Vehicle automatic parking method, device, storage medium and computer equipment
By constructing a digital twin model for path planning and safety detection, the limitations and error rates of existing automatic parking technologies have been addressed, enabling precise vehicle parking and efficient use of parking lots.
Patent Information
- Application Number
- CN202411186760.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-28
AI Technical Summary
Existing automatic parking technology is limited to the identification and parking of a single parking space, resulting in limitations and error rates in automatic parking operations, which affects the utilization rate of parking lots.
By acquiring the target vehicle's location information and real-time parking data, a digital twin model is constructed to perform path planning and parking simulation. The digital twin model is then used for global information perception, optimizing parking paths, and performing safe distance detection and fine-tuning to ensure precise vehicle parking.
It improves the accuracy and efficiency of automatic parking, reduces limitations and error rates, and increases the utilization rate of parking lots.
Smart Images

Figure CN118701036B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a vehicle automatic parking method, device, storage medium and computer equipment. Background Art
[0002] In structured indoor parking lots, the parking and exit process is relatively smooth. However, in outdoor areas that are not fully planned, the situation is quite different; due to the lack of clear parking line guidance, drivers may find it difficult to fully assess the surrounding environment due to limited vision when looking for parking spaces, which may lead to improper parking and even be blocked by following vehicles, making it difficult for vehicles to enter and exit, greatly reducing the convenience and efficiency of parking.
[0003] In this context, automatic parking technologies have been gradually introduced on the market, which mainly rely on sensors such as car radars and cameras to identify empty parking spaces and assist vehicles in parking. However, traditional automatic parking technologies are often limited to the identification and parking operations of a single parking space, resulting in certain limitations and error rates in the automatic parking operations of vehicles, which in turn affects the utilization rate of parking lots. Summary of the invention
[0004] The purpose of the present application is to solve at least one of the above-mentioned technical defects, especially the technical defect that the automatic parking technology in the prior art is often limited to the identification and parking operation of a single parking space, resulting in certain limitations and error rates in the automatic parking operation of the vehicle, thereby affecting the utilization rate of the parking lot.
[0005] The present application provides a vehicle automatic parking method, the method comprising:
[0006] Acquire the positioning information of the target vehicle, and collect the real-time parking data of the target vehicle based on the positioning information;
[0007] Determine a digital twin model of the target vehicle, and use the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space;
[0008] Controlling the target vehicle to go to the target parking space to perform a parking operation according to the optimal path, and obtaining a vehicle parking position of the target vehicle in the target parking space after completing the parking operation;
[0009] A safety distance detection is performed on the parking position of the vehicle, and the parking position of the vehicle is fine-tuned in the target parking space according to the detection result to complete the parking action of the target vehicle.
[0010] Optionally, the collecting the real-time parking data of the target vehicle based on the positioning information includes:
[0011] Determine a parking place for the target vehicle according to the positioning information; the parking place includes an indoor parking lot and an outdoor parking lot;
[0012] When the parking lot is an indoor parking lot, the environmental information of the location of the target vehicle is collected by a detection sensor as the real-time parking data of the target vehicle;
[0013] When the parking place is an outdoor parking lot, the environmental information of the location of the target vehicle is collected by the detection sensor, and a remote sensing image of the outdoor parking lot is collected by a remote sensing satellite, and real-time parking data of the target vehicle is formed based on the environmental information and the remote sensing image.
[0014] Optionally, determining the digital twin model of the target vehicle includes:
[0015] Acquiring historical operation data and historical parking data of the target vehicle, and performing preprocessing operations on the historical operation data and the historical parking data to obtain training data;
[0016] Iteratively training a preset pre-trained model using the training data, so as to optimize the parameters of the pre-trained model using an optimization algorithm during the training process;
[0017] When the pre-trained model meets the preset training end condition, the pre-trained model that has completed training is used as the digital twin model of the target vehicle.
[0018] Optionally, the using the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space includes:
[0019] Using the digital twin model to perform parking simulation on the target vehicle based on the real-time parking data, and outputting parking scores of multiple vacant parking spaces;
[0020] Sorting the parking scores from low to high, and taking the vacant parking space with the lowest parking score in the sorting result as the target parking space for the target vehicle;
[0021] The position information of the target parking space is extracted, and the optimal path of the target vehicle is planned according to the position information and the positioning information.
[0022] Optionally, the using the digital twin model to perform parking simulation on the target vehicle based on the real-time parking data and outputting parking scores of multiple vacant parking spaces includes:
[0023] The real-time parking data is input into the digital twin model, so that the digital twin model determines a plurality of vacant parking spaces according to the real-time parking data, and simulates the parking operation of the target vehicle in each vacant parking space based on the real-time parking data, obtains the influence coefficient corresponding to each parking influencing factor in each vacant parking space, and determines the parking score of each vacant parking space according to the influence coefficient corresponding to each vacant parking space.
[0024] Optionally, the detecting a safe distance of the parking position of the vehicle includes:
[0025] Collecting the parking space line of the target parking space, and determining the parking distance of the target vehicle according to the parking space line and the parking position of the vehicle;
[0026] The parking space parking distance is matched with a preset safety distance range, and the matching result is used as the detection result of the safety distance detection of the target vehicle.
[0027] Optionally, the fine-tuning of the parking position of the vehicle in the target parking space according to the detection result includes:
[0028] When the detection result is failure, the parking space environment information of the target vehicle is collected, and the fine-tuning action of the target vehicle is determined based on the parking space environment information through the digital twin model, and after the target vehicle is controlled to perform a parking operation in the target parking space according to the fine-tuning action, the process returns to execute the step of obtaining the vehicle parking position of the target vehicle in the target parking space after completing the parking operation and subsequent steps;
[0029] When the detection result is passed, it is confirmed that the parking action of the target vehicle is completed.
