Parking path planning method and path node prediction model training method

Abstract

The application provides a parking path planning method and a path node prediction model training method, and relates to the technical field of vehicle path planning. The planning method comprises the following steps: acquiring scene information of a target parking space of a to-be-parked vehicle; the scene information of the target parking space is used for describing the target parking space; inputting the scene information of the target parking space into a path node prediction model to determine parking path node information of the to-be-parked vehicle; wherein the path node prediction model is used for representing a corresponding relationship between the parking path node information and the scene information of the parking space, and the parking path node information comprises position information and attitude information of the to-be-parked vehicle at the parking path node; and planning a path for the to-be-parked vehicle to enter the target parking space according to the parking path node information. The method is suitable for the path planning process when the vehicle is parked, and is used for reducing the calculation amount and improving the parking efficiency.

Description

Technical Field

[0001] This application relates to the field of vehicle path planning technology, and in particular to a parking path planning method and a training method for a path node prediction model. Background Technology

[0002] Automatic parking assist (APA) is a driver assistance system that enables vehicles to park automatically by controlling the vehicle's acceleration, deceleration, steering angle, and forward / reverse gears.

[0003] When planning the automatic parking path, automatic parking systems typically divide the entire path planning process into multiple segments due to limitations such as perception errors, vehicle control errors, and computational load. For example, an automatic parking system might first plan the forward path, controlling the vehicle to advance to a shift position, stopping at that position, shifting the forward gear to reverse, and then planning the reverse path to control the vehicle to back into the parking space.

[0004] However, current automatic parking systems have low parking efficiency. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a parking path planning method and a training method for a path node prediction model. This planning method can utilize the parking path node information and the path node prediction model to predict parking path node information that matches the scene information of the target parking space. The parking path planned based on this parking path node information is of high quality, thereby improving the efficiency of automatic parking.

[0006] In a first aspect, this application provides an automatic parking path planning method, which includes: acquiring scene information of a target parking space for a vehicle to be parked; the scene information of the target parking space is used to describe the target parking space; inputting the scene information of the target parking space into a path node prediction model to determine the parking path node information of the vehicle to be parked; wherein, the path node prediction model is used to characterize the correspondence between the parking path node information and the scene information of the parking space, and the parking path node information includes the position information and attitude information of the vehicle to be parked at the parking path node; and planning the path for the vehicle to be parked to enter the target parking space based on the parking path node information.

[0007] The parking path planning method provided in this application can obtain scene information of the target parking space and predict the target parking path node information corresponding to the scene information of the target parking space based on the path node prediction model used to characterize the correspondence between parking path node information and scene information of the parking space. The target parking path node information and the scene information of the target parking space have a pre-trained matching relationship. The parking route planned from the position corresponding to the parking path node information and according to the posture corresponding to the parking path node information has higher path quality, thus improving the overall efficiency of vehicle parking.

[0008] Optionally, the scene information of the target parking space includes at least one of the following: boundary information of obstacles around the target parking space, boundary information of the target parking space, or final pose information of the vehicle corresponding to the target parking space; the boundary information of obstacles around the target parking space is used to describe the drivable area around the target parking space; the boundary information of the target parking space and the final pose information of the vehicle corresponding to the target parking space are used to describe the location of the target parking space.

[0009] It should be understood that the parking path planning method provided in this application can also take into account the boundaries of obstacles around the target parking space when using the path node prediction model to predict parking path node information. Different obstacle boundaries can correspond to different parking path node information and different parking paths, which enriches the types of planned parking paths and improves the versatility of the path node prediction model.

[0010] Optionally, the parking path node includes the gear shift node of the vehicle to be parked. The method further includes: controlling the vehicle to be parked to drive to the position corresponding to the parking path node information, and controlling the vehicle to be parked to be in reverse gear and in the posture corresponding to the parking path node information at that position; while in reverse gear and in the posture corresponding to the parking path node information, controlling the vehicle to be parked to drive into the target parking space according to the planned path.

[0011] Optionally, the path node prediction model is obtained by training a preset model based on a training sample set. The training sample set includes multiple training samples, each of which includes scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes the parking path node information of the vehicle.

[0012] Optionally, the label corresponding to each parking space is obtained by filtering from multiple candidate parking path node information for each parking space based on the parking planning path corresponding to each candidate parking path node information; the parking planning path corresponding to each candidate parking path node information is the path for the vehicle to enter the parking space planned based on each candidate parking path node information; the multiple candidate parking path node information for each parking space includes multiple location information within the candidate area and preset posture information of the vehicle in the parking state corresponding to each location information.

[0013] Optionally, the label for each parking space includes candidate parking path node information when the quality of the parking planning path is greater than a preset quality threshold or the highest quality. The quality of the parking planning path is determined based on at least one of the following: the number of gear shifts, path length, and path curvature during the process of driving into the parking space according to the parking planning path.

[0014] It should be understood that when obtaining labels from training samples, this application may select candidate parking path node information corresponding to parking planning paths with higher path quality as labels. The parking path node information predicted by the path node prediction model trained with the training samples with the labels will also have the characteristics of the labels, that is, the parking path corresponding to the predicted parking path node information has higher path quality (fewer gear shifts, shorter parking path and smaller parking path curvature).

[0015] Optionally, the path node prediction model is a neural network model; the neural network model includes multiple network modules and an output layer, each network module including a fully connected layer, an activation layer, and a normalization layer; the fully connected layer is used to extract features of the scene information of the target parking space; the activation layer is used to perform nonlinear processing on the features of the scene information of the target parking space extracted by the fully connected layer to obtain nonlinear features; the normalization layer is used to normalize the nonlinear features; the output layer is used to output the parking path node information of the vehicle to be parked based on the nonlinear features normalized by the normalization layer.

[0016] Secondly, this application provides a parking route planning device, which includes various modules for the method described in the first aspect above.

[0017] Thirdly, this application provides a training method for a path node prediction model. The method includes: acquiring a training sample set, which includes multiple training samples. Each training sample includes scene information of a parking space and a label corresponding to the scene information of the parking space. The label includes parking path node information during the parking process. The parking path node information includes the position information and attitude information of the vehicle at the parking path node. Based on the training sample set, a preset model is trained to obtain a path node information prediction model.

