value, based on the target cost table of the target offline parking scenario, using the Dijkstra algorithm to determine multiple current costs from the endpoint to each reachable point in the current parking scenario, and updating the multiple current costs to the target cost table to obtain the current cost table corresponding to the current parking scenario.

[0008] Optionally, determining the matching degree between the current parking scenario and each of the offline parking scenarios includes: extracting feature vectors of the current parking scenario and each of the offline parking scenarios using a neural network model; and determining the matching degree between the current parking scenario and each of the offline parking scenarios based on the feature vectors of the current parking scenario and the feature vectors of each of the offline parking scenarios.

[0009] Optionally, before extracting feature vectors for the current parking scenario and each of the offline parking scenarios using a neural network model, the method further includes: constructing an initial neural network model; constructing a training dataset based on the offline parking scenarios; training the initial neural network model using the training dataset to obtain the neural network model, wherein the training dataset includes multiple sets of positive samples and multiple sets of negative samples.

[0010] Optionally, constructing a training dataset based on the offline parking scenario includes: a first determination step: randomly selecting one of the offline parking scenarios as the original preset scenario, performing a first preprocessing operation on the obstacles in the original preset scenario to obtain a first preprocessed scenario, and determining the original preset scenario and the first preprocessed scenario as a group of positive samples, wherein the first preprocessing operation includes obstacle translation processing and obstacle rotation processing; a second determination step: determining the first preprocessed scenario and other offline parking scenarios as a group of negative samples; a third determination step: performing a second preprocessing operation on the obstacles in the original preset scenario to obtain a second preprocessed scenario, and determining the original preset scenario and the second preprocessed scenario as a group of negative samples, wherein the second preprocessing operation includes obstacle addition operation and obstacle deletion operation; and repeating the first determination step, the second determination step, and the third determination step at least once in sequence to construct the training dataset.

[0011] Optionally, determining multiple costs from the endpoint to each reachable point in each of the offline parking scenarios includes: using Dijkstra's algorithm to determine multiple costs from the endpoint to each reachable point in each of the offline parking scenarios.

[0012] Optionally, constructing multiple offline parking scenarios for the vehicle includes: obtaining sampling rules, and constructing multiple offline parking scenarios for the vehicle based on the sampling rules, wherein the sampling rules include at least one of the following: the distance to obstacles on the left and right sides of the parking space, the distance to obstacles in front of the parking space, the width of obstacles in front of the parking space, and the length of the road in front of the parking space.

[0013] According to another aspect of this application, a vehicle is provided, comprising: a controller, the controller being configured to perform any of the aforementioned hybrid A-based... Parking optimization methods based on algorithms.

[0014] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to execute any of the aforementioned hybrid A-based... Parking optimization methods based on algorithms.

[0015] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing any of the hybrid A-based methods described above. Parking optimization methods based on algorithms.

[0016] Applying the technical solution of this application, multiple offline parking scenarios containing endpoints and reachable points are constructed. For each offline parking scenario, the cost from the endpoint to the reachable point is calculated, generating a three-dimensional heuristic cost table. When a vehicle needs to park, the current parking scenario is obtained and matched with offline scenarios to determine the target offline parking scenario with the highest matching degree. Subsequently, the heuristic cost table matching the scenario is downloaded from the cloud, and a hybrid A / B algorithm is used. The algorithm adjusts its search strategy based on real-time cost information, generates a cost table for the current parking scenario, and ultimately determines the optimal parking path. This method reduces the computational load of online search by constructing offline parking scenarios, while ensuring the algorithm's adaptability to new scenarios through online cost updates, effectively solving the problem of hybrid A / B parking scenarios. The algorithm addresses the issues of low search efficiency and poor path quality in parking path planning, improving the accuracy and robustness of parking path planning. This solves the problem of existing solutions using a hybrid A / B algorithm. The algorithm suffers from low search efficiency in determining parking routes. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 An embodiment of the present application is shown, illustrating an implementation based on hybrid A. Hardware structure diagram of the mobile terminal for the parking optimization algorithm;

[0019] Figure 2 An embodiment of this application is shown, based on a hybrid A... A flowchart illustrating the parking optimization method using the algorithm;

[0020] Figure 3 An embodiment of this application is shown, based on a hybrid A... Block diagram of the parking optimization device based on the algorithm. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] As described in the background section, existing solutions employ a hybrid A The algorithm suffers from low search efficiency in determining the parking path. To address this, a hybrid A / B algorithm is adopted in existing solutions. The algorithm suffers from low search efficiency in determining parking paths. Embodiments of this application provide a hybrid A / B algorithm. Parking optimization methods based on algorithms, vehicles, computer-readable storage media, and electronic devices.