[0030] The present application also provides a vehicle automatic parking device, comprising:
[0031] A data acquisition module, used to acquire the positioning information of the target vehicle, and collect the real-time parking data of the target vehicle based on the positioning information;
[0032] A path planning module, used to determine a digital twin model of the target vehicle, and use the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space;
[0033] a parking operation module, configured to control the target vehicle to go to the target parking space to perform a parking operation according to the optimal path, and obtain a vehicle parking position of the target vehicle in the target parking space after completing the parking operation;
[0034] The position fine-tuning module is used to detect the safe distance of the parking position of the vehicle and fine-tune the parking position of the vehicle in the target parking space according to the detection result to complete the parking action of the target vehicle.
[0035] The present application also provides a storage medium, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the vehicle automatic parking method as described in any one of the above embodiments.
[0036] The present application also provides a computer device, comprising: one or more processors, and a memory;
[0037] The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the vehicle automatic parking method as described in any one of the above embodiments are performed.
[0038] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:
[0039] The vehicle automatic parking method, device, storage medium and computer equipment provided by the present application can obtain the positioning information of the target vehicle when a target vehicle arrives at the parking lot, and collect the real-time parking data of the target vehicle based on the positioning information. Since the real-time parking data can provide global information of the target vehicle and its surrounding environment, the digital twin model of the target vehicle can be determined, and the digital twin model is used to perform path planning based on the real-time parking data to obtain the target parking space and the optimal path to the target parking space, so as to map and predict the dynamic changes of the parking lot, provide comprehensive parking decision support for the vehicle, and reduce the limitations of automatic parking; then, the target vehicle can be controlled to go to the target parking space for parking according to the optimal path, and the vehicle parking position of the target vehicle in the target parking space is obtained after the parking operation is completed, and then the safety distance detection of the vehicle parking position is performed, and the vehicle parking position in the target parking space is fine-tuned according to the detection result, so as to improve the accuracy of the parking operation of the target vehicle, so as to facilitate the smooth passage of subsequent vehicles and improve the utilization rate of the parking lot. This application uses a digital twin model to perform real-time perception of the global information of the parking lot and parking analysis, and then controls the target vehicle to perform precise parking, thereby reducing the limitations and error rate of automatic parking and improving the utilization rate of the parking lot. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0041] Figure 1 A schematic diagram of a process flow of an automatic parking method for a vehicle provided in an embodiment of the present application;
[0042] Figure 2 A schematic diagram of a vehicle automatic parking scenario provided in an embodiment of the present application;
[0043] Figure 3 A schematic diagram of the structure of an automatic parking device for a vehicle provided in an embodiment of the present application;
[0044] Figure 4 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0045] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0046] Automatic parking technologies have been gradually introduced on the market, which mainly rely on sensors such as car radars and cameras to identify empty parking spaces and assist vehicles in parking. However, traditional automatic parking technologies are often limited to the identification and parking operations of a single parking space, resulting in certain limitations and error rates in the automatic parking operations of vehicles, which in turn affects the utilization rate of parking lots.
[0047] Based on this, this application proposes the following technical solutions, see below for details:
[0048] In some embodiments, the vehicle automatic parking method provided by the present application can be applied to different types of target vehicles, each of which is provided with a data acquisition module, a vehicle control system and a data transmission module, and can be communicatively connected to a host computer. Among them, each target vehicle can collect relevant data such as the target vehicle and the surrounding environment in real time through the data acquisition module, and transmit the collected data to the host computer through the data transmission module, so that the host computer automatically plans the target parking space and the optimal path of the parking lot according to the collected data and sends it to the vehicle control system, and finally the vehicle control system can control the target vehicle to drive to the target parking space according to the optimal path for automatic parking.
[0049] Specifically, the data acquisition module set in the target vehicle may include direct acquisition and indirect acquisition, such as direct data acquisition through millimeter wave radars and vehicle-mounted cameras mounted on the target vehicle, and indirect data acquisition through remote acquisition tools such as remote sensing satellite systems. The host computer may include modules such as data twin models, databases, data processing modules, image processing modules, and data conversion modules. The data collected by the data acquisition module can be stored in the database, and data processing operations can also be performed through modules such as data processing modules, image processing modules, and data conversion modules, as well as input into the data twin model to perform parking simulation of the target vehicle, so as to plan the target parking space and optimal path of the planned parking lot.
[0050] In one embodiment, Figure 1 As shown, Figure 1 A schematic diagram of a process flow of a vehicle automatic parking method provided in an embodiment of the present application; the present application also provides a vehicle automatic parking method, which specifically includes the following:
[0051] S110: Acquire the positioning information of the target vehicle, and collect the real-time parking data of the target vehicle based on the positioning information.
[0052] In this step, when a target vehicle arrives at a parking place, the vehicle control system can obtain the positioning information of the target vehicle, so as to determine the location of the parking place based on the positioning information, and then collect the real-time parking data of the target vehicle in the parking place. The real-time parking data here may include vehicle operation data, ground remote sensing data, surrounding environment data and other data, which are used to provide global information of the target vehicle and the parking place.
[0053] It can be understood that the target vehicle of the present application is pre-installed with a positioning device, such as a GPS locator, so when the target vehicle arrives at the parking place and starts the automatic parking mode, the vehicle control system can obtain the positioning information of the target vehicle in real time through the positioning device, and then based on the positioning information, the real-time parking data of the target vehicle can be collected through the data acquisition module.
[0054] The positioning information may include the real-time position and direction of the target vehicle in the parking space, so the real-time parking data collected under different positioning information will be different. Before the target vehicle starts parking, when the positioning information changes, the vehicle control system will re-collect the real-time parking data of the target vehicle to improve the accuracy of subsequent automatic parking.