[0018] In one possible implementation, obtaining the training sample set includes: acquiring scene information for multiple parking spaces and determining candidate regions for each parking space; selecting multiple location information and preset posture information of the vehicle in a parking state corresponding to each location information from the candidate regions of each parking space, as candidate parking path node information for each parking space; planning a parking planning path corresponding to each candidate parking path node information based on each candidate parking path node information; and filtering out the label corresponding to each parking space from the multiple candidate parking path node information corresponding to each parking space based on the parking planning path corresponding to each candidate parking path node information.

[0019] Optionally, after planning the parking planning path corresponding to each candidate parking path node based on each candidate parking path node information, the method further includes: determining the quality of the parking planning path corresponding to each candidate parking path node based on at least one of the following during the process of driving into the parking space according to the parking planning path corresponding to each candidate parking path node information: the number of gear shifts, the path length, and the path curvature. Based on the parking planning path corresponding to each candidate parking path node information, a label corresponding to each parking space is obtained by filtering from multiple candidate parking path node information corresponding to each parking space, including: determining the candidate parking path node information corresponding to the case where the quality of the parking planning path is greater than a preset quality threshold or the highest quality from multiple candidate parking path node information corresponding to each parking space as the label of each parking space.

[0020] Optionally, the preset model is a neural network model; the neural network model includes multiple network modules and an output layer. Each network module includes a fully connected layer, an activation layer, and a normalization layer; the fully connected layer is used to extract features of the parking space scene information in the training samples; the activation layer is used to perform nonlinear processing on the features of the parking space scene information extracted by the fully connected layer in the training samples to obtain nonlinear features; the normalization layer is used to normalize the nonlinear features; the output layer is used to output parking path node information based on the nonlinear features normalized by the normalization layer.

[0021] Fourthly, this application provides a training apparatus for a path node prediction model, the apparatus comprising various modules for the method described in the third aspect above.

[0022] Fifthly, this application provides a computer program product that, when run on a computer, causes the computer to perform the steps of the related methods described in the first aspect above, so as to implement the methods described in the first or third aspect above.

[0023] In a sixth aspect, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, causing the electronic device to perform the method described in the first or third aspect above.

[0024] In a seventh aspect, this application provides a computer-readable storage medium comprising: computer software instructions; which, when executed in an electronic device, cause the electronic device to perform the method described in the first or third aspect above.

[0025] Eighthly, this application provides a chip including a processor and an interface. The processor is coupled to a memory through the interface. When the processor executes a computer program in the memory or an electronic device executes instructions, the methods described in the first or third aspect are executed.

[0026] The beneficial effects of aspects two through eight mentioned above can be referred to in aspect one, and will not be repeated here. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of a parallel parking scenario;

[0028] Figure 2 This is a schematic diagram illustrating a scenario of reversing into a parking space.

[0029] Figure 3 A schematic diagram of a scenario where parking is done at an angle;

[0030] Figure 4 This is a schematic diagram of the composition of the automatic parking system provided in the embodiments of this application;

[0031] Figure 5 A schematic diagram illustrating the composition of the electronic device provided in the embodiments of this application;

[0032] Figure 6 A flowchart illustrating the parking path planning method provided in this application embodiment;

[0033] Figure 7 A schematic diagram illustrating scene information of a target parking space provided in an embodiment of this application;

[0034] Figure 8 A schematic diagram illustrating scene information for another target parking space provided in an embodiment of this application;

[0035] Figure 9 A schematic diagram illustrating scene information for another target parking space provided in an embodiment of this application;

[0036] Figure 10 A flowchart illustrating the training method of the path node prediction model provided in this application embodiment;

[0037] Figure 11 A schematic diagram of a candidate region provided in an embodiment of this application;

[0038] Figure 12 This is a schematic diagram of another candidate region provided in an embodiment of this application;

[0039] Figure 13 This is another schematic diagram of a candidate region provided in an embodiment of this application;

[0040] Figure 14 A schematic diagram illustrating a method for obtaining labels for training samples provided in an embodiment of this application;

[0041] Figure 15 This is a schematic diagram of the composition of the parking path planning device provided in the embodiments of this application;

[0042] Figure 16 This is a schematic diagram of the composition of the training device for the path node prediction model provided in the embodiments of this application. Detailed Implementation

[0043] Currently, when planning parking routes, automatic parking systems typically divide the entire path planning process into multiple segments due to limitations such as perception errors, vehicle control errors, and computational load. However, current automatic parking systems are relatively inefficient during automatic parking.

[0044] Based on this, embodiments of this application provide a parking path planning method, apparatus, device, and storage medium, which can use a trained path node prediction model to predict parking space map scenarios, obtain target shift nodes adapted to the parking space map scenarios, improve the quality of parking paths, and thus improve parking efficiency.

[0045] Figure 1 This is a schematic diagram of a parallel parking scenario. Figure 1 As shown, taking a parallel parking space next to a non-motorized vehicle lane as an example, the vehicle waiting to park can drive in the non-motorized vehicle lane using a forward gear to a certain shift position, and then shift from forward to reverse gear at that shift position, finally reversing into the target parking space from that shift position. This position can also be understood as the (starting) entry position for parallel parking, or the position where the vehicle begins to park in the target parking space. Figure 1 The black arrow in the image illustrates the movement trajectory of a vehicle waiting to be parked.

[0046] Figure 2 This is a schematic diagram illustrating a scenario of reversing into a parking space. (Example:) Figure 2As shown, taking a parking space in a garage as an example, a vehicle waiting to be parked can drive in the garage aisle using forward gear to a certain shift position, and then shift from forward gear to reverse gear at that shift position, finally reversing into the target parking space from that shift position. This position can also be understood as the (starting) parking position, or the position where the vehicle begins to park in the target parking space. Figure 2 The text also uses black arrows as an example to illustrate the movement trajectory of vehicles waiting to be parked.

