Vehicle travel control method, device, computer program product, and vehicle

Abstract

The application provides a vehicle driving control method and device, a computer program product and a vehicle. The method comprises: acquiring road information of a current driving road of the vehicle, wherein the road information is whether a water accumulation area exists in front of the vehicle and whether a traffic participant exists in an area where the vehicle is located, the traffic participant comprising at least one of a pedestrian and a non-motor vehicle; and in a case where the water accumulation area exists in front of the vehicle and the traffic participant exists in the area where the vehicle is located, controlling the vehicle to perform an avoidance operation, the avoidance operation comprising at least one of controlling the vehicle to slow down and controlling the vehicle to bypass the water accumulation area. The scheme solves the problem that in the prior art, water splashes are easily caused when a vehicle drives over a water pit in a driving process, and water splashes can splash onto surrounding pedestrians or non-motor vehicles.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and more specifically, to a vehicle driving control method, a vehicle driving control device, a computer program product, and a vehicle. Background Technology

[0002] In the environmental perception and trajectory planning systems of autonomous vehicles, puddles are a common dynamic obstacle scenario, especially in rainy weather or poorly drained road sections. When a vehicle drives over a puddle, water can easily splash, potentially hitting pedestrians or non-motorized vehicles nearby. Summary of the Invention

[0003] The main objective of this application is to provide a vehicle driving control method, a vehicle driving control device, a computer program product, and a vehicle, so as to at least solve the problem in the prior art that when a vehicle drives through a puddle, water splashes easily and the water splashes onto the surrounding pedestrians or non-motorized vehicles.

[0004] To achieve the above objectives, according to one aspect of this application, a vehicle driving control method is provided, comprising: acquiring road information of the road on which the vehicle is currently traveling, wherein the road information includes whether there is a water accumulation area ahead of the vehicle and whether there are traffic participants in the area where the vehicle is located, wherein the traffic participants include at least one of the following: pedestrians and non-motorized vehicles; and, when there is a water accumulation area ahead of the vehicle and the traffic participants are in the area where the vehicle is located, controlling the vehicle to perform an avoidance operation, wherein the avoidance operation includes at least one of the following: controlling the vehicle to decelerate and controlling the vehicle to bypass the water accumulation area.

[0005] Optionally, obtaining road information of the vehicle's current driving route includes: acquiring an image of the vehicle's current driving route to obtain a driving image; calling a recognition model, wherein the recognition model is one of a CNN model, an RNN model, or a Transformer model; forming a training set by combining historical driving images and corresponding water accumulation labels, and training the recognition model using the training set to obtain a water accumulation area recognition model, wherein the water accumulation labels are whether a water accumulation area exists in the historical driving images of the training set and a water accumulation area selection box; inputting the driving image into the water accumulation area recognition model to obtain a recognition result corresponding to the driving image, wherein the recognition result is whether the water accumulation area exists in front of the vehicle.

[0006] Optionally, controlling the vehicle to perform an avoidance maneuver includes at least one of the following: controlling the vehicle to decelerate based on the area of ​​the water accumulation area, wherein the size of the water accumulation area is negatively correlated with the vehicle's speed; or controlling the vehicle to bypass the water accumulation area based on the area of ​​the water accumulation area, wherein the size of the water accumulation area is positively correlated with the vehicle's lateral avoidance offset.

[0007] Optionally, controlling the vehicle to decelerate includes: acquiring a safe speed and a safe distance, wherein the safe speed is a preset upper limit for safely passing through the waterlogged area, and the safe distance is a preset minimum braking distance; acquiring the speed difference between the vehicle and the traffic participant to obtain a relative speed difference; acquiring the distance between the vehicle and the waterlogged area to obtain a longitudinal distance; calculating the vehicle speed based on the safe speed, the safe distance, the relative speed difference, and the longitudinal distance; and controlling the vehicle to decelerate based on the vehicle speed, wherein the vehicle speed is the speed required to suppress water splashing.

[0008] Optionally, controlling the vehicle to bypass the flooded area includes: extracting the polygonal outline of the flooded area to obtain the outline of the flooded area; designating the area enclosed by the outline of the flooded area as a no-driving area, replanning the vehicle's driving path to obtain a planned driving path; and controlling the vehicle to drive along the planned driving path so that the vehicle bypasses the flooded area.