[0055] Specifically, since parking lots include many different types of parking environments, for example, there are well-managed closed environments and inadequately planned open-air environments, the data that can be collected in each type of parking environment is also different. Therefore, in the face of different parking environments, this application can take different data collection measures according to the complexity of the parking environment to collect the real-time parking data of the target vehicle, so as to solve the optimal solution for automatic parking based on the real-time parking data, and then realize the automatic parking of the target vehicle in parking lots with different parking environments, so as to reduce the limitations of automatic parking.
[0056] S120: Determine a digital twin model of the target vehicle, and use the digital twin model to perform path planning based on real-time parking data to obtain a target parking space and an optimal path to the target parking space.
[0057] In this step, after collecting the real-time parking data of the target vehicle through step S110, the vehicle control system can also determine the digital twin model of the target vehicle, so that the digital twin model can be used to perform path planning based on the real-time parking data to obtain the target parking space and the optimal path to the target parking space, so as to reduce the degree of vehicle congestion in complex parking environments and improve the efficiency of automatic parking.
[0058] Among them, the digital twin model refers to the virtual model of the target vehicle in a digital environment. It is obtained by integrating multiple data sources, such as vehicle basic performance data, vehicle operation data, vehicle operating environment data, etc., and then using machine learning and other data analysis technologies for model training and parameter updates. It can accurately reflect and simulate the status, behavior and performance of the target vehicle in the actual parking scenario in a digital environment.
[0059] Therefore, before controlling the target vehicle to perform automatic parking, the vehicle control system can first input the collected real-time parking data into the digital twin model of the target vehicle, so that the digital twin model can simulate the parking operation of the target vehicle in the parking space in a digital environment, and then perform route planning based on the simulation results to determine the most suitable target parking space and the optimal path between the target vehicle and the target parking space. Through the parking simulation of the digital twin model, this application can avoid the situation where the analysis results are incomplete due to the limitations of data collection, resulting in wrong parking decisions, and even being blocked by subsequent vehicles, resulting in a dilemma for the vehicle.
[0060] It should be noted that since the basic performance of different vehicles is different, when facing different target vehicles, the present application can pre-acquire the training data of each target vehicle, and then train the digital twin model corresponding to each target vehicle and store it in the host computer connected to the vehicle control system. When the target vehicle needs to perform automatic parking, the vehicle control system can directly obtain the corresponding digital twin model from the host computer and perform parking simulation.
[0061] S130: Controlling the target vehicle to go to the target parking space to perform a parking operation according to the optimal path, and obtaining the vehicle parking position of the target vehicle in the target parking space after completing the parking operation.
[0062] In this step, after obtaining the target parking space and the optimal path to the target parking space through step S120, the vehicle control system can control the target vehicle to go to the target parking space for parking according to the optimal path, and obtain the vehicle parking position of the target vehicle in the target parking space after completing the parking operation, so as to determine whether the target vehicle is successfully parked through the vehicle parking position.
[0063] Specifically, when the target vehicle arrives at the target parking space and completes the parking operation, in order to prevent the target vehicle from being parked improperly, the vehicle control system can also measure the parking position of the target vehicle in the target parking space through on-board tools such as radar or sensors, so as to judge whether the target vehicle is parked properly based on the parking position of the vehicle, and then determine whether adjustments are needed to avoid affecting the entry, exit and parking of other vehicles.
[0064] Furthermore, the vehicle control system can also perform real-time detection of the optimal path and the target parking space while controlling the target vehicle to move to the target parking space. If it is detected that there are driving obstacles on the optimal path and / or the target parking space is preempted by other vehicles, the vehicle control system can re-collect the real-time parking data of the target vehicle and re-plan the path based on the new real-time parking data.
[0065] S140: Perform safety distance detection on the parking position of the vehicle, and fine-tune the parking position of the vehicle in the target parking space according to the detection result to complete the parking action of the target vehicle.
[0066] In this step, after obtaining the parking position of the target vehicle through step S130, the vehicle control system can perform a safety distance detection on the parking position of the vehicle, and fine-tune the parking position of the vehicle in the target parking space according to the detection result to complete the parking action of the target vehicle. This can improve the accuracy of the parking operation of the target vehicle, facilitate the smooth passage of subsequent vehicles, and improve the utilization rate of the parking lot.
[0067] It is understandable that the vehicle control system can perform a safety distance detection based on the parking position of the vehicle and the target parking space to determine whether the target vehicle has completed the parking action according to the detection result. For example, when the detection result passes, such as the parking position of the target vehicle is located in the middle of the target parking space, it can be regarded as the parking standard of the target vehicle, and no position adjustment is required, which means that the target vehicle has completed the parking action; when the detection result fails, such as the line is pressed, the angle is skewed, etc., it can be regarded as the parking of the target vehicle is not standardized, so it is still necessary to fine-tune the position of the target vehicle in the target parking space until it is parked in accordance with the standard.
[0068] Furthermore, when the vehicle control system fine-tunes the parking position of the vehicle in the target parking space based on the detection results, it can collect relevant data of the target vehicle in the target parking space in real time, and input the collected data into the digital twin model of the target vehicle, so that the digital twin can simulate operations based on the data to obtain the optimal adjustment method, such as forward distance, backward distance, steering angle, etc.