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] The methods and embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is an embodiment of the present invention based on hybrid A The hardware structure diagram of the mobile terminal for the parking optimization algorithm. (Example:) Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0027] Memory 104 can be used to store computer programs, such as application software programs and modules, as in the hybrid A-based embodiment of the present invention. The computer program corresponding to the parking optimization method of the algorithm is executed by the processor 102, which runs the computer program stored in the memory 104, thereby performing various functional applications and data processing, i.e., implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one instance, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one instance, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0028] This embodiment provides a hybrid A-based system that runs on a mobile terminal, computer terminal, or similar computing device. Regarding the parking optimization method of the algorithm, it should be noted that the steps shown in the flowchart in the attached figure can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than that shown here.

[0029] Figure 2 Based on the embodiments of this application, it is a hybrid A A flowchart of the parking optimization method of the algorithm. (e.g.) Figure 2 As shown, the method includes the following steps:

[0030] Step S201: Construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenarios include an endpoint, multiple reachable points and obstacles, the endpoint represents the parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0031] In the offline parking scenario, obstacles are represented by images, showing their range and outline.

[0032] Step S202: Determine multiple costs from the destination to the reachable points in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0033] Specifically, by mixing A The algorithm determines multiple costs from the endpoint to each reachable point in each of the aforementioned offline parking scenarios. These costs include heuristic costs and actual costs. Heuristic costs are estimated costs from the current node to the target node, based on distance metrics (such as Euclidean distance, Manhattan distance, etc.), aiming to guide the search towards the target direction and accelerate the process of finding the optimal path. Actual costs are the actual costs from the starting point to the current node, taking into account the path length, number of turns, and any costs related to vehicle kinematic constraints (such as acceleration, deceleration, steering, etc.).

[0034] Step S203: Obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree. The target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0035] Step S204: Obtain the target cost table and use hybrid A. The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0036] Among them, mixture A The algorithm refers to a path planning method that combines the traditional A-star (A-search) algorithm with a vehicle kinematics model, and is suitable for automatic parking scenarios. The algorithm is a widely used algorithm for finding the shortest path in a graph. It combines heuristic search with breadth-first search or depth-first search, and guides the search process by evaluating the estimated cost from the current node to the target node (usually called h(n)) and the known cost from the starting point to the current node (usually called g(n)).

[0037] In this embodiment, by applying steps S201, S202, S203, and S204, multiple offline parking scenarios containing destinations and reachable points are constructed. For each offline parking scenario, the cost from the destination to the reachable point is calculated, generating a three-dimensional heuristic cost table. When a vehicle needs to park, the current parking scenario is obtained and matched with the offline scenarios to determine the target offline parking scenario with the highest matching degree. Subsequently, the heuristic cost table matching the scenario is downloaded from the cloud, and a hybrid A / B algorithm is used. The algorithm adjusts its search strategy based on real-time cost information, generates a cost table for the current parking scenario, and ultimately determines the optimal parking path. This method reduces the computational load of online search by constructing offline parking scenarios, while ensuring the algorithm's adaptability to new scenarios through online cost updates, effectively solving the problem of hybrid A / B parking scenarios. The algorithm addresses the issues of low search efficiency and poor path quality in parking path planning, improving the accuracy and robustness of parking path planning. This solves the problem of existing solutions using a hybrid A / B algorithm. The algorithm suffers from low search efficiency in determining parking routes.

[0038] In the specific implementation process, after determining the target offline parking scenario based on the matching degree, the method further includes: determining whether the matching degree between the current parking scenario and the target offline parking scenario is greater than a preset matching degree value; if the matching degree between the current parking scenario and the target offline parking scenario is less than or equal to the preset matching degree value, based on the target cost table of the target offline parking scenario, using the Dijkstra algorithm to determine multiple current costs from the endpoint to each reachable point in the current parking scenario, and updating the multiple current costs to the target cost table to obtain the current cost table corresponding to the current parking scenario.

[0039] The matching degree preset value can be set according to the accuracy of the test results. For example, the matching degree preset value can be set to 90% or 95%.