[0047] Figure 3 This is a schematic diagram of a scenario where parking is done at an angle. Figure 3 As shown, taking a slanted parking space in a parking lot as an example, a vehicle waiting to park can drive in the parking lot aisle using forward gear to a certain shift position, and then shift from forward gear to reverse gear at that shift position, finally reversing diagonally into the target parking space from that shift position. This position can also be understood as the (starting) parking position for slanted parking, or the position where the vehicle begins to park in the target parking space. Figure 3 The text also uses black arrows as an example to illustrate the movement trajectory of vehicles waiting to be parked.

[0048] For example, Figure 4 This is a schematic diagram illustrating the composition of an automatic parking system provided in an embodiment of this application. Figure 4 As shown, the automatic parking system may include an image acquisition device 100 and a computing processing device 200. The image acquisition device 100 and the computing processing device 200 may be connected via a wired network or a wireless network.

[0049] The image acquisition device 100 may be disposed in the above-mentioned Figures 1 to 3 The image acquisition device 100 can be one or more vehicle-mounted cameras on the vehicle to be parked. Examples include a reversing camera or a 360-degree panoramic camera. Optionally, the image acquisition device 100 can also be an ultrasonic sensing device. For example, the ultrasonic sensing device may include an ultrasonic radar module and an imaging module; the ultrasonic radar module may include an APA ultrasonic radar or an ultrasonic parking assist (UPA) ultrasonic radar. Optionally, the image acquisition device 100 can also be a lidar.

[0050] The image acquisition device 100 can be used to acquire monitoring images around a vehicle to be parked.

[0051] The computing processing device 200 can be an electronic device with computing processing capabilities, such as a computer or a server. The server can be a single server or a server cluster consisting of multiple servers. In some embodiments, the server cluster can also be a distributed cluster. Optionally, the server can also be implemented on a cloud platform, which can include, for example, private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, inter-cloud, and multi-cloud, or any combination thereof.

[0052] Optionally, the aforementioned electronic device may also be disposed in the aforementioned Figures 1 to 3 The vehicle's onboard computer, also known as the electronic control unit (ECU), is located in the vehicle waiting to be parked. In this case, the ECU can be connected to the aforementioned vehicle camera via automotive wiring harnesses (such as the CAN bus).

[0053] Optionally, the computing processing device 200 may also be an application (APP) providing automatic parking function installed in the aforementioned electronic device; or, the computing processing device 200 may be a central processing unit (CPU) in the aforementioned electronic device; or, the computing processing device 200 may be a functional module in the aforementioned electronic device for executing the parking path planning method. This application embodiment does not impose any limitations on this.

[0054] The computing processing unit 200 can determine the scene information of the target parking space for the vehicle to be parked based on the monitoring images around the vehicle to be parked acquired by the image acquisition device 100. The parking path node information prediction model inputs the scene information of the target parking space into the path node prediction model, and plans the path for the vehicle to enter the target parking space based on the parking path node information output by the path node prediction model. The specific process can be referred to in the following method embodiments, and will not be repeated here.

[0055] For simplicity, the above-mentioned electronic device, computing processing device 200, will be used as the example for the following description.

[0056] Figure 5 This is a schematic diagram illustrating the composition of an electronic device provided in an embodiment of this application. For example... Figure 5 As shown, the electronic device may include a processor 10, a memory 20, a communication line 30, a communication interface 40, and an input / output interface 50.

[0057] The processor 10, memory 20, communication interface 40, and input / output interface 50 can be connected via communication line 30.

[0058] Processor 10 is used to execute instructions stored in memory 20 to implement the parking path planning method provided in the following embodiments of this application. Processor 10 may be a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. Processor 10 may also be any other device with processing capabilities, such as a circuit, device, or software module; this application embodiment does not limit this. In one example, processor 10 may include one or more CPUs, for example... Figure 5 CPU0 and CPU1 are mentioned. As an optional implementation, the electronic device may include multiple processors; for example, in addition to processor 10, it may also include processor 60. Figure 5 (The example shown is a dashed line).

[0059] The memory 20 is used to store instructions. For example, the instructions may be computer programs. Optionally, the memory 20 may be a read-only memory (ROM) or other types of static storage devices that can store static information and / or instructions; it may also be a random access memory (RAM) or other types of dynamic storage devices that can store information and / or instructions; it may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, etc. The embodiments of this application do not limit this.

[0060] It should be noted that the memory 20 can exist independently of the processor 10 or it can be integrated with the processor 10. The memory 20 can be located inside or outside the electronic device, and this application embodiment does not impose any restrictions on this.

[0061] Communication line 30 is used to transmit information between the components included in the electronic device.

[0062] The communication interface 40 is used to communicate with other devices (such as the aforementioned vehicle-mounted camera) or other communication networks. These other communication networks can be Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc. The communication interface 40 can be a module, circuit, transceiver, or any device capable of enabling communication.

[0063] Input / output interface 50 is used to enable human-computer interaction between users and electronic devices. This includes, for example, gesture-based, text-based, or voice-based interactions between users and electronic devices.

[0064] For example, the input / output interface 50 can be a touch screen, keyboard, mouse, physical buttons on the vehicle's center console, or the vehicle's center console screen, etc. Through the touch screen, keyboard, mouse, physical buttons on the vehicle's center console, or the vehicle's center console screen, the user and the electronic device can achieve action interaction or text interaction.

[0065] For example, the input / output interface 50 can also be an audio module, which may include speakers and microphones, etc., through which voice interaction between the user and the electronic device can be realized.

[0066] It should be noted that, Figure 5 The structures shown do not constitute a limitation on electronic devices, except... Figure 5 In addition to the components shown, electronic devices may include more or fewer components than illustrated, or combinations of certain components, or different component arrangements.

[0067] The parking path planning method provided in the embodiments of this application is described below with reference to the accompanying drawings.

[0068] Figure 6 This is a flowchart illustrating the parking path planning method provided in an embodiment of this application. Optionally, this method can be implemented by someone with... Figure 5 The electronic device with the hardware structure shown performs, such as Figure 6 As shown, the method may include S101 to S103.

[0069] S101. The electronic device acquires scene information of the target parking space for the vehicle to be parked.

[0070] The target parking space can be any parking space that is not occupied (or is vacant). The scene information for the target parking space is used to describe the target parking space.

[0071] Optionally, the scene information of the target parking space may include, but is not limited to, at least one of the following: boundary information of obstacles around the target parking space, information of the target parking space, or the final pose information of the vehicle corresponding to the target parking space.