[0069] In the above embodiment, when a target vehicle arrives at the parking lot, the positioning information of the target vehicle can be obtained, and the real-time parking data of the target vehicle can be collected based on the positioning information. Since the real-time parking data can provide global information of the target vehicle and its surrounding environment, the digital twin model of the target vehicle can be determined, and the digital twin model is used to perform path planning based on the real-time parking data to obtain the target parking space and the optimal path to the target parking space, so as to map and predict the dynamic changes of the parking lot, provide comprehensive parking decision support for the vehicle, and reduce the limitations of automatic parking; then the target vehicle can be controlled to go to the target parking space for parking according to the optimal path, and the vehicle parking position of the target vehicle in the target parking space is obtained after the parking operation is completed, and then the safety distance detection of the vehicle parking position is performed, and the vehicle parking position in the target parking space is fine-tuned according to the detection result, so as to improve the accuracy of the parking operation of the target vehicle, so as to facilitate the smooth passage of subsequent vehicles and improve the utilization rate of the parking lot. This application uses a digital twin model to perform real-time perception of the global information of the parking lot and parking analysis, and then controls the target vehicle to perform precise parking, thereby reducing the limitations and error rate of automatic parking and improving the utilization rate of the parking lot.
[0070] In one embodiment, collecting the real-time parking data of the target vehicle based on the positioning information in step S110 may include:
[0071] S111: Determine a parking place for the target vehicle according to the positioning information; the parking place includes an indoor parking lot and an outdoor parking lot.
[0072] S112: When the parking lot is an indoor parking lot, the environmental information of the location of the target vehicle is collected by the detection sensor as the real-time parking data of the target vehicle.
[0073] S113: When the parking place is an outdoor parking lot, the environmental information of the location of the target vehicle is collected by a detection sensor, and a remote sensing image of the outdoor parking lot is collected by a remote sensing satellite, and real-time parking data of the target vehicle is formed based on the environmental information and the remote sensing image.
[0074] In this embodiment, after obtaining the positioning information of the target vehicle, the vehicle control system can determine the parking place where the target vehicle is located according to the positioning information, and the parking place here can include indoor parking lots and outdoor parking lots. Since the layout of indoor parking lots and outdoor parking lots are different and the complexity of the environment is also different, the application can adopt corresponding data collection measures according to the type of parking place to collect real-time parking data of the target vehicle.
[0075] It is understandable that since indoor parking lots are closed and well-managed, and parking and driving areas are pre-planned, it is relatively convenient for the target vehicle to park and drive out of indoor parking lots. Therefore, the vehicle control system can only collect environmental information about the location of the target vehicle, and then plan the route based on the environmental information according to the planning instructions of the indoor parking lot. In some outdoor parking lots, such as open-air parking lots and street parking spaces that are not clearly planned, these areas often lack clear parking instructions, resulting in random parking of vehicles and causing traffic congestion. Therefore, after collecting environmental information about the location of the target vehicle, the vehicle control system can also collect remote sensing images of the parking space to comprehensively analyze the global distribution of the parking space, and then plan the route for the parking action of the target vehicle.
[0076] Specifically, when the parking place is an indoor parking lot, the vehicle control system can collect environmental information of the target vehicle's location through detection sensors as real-time parking data of the target vehicle; when the parking place is an outdoor parking lot, the vehicle control system can collect environmental information of the target vehicle's location through detection sensors, and use remote sensing satellites to collect remote sensing images of the outdoor parking lot, and form real-time parking data of the target vehicle based on the environmental information and remote sensing images.
[0077] Among them, the environmental information of this application refers to the data of physical and dynamic conditions directly related to the location of the target vehicle, which are usually collected by detection sensors, radars or cameras installed on the vehicle. For example, the target vehicle equipment can use detection sensors to collect information on the road conditions of the parking lot, or use a vehicle-mounted camera to collect images of the environment around the target vehicle, and can also use a vehicle-mounted millimeter-wave radar to collect location information of vehicles, obstacles and other objects around the target vehicle. Remote sensing images refer to image data of outdoor parking lots and their surrounding environments acquired from a long distance by remote sensing satellites, drones or other remote sensors. These data usually have higher spatial resolution and temporal resolution, and can provide a comprehensive view of the parking lot and its surrounding environment.
[0078] Furthermore, after acquiring the remote sensing image of the outdoor parking lot, the vehicle control system can first perform noise reduction and smoothing on the remote sensing image. Noise reduction here refers to the process of reducing or eliminating noise in the image, aiming to improve the clarity and signal-to-noise ratio of the image, while smoothing refers to the process of reducing the high-frequency components in the image and enhancing the low-frequency components to improve the visual effect of the image. Then, the vehicle control system can also perform image correction and restoration on the processed remote sensing image, accurately locate the remote sensing data in the image with the geographic coordinates by registering the remote sensing image, and correct the deformed image data so that the image is restored and has accurate geographic coordinates. Finally, the vehicle control system can also perform enhancement processing on the remote sensing image, including contrast transformation, color transformation and image operation, to highlight the vehicles and parking spaces in the remote sensing image, distinguish them from the background environment, highlight the key information in the remote sensing image, and thus improve the accuracy of subsequent global image analysis.
[0079] In one embodiment, determining the digital twin model of the target vehicle in step S120 may include:
[0080] S121: Acquire historical operation data and historical parking data of the target vehicle, and perform preprocessing operations on the historical operation data and historical parking data to obtain training data.
[0081] S122: Iteratively train the preset pre-trained model using the training data, and optimize the parameters of the pre-trained model using an optimization algorithm during the training process.
[0082] S123: When the pre-trained model meets the preset training end condition, the pre-trained model that has been trained is used as the digital twin model of the target vehicle.
[0083] In this embodiment, when constructing the digital twin model of the target vehicle, the application may first obtain the historical operation data and historical parking data of the target vehicle, and perform preprocessing operations on the historical operation data and historical parking data to obtain training data, and then use the training data to iteratively train the preset pre-trained model, so as to use the optimization algorithm to optimize the parameters of the pre-trained model during the training process. When the pre-trained model meets the preset training end condition, the pre-trained model that has been trained is used as the digital twin model of the target vehicle.