[0040] In addition, after obtaining the current cost table corresponding to the current parking scenario, the current parking scenario and the corresponding current cost table are uploaded to the cloud as new offline parking scenarios to enrich the offline parking scenario library, so that they can be directly matched and used when the vehicle is parked next time.

[0041] This method addresses the issue where, when the matching degree between the current parking environment matched online and the target parking scenario in the offline preset scenarios falls below a preset threshold, the system uses the offline scenario with the highest matching degree as the basis for the target cost table. It then employs Dijkstra's algorithm to calculate the current cost from the destination to each reachable point in the current parking scenario in real time and updates it to the target cost table, thus forming the updated current cost table corresponding to the current parking scenario. This process ensures that heuristic costs can be quickly adjusted even in new environments, guaranteeing the accuracy and efficiency of path planning. The new cost data generated through online reverse search not only allows for immediate correction of the current search path but also enables the data to be uploaded to the cloud as new preset scenarios after the parking task is completed, continuously optimizing the scenario library and enhancing the algorithm's adaptability and generalization performance to unknown environments. The combination of online cost updates and scenario addition mechanisms achieves dynamic improvement in heuristic search capabilities, providing a more flexible and efficient technical solution for applications such as autonomous parking and robot navigation.

[0042] Specifically, determining the matching degree between the current parking scenario and each of the offline parking scenarios includes: using a neural network model to extract feature vectors of the current parking scenario and each of the offline parking scenarios respectively; and determining the matching degree between the current parking scenario and each of the offline parking scenarios based on the feature vectors of the current parking scenario and the feature vectors of each of the offline parking scenarios.

[0043] Among them, the neural network model can be the CNN model.

[0044] This method uses a neural network model to extract feature vectors when determining the matching degree between the current parking scene and each offline parking scene. Specifically, by extracting features from the current parking scene and all preset offline parking scene images, the model can quantitatively capture the key obstacle layout and environmental characteristics of each scene. Subsequently, based on the extracted feature vectors, the similarity distance or correlation coefficient between the vectors can be calculated to objectively measure the matching degree between the current parking scene and the preset offline scenes. This method not only improves the accuracy of scene recognition but also ensures the automation and efficiency of the matching process, enabling online search to quickly locate the closest offline heuristic cost table, thereby guiding A... The algorithm performs path planning in the new environment. Of course, if the online scene cannot reach the preset matching threshold with any offline scene, it will trigger an online reverse search to generate a real-time heuristic cost, or upload the new scene and its cost table to the cloud after the task is completed, further enriching and optimizing the coverage and matching capability of the offline scene library, and realizing the algorithm's self-iteration and improvement.

[0045] Furthermore, before extracting the feature vectors of the current parking scenario and each of the offline parking scenarios using a neural network model, the method further includes: constructing an initial neural network model; constructing a training dataset based on the offline parking scenarios; training the initial neural network model using the training dataset to obtain the neural network model, wherein the training dataset includes multiple sets of positive samples and multiple sets of negative samples.

[0046] This method aims to improve A To improve the performance of the heuristic search algorithm in autonomous driving parking scenarios, the solution incorporates a combination of offline preset scenarios and online correction. First, an initial neural network model is constructed. This model is used for feature extraction and matching of the current parking scenario during the online phase, quickly identifying the closest offline preset scenario. A training dataset is built based on the offline parking scenarios, encompassing multiple sets of positive and negative samples. Through training with a large number of samples, the neural network model learns subtle differences in features between different scenarios, thereby improving the accuracy of online scenario matching. The generation strategies for positive and negative samples include translating and rotating obstacles, as well as adding new obstacles. These operations simulate various changes that may occur in the actual parking environment. When encountering a new scenario that differs significantly from the offline preset scenario during the online phase, the solution dynamically adjusts the heuristic cost by performing a reverse search and online cost update, ensuring the real-time performance and flexibility of the search strategy. The reverse search is based on the scenario with the highest offline matching degree, gradually correcting the heuristic cost. The online cost update, after the parking task is completed, uploads the corrected cost table as a new preset scenario to the cloud, enriching the scenario library and enhancing the algorithm's generalization ability. This series of operations not only reduces the complexity of online search, but also improves the efficiency and quality of route planning, especially when dealing with complex or dynamically changing parking environments.