[0072] The boundary information of obstacles around the target parking space can be used to describe the drivable area around the target parking space. The boundary information of the target parking space and the vehicle's final state pose information corresponding to the target parking space can be used to describe the location of the target parking space. For example, the vehicle's final state pose information corresponding to the target parking space can include the vehicle's position coordinates (x, y) when it enters the target parking space and the vehicle's attitude information theta (or the vehicle's heading information), that is, the vehicle's final state information can be represented as (x, y, theta), or the vehicle's final state pose information can only include the vehicle's position coordinates (x, y) when it enters the target parking space, with the xy plane parallel to the ground. When the vehicle to be parked is in the aforementioned vehicle final state pose information, it can be considered that the vehicle to be parked has entered the target parking space. The vehicle position coordinates can be the coordinates of a feature point on the vehicle. For example, the coordinates of the vehicle's geometric center or the center of the rear axle (or rear differential). Alternatively, the vehicle position coordinates can also be the coordinates of multiple feature points on the vehicle. For example, the coordinates of the center points of the four wheels, the coordinates of the four wheel contact points with the ground, or the coordinates of points on the vehicle's boundary line. Depending on the chosen coordinate system, (x, y, theta) can have different representations. This application does not specifically limit this; for example, the vehicle's own coordinate system or the world coordinate system can be selected.

[0073] For example, obstacles around the target parking space may include, but are not limited to, vehicles parked in adjacent parking spaces, curbs, warning signs, trash cans, walls, and road markings.

[0074] For example, Figure 7 This is a schematic diagram illustrating scene information of a target parking space provided in an embodiment of this application. Figure 7 As shown above, Figure 1 Taking the side parking space scenario shown as an example, the scene information of the target parking space obtained by the electronic device can include the boundary information of the target parking space. Figure 7 (Example shown in the dashed box) Boundary information of obstacles around the target parking space ( Figure 7 (The example shown is a thick black line) and the final pose information of the vehicle corresponding to the target parking space. Figure 7(The coordinates of the points with arrows are shown as an example). The boundary information of the target parking space can be obtained from the road markings around it. Obstacles around the target parking space can include vehicles parked in the surrounding parking spaces, and the boundary lines between non-motorized vehicle lanes and motorized vehicle lanes. The final pose information of the vehicle corresponding to the target parking space can include the coordinates of the vehicle's rear axle center on a plane parallel to the ground (this coordinate includes the x-coordinate and y-coordinate), and the vehicle's heading theta.

[0075] For example, Figure 8 This is a schematic diagram illustrating scene information for another target parking space provided in an embodiment of this application. For example... Figure 8 As shown above, Figure 2 Taking the parking space scenario shown as an example, the electronic device can also obtain the boundary information of the target parking space, the boundary information of the obstacles around the target parking space, and the final pose information of the vehicle corresponding to the target parking space. For details, please refer to the above. Figure 7 As mentioned above, it will not be repeated here.

[0076] For example, Figure 9 This is a schematic diagram illustrating another scenario of a target parking space provided in an embodiment of this application. For example... Figure 9 As shown above, Figure 3 Taking the angled parking space scenario shown as an example, the electronic device can also obtain the boundary of the target parking space, the boundary of the obstacles around the target parking space, and the final pose information of the vehicle to be parked. For details, please refer to the above. Figure 7 As mentioned above, it will not be repeated here.

[0077] It should be understood that the parking path planning method provided in this application, when predicting parking path node information using a path node prediction model, can also take into account the boundaries of obstacles around the target parking space. Different obstacle boundaries can correspond to different parking path node information and different parking paths, enriching the types of planned parking paths and improving the versatility of the path node prediction model. Furthermore, this solution is applicable to various types of parking space scenarios, such as lateral parking spaces, longitudinal parking spaces, or angled parking spaces. Compared to existing path node search strategies, the method provided in this application can also significantly reduce the computational load.

[0078] In one possible implementation, as described above, the electronic device can determine the scene information of the target parking space based on the monitoring images around the vehicle to be parked acquired by the image acquisition device 100. For example, the electronic device can use machine vision algorithms to recognize the monitoring images around the vehicle to be parked acquired by the image acquisition device 100 to determine the scene information of the target parking space.

[0079] In some possible embodiments, the electronic device acquires the scene information of the target parking space only after it has found (or determined) the target parking space.

[0080] In one possible implementation, as described above, the electronic device may include an input / output interface 50, which may be a physical button on the vehicle's center console or a central control screen of the vehicle. The electronic device may receive click operations from the user on the physical button on the vehicle's center console or the central control screen of the vehicle, and in response to the click operation, search for available parking spaces and determine the target parking space based on the available parking spaces.

[0081] In another possible implementation, as described above, the input / output interface 50 in the electronic device can also be an audio module. The electronic device can receive the user's voice command through the audio module and, in response to the voice command, search for an available parking space and determine the target parking space based on the available parking space.

[0082] In another possible implementation, the electronic device can be connected to the vehicle speed sensor of the vehicle to be parked. The electronic device can obtain the current speed of the vehicle to be parked through the vehicle speed sensor. When the current speed is lower than a preset speed threshold, the electronic device is automatically triggered to find an empty parking space and determine the target parking space based on the empty parking space.

[0083] Optionally, when the electronic device detects multiple vacant parking spaces, it can also display these multiple vacant parking spaces on the vehicle's central control screen, receive the user's selection operation on the multiple vacant parking spaces displayed on the vehicle's central control screen, and in response to the selection operation, determine the target parking space from the multiple vacant parking spaces.

[0084] S102. The electronic device inputs the scene information of the target parking space into the path node prediction model to determine the parking path node information of the vehicle to be parked.

[0085] The parking path node information can include the position and attitude information of the vehicle to be parked at that node. A parking path node is a node the vehicle passes through when parking, such as the starting point of the parking path. The position information of the vehicle at that node can be represented by position coordinates. The attitude information theta of the vehicle at that node can include the deflection angles of the vehicle in the x, y, and z directions in a preset three-dimensional coordinate system (e.g., a three-dimensional world coordinate system). (For example, the xy plane can be a plane parallel to the ground, and the z-axis perpendicular to the xy plane). This deflection angle can also be understood as heading. The path node prediction model is pre-installed in the electronic device. This model can be used to characterize the correspondence between parking path node information and parking space scene information; in other words, the model can be used to predict the parking path node information corresponding to the scene information of the target parking space based on the scene information of the target parking space. The training process of this model can be referred to the following... Figure 10 As mentioned above, it will not be repeated here.