[0084] Among them, historical operation data refers to various data during the operation of the target vehicle within the historical time period, including but not limited to speed information, energy consumption, engine status, etc.; and historical parking data refers to the data recorded when the target vehicle was parked within the historical time period, including parking location, parking environment, parking time, parking mode, etc.
[0085] It should be noted that since the data types collected by multiple different acquisition devices in the historical parking data are different, such as the point cloud data collected by the on-board millimeter-wave radar and the image data collected by the on-board camera, the present application can fuse the point cloud data and the image data according to the bounding box and features before pre-processing the historical parking data to obtain the scene data of the parking scene. Finally, the remote sensing images in the historical parking data and the fused scene data and other images can be converted into a data format that can be recognized by the digital twin model.
[0086] It is understood that the preprocessing operations of the present application include but are not limited to denoising, data cleaning, feature extraction and dimensionality reduction, and data normalization. Among them, denoising refers to removing noise and interference in the data to improve the quality and availability of the data; data cleaning refers to removing outliers, supplementing missing values, and smoothing the data to improve the accuracy and completeness of the data; feature extraction and dimensionality reduction refer to extracting effective information from the data and converting the extracted effective information into a low-dimensional vectorized form for subsequent mathematical operations and analysis; and data normalization refers to normalizing the data to ensure that the weights of data with different features are the same to avoid data bias and errors.
[0087] Specifically, the present application can obtain the initial digital model pre-built by the digital modeling tool as the pre-trained model of the target vehicle, and divide the training data into a training set, a validation set, and a test set. Among them, the training set can be used to iteratively train the model and minimize the loss function by adjusting the model parameters; the validation set can be used to verify the trained model to evaluate the generalization ability of the model and adjust the parameters of the model according to actual needs; the test set can be used to test the verified model to ensure the performance of the model in practical applications. The digital twin model trained in this way can simulate and predict the actual operation of the target vehicle in various parking scenarios to provide comprehensive parking decision support.
[0088] In one embodiment, in step S120, the digital twin model is used to perform path planning based on real-time parking data to obtain a target parking space and an optimal path to the target parking space, which may include:
[0089] S124: Using the digital twin model to perform parking simulation on the target vehicle based on the real-time parking data, and outputting parking scores for multiple vacant parking spaces.
[0090] S125: sorting the parking scores from low to high, and taking the vacant parking space with the lowest parking score in the sorting result as the target parking space for the target vehicle.
[0091] S126: extracting the location information of the target parking space, and planning an optimal path for the target vehicle based on the location information and positioning information.
[0092] In this embodiment, when planning the path for the target vehicle, the vehicle control system can first use the digital twin model to perform a parking simulation on the target vehicle based on real-time parking data, and output the parking scores of multiple vacant parking spaces, and then sort the parking scores in order from low to high, and use the vacant parking space with the lowest parking score in the sorting result as the target parking space for the target vehicle. Finally, the vehicle control system can extract the location information of the target parking space, and plan the optimal path for the target vehicle based on the location information and positioning information.
[0093] It is understood that the parking score of this application refers to the comprehensive score of the impact of various factors in the process of the target vehicle going to each parking space. Therefore, the higher the parking score, the more obstacles the target vehicle encounters in the process of going to the corresponding parking space, and the lower the smoothness; the lower the parking score, the fewer obstacles the target vehicle encounters in the process of going to the corresponding parking space, and the higher the smoothness.
[0094] Specifically, the vehicle control system can globally perceive and analyze the real-time parking data to determine the influencing factors in the parking space based on the analysis results. Then, the vehicle control system can use the digital twin model to simulate parking for each free parking space, thereby obtaining the parking score of each free parking space. Finally, the vehicle control system uses the free parking space with the lowest parking score as the target parking space for the target vehicle, and then plans the optimal path for the target vehicle based on the location information and positioning information of the target parking space, so that the target vehicle can park in the most convenient way, avoiding the dilemma of the target vehicle during parking due to wrong decisions, thereby improving the efficiency of automatic parking.
[0095] In one embodiment, in step S124, using the digital twin model to perform parking simulation on the target vehicle based on real-time parking data and outputting parking scores of multiple vacant parking spaces may include:
[0096] S1241: Input the real-time parking data into the digital twin model, so that the digital twin model determines multiple vacant parking spaces according to the real-time parking data, and simulates the parking operation of the target vehicle in each vacant parking space based on the real-time parking data, obtains the influence coefficient corresponding to each parking influencing factor in each vacant parking space, and determines the parking score of each vacant parking space according to the influence coefficient corresponding to each vacant parking space.
[0097] In this embodiment, when performing parking simulation, the digital twin model can first perform a global analysis of the real-time parking data to determine multiple parking influencing factors in the parking lot and multiple vacant parking spaces that have not been parked according to the analysis results. For each vacant parking space, the digital twin model can simulate the parking operation of the target vehicle in the vacant parking space based on the real-time parking data, thereby obtaining the influence coefficient corresponding to each parking influencing factor in the vacant parking space. Finally, the digital twin model can determine the parking score of each vacant parking space according to the influence coefficient corresponding to each vacant parking space.
[0098] Among them, the parking influencing factors of a parking lot may include the number of entrances and exits, the probability of being blocked, the distance from the entrances and exits, the difficulty of parking, etc. Different parking lots have different corresponding parking influencing factors, which are not limited here.
[0099] Specifically, the calculation formula of the parking score can be expressed as follows:
[0100]
[0101] In the formula, represents the parking score of the jth free parking space; n represents the total number of parking influencing factors in the parking lot; represents the influence weight of the i-th parking influencing factor; Represents the influence coefficient of the i-th parking influencing factor.