[0047] Furthermore, the training dataset is constructed based on the aforementioned offline parking scenarios, including: a first determination step: randomly selecting one of the aforementioned offline parking scenarios as the original preset scenario, and performing a first preprocessing operation on the obstacles in the original preset scenario to obtain a first preprocessed scenario, and determining the original preset scenario and the first preprocessed scenario as a group of positive samples, wherein the first preprocessing operation includes obstacle translation processing and obstacle rotation processing; a second determination step: determining the first preprocessed scenario and other offline parking scenarios as a group of negative samples; a third determination step: performing a second preprocessing operation on the obstacles in the original preset scenario to obtain a second preprocessed scenario, and determining the original preset scenario and the second preprocessed scenario as a group of negative samples, wherein the second preprocessing operation includes obstacle addition operation and obstacle deletion operation; the first determination step, the second determination step, and the third determination step are repeated at least once in sequence to construct the aforementioned training dataset.

[0048] In the offline phase, this method generates a set of positive samples closely related to the original scene by translating and rotating obstacles in a pre-defined scene. Then, it optimizes the CNN Siamese network using contrastive loss to effectively extract and match scene features. Subsequently, by adding or removing obstacles from the original scene, a set of negative samples is constructed, further enhancing the model's ability to recognize new scenes. In the online phase, when encountering a new scene not pre-defined, the system uses Dijkstra's algorithm for real-time reverse search, dynamically updating the heuristic cost table to ensure that cost information during the search process reflects environmental changes in a timely manner, improving search efficiency and path planning accuracy.

[0049] Dijkstra's algorithm is a classic algorithm for calculating the single-source shortest path in a weighted graph.

[0050] Furthermore, after a parking task is completed, the heuristic cost table generated online is uploaded to the cloud as a new preset scenario, continuously enriching the preset scenario library and improving search speed and decision quality in similar future scenarios. This series of operations effectively balances the algorithm's preprocessing work with real-time adaptability, achieving efficient path planning in diverse parking environments. When the matching degree between the online scenario and any offline preset scenario is lower than a set threshold, the system can automatically identify and generate a new heuristic cost table, indicating that the solution has good adaptability to new scenarios and a self-improvement mechanism. By continuously updating and adding preset scenarios, the system can better cope with various parking challenges, improving search success rate and robustness of path planning. Real-time updates of online costs ensure that the algorithm can respond instantly to environmental changes, reduce unnecessary search node expansion, and accelerate the search process, thereby significantly improving the efficiency of A heuristic search while ensuring path quality.

[0051] Specifically, determining multiple costs from the endpoint to the reachable point in each of the aforementioned offline parking scenarios includes: using Dijkstra's algorithm to determine multiple costs from the endpoint to the reachable point in each of the aforementioned offline parking scenarios.

[0052] This method uses Dijkstra's algorithm, starting from the end of the parking space, to calculate and store the cost of each reachable location, forming a three-dimensional heuristic cost table to provide a reference for subsequent online searches. This process is completed in an offline environment, and the generated cost table can cover various parking scenarios. Therefore, during online searches, it can quickly find the closest preset situation through scenario matching, significantly improving A / B performance. The efficiency of the search algorithm and the accuracy of the path planning.

[0053] More specifically, constructing multiple offline parking scenarios for the vehicle includes: obtaining sampling rules, and constructing multiple offline parking scenarios for the vehicle based on the sampling rules, wherein the sampling rules include at least one of the following: the distance to obstacles on the left and right sides of the parking space, the distance to obstacles in front of the parking space, the width of obstacles in front of the parking space, and the length of the road in front of the parking space.

[0054] Among them, the preset scenarios are distinguished by vertical / horizontal parking. For each parking scenario, equal-interval sampling is performed according to the following dimensional combinations based on the sampling rules (taking vertical as an example):

[0055] 1) Distance to obstacles on the left and right sides of the parking space;

[0056] 2) Distance to obstacles in front of the parking space;

[0057] 3) Width of obstacles in front of the parking space;

[0058] 4) Length of the road in front of the parking space;

[0059] This method obtains multiple different offline parking scenarios for vehicles by sampling the sampling rules at equal intervals according to different dimensions; and constructs an offline scenario library that can cover a wide range of parking environments. The core of this design lies in utilizing a heuristic cost table from offline computation to provide a basis for online hybrid A / B testing. Search provides initial guidance information, significantly improving search efficiency.

[0060] To enable those skilled in the art to better understand the technical solution of this application, the following will describe the application based on hybrid A in conjunction with specific embodiments. The implementation process of the parking optimization method of the algorithm is explained in detail.