[0086] Optionally, the path node prediction model can be a neural network model. This neural network model may include multiple network modules and an output layer. For example, the multiple network modules and the output layer may be connected in series.

[0087] Each network module includes a fully connected layer, an activation layer, and a normalization layer. The fully connected layer extracts features from the scene information of the target parking space. The activation layer performs non-linear processing on the scene information extracted by the fully connected layer to obtain non-linear features. The normalization layer normalizes these non-linear features. The output layer outputs the parking path node information of the vehicle to be parked, based on the normalized non-linear features obtained from the normalization layer.

[0088] For example, the correspondence between parking path node information and parking space scene information can be shown in Table 1 below.

[0089] Table 1

[0090]

[0091]

[0092] As shown in Table 1, this table can include scene information items for parking spaces, parking path node information items, location information items, and attitude information items. Specifically, the scene information items for parking spaces can include information such as Scene 1, Scene 2, and Scene 3; the parking path node information items can include parking path node information such as Information 1, Information 2, and Information 3; the location information items can include location information such as Location 1, Location 2, and Location 3; and the attitude information items can include attitude information such as Direction 1, Direction 2, and Direction 3. Scene 1, Information 1, Location 1, and Direction 1 have a corresponding relationship. Scene 2, Information 2, Location 2, and Direction 2 have a corresponding relationship. Scene 3, Information 3, Location 3, and Direction 3 have a corresponding relationship. Scene 4, Information 4, Location 4, and Direction 4 have a corresponding relationship.

[0093] S103. The electronic equipment plans the path for the vehicle to be parked to enter the target parking space based on the parking path node information.

[0094] Optionally, the electronic device can use a planning algorithm to plan the path for the vehicle to enter the target parking space. The planning algorithm can be found in related technical documents and will not be elaborated upon here.

[0095] In one possible implementation, the parking path node includes a gear shift node for the vehicle to be parked. In this case, the method may further include: electronic devices controlling the vehicle to be parked to travel to the location corresponding to the parking path node information, and controlling the vehicle to be parked in reverse gear and in the posture corresponding to the parking path node information at that location; while in reverse gear and in the posture corresponding to the parking path node information, the electronic devices controlling the vehicle to be parked to enter the target parking space according to the planned path.

[0096] Using the parking path planning method provided in this application, scene information of the target parking space can be obtained, and the target parking path node information corresponding to the scene information of the target parking space can be predicted according to the path node prediction model used to characterize the correspondence between parking path node information and scene information of the parking space. The path node can be a shift node (e.g., the first shift node in the parking process). The target parking path node information and the scene information of the target parking space have a pre-trained matching relationship. The parking route planned from the position corresponding to the parking path node information and according to the posture corresponding to the parking path node information has higher path quality, which improves the overall efficiency of vehicle parking.

[0097] In some possible embodiments, prior to S101, the electronic device may also acquire a path node prediction model.

[0098] The path node prediction model can be obtained by the electronic device training the neural network in advance based on training samples. Optionally, the path node prediction model can also be obtained by any computing device with computing power training the neural network in advance based on training samples and then sending it to the electronic device, or it can be transferred to the electronic device through an intermediate storage medium. This application embodiment does not impose any limitations on this.

[0099] Optionally, taking the path node prediction model obtained by pre-training a neural network on training samples by an electronic device as an example, refer to... Figure 10 , Figure 10 This is a flowchart illustrating the training method of the path node prediction model provided in an embodiment of this application. Figure 10 As shown, before S101, the method may also include S201 to S202.

[0100] S201. Electronic equipment acquires training sample set.

[0101] The training sample set may include multiple training samples. Each training sample may include scene information of the parking space and a label corresponding to that scene information. The label includes parking path node information during the vehicle parking process. The parking path node information includes the vehicle's position and attitude information at that parking path node.

[0102] In one possible implementation, S201 can be specifically implemented as follows: the electronic device acquires scene information of multiple parking spaces and determines candidate areas for each parking space; the electronic device selects multiple location information and preset posture information of the vehicle in a parking state corresponding to each location information from the candidate areas of each parking space, as candidate parking path node information for each parking space; the electronic device plans a parking planning path corresponding to each candidate parking path node information based on each candidate parking path node information; the electronic device filters the label corresponding to each parking space from the multiple candidate parking path node information corresponding to each parking space based on the parking planning path corresponding to each candidate parking path node information.

[0103] The candidate areas can be set by managers based on their experience.

[0104] For example, Figure 11 This is a schematic diagram of a candidate region provided for an embodiment of this application. For example... Figure 11 As shown above, Figure 7 Taking the side parking scenario shown as an example, the candidate area can be set at the same level as the parking space preceding the target parking space. This candidate area can include multiple locations. x0 represents the horizontal coordinate of a location within the candidate area, y0 represents the vertical coordinate of that location, and theta0 represents the preset attitude information (heading) of the vehicle corresponding to that location when it is in a parking state.

[0105] It should be noted that for a certain position in the candidate region (i.e. Figure 11 For a dot in the diagram, the position and the different preset pose information corresponding to that position (i.e., ...) Figure 11 The arrows connecting the center point and the dot can form different candidate parking path node information. Figure 11 (The example shows different arrows connecting the same dot).

[0106] For example, Figure 12 This is a schematic diagram of another candidate region provided for an embodiment of this application. For example... Figure 12 As shown, Figure 12 The image shows a schematic diagram of the candidate areas in a garage parking space scenario. For more details, please refer to [link / reference]. Figure 11 As previously mentioned, this will not be repeated here.

[0107] For example, Figure 13 This is a schematic diagram of yet another candidate region provided in an embodiment of this application. For example... Figure 13 As shown, Figure 13 The image shows a schematic diagram of the candidate area in a slanted parking space scenario. Further details can be found in [the image / document / reference]. Figure 11 As previously mentioned, this will not be repeated here.