[0009] Optionally, after extracting the polygonal outline of the waterlogged area to obtain the waterlogged area outline, the method further includes: obtaining the drivable boundary of the road; obtaining the distance between the leftmost edge point of the waterlogged area outline and the drivable boundary to obtain the edge spacing, or obtaining the distance between the rightmost edge point of the waterlogged area outline and the drivable boundary to obtain the edge spacing; if the edge spacing is less than the width of the vehicle, cropping the waterlogged area outline in an outward direction to obtain the cropped waterlogged area outline; using the area enclosed by the cropped waterlogged area outline as the prohibited driving area, and replanning the vehicle's driving path to obtain the planned driving path.

[0010] Optionally, the vehicle's driving path is replanned to obtain the planned driving path, including: obtaining the minimum lateral displacement of the vehicle's driving path relative to the centerline to obtain the lateral offset; and using the lateral offset as a constraint, the vehicle's driving path is replanned to obtain the planned driving path.

[0011] According to another aspect of this application, a vehicle driving control device is provided, comprising: a first acquisition unit, configured to acquire road information of the road on which the vehicle is currently traveling, wherein the road information includes whether there is a water accumulation area in front of the vehicle and whether there are traffic participants in the area where the vehicle is located, wherein the traffic participants include at least one of the following: pedestrians and non-motorized vehicles; and a control unit, configured to control the vehicle to perform an avoidance operation when there is a water accumulation area in front of the vehicle and the traffic participants are in the area where the vehicle is located, wherein the avoidance operation includes at least one of the following: controlling the vehicle to decelerate and controlling the vehicle to bypass the water accumulation area.

[0012] According to another aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the steps of any of the vehicle driving control methods.

[0013] According to another aspect of this application, a vehicle is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing any of the described vehicle driving control methods.

[0014] By applying the technical solution of this application, when a puddle is detected in front of the vehicle and there are pedestrians or non-motorized vehicles within the lateral influence range of the puddle, the vehicle actively initiates an avoidance operation. Deceleration can reduce the kinetic energy of the vehicle when passing through the puddle and the speed at which the tires throw water, thereby significantly reducing the height and range of water splash. By detouring can allow the main body of the vehicle to avoid the water accumulation area, further reducing the splash direction, thereby preventing water from splashing onto surrounding pedestrians or non-motorized vehicles. Attached Figure Description

[0015] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0016] Figure 1 A hardware structure block diagram of a mobile terminal for executing a vehicle driving control method according to an embodiment of this application is shown;

[0017] Figure 2 A schematic flowchart of a vehicle driving control method according to an embodiment of this application is shown;

[0018] Figure 3 A schematic diagram of the splash guard location is shown;

[0019] Figure 4A schematic diagram of region clipping is shown;

[0020] Figure 5 A schematic diagram illustrating the constraint on vehicle avoidance is shown;

[0021] Figure 6 A structural block diagram of a vehicle driving control device according to an embodiment of this application is shown.

[0022] The above figures include the following reference numerals:

[0023] 102. Processor; 104. Memory; 106. Transmission device; 108. Input / output device. Detailed Implementation

[0024] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

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

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

[0027] In real-world driving, autonomous vehicles frequently encounter puddles formed by rainwater accumulation. Traditional autonomous driving systems typically treat puddles as ordinary obstacles and attempt to avoid them, or ignore them and drive through them at high speed. The former can easily lead to excessive lateral movement of the vehicle, encroaching on oncoming lanes or causing a side collision risk; the latter can cause the vehicle to wade through too deep water and lose traction, or splash water onto pedestrians or non-motorized vehicles while passing through at high speed, lacking both safety and social friendliness.

[0028] In existing autonomous driving systems, handling road flooding is typically quite simple. Most current technologies treat deep water as a static obstacle and bypass it, or implement a globally uniform fixed speed limit in shallow water. This simplistic approach ignores the interaction between autonomous vehicles and their surroundings.

[0029] When vehicles drive at high speeds through flooded roads, they can easily splash water onto pedestrians or cyclists on the roadside, causing serious social discomfort and safety hazards. Existing systems lack an intelligent dynamic decision-making and alert mechanism that can comprehensively consider the coupling relationship between "road flooding," "vehicle movement status," and "surrounding traffic participants."