[0061] This embodiment relates to a specific hybrid A-based approach. The core of the parking optimization algorithm is to replace the original hybrid A algorithm by offline pre-setting heuristic cost tables for multiple scenarios and online scene matching. Heuristic costs; and improving functionality versatility by updating heuristic costs and enriching the cost table in real time when new scenarios emerge.

[0062] I. Offline Preset Scenes:

[0063] 1. The preset scenarios are categorized into vertical and horizontal parking scenarios. For each parking scenario, the following dimensions are combined and sampled at equal intervals (taking vertical as an example):

[0064] a. Distance to obstacles on the left and right sides of the parking space;

[0065] b. Distance to obstacles in front of the parking space;

[0066] c. Width of obstacles in front of the parking space;

[0067] d. Length of the road in front of the parking space;

[0068] 2. Calculate the cost from the destination to each reachable point for each preset scenario using Dijkstra's algorithm. The cost calculation method is similar to that of hybrid A / B algorithm. The gcost is the same. The calculation results are used to generate a 3D heuristic cost table of x, y, and theta around the parking space and stored in the cloud for use by all vehicles. At the same time, an obstacle image of the preset scene is generated for subsequent online matching.

[0069] II. Online Scene Matching:

[0070] The obstacles in the target task scene are constructed as images, and the features of the images are extracted using a CNN network. The matching degree is calculated with all offline preset scenes, and the offline scene with the highest matching degree is found.

[0071] Training methods:

[0072] Construct a CNN twin network to extract feature vectors from two scene images and optimize the contrast loss.

[0073] Training data:

[0074] Randomly translate and rotate obstacles in an offline preset scene by a small amount, and generate positive samples with the original preset scene;

[0075] Randomly translate and rotate obstacles in an offline preset scene by a small amount, and generate negative samples with other offline preset scenes;

[0076] Randomly add new obstacles to an offline preset scene and generate negative samples with the original preset scene;

[0077] Use the above method to generate multiple sets of positive and negative samples for training.

[0078] How to use:

[0079] The feature vectors of all preset scenes are calculated and pre-stored offline. Online, only a single CNN network is used to calculate the feature vector of the target scene. After the target scene successfully matches the offline scene, the corresponding heuristic cost table is downloaded from the cloud. Hybrid A When heuristic costs need to be calculated, a direct match query is performed from the table.

[0080] III. Online Cost Updates:

[0081] Set a matching threshold. When the matching degree between the target scene and all offline scenes is lower than the threshold, the current scene is considered a new scene, and the following strategy is executed:

[0082] 1. Online reverse search:

[0083] Using the cost table with the highest matching degree as a base, a reverse search is continuously performed during the task. The reverse search method involves using Dijkstra's algorithm to calculate the cost from the destination to each reachable point, and directly replacing the cost value in the base table with the result. At this point, a hybrid A... The search uses a cost table that is updated in real time to calculate the heuristic cost.

[0084] 2. Add preset scenes:

[0085] After the parking task is completed, the heuristic cost table generated by the online reverse search is uploaded to the cloud as a new preset scenario, and a new image of the preset scenario obstacles is generated for subsequent online matching.

[0086] This embodiment pre-sets multiple typical scenarios and obstacle configurations offline, covering diverse situations such as static obstacles, dynamic components, and complex terrain. By using A... Using the same parameters, a reverse search is performed from the target point, calculating and storing the cost of each location relative to the target, as an online hybrid A. Heuristic cost is used to improve the guidance quality and pruning efficiency of online search.

[0087] In the online phase, this embodiment uses machine learning methods to map / match the current environmental perception or map information to the most similar scene in the offline scene set, and reads the pre-computed heuristic cost corresponding to the offline scene as the online hybrid A. The initial heuristic information quickly guides the search direction.

[0088] Because the online environment differs from the offline preset, this embodiment continuously performs a reverse search while the online search is underway to constantly update and enrich the heuristic cost. The new costs obtained from the reverse search are updated in real time to the online heuristic cost, ensuring that the heuristic information remains consistent with the current environment. Alternatively, the results of the online search can be used to construct new scenarios to expand the richness of the offline scenario library.

[0089] This embodiment leverages existing high-quality offline heuristic costs to reduce the expansion depth of online search, significantly improving search efficiency. Continuous updates to the heuristic costs via online reverse search enhance robustness to scenario changes. This method is applicable not only to autonomous driving path planning but also to scenarios requiring efficient and interpretable heuristic search, such as robot navigation.