[0102] Indicatively, Figure 2 As shown, Figure 2 A schematic diagram of a vehicle automatic parking scenario provided in an embodiment of the present application; Figure 2 In the order of parking spaces from left to right, the first row is X1~X6, the second row is Y1~Y8, and the third row is Z1~Z7. The first row of free parking spaces is 3, the second row is 5, and the third row is 5, for a total of 13 free parking spaces. Figure 2 In the parking lot shown, there are four parking influencing factors, namely the number of entrances and exits, the probability of being blocked, the distance from the entrances and exits, and the difficulty of parking. The influence weights of each parking influencing factor are 20%, 25%, 30%, and 25%, respectively; the factor weights here can be set according to the training situation of the digital twin model or the environment of each parking lot, and there is no restriction here.
[0103] To expand on this, Figure 2 In the parking scenario shown, for the number of entrances and exits, all the free parking spaces are 2. If the second row of parking spaces is full, then the number of entrances and exits for the first row of free parking spaces is 1. For the probability of being blocked, the free parking spaces in the first row are more likely to be blocked as they move to the right. The probability of the two free parking spaces in the middle of the second row being blocked is greater than that on the left and right sides. The probability of the free parking spaces in the middle of the third row being blocked is greater than that on the left and right sides, and the probability decreases as they move to the sides. For the distance from the entrances and exits, the distances between the free parking spaces and the two entrances and exits can be calculated, and the shortest distance can be taken as the calculation value. When the parking lot specifies the exits and entrances, only the distance from the free parking spaces to the exits is calculated, and then one free parking space is used as the minimum unit of distance when calculating. Regarding the difficulty of parking, when there are no vehicles on both sides of the vacant parking space, the parking impact coefficient is 0.5; when there is a vehicle on one side of the vacant parking space and the other side is an empty parking space, the parking impact coefficient is 0.75; when there are vehicles on both sides of the parking space, the parking impact coefficient is 1.2; when there is a building on one side of the vacant parking space and an empty parking space on the other side, the parking impact coefficient is 0.9; when there is a building on one side of the vacant parking space and a vehicle parked on the other side, the parking impact coefficient is 1.5; when there are buildings on both sides of the vacant parking space, the parking impact coefficient is 1.8. Figure 2 In the figure, the influence coefficients of the four parking influencing factors of Y1 are 2, 2 / 12, 1, and 0.5 respectively, and its parking score is 0.87, which is the vacant parking space with the lowest parking score in the parking lot. Therefore, it can be used as the target parking space for the target vehicle.
[0104] In one embodiment, the safety distance detection of the parking position of the vehicle in step S140 may include:
[0105] S141: Collect the parking space line of the target parking space, and determine the parking distance of the target vehicle according to the parking space line and the parking position of the vehicle.
[0106] S142: Match the parking distance of the parking space with a preset safety distance range, and use the matching result as the detection result of the safety distance detection of the target vehicle.
[0107] In this embodiment, when the vehicle control system performs safety distance detection on the parking position of the vehicle, it can first collect the parking space lines of the target parking space, and determine the parking distance of the target vehicle based on the parking space lines and the parking position of the vehicle. Then, the parking distance can be matched with the preset safety distance range, and the matching result can be used as the detection result of the safety distance detection of the target vehicle.
[0108] Specifically, the vehicle control system can use acquisition equipment, such as on-board cameras, radars, etc. to capture data of the target parking space, and then use edge detection algorithms, color recognition algorithms, or shape matching algorithms to identify parking space lines in the collected data. After identifying the parking space line of the target parking space, the vehicle control system can measure the relative distance between the parking space line and the parking position of the vehicle, that is, the parking distance of the parking space, and match the parking distance of the parking space with a preset safety distance range to determine whether the target vehicle is accurately parked in the target parking space based on the matching result.
[0109] The safety distance range of the present application can be determined based on multiple dimensions such as the size of the target vehicle, the size of the target parking space, the accuracy of the acquisition equipment, and safety standards. The safety distance range usually includes the minimum and maximum allowable distances in the four directions of front, back, left, and right. If the distances in all directions are within the safety distance range, the vehicle control system can determine that the target vehicle is parked accurately; otherwise, the vehicle control system needs to readjust the position of the target vehicle in the target parking space.
[0110] In one embodiment, fine-tuning the parking position of the vehicle in the target parking space according to the detection result in step S140 may include:
[0111] S143: When the detection result is failure, the parking space environment information of the target vehicle is collected, and the fine-tuning action of the target vehicle is determined based on the parking space environment information through the digital twin model, and after the target vehicle is controlled to perform a parking operation in the target parking space according to the fine-tuning action, it returns to execute the vehicle parking position of the target vehicle in the target parking space after completing the parking operation and its subsequent steps.
[0112] S144: When the detection result is passed, confirm that the parking action of the target vehicle is completed.
[0113] In this embodiment, when the detection result fails, the vehicle control system needs to continue to collect parking environment information of the target vehicle in the target parking space, and input the parking environment information into the digital twin model, so that the digital twin model can perform fine-tuning simulation based on the parking environment information, and output corresponding fine-tuning actions, including command actions such as turning angle, forward distance or backward distance. In this way, the vehicle control system can directly control the target vehicle to fine-tune the position in the target parking space according to the fine-tuning action, and re-detect the safety distance of the target vehicle after the fine-tuning is completed. Until the detection result is passed, the vehicle control system can confirm that the target vehicle has completed the parking action.
[0114] The following is a description of the vehicle automatic parking device provided in an embodiment of the present application. The vehicle automatic parking device described below and the vehicle automatic parking method described above can be referenced to each other.
[0115] In one embodiment, Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of an automatic parking device for a vehicle provided in an embodiment of the present application; the present application also provides an automatic parking device for a vehicle, including a data acquisition module 210, a path planning module 220, a parking operation module 230 and a position fine-tuning module 240, which specifically include the following:
[0116] The data acquisition module 210 is used to obtain the positioning information of the target vehicle and collect the real-time parking data of the target vehicle based on the positioning information.