[0030] This solution aims to address the shortcomings of existing technologies by providing an autonomous driving water-wading strategy that balances vehicle safety with social anthropomorphism (avoiding splashing groups). This method can dynamically identify waterlogged areas and, upon sensing the presence of pedestrians or non-motorized vehicles around the waterlogged area, adaptively apply stricter longitudinal motion constraints (such as a stronger lower limit for deceleration). Simultaneously, it manages prompts through state memory and a cooling mechanism, achieving more elegant and defensive intelligent planning.

[0031] As described in the background section, in the prior art, when a vehicle drives through a puddle, water splashes easily, and the accumulated water can splash onto pedestrians or non-motorized vehicles. To solve the above problems, embodiments of this application provide a vehicle driving control method, a vehicle driving control device, a computer program product, and a vehicle.

[0032] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0033] The methods and embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a vehicle driving control method according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0034] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the vehicle driving control method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0035] This embodiment provides a vehicle driving control method that runs on a mobile terminal, computer terminal, or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0036] Figure 2 This is a schematic flowchart of a vehicle driving control method according to an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0037] Step S201: Obtain road information of the road on which the vehicle is currently traveling. The road information includes whether there is a waterlogged area in front of the vehicle and whether there are traffic participants in the area where the vehicle is located. The traffic participants include at least one of the following: pedestrians and non-motorized vehicles.

[0038] Specifically, onboard sensors, including a forward-facing camera and a lidar, can acquire road environment data in front of the vehicle. The camera captures road images, while the lidar is used for point cloud modeling and obstacle localization. The images are analyzed to determine if there are any puddles or water accumulation areas. Simultaneously, the lidar detects dynamic obstacles around the vehicle to determine if pedestrians or non-motorized vehicles are present, and uses spatial coordinates to determine if they are within the lateral influence range of the vehicle's path.

[0039] Step S202: When there is a waterlogged area in front of the vehicle and traffic participants are present in the area where the vehicle is located, the vehicle is controlled to perform an avoidance operation, wherein the avoidance operation includes at least one of the following: controlling the vehicle to decelerate, or controlling the vehicle to bypass the waterlogged area.

[0040] Specifically, when it is determined that there is a water accumulation area ahead, and a pedestrian or non-motorized vehicle is detected within the lateral influence range of the water accumulation area (e.g., within 10 meters), an avoidance decision is triggered, and longitudinal deceleration control is executed. If there is sufficient lateral driving space (the remaining width of the left or right driving boundary is greater than 1.2 times the width of the vehicle), lateral detour is initiated at the same time to bypass the water pit.

[0041] In this embodiment, when a puddle is detected in front of the vehicle and there are pedestrians or non-motorized vehicles within the lateral influence range of the puddle, an avoidance operation is actively initiated. Deceleration can reduce the kinetic energy of the vehicle when passing through the puddle and the speed at which the tires throw water, thereby significantly reducing the height and range of water splash. Bypassing the puddle allows the vehicle body to avoid the water accumulation area, further reducing the splash direction, thereby preventing water from splashing onto surrounding pedestrians or non-motorized vehicles.

[0042] Specifically, the solution proposed in this application aims to provide a method for dealing with water puddles for autonomous vehicles, which comprehensively considers the vehicle's lateral safety boundary constraints and longitudinal splash prevention social interactions, thereby improving the safety and anthropomorphism of autonomous driving.

[0043] In the specific implementation process, obtaining road information of the vehicle's current driving route can be achieved through the following steps: acquiring an image of the vehicle's current driving route to obtain a driving image; calling a recognition model, wherein the recognition model is one of a CNN model, an RNN model, or a Transformer model; forming a training set with historical driving images and corresponding water accumulation labels for the historical driving images, and training the recognition model using the training set to obtain a water accumulation area recognition model, wherein the water accumulation labels are the presence or absence of water accumulation areas and water accumulation area selection boxes in the historical driving images of the training set; inputting the driving image into the water accumulation area recognition model to obtain the recognition result corresponding to the driving image, wherein the recognition result is whether there is a water accumulation area in front of the vehicle.