[0090] This application also provides a hybrid A-based embodiment. The parking optimization device based on the algorithm should be noted that the embodiments of this application are based on hybrid A. The parking optimization device of the algorithm can be used to execute the parking optimization device based on hybrid A provided in the embodiments of this application. A parking optimization method based on an algorithm. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0091] This embodiment also relates to a vehicle, characterized in that it includes: a controller, the controller being used to execute any of the above-described hybrid A-based... The algorithm describes a parking optimization method. The controller is the vehicle's ECU.

[0092] The following describes the hybrid A-based embodiments provided in this application. The parking optimization device based on the algorithm will be introduced.

[0093] Figure 3 Based on the embodiments of this application, it is a hybrid A A schematic diagram of the parking optimization device based on the algorithm. (Example) Figure 3 As shown, the device includes:

[0094] Construction unit 31 is used to construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenario includes an endpoint, multiple reachable points and obstacles, the endpoint represents a parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0095] The first determining unit 32 is used to determine multiple costs from the endpoint to the reachable point in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0096] The second determining unit 33 is used to obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree, wherein the target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0097] The third determining unit 34 is used to obtain the target cost table, employing a hybrid A... The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0098] In this embodiment, the construction unit is used to construct multiple offline parking scenarios for the vehicle, wherein each offline parking scenario includes a destination, multiple reachable points, and obstacles. The destination represents a parking location point, and the reachable points represent the points the vehicle passes through to reach the destination. The first determining unit is used to determine multiple costs from the destination to each reachable point in each offline parking scenario, and generate a three-dimensional heuristic cost table for the vehicle in each offline parking scenario based on the costs. The second determining unit is used to obtain the current parking scenario of the vehicle, determine the matching degree between the current parking scenario and each offline parking scenario, and determine a target offline parking scenario based on the matching degree, wherein the target offline parking scenario represents the offline parking scenario with the highest matching degree to the current parking scenario. The third determining unit is used to obtain the target cost table, using a hybrid A The algorithm determines the current cost table for the current parking scenario based on the aforementioned target cost table, and then determines the parking path of the vehicle within that scenario based on the current cost table. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario. Multiple offline parking scenarios containing endpoints and reachable points are constructed. For each offline parking scenario, the cost from the endpoint to the reachable point is calculated, generating a three-dimensional heuristic cost table. When a vehicle needs to park, the current parking scenario is obtained and matched with the offline scenarios to determine the target offline parking scenario with the highest matching degree. Subsequently, the heuristic cost table matching this scenario is downloaded from the cloud and a hybrid A / B algorithm is used. The algorithm adjusts its search strategy based on real-time cost information, generates a cost table for the current parking scenario, and ultimately determines the optimal parking path. This method reduces the computational load of online search by constructing offline parking scenarios, while ensuring the algorithm's adaptability to new scenarios through online cost updates, effectively solving the problem of hybrid A / B parking scenarios. The algorithm addresses the issues of low search efficiency and poor path quality in parking path planning, improving the accuracy and robustness of parking path planning. This solves the problem of existing solutions using a hybrid A / B algorithm. The algorithm suffers from low search efficiency in determining parking routes.

[0099] As an optional solution, the device further includes a fourth determining unit and a fifth determining unit; the fourth determining unit is used to determine whether the matching degree between the current parking scene and the target offline parking scene is greater than a preset matching degree value after determining the target offline parking scene based on the matching degree; the fifth determining unit is used to determine multiple current costs from the endpoint to each reachable point in the current parking scene based on the target cost table of the target offline parking scene when the matching degree between the current parking scene and the target offline parking scene is less than or equal to the preset matching degree value, and update the multiple current costs to the target cost table to obtain the current cost table corresponding to the current parking scene.

[0100] In one optional scheme, the second determining unit includes an extraction module and a first determining module; the extraction module is used to extract feature vectors of the current parking scenario and each of the offline parking scenarios using a neural network model; the first determining module is used to determine the matching degree between the current parking scenario and each of the offline parking scenarios based on the feature vectors of the current parking scenario and the feature vectors of each of the offline parking scenarios.

[0101] In one optional scheme, the second determining unit further includes a construction module and a training module. The construction module is used to construct an initial neural network model before extracting feature vectors of the current parking scenario and each of the offline parking scenarios using a neural network model. The training module is used to construct a training dataset based on the offline parking scenarios and to train the initial neural network model using the training dataset to obtain the neural network model. The training dataset includes multiple sets of positive samples and multiple sets of negative samples.