[0117] The path planning module 220 is used to determine the digital twin model of the target vehicle, and use the digital twin model to perform path planning based on real-time parking data to obtain the target parking space and the optimal path to the target parking space.
[0118] The parking operation module 230 is used to control the target vehicle to go to the target parking space to perform the parking operation according to the optimal path, and obtain the vehicle parking position of the target vehicle in the target parking space after completing the parking operation.
[0119] The position fine-tuning module 240 is used to detect the safe distance of the parking position of the vehicle and fine-tune the parking position of the vehicle in the target parking space according to the detection result to complete the parking action of the target vehicle.
[0120] In the above embodiment, when a target vehicle arrives at the parking lot, the positioning information of the target vehicle can be obtained, and the real-time parking data of the target vehicle can be collected based on the positioning information. Since the real-time parking data can provide global information of the target vehicle and its surrounding environment, the digital twin model of the target vehicle can be determined, and the digital twin model is used to perform path planning based on the real-time parking data to obtain the target parking space and the optimal path to the target parking space, so as to map and predict the dynamic changes of the parking lot, provide comprehensive parking decision support for the vehicle, and reduce the limitations of automatic parking; then the target vehicle can be controlled to go to the target parking space for parking according to the optimal path, and the vehicle parking position of the target vehicle in the target parking space is obtained after the parking operation is completed, and then the safety distance detection of the vehicle parking position is performed, and the vehicle parking position in the target parking space is fine-tuned according to the detection result, so as to improve the accuracy of the parking operation of the target vehicle, so as to facilitate the smooth passage of subsequent vehicles and improve the utilization rate of the parking lot. This application uses a digital twin model to perform real-time perception of the global information of the parking lot and parking analysis, and then controls the target vehicle to perform precise parking, thereby reducing the limitations and error rate of automatic parking and improving the utilization rate of the parking lot.
[0121] In one embodiment, the data acquisition module 210 may include:
[0122] The parking place determination submodule is used to determine the parking place of the target vehicle according to the positioning information; the parking place includes indoor parking lots and outdoor parking lots.
[0123] The first acquisition submodule is used to acquire environmental information of the location of the target vehicle through detection sensors when the parking place is an indoor parking lot, as real-time parking data of the target vehicle.
[0124] The second acquisition submodule is used to obtain environmental information of the location of the target vehicle through detection sensors when the parking place is an outdoor parking lot, and to obtain remote sensing images of the outdoor parking lot using remote sensing satellites, and to form real-time parking data of the target vehicle based on the environmental information and remote sensing images.
[0125] In one embodiment, the path planning module 220 may include:
[0126] The training data acquisition submodule is used to acquire the historical operation data and historical parking data of the target vehicle, and perform preprocessing operations on the historical operation data and historical parking data to obtain training data.
[0127] The model training submodule is used to iteratively train the preset pre-trained model using the training data, so as to optimize the parameters of the pre-trained model using an optimization algorithm during the training process.
[0128] The model generation submodule is used to use the pre-trained model as the digital twin model of the target vehicle when the pre-trained model meets the preset training end conditions.
[0129] In one embodiment, the path planning module 220 may further include:
[0130] The parking simulation submodule is used to use the digital twin model to perform parking simulation on the target vehicle based on real-time parking data, and output the parking scores of multiple vacant parking spaces.
[0131] The target parking space selection submodule is used to sort the parking scores from low to high, and select the vacant parking space with the lowest parking score in the sorting result as the target parking space for the target vehicle.
[0132] The optimal path planning submodule is used to extract the location information of the target parking space and plan the optimal path for the target vehicle based on the location information and positioning information.
[0133] In one embodiment, in step S124, using the digital twin model to perform parking simulation on the target vehicle based on real-time parking data and outputting parking scores of multiple vacant parking spaces may include:
[0134] A parking score determination unit is used to input real-time parking data into the digital twin model so that the digital twin model determines multiple vacant parking spaces according to the real-time parking data, and simulates the parking operation of the target vehicle in each vacant parking space based on the real-time parking data, obtains the influence coefficient corresponding to each parking influencing factor in each vacant parking space, and determines the parking score of each vacant parking space according to the influence coefficient corresponding to each vacant parking space.
[0135] In one embodiment, the position fine-tuning module 240 may include:
[0136] The distance determination submodule is used to collect the parking space lines of the target parking space and determine the parking distance of the target vehicle according to the parking space lines and the parking position of the vehicle.
[0137] The safety distance matching submodule is used to match the parking distance of the parking space with a preset safety distance range, and use the matching result as the detection result of the safety distance detection of the target vehicle.
[0138] In one embodiment, the position fine-tuning module 240 may further include:
[0139] The parking fine-tuning submodule is used to collect the parking space environment information of the target vehicle when the detection result is failure, and determine the fine-tuning action of the target vehicle based on the parking space environment information through the digital twin model, and after controlling the target vehicle to perform a parking operation in the target parking space according to the fine-tuning action, return to execute the vehicle parking position of the target vehicle in the target parking space after completing the parking operation and its subsequent steps.
[0140] The parking completion submodule is used to confirm that the parking action of the target vehicle is completed when the detection result is passed.
[0141] In one embodiment, the present application further provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the vehicle automatic parking method as described in any one of the above embodiments.
[0142] In one embodiment, the present application further provides a computer device having computer-readable instructions stored therein. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the vehicle automatic parking method as described in any one of the above embodiments.
[0143] Indicatively, Figure 4 As shown, Figure 4 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 may be provided as a server. Figure 4 The computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301, for storing instructions executable by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the vehicle automatic parking method of any of the above embodiments.