[0044] This solution employs a multi-task-based deep learning model to model the local details of the water accumulation area, thereby enabling more accurate identification of the water accumulation area and more accurately determining whether there are puddles in front of the vehicle.

[0045] Specifically, the system acquires road environment information ahead of the autonomous vehicle, using a forward-facing camera to capture the current RGB image as the driving image for subsequent water accumulation area identification. The model primarily employs a joint multi-task network based on candidate boxes and local image cropping, with a convolutional neural network (CNN) as its backbone, supporting the extraction of local features from image patches. Alternatively, it can be replaced with structures such as Swin Transformer or ConvNeXt, all belonging to any of the CNN, RNN, or Transformer model types, to achieve end-to-end perception of water accumulation areas.

[0046] The training set consists of labeled driving images. Each image contains two types of labels: a binary classification label (whether there is standing water) and rectangular candidate bounding boxes (selection boxes) for locating water-filled areas. During training, the model simultaneously learns classification, regression, and segmentation tasks. The loss function includes classification cross-entropy, bounding box regression loss, and pixel-level mask loss. The training data covers multiple scenarios, including sunny days, rainy days, nighttime, glare, and shadows, to enhance the model's generalization ability. The current driving image is input into the trained water-filled area recognition model, which outputs four dimensions of recognition results: classification confidence (whether there is standing water), pixel-level mask (outline of the puddle area), area percentage (percentage of the pre-selected box), and uncertainty score. Only when the classification confidence is higher than 0.85 and the uncertainty is lower than 0.15 is a water-filled area determined to exist ahead, serving as the input for subsequent avoidance decisions.

[0047] Specifically, the environmental perception and target recognition mentioned above in this solution are as follows: obtain road environment information in front of the autonomous vehicle, identify polygonal areas of puddles, and simultaneously detect traffic participants (pedestrians and non-motorized vehicles) around the vehicle.

[0048] Specifically, the model primarily employs a joint multi-task network based on candidate bounding boxes and local image cropping. The specific calculation process is as follows:

[0049] 1. Multimodal feature input: In addition to the current RGB image, the model also combines the temporal semantic segmentation results (the historical feature maps at times t, t-2, and t-4 are extracted and stitched together).

[0050] 2. ROI cropping: Based on the given pre-selected bounding box, a local patch is extracted from the feature map and the boundary is expanded outward to a certain extent;

[0051] 3. Multi-task prediction branch: After a series of convolutions and pooling processes, the model outputs four dimensions for each patch: a. Classification: Determines whether there is a puddle in the bounding box (2-class classification). b. Segmentation: Generates pixel-level masks inside the bounding box, marking which pixels are puddles. c. Regression: Predicts the percentage of the local bounding box area occupied by the puddle. d. Uncertainty: Evaluates the model's confidence in the current prediction.

[0052] Specifically, the model captures puddles of varying sizes by generating candidate boxes of multiple scales across the entire image; furthermore, data augmentation mechanisms such as cropping and scaling are introduced during training to enhance robustness to size variations. During prediction, it not only provides the category but also accurately regresses the proportion of the puddle area.

[0053] Specifically, in terms of network architecture, all puddles are processed through the same convolutional network branch without any hard splitting. However, to better handle the dynamic changes of puddles (such as splashing water), temporal feature map fusion can be added. The model input intentionally concatenates three consecutive historical segmentation maps (t, t-2, t-4, where t represents time). Through this superposition of multi-frame temporal information, the model can capture the dynamic temporal information features of "splashing water and water spreading" when a vehicle drives over a puddle, thereby better identifying which areas have the physical characteristics of puddles. Even if the size of the splashes differs, the model can use the differences in temporal features to make a comprehensive judgment.

[0054] Specifically, the above solution goes beyond simply outlining; it employs a more refined "local pixel-level masking" approach. While the first step provides a rough rectangular pre-selection of the puddle, the final output branch performs segmentation (by amplifying the features through transposed convolution). This is similar to instance segmentation, classifying each pixel within the local bounding box and determining pixel-by-pixel where the puddle is and where the puddle is, rather than simply outlining with quadrilaterals.