[0102] An optional scheme, the training module includes a first determining submodule, a second determining submodule, a third determining submodule, and a repetitive execution submodule; the first determining submodule is used to perform a first determining step: randomly select one of the above-mentioned offline parking scenarios as the original preset scenario, and perform a first preprocessing operation on the obstacles in the original preset scenario to obtain a first preprocessed scenario, and determine the original preset scenario and the first preprocessed scenario as a group of positive samples, wherein the first preprocessing operation includes obstacle translation processing and obstacle rotation processing; the second determining submodule is used to perform a second determining step: determine the first preprocessed scenario and other offline parking scenarios as a group of negative samples; the third determining submodule is used to perform a third determining step: perform a second preprocessing operation on the obstacles in the original preset scenario to obtain a second preprocessed scenario, and determine the original preset scenario and the second preprocessed scenario as a group of negative samples, wherein the second preprocessing operation includes obstacle addition operation and obstacle deletion operation; the repetitive execution submodule is used to sequentially repeat the first determining step, the second determining step, and the third determining step at least once to construct the above-mentioned training dataset.

[0103] In one alternative, the first determining unit includes a second determining module, used to determine multiple costs from the endpoint to the reachable points in each of the aforementioned offline parking scenarios using the Dijkstra algorithm.

[0104] In one optional scheme, the acquisition unit includes an acquisition module for acquiring sampling rules and constructing multiple offline parking scenarios of the vehicle based on the sampling rules, wherein the sampling rules include at least one of the following: the distance to obstacles to the left and right of the parking space, the distance to obstacles in front of the parking space, the width of obstacles in front of the parking space, and the length of the road in front of the parking space.

[0105] The above is based on hybrid A The parking optimization device of the algorithm includes a processor and a memory. The aforementioned acquisition unit, first determination unit, second determination unit, and third determination unit are all stored as program units in the memory. The processor executes the aforementioned program units stored in the memory to implement the corresponding functions. All of the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0106] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured; the solution to the problem can be achieved by adjusting kernel parameters. (The last sentence appears to be incomplete and possibly refers to a different processor using A.) The algorithm suffers from low search efficiency in determining parking routes.

[0107] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0108] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the above-described hybrid A-based... Parking optimization methods based on algorithms.

[0109] Specifically, based on hybrid A The parking optimization methods of the algorithm include:

[0110] Step S201: Construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenarios include an endpoint, multiple reachable points and obstacles, the endpoint represents the parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0111] Step S202: Determine multiple costs from the destination to the reachable points in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0112] Step S203: Obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree. The target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0113] Step S204: Obtain the target cost table and use hybrid A. The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0114] This invention provides a processor for running a program, wherein the program executes the above-mentioned hybrid A-based... Parking optimization methods based on algorithms.

[0115] Specifically, based on hybrid A The parking optimization methods of the algorithm include:

[0116] Step S201: Construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenarios include an endpoint, multiple reachable points and obstacles, the endpoint represents the parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0117] Step S202: Determine multiple costs from the destination to the reachable points in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0118] Step S203: Obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree. The target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0119] Step S204: Obtain the target cost table and use hybrid A. The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0120] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0121] Step S201: Construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenarios include an endpoint, multiple reachable points and obstacles, the endpoint represents the parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0122] Step S202: Determine multiple costs from the destination to the reachable points in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0123] Step S203: Obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree. The target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0124] Step S204: Obtain the target cost table and use hybrid A. The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0125] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0126] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps:

[0127] Step S201: Construct multiple offline parking scenarios for the vehicle, wherein the offline parking scenarios include an endpoint, multiple reachable points and obstacles, the endpoint represents the parking location point, and the reachable points represent the points through which the vehicle passes to reach the endpoint.

[0128] Step S202: Determine multiple costs from the destination to the reachable points in each of the above-mentioned offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the above-mentioned offline parking scenarios based on the costs.

[0129] Step S203: Obtain the current parking scene of the vehicle, determine the matching degree between the current parking scene and each of the offline parking scenes, and determine the target offline parking scene based on the matching degree. The target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene.

[0130] Step S204: Obtain the target cost table and use hybrid A. The algorithm determines the current cost table of the current parking scenario based on the target cost table mentioned above, and determines the parking path of the vehicle in the current parking scenario based on the current cost table mentioned above. The target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario mentioned above.