[0144] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server TM, Mac OS X TM, Unix TM, Linux TM, Free BSD TM, or the like.
[0145] Those skilled in the art will understand that Figure 4The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0146] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0147] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.
[0148] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A vehicle automatic parking method, characterized in that: The method comprises: Acquire the positioning information of the target vehicle, and collect the real-time parking data of the target vehicle based on the positioning information; Determine a digital twin model of the target vehicle, and use the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space; Controlling the target vehicle to go to the target parking space to perform a parking operation according to the optimal path, and obtaining a vehicle parking position of the target vehicle in the target parking space after completing the parking operation; Performing a safety distance detection on the parking position of the vehicle, and fine-tuning the parking position of the vehicle in the target parking space according to the detection result to complete the parking action of the target vehicle; Wherein, fine-tuning the parking position of the vehicle in the target parking space according to the detection result includes: When the detection result is failure, the parking space environment information of the target vehicle is collected, and the fine-tuning action of the target vehicle is determined based on the parking space environment information through the digital twin model, and after the target vehicle is controlled to perform a parking operation in the target parking space according to the fine-tuning action, the process returns to execute the step of obtaining the vehicle parking position of the target vehicle in the target parking space after completing the parking operation and subsequent steps; When the detection result is passed, it is confirmed that the parking action of the target vehicle is completed.
2. The vehicle automatic parking method according to claim 1, characterized in that: The collecting of the real-time parking data of the target vehicle based on the positioning information includes: Determine a parking place for the target vehicle according to the positioning information; the parking place includes an indoor parking lot and an outdoor parking lot; When the parking lot is an indoor parking lot, the environmental information of the location of the target vehicle is collected by a detection sensor as the real-time parking data of the target vehicle; When the parking place is an outdoor parking lot, the environmental information of the location of the target vehicle is collected by the detection sensor, and a remote sensing image of the outdoor parking lot is collected by a remote sensing satellite, and real-time parking data of the target vehicle is formed based on the environmental information and the remote sensing image.
3. The vehicle automatic parking method according to claim 1, characterized in that: Determining the digital twin model of the target vehicle includes: Acquiring historical operation data and historical parking data of the target vehicle, and performing preprocessing operations on the historical operation data and the historical parking data to obtain training data; Iteratively training a preset pre-trained model using the training data, so as to optimize the parameters of the pre-trained model using an optimization algorithm during the training process; When the pre-trained model meets the preset training end condition, the pre-trained model that has completed training is used as the digital twin model of the target vehicle.
4. The vehicle automatic parking method according to claim 1, characterized in that: The using the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space includes: Using the digital twin model to perform parking simulation on the target vehicle based on the real-time parking data, and outputting parking scores of multiple vacant parking spaces; Sorting the parking scores from low to high, and taking the vacant parking space with the lowest parking score in the sorting result as the target parking space for the target vehicle; The position information of the target parking space is extracted, and the optimal path of the target vehicle is planned according to the position information and the positioning information.
5. The vehicle automatic parking method according to claim 4, characterized in that: The using the digital twin model to perform parking simulation on the target vehicle based on the real-time parking data and outputting parking scores of multiple vacant parking spaces includes: The real-time parking data is input into the digital twin model, so that the digital twin model determines a plurality of vacant parking spaces according to the real-time parking data, and simulates the parking operation of the target vehicle in each vacant parking space based on the real-time parking data, obtains the influence coefficient corresponding to each parking influencing factor in each vacant parking space, and determines the parking score of each vacant parking space according to the influence coefficient corresponding to each vacant parking space.
6. The vehicle automatic parking method according to claim 1, characterized in that: The detecting of the safe distance of the parking position of the vehicle comprises: Collecting the parking space line of the target parking space, and determining the parking distance of the target vehicle according to the parking space line and the parking position of the vehicle; The parking space parking distance is matched with a preset safety distance range, and the matching result is used as the detection result of the safety distance detection of the target vehicle.
7. A vehicle automatic parking device, characterized in that: include: A data acquisition module, used to acquire the positioning information of the target vehicle, and collect the real-time parking data of the target vehicle based on the positioning information; A path planning module, used to determine a digital twin model of the target vehicle, and use the digital twin model to perform path planning based on the real-time parking data to obtain a target parking space and an optimal path to the target parking space; a parking operation module, configured to control the target vehicle to go to the target parking space to perform a parking operation according to the optimal path, and obtain a vehicle parking position of the target vehicle in the target parking space after completing the parking operation; A position fine-tuning module, used to detect the safe distance of the parking position of the vehicle, and fine-tune the parking position of the vehicle in the target parking space according to the detection result, so as to complete the parking action of the target vehicle; Among them, the position fine-tuning module includes: When the detection result is failure, the parking space environment information of the target vehicle is collected, and the fine-tuning action of the target vehicle is determined based on the parking space environment information through the digital twin model, and after the target vehicle is controlled to perform a parking operation in the target parking space according to the fine-tuning action, the process returns to execute the step of obtaining the vehicle parking position of the target vehicle in the target parking space after completing the parking operation and subsequent steps; When the detection result is passed, it is confirmed that the parking action of the target vehicle is completed.
8. A storage medium, characterized in that: The storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the vehicle automatic parking method as described in any one of claims 1 to 6.
9. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the vehicle automatic parking method as claimed in any one of claims 1 to 6 are performed.
Citation Information
Patent Citations
Management and control method of hybrid electric vehicle based on digital twinning technology
CN110488629A
Autonomous parking system and method based on visual synchronous positioning and mapping
CN117218892A
Automatic parking method and device, vehicle and storage medium
CN118004154A
Automatic driving behavior decision optimization system based on digital twinning
CN118289035A
Digital twinning-oriented intelligent cloud parking management method and management platform
CN118334900A