[0055] Specifically, the model is designed with a safety mechanism based on uncertainty learning:

[0056] (1) Uncertainty prediction branch: The network outputs a single value (between 0 and 1 after Sigmoid activation) to predict the unreliability of the current judgment;

[0057] (2) Loss weight adjustment: When calculating the classification loss, if the rainy day scene is blurry, the network can output a very high "uncertain" classification, thereby dynamically reducing the penalty weight of this part of the occluded / blurred samples during training, and avoiding the model from "guessing randomly" and learning bad features;

[0058] (3) High-confidence filtering: During the verification and final evaluation output stages, the code not only measures conventional metrics, but also specifically evaluates the precision-recall curves under different confidence levels. This ensures that downstream modules only use water puddle predictions with extremely high model certainty (extremely low uncertainty) to cope with low visibility conditions in rainy weather.

[0059] Specifically, traffic participants are detected using traditional LiDAR (Light Detection and Ranging) technology. LiDAR can detect the position, speed, and shape of obstacles, which is sufficient for accurate obstacle classification. Existing detection methods can be used, or the aforementioned identification model can be employed for analysis. The specific analysis process is similar to the analysis of whether water accumulation areas exist, and will not be elaborated upon here.

[0060] In some embodiments, controlling the vehicle to perform an avoidance maneuver can be achieved through at least one of the following steps: controlling the vehicle to decelerate based on the area of ​​the water accumulation area, wherein the size of the water accumulation area is negatively correlated with the speed of the vehicle; controlling the vehicle to bypass the water accumulation area based on the area of ​​the water accumulation area, wherein the size of the water accumulation area is positively correlated with the lateral avoidance offset of the vehicle.

[0061] In this solution, the deceleration intensity and lateral avoidance range are adaptively adjusted according to the size of the water accumulation area, so as to achieve more precise control of vehicle movements and optimize ride comfort.

[0062] Specifically, when an autonomous vehicle approaches a puddle, the longitudinal deceleration intensity is dynamically adjusted based on the area ratio of the puddle (the percentage of the local candidate box) output by the perception module. The larger the area ratio, the wider the water coverage and the higher the water volume, and the stronger the potential splash risk, so a stronger deceleration is applied.

[0063] When conditions for lateral detour are determined to be met, the lateral avoidance offset is dynamically adjusted based on the area percentage of the waterlogged area. The smaller the area percentage (e.g., <10%), only a basic offset (0.8m) is applied. For a medium area percentage (10%–30%), the offset increases linearly to 1m. When the area percentage is greater than 30%, the offset is increased to the maximum allowable value of 2.5m. This offset serves as the target displacement input for the lateral repulsive field. The path planner, combined with vehicle dynamics constraints (lateral acceleration ≤1.5m / s², jerkiness ≤2.0m / s³), generates a smooth trajectory. The positive correlation between area and offset ensures that large areas of waterlogged area are avoided more thoroughly, while small areas of waterlogged area are only slightly adjusted to avoid unnecessary lateral movements.

[0064] In the specific implementation process, controlling the deceleration of the aforementioned vehicle can be achieved through the following steps: obtaining a safe speed and a safe distance, wherein the safe speed is a preset upper limit for safely passing through the waterlogged area, and the safe distance is a preset minimum braking distance; obtaining the speed difference between the aforementioned vehicle and the aforementioned traffic participants to obtain a relative speed difference; obtaining the distance between the aforementioned vehicle and the aforementioned waterlogged area to obtain a longitudinal distance; calculating the vehicle speed based on the aforementioned safe speed, the aforementioned safe distance, the aforementioned relative speed difference, and the aforementioned longitudinal distance; and controlling the deceleration of the aforementioned vehicle based on the aforementioned vehicle speed, wherein the aforementioned vehicle speed is the speed required to suppress water splashing.

[0065] This solution utilizes the spatial-velocity coupling relationship between traffic participants and the flooded area to achieve dynamic and adaptive longitudinal deceleration control, ensuring that unnecessary deceleration is minimized while maintaining safety and improving passenger comfort.

[0066] Specifically, the splash prevention longitudinal control decision of this scheme is as follows: when a puddle is detected and the aforementioned traffic participants are within the set radiation threshold range of the puddle, the splash prevention longitudinal deceleration (i.e., vehicle speed) is calculated based on the lateral distance and relative speed between the traffic participants and the puddle; the splash prevention longitudinal deceleration is used to replace the default comfort deceleration to generate a longitudinal speed planning instruction.