[0108] Optionally, the electronic device acquiring scene information for multiple parking spaces may include: the electronic device randomly generating scene information for multiple parking spaces. For example, the electronic device may randomly generate scene information for multiple parking spaces based on human experience information input by management personnel.

[0109] Optionally, after planning the parking path corresponding to each candidate parking path node based on the candidate parking path node information, the electronic device can also evaluate the quality of each parking path. That is, the electronic device can determine the quality of the parking path corresponding to each candidate parking path node based on at least one of the following information related to path quality: the number of gear shifts, path length, and path curvature during the process of driving into the parking space according to the parking path corresponding to each candidate parking path node information. In this case, the electronic device, based on the parking path corresponding to each candidate parking path node information, filters the label corresponding to each parking space from multiple candidate parking path node information corresponding to each parking space. This can include: the electronic device determining the candidate parking path node information corresponding to the parking path with a quality greater than a preset quality threshold or the highest quality from multiple candidate parking path node information corresponding to each parking space as the label for each parking space.

[0110] Optionally, the quality of the parking planning path corresponding to each candidate parking path node can be determined based on the number of gear shifts, path length, and path curvature of the vehicle during the process of entering the parking space, as well as the corresponding weights for the number of gear shifts, path length, and path curvature. For example, the relationship between the path quality of the parking planning path and the number of gear shifts, path length, and path curvature of the vehicle during the process of entering the parking space according to the parking planning path can satisfy the following formula (1).

[0111] Q=α*(-T)+β*(-L)+γ*(-K) Formula (1)

[0112] In formula (1), Q represents the path quality of the parking path. T represents the number of gear shifts in the parking path. α represents the weight of the number of gear shifts in the path quality of the parking path. L represents the length of the parking path. β represents the weight of the length of the parking path in the path quality of the parking path. K represents the curvature of the parking path. γ represents the weight of the curvature of the parking path in the path quality of the parking path. α, β, and γ can be preset in the electronic device by the administrator. α, β, and γ are all greater than 0 and less than 1, and the sum of α, β, and γ is 1.

[0113] It should be understood that when obtaining labels from training samples in this application embodiment, candidate parking path node information corresponding to parking planning paths with higher path quality can be selected as labels. The parking path node information predicted by the path node prediction model trained with the training samples with the labels will also have the characteristics of the labels, that is, the parking path corresponding to the predicted parking path node information has higher path quality (for example, high path quality may include at least one of the following: fewer gear shifts, shorter parking path, and smaller parking path curvature).

[0114] In this embodiment of the application, a path node prediction model is obtained through training and used to predict parking path node information on the parking path. Compared with the existing path node search strategy, it can significantly reduce the amount of computation and enrich the types of plannable parking paths, which is conducive to improving the quality of path planning (including reducing the number of gear shifts and shortening the path length, etc.). Moreover, the model has strong versatility for various parking space scenarios.

[0115] For example, Figure 14 This is a schematic diagram illustrating a method for obtaining labels for training samples provided in an embodiment of this application. For example... Figure 14 As shown, the method may include six parts: generating scene information of parking spaces, setting candidate areas, randomly generating candidate parking path node information, planning parking paths, evaluating the path quality of parking paths, and outputting labels. These six parts can be referred to in S201 above, and will not be repeated here.

[0116] S202. The electronic equipment trains a preset model based on the training sample set to obtain a path node prediction model.

[0117] The preset model can also be a neural network model. The structure of the neural network model can be referred to the path node information prediction model structure described above, and will not be repeated here.

[0118] Optionally, as described above, the training sample set may include multiple training samples. The electronic device can input one or more training samples into the neural network each time to obtain predicted values ​​(parking path node information corresponding to the scene information of a certain parking space predicted by the neural network). Based on the predicted values ​​and the labels in the training samples (parking path node information corresponding to the scene information of that particular parking space), a loss function is calculated, and the parameters of the preset model are adjusted. This process is repeated, inputting multiple training samples from the training sample set into the preset model one by one, iteratively training until the preset model converges.

[0119] Optionally, the preset conditions for model convergence (end of training) may include: the electronic device inputs training samples into the preset model a preset number of times, or the error between the predicted value and the label is less than a preset error threshold.

[0120] The foregoing primarily describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the aforementioned functions, it includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0121] In an exemplary embodiment, this application also provides an automatic parking device, which can be applied to the above-mentioned... Figure 5 The electronic device shown. Figure 15 This is a schematic diagram illustrating the composition of the parking path planning device provided in an embodiment of this application. Figure 15 As shown, the device may include an acquisition module 1501 and a processing module 1502. The acquisition module 1501 and the processing module 1502 are connected.

[0122] The acquisition module 1501 is used to acquire scene information of the target parking space for the vehicle to be parked; the scene information of the target parking space is used to describe the target parking space. The processing module 1502 is used to input the scene information of the target parking space into the path node prediction model to determine the parking path node information of the vehicle to be parked; wherein, the path node prediction model is used to characterize the correspondence between the parking path node information and the scene information of the parking space, and the parking path node information includes the position information and attitude information of the vehicle to be parked at the parking path node; based on the parking path node information, the path for the vehicle to be parked to enter the target parking space is planned.

[0123] In some possible embodiments, the scene information of the target parking space includes at least one of the following: boundary information of obstacles around the target parking space, boundary information of the target parking space, or final pose information of the vehicle corresponding to the target parking space; the boundary information of obstacles around the target parking space is used to describe the drivable area around the target parking space; the boundary information of the target parking space and the final pose information of the vehicle corresponding to the target parking space are used to describe the location of the target parking space.

[0124] In other possible embodiments, the parking path node includes a gear shift node for the vehicle to be parked. The processing module 1502 is further configured to control the vehicle to be parked to drive to the position corresponding to the parking path node information, and control the vehicle to be parked to be in reverse gear and in the posture corresponding to the parking path node information at that position; while in reverse gear and in the posture corresponding to the parking path node information, control the vehicle to be parked to drive into the target parking space according to the planned path.

[0125] In some other possible embodiments, the path node prediction model is obtained by training a preset model based on a training sample set. The training sample set includes multiple training samples, each of which includes scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes parking path node information of the vehicle.