[0131] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0132] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0133] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0134] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0135] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0136] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0137] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0138] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0139] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0140] It should also be noted that 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 process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0141] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A hybrid A-based The parking optimization method of the algorithm is characterized by, include: Construct multiple offline parking scenarios for a vehicle, wherein each offline parking scenario includes a destination, multiple reachable points, and obstacles, the destination represents a parking location point, and the reachable points represent the points the vehicle passes through to reach the destination; Determine multiple costs from the destination to each reachable point in each of the offline parking scenarios, and generate a three-dimensional heuristic cost table for the vehicle in each of the offline parking scenarios based on the costs; The current parking scene of the vehicle is obtained, the matching degree between the current parking scene and each of the offline parking scenes is determined, and a target offline parking scene is determined based on the matching degree, wherein the target offline parking scene represents the offline parking scene with the highest matching degree with the current parking scene; Obtain the target cost table and use hybrid A The algorithm determines the current cost table of the current parking scenario based on the target cost table, and determines the parking path of the vehicle in the current parking scenario based on the current cost table, wherein the target cost table is the three-dimensional heuristic cost table corresponding to the target offline parking scenario.

2. The method according to claim 1, characterized in that, After determining the target offline parking scenario based on the matching degree, the method further includes: Determine whether the matching degree between the current parking scenario and the target offline parking scenario is greater than a preset matching degree value; If the matching degree between the current parking scenario and the target offline parking scenario is less than or equal to the preset matching degree value, based on the target cost table of the target offline parking scenario, the Dijkstra algorithm is used to determine multiple current costs from the endpoint to each reachable point in the current parking scenario, and the multiple current costs are updated to the target cost table to obtain the current cost table corresponding to the current parking scenario.

3. The method according to claim 1, characterized in that, Determining the matching degree between the current parking scenario and each of the offline parking scenarios includes: A neural network model is used to extract feature vectors for the current parking scenario and each of the offline parking scenarios. The matching degree between the current parking scenario and each offline parking scenario is determined based on the feature vector of the current parking scenario and the feature vector of each offline parking scenario.

4. The method according to claim 3, characterized in that, Before using a neural network model to extract feature vectors for the current parking scenario and each of the offline parking scenarios, the method further includes: Construct the initial neural network model; A training dataset is constructed based on the offline parking scenario, and the initial neural network model is trained using the training dataset to obtain the neural network model. The training dataset includes multiple sets of positive samples and multiple sets of negative samples.

5. The method according to claim 4, characterized in that, A training dataset is constructed based on the aforementioned offline parking scenario, including: First determination step: Randomly select one of the offline parking scenarios as the original preset scenario, and perform a first preprocessing operation on the obstacles in the original preset scenario to obtain a first preprocessed scenario, and determine the original preset scenario and the first preprocessed scenario as a set of positive samples, wherein the first preprocessing operation includes obstacle translation processing and obstacle rotation processing; The second determination step: The first preprocessed scenario and other offline parking scenarios are determined as a group of negative samples; The third determination step is to perform a second preprocessing operation on the obstacles in the original preset scene to obtain a second preprocessed scene, and to determine the original preset scene and the second preprocessed scene as a set of negative samples, wherein the second preprocessing operation includes an obstacle addition operation and an obstacle deletion operation; The first determination step, the second determination step, and the third determination step are executed at least once in sequence to construct the training dataset.

6. The method according to claim 1, characterized in that, Determine multiple costs from the destination to each reachable point in each of the aforementioned offline parking scenarios, including: Dijkstra's algorithm is used to determine multiple costs from the destination to each reachable point in each of the offline parking scenarios.

7. The method according to claim 1, characterized in that, Construct multiple offline parking scenarios for the vehicle, including: Obtain sampling rules, and construct multiple offline parking scenarios for the vehicle based on the sampling rules, wherein the sampling rules include at least one of the following: distance to obstacles on the left and right sides of the parking space, distance to obstacles in front of the parking space, width of obstacles in front of the parking space, and length of road in front of the parking space.

8. A vehicle, characterized in that, include: Controller, the controller being configured to perform the hybrid A-based method as described in any one of claims 1 to 7 Parking optimization methods based on algorithms.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the hybrid A-based method according to any one of claims 1 to 7. Parking optimization methods based on algorithms.

10. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising methods for performing the hybrid A-based method according to any one of claims 1 to 7. Parking optimization methods based on algorithms.

Images