[0067] Specifically, the splash-avoidance longitudinal control formula based on traffic participants is described below. When an autonomous vehicle approaches a puddle, if the system detects traffic participants (pedestrians or non-motorized vehicles) within a predetermined lateral range (e.g., within 10 meters) and they are located in front of the vehicle (relative to longitudinal distance)... If the value is >0, the system will trigger a uniform deceleration calculation to prevent splashing. For example... Figure 3 As shown, the specific calculation formula and steps are as follows:

[0068] Calculate the relative velocity difference between the vehicle and the obstacle (i.e., the longitudinal velocity difference vector between the vehicle and other road users, representing the longitudinal velocity difference vector (initial velocity) between the vehicle and other road users)). , , This represents the velocity vector of the vehicle. This represents the projection of the traffic participant's velocity vector onto the ADV velocity vector.

[0069] Calculate the uniform deceleration of the anti-splash theory Based on kinematic formulas To reduce the relative velocity to the safe speed limit of the puddle before approaching it, the theoretical deceleration is calculated:

[0070] ;

[0071] in, The preset maximum speed limit for passing through water pits (the upper limit of the difference between the maximum speed of a vehicle and the maximum speed of other road users passing through water pits (final speed)). A longitudinal safety buffer distance is reserved (e.g., 5.0m). Indicates the final velocity. Indicates the initial velocity, Indicates acceleration. Indicates the distance traveled by an object. This indicates the distance from the puddle to the vehicle. To represent a very small positive number, such as 10 -9 The purpose is to avoid division by zero errors. Applying the kinematic formulas above, we can calculate... To reduce the relative speed to the safe speed limit of the puddle before approaching it, the theoretical deceleration required is the vehicle speed. Then, the vehicle is controlled to gradually reduce its current speed to the vehicle speed.

[0072] Safety acceleration cutoff and activation (The theoretical deceleration required to reduce the relative speed to the safe speed limit of the puddle before approaching it), the calculated theoretical deceleration is truncated at upper and lower limits to ensure smooth deceleration without exceeding the vehicle's physical limits:

[0073] ,in, This is the lower limit of strong deceleration (e.g., -4.0 m / s²). This is the lower limit for weak deceleration (e.g., -2.0 m / s²). When When the absolute value of the acceleration is greater than the currently planned minimum acceleration, and the vehicle's trajectory does indeed intersect with the edge of the puddle, use... Replace the original planned acceleration to generate the final longitudinal control command. We will... The result is truncated at [ , Between (for example) The calculated value is -5 m / s². If it is -4m / s², then (with a speed of -4m / s²), ensuring smooth deceleration without exceeding the vehicle's physical limits.

[0074] In some embodiments, controlling the vehicle to bypass the flooded area can be achieved through the following steps: extracting the polygonal outline of the flooded area to obtain the outline of the flooded area; designating the area enclosed by the outline of the flooded area as a no-driving zone, replanning the driving path of the vehicle to obtain the planned driving path; and controlling the vehicle to drive along the planned driving path so that the vehicle bypasses the flooded area.

[0075] This solution uses precise polygonal contour modeling and guidance to achieve safe and comfortable detours around waterlogged areas, avoiding path abrupt changes, boundary crossings, or excessive deviations caused by rigid avoidance.

[0076] Specifically, the water accumulation area identification result includes a pixel-level mask. This mask is converted into a polygonal contour in two-dimensional space. A contour approximation algorithm (such as Douglas-Peucker) is used to simplify the original pixel boundaries, generating a convex hull polygon composed of 5–12 vertices as the accurate geometric representation of the water accumulation area. The traverser combines vehicle dynamics constraints and road drivable boundaries, and uses a sampling optimization algorithm (such as RRT) to generate a continuous, smooth, collision-free lateral offset trajectory.

[0077] The output trajectory consists of a set of discrete path points, with the lateral offset changing smoothly with the longitudinal distance. This trajectory is input into the vehicle's underlying controller (such as MPC), and the steering actuator drives the front wheels to follow the trajectory, ensuring that the vehicle's center of gravity is always within ±0.15m of the planned path, completing a contactless detour around the waterlogged area, and maintaining continuous lateral acceleration throughout the detour to avoid passenger discomfort.