[0126] In some other possible embodiments, the label corresponding to each parking space is obtained by filtering from multiple candidate parking path node information corresponding to each parking space based on the parking planning path corresponding to each candidate parking path node information; the parking planning path corresponding to each candidate parking path node information is the path for the vehicle to enter the parking space planned based on each candidate parking path node information; the multiple candidate parking path node information corresponding to each parking space includes multiple location information within the candidate area and preset posture information of the vehicle in the parking state corresponding to each location information.

[0127] In some other possible embodiments, the label corresponding to each parking space includes candidate parking path node information when the quality of the parking planning path is greater than a preset quality threshold or the highest quality; the quality of the parking planning path is determined based on at least one of the following when the vehicle enters the parking space according to the parking planning path: the number of gear shifts, the path length, and the path curvature.

[0128] In some other possible embodiments, the path node prediction model is a neural network model; the neural network model includes multiple network modules and an output layer, each network module including a fully connected layer, an activation layer, and a normalization layer; the fully connected layer is used to extract features of the scene information of the target parking space; the activation layer is used to perform nonlinear processing on the features of the scene information of the target parking space extracted by the fully connected layer to obtain nonlinear features; the normalization layer is used to normalize the nonlinear features; the output layer is used to output the parking path node information of the vehicle to be parked based on the nonlinear features normalized by the normalization layer.

[0129] In an exemplary embodiment, this application also provides a training apparatus for a path node prediction model, which can be applied to the above-mentioned... Figure 5 The electronic device shown. Figure 16 This is a schematic diagram illustrating the composition of the training apparatus for the path node prediction model provided in an embodiment of this application. Figure 16 As shown, the device includes an acquisition module 1601 and a processing module 1602.

[0130] The acquisition module 1601 is used to acquire a training sample set, which includes multiple training samples. Each training sample includes scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes parking path node information during the parking process. The parking path node information includes the position and attitude information of the vehicle at the parking path node.

[0131] The processing module 1602 is used to train a preset model based on a training sample set to obtain a path node information prediction model.

[0132] In some possible embodiments, the acquisition module 1601 is specifically used to acquire scene information of multiple parking spaces and determine candidate areas for each parking space; select multiple location information and preset posture information of the vehicle in parking state corresponding to each location information from the candidate areas of each parking space as candidate parking path node information for each parking space; plan a parking planning path corresponding to each candidate parking path node information based on each candidate parking path node information; and filter out the label corresponding to each parking space from the multiple candidate parking path node information corresponding to each parking space based on the parking planning path corresponding to each candidate parking path node information.

[0133] In other possible embodiments, the acquisition module 1601 is further configured to determine the quality of the parking planning path corresponding to each candidate parking path node based on at least one of the following: the number of gear shifts, the path length, and the path curvature during the process of driving into the parking space according to the parking planning path corresponding to each candidate parking path node information. Specifically, the acquisition module 1601 is configured to determine, from the multiple candidate parking path node information corresponding to each parking space, the candidate parking path node information corresponding to the case where the quality of the parking planning path is greater than a preset quality threshold or the highest quality, as the label for each parking space.

[0134] In some other possible embodiments, the preset model is a neural network model; the neural network model includes multiple network modules and an output layer, each network module including a fully connected layer, an activation layer, and a normalization layer; the fully connected layer is used to extract features of parking space scene information in the training samples; the activation layer is used to perform nonlinear processing on the features of parking space scene information extracted by the fully connected layer in the training samples to obtain nonlinear features; the normalization layer is used to normalize the nonlinear features; the output layer is used to output parking path node information based on the nonlinear features normalized by the normalization layer.

[0135] It should be noted that, Figure 15 and Figure 16The module division shown is illustrative and represents only one logical functional division; in actual implementation, other division methods are possible. For example, two or more functions can be integrated into a single processing module. These integrated modules can be implemented either in hardware or as software functional modules.

[0136] In an exemplary embodiment, this application also provides a computer-readable storage medium including computer-executable instructions that, when run on an electronic device, cause the electronic device to perform any of the methods provided in the above embodiments.

[0137] In an exemplary embodiment, this application also provides a computer program product containing computer execution instructions, which, when run on an electronic device, causes the electronic device to perform any of the methods provided in the above embodiments.

[0138] In an exemplary embodiment, this application also provides a chip, including: a processor and an interface, wherein the processor is coupled to a memory through the interface, and when the processor executes computer programs in the memory or electronic devices execute instructions, any of the methods provided in the above embodiments are executed.

[0139] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer-executable instructions. When these computer-executable instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer-executable instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer-executable instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks, SSDs).

[0140] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple components. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0141] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of this application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from the spirit and scope of this application. Thus, if such modifications and modifications of this application fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and modifications.

[0142] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A parking path planning method characterized by comprising: The method includes: Obtain scene information of the target parking space for the vehicle to be parked; the scene information of the target parking space is used to describe the target parking space; The scene information of the target parking space is input into the path node prediction model to determine the parking path node information of the vehicle to be parked. The path node prediction model characterizes the correspondence between the parking path node information and the scene information of the parking space. The parking path node information includes the position and attitude information of the vehicle to be parked at that parking path node. The path node prediction model is trained on a preset model based on a training sample set. The training sample set includes multiple training samples, each of which includes the scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes the parking path node information of the vehicle. The label corresponding to each parking space is obtained by filtering from multiple candidate parking path node information for each parking space based on the parking planning path corresponding to each candidate parking path node information; the parking planning path corresponding to each candidate parking path node information is the path for the vehicle to enter the parking space planned based on each candidate parking path node information; the multiple candidate parking path node information for each parking space includes multiple location information within the candidate area and preset posture information of the vehicle in the parking state corresponding to each location information. The label corresponding to each parking space includes candidate parking path node information when the quality of the parking planning path is greater than a preset quality threshold or the highest quality. The quality of the parking planning path is determined based on at least one of the following: the number of gear shifts, path length, and path curvature of the vehicle during the process of entering the parking space according to the parking planning path. Based on the parking path node information, the path for the vehicle to be parked to enter the target parking space is planned.