[0078] In the specific implementation process, after extracting the polygonal outline of the waterlogged area to obtain the waterlogged area outline, the method further includes the following steps: obtaining the drivable boundary of the road; obtaining the distance between the leftmost edge point of the waterlogged area outline and the drivable boundary to obtain the edge spacing, or obtaining the distance between the rightmost edge point of the waterlogged area outline and the drivable boundary to obtain the edge spacing; if the edge spacing is less than the width of the vehicle, the waterlogged area outline is clipped in the direction from the outside to the inside to obtain the clipped waterlogged area outline; the area enclosed by the clipped waterlogged area outline is taken as the prohibited driving area, and the driving path of the vehicle is replanned to obtain the planned driving path.

[0079] This solution effectively prevents vehicles from overshooting the boundary when avoiding floodwaters by using a dynamic trimming mechanism based on drivable boundaries, significantly improving the safety and reliability of vehicles in narrow or congested roads.

[0080] Specifically, the safety boundary lateral clipping decision of this scheme is as follows: obtain the drivable area boundary of the current road; perform intersection clipping between the above-mentioned puddle polygon area and the drivable area boundary, remove the water accumulation area outside the boundary, and obtain the effective blocking polygon.

[0081] Specifically, this solution is not simply about "finding a puddle-free area and leaving," but rather about achieving this through a "flexible avoidance" mechanism based on safety buffers and artificial potential fields.

[0082] The handling of puddles involves the following judgments and strategies:

[0083] Filtering valid puddles: First, filter the puddles. Only shallow puddles (the output of Perception will have a depth indicator) will be entered into the avoidance logic; otherwise, they will be ignored.

[0084] Determining direction: The current code primarily supports avoiding obstacles to the left. In scenarios where avoiding obstacles to the right is necessary (such as when the left side of a puddle is much wider than the right side), we do not perform any lateral avoidance maneuvers, but only apply longitudinal braking.

[0085] Repulsive force avoidance: In decision-making, two layers of repulsive forces are applied to puddles that are determined to need to be avoided.

[0086] First safety layer: Ensures the minimum safe buffer distance between the main vehicle and the puddle, allowing for a relatively large degree of steering;

[0087] The second ideal level: It is hoped that the main vehicle can achieve a more ideal distance to avoid water pits without making sudden steering wheel movements, and ensure the comfort of passengers in the vehicle under the constraints of lateral acceleration and abrupt movement.

[0088] Therefore, it doesn't rigidly prohibit passage through a certain area, but rather applies a lateral offset force, causing the vehicle to try to avoid it within force and comfort constraints. Ultimately, the repulsive force we apply needs to pass through a path solver (which comprehensively considers the repulsive forces applied by all decision-makers as well as the vehicle's own steering limitations) to determine the final path the vehicle will take.

[0089] Specifically, if the road is entirely covered in puddles, or if the puddles occupy a very large horizontal area, making it impossible for the vehicle to avoid them, the system will choose "not to avoid" and drive directly through. This is because there are several logical constraints in the code specifically designed to prevent "driving off the road due to excessive avoidance":

[0090] To determine the drivable area, we will read the distance between the current lane and the left boundary of the drivable area. If the maximum lateral boundary of the puddle... It is very large, and the space left on the left is even smaller than the width of the main vehicle itself, which means that going to the left will definitely exceed the driving boundary (go off the road).

[0091] Conflict clipping: When a puddle overlaps with the drivable boundary of the lane, this function will clip the polygon of the puddle that extends beyond the road boundary.

[0092] The safety clamp adds a lateral offset limit to prevent the lateral offset from being calculated to be too large, which would make the evasion action too violent.

[0093] Specifically, such as Figure 4As shown, to prevent vehicles from driving off the road boundary when avoiding puddles, this solution performs joint intersection calculation and bidirectional clipping of the puddle's geometric polygon with the left and right drivable boundaries of the current road.

[0094] 1. Determine the path sampling points Lateral offset limits of the left and right boundaries:

[0095] Set the center line of the vehicle path as the reference (lateral offset L=0, positive on the left and negative on the right).