2. The method of claim 1, wherein, The scene information of the target parking space includes at least one of the following: boundary information of obstacles around the target parking space, boundary information of the target parking space, or final state pose information of the vehicle corresponding to the target parking space; The boundary information of the obstacles around the target parking space is used to describe the drivable area around the target parking space; The boundary information of the target parking space and the final pose information of the vehicle corresponding to the target parking space are used to describe the location of the target parking space.

3. The method according to claim 1, characterized in that, The parking path node includes the gear shift node of the vehicle to be parked; the method further includes: Control the vehicle to be parked to drive to the position corresponding to the parking path node information, and control the vehicle to be parked to be in reverse gear and in the posture corresponding to the parking path node information at that position; In reverse gear and in the posture corresponding to the parking path node information, the vehicle to be parked is controlled to drive into the target parking space according to the planned path.

4. The method according to any one of claims 1-3, characterized in that, The path node prediction model is a neural network model; The neural network model includes multiple network modules and an output layer. Each network module includes a fully connected layer, an activation layer, and a normalization layer. The fully connected layer is used to extract features of the scene information of the target parking space; The activation layer is used to perform nonlinear processing on the features of the scene information of the target parking space extracted by the fully connected layer to obtain nonlinear features; The normalization layer is used to normalize the nonlinear features; The output layer is used to output the parking path node information of the vehicle to be parked based on the nonlinear characteristics after normalization processing by the normalization layer.

5. A training method for a path node prediction model, characterized in that, The method includes: Obtain a training sample set, which includes multiple training samples. Each training sample includes scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes parking path node information during the parking process. The parking path node information includes the position information and attitude information of the vehicle at the parking path node. The acquisition of the training sample set includes: Acquire scene information for multiple parking spaces and determine candidate areas for each parking space; from the candidate areas of each parking space, select multiple location information and preset posture information of the vehicle in parking state corresponding to each location information as candidate parking path node information for each parking space; based on each candidate parking path node information, plan a parking planning path corresponding to each candidate parking path node information; based on at least one of the following during the process of driving into the parking space according to the parking planning path corresponding to each candidate parking path node information, the quality of the parking planning path corresponding to each candidate parking path node information is determined; from the multiple candidate parking path node information corresponding to each parking space, determine the candidate parking path node information corresponding to the case where the quality of the parking planning path is greater than a preset quality threshold or the highest quality, and use it as the label for each parking space; Based on the training sample set, the preset model is trained to obtain the path node information prediction model.

6. The method according to claim 5, characterized in that, The preset model is a neural network model; The neural network model includes multiple network modules and an output layer. Each network module includes a fully connected layer, an activation layer, and a normalization layer. The fully connected layer is used to extract features of the parking space scene information in the training samples; The activation layer is used to perform nonlinear processing on the features of the parking space scene information extracted by the fully connected layer from the training samples to obtain nonlinear features; The normalization layer is used to normalize the nonlinear features; The output layer is used to output parking path node information based on the nonlinear characteristics after normalization processing by the normalization layer.

7. A parking path planning device, characterized in that, The device includes: an acquisition module and a processing module; The acquisition module is used to acquire scene information of the target parking space for the vehicle to be parked; the scene information of the target parking space is used to describe the target parking space. The processing module is used to input the scene information of the target parking space into the path node prediction model to determine the parking path node information of the vehicle to be parked. The path node prediction model is used to characterize the correspondence between the parking path node information and the scene information of the parking space. The parking path node information includes the position and attitude information of the vehicle to be parked at the parking path node. The path node prediction model is obtained by training a preset model based on a training sample set. The training sample set includes multiple training samples, each of which includes scene information of the parking space and a label corresponding to the scene information of the parking space. The label includes the parking path node information of the vehicle. The label corresponding to each parking space is selected from multiple candidate parking path node information corresponding to each parking space, based on the information of each candidate parking path node. The parking planning path is obtained by filtering the corresponding information; the parking planning path corresponding to each candidate parking path node information is the path for the vehicle to enter the parking space planned based on each candidate parking path node information; the multiple candidate parking path node information corresponding to each parking space includes multiple location information within the candidate area and preset posture information of the vehicle in the parking state corresponding to each location information; the label corresponding to each parking space includes the candidate parking path node information corresponding to the case where the quality of the parking planning path is greater than a preset quality threshold or the highest quality; the quality of the parking planning path is determined based on at least one of the following: the number of gear shifts, path length, and path curvature of the vehicle during the process of entering the parking space according to the parking planning path; based on the parking path node information, the path for the vehicle to be parked to enter the target parking space is planned.

8. A training device for a path node prediction model, characterized in that, The device includes: The acquisition module is used to acquire a training sample set, which includes multiple training samples. Each training sample includes scene information of a parking space and a label corresponding to the scene information of the parking space. The label includes parking path node information during the vehicle parking process, and the parking path node information includes the vehicle's position and attitude information at that parking path node. Specifically, the acquisition module is used to acquire scene information of multiple parking spaces and determine a candidate area for each parking space. From the candidate areas of each parking space, multiple position information and preset attitude information of the vehicle in a parking state corresponding to each position information are selected as the corresponding area for each parking space. Candidate parking path node information; based on each candidate parking path node information, plan a parking planning path corresponding to each candidate parking path node information; based on at least one of the following during the process of driving into the parking space according to the parking planning path corresponding to each candidate parking path node information, the number of gear shifts, the path length, and the path curvature, determine the quality of the parking planning path corresponding to each candidate parking path node information; from the multiple candidate parking path node information corresponding to each parking space, determine the candidate parking path node information corresponding to the case where the quality of the parking planning path is greater than a preset quality threshold or the highest quality, and use it as the label of each parking space; The processing module is used to train a preset model based on the training sample set to obtain a path node information prediction model.

9. An electronic device, characterized in that, The electronic device includes: a processor and a memory; The memory stores instructions that the processor can execute; When the processor is configured to execute the instructions, it causes the electronic device to implement the method as claimed in any one of claims 1-4 or 5 or 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes: computer software instructions; When the computer software instructions are executed in an electronic device, the electronic device causes the electronic device to perform the method as claimed in any one of claims 1-4 or 5 or 6.

Images