[0096] Left safety boundary limit : ,in:

[0097] This represents the physical lateral offset of the left edge of the road at this location. A safety buffer is reserved on the left side (usually taken as half or the full width of the vehicle).

[0098] Right-side safety boundary limit : ,in:

[0099] This is the physical lateral offset (usually a negative value) of the right edge of the road at this location. A safety buffer is reserved on the right side.

[0100] 2. Calculate the lateral offset of the intersection point of the puddle polygon and the normal direction, and calculate the boundary line segment of the puddle polygon and the normal direction of the vehicle path ( The intersection of ( ). Same sampling point Find the maximum lateral displacement of the puddle on that cross section. (Leftmost side of the puddle) and minimum lateral offset (The far right of the puddle).

[0101] 3. Two-way out-of-bounds clipping and intersection reconstruction:

[0102] Perform boundary checks and clipping on both sides of the puddle polygon: Clipping for left-side boundary violations: If... And the main body of the water pit is still within the boundary (i.e.) This indicates that the puddle has spread to the left and crossed the boundary. At this point, the original leftmost intersection point is discarded, and a new boundary control point is generated at the left safe boundary. :

[0103] ;

[0104] Right-side out-of-bounds cropping: If And the main body of the water pit is still within the boundary (i.e.) This indicates that the puddle has spread to the right and crossed the boundary. At this point, the original rightmost intersection point is discarded, and a new boundary control point is generated at the rightmost safe boundary. :

[0105] ;

[0106] Safe zone reserved: if and This indicates that the puddle is entirely within the safety boundary, preserving the original intersection point. By traversing the entire path, the resulting cropped path is... , Connect the original intersection points in sequence with the compliant ones to form a new "effective puddle polygon that has been safely clipped by the left and right boundaries in both directions", which serves as the sole geometric input for subsequent obstacle repulsion force calculations.

[0107] In some embodiments, the driving path of the vehicle is replanned to obtain the replanned driving path. This can be achieved through the following steps: obtaining the minimum lateral displacement of the vehicle's driving path relative to the centerline to obtain the lateral offset; using the lateral offset as a constraint, the driving path of the vehicle is replanned to obtain the replanned driving path.

[0108] This solution uses a dynamic lateral offset truncation mechanism based on longitudinal distance to achieve "loose at a distance and tight at a close distance" control of avoidance actions. This improves ride comfort and path smoothness while ensuring safety, and effectively prevents loss of control or passenger motion sickness caused by sudden close-range avoidance.

[0109] Specifically, the layered repulsive force detour planning of this scheme is as follows: Based on the above effective blocking polygons, a double-layered lateral repulsive force field of safety layer and ideal layer is constructed, and the upper limit of lateral offset is used for truncation (to avoid excessive lateral avoidance and resulting in unsafety), generating the final lateral avoidance trajectory.

[0110] Specifically, there are drivable area judgment (to prevent encroachment on the oncoming lane), conflict trimming (if encroachment on the oncoming lane is caused by a puddle, we will trim the puddle to a smaller size to prevent encroachment on the oncoming lane due to the puddle), and safety clamping (even if encroachment on the oncoming lane is caused, we will not encroach too much because we have a lateral offset limit).

[0111] Specifically, before initiating a water pit detour decision, it is necessary to make a two-way judgment on the current detour space in combination with the left and right drivable boundaries to determine whether the detour conditions are met and the optimal detour direction.

[0112] 1. Insufficient space for detour to the left:

[0113] If the system tends to veer to the left around the puddle, the space from the right edge of the puddle to its left boundary needs to be checked. This includes the maximum lateral offset of the puddle. Maximum lateral offset from the left physical boundary satisfy:

[0114] , The width of the vehicle is indicated, meaning the puddle almost completely covers the drivable area on the left, leaving insufficient space for the vehicle. In this situation, the system prohibits the vehicle from navigating to the left.

[0115] 2. Insufficient space for detour to the right:

[0116] If the system tends to veer to the right around the puddle, the space between the left edge and right boundary of the puddle needs to be checked. The minimum lateral offset of the puddle... (Usually a negative number or a small positive number) Minimum lateral offset from the right physical boundary (Negative numbers) satisfy:

[0117] This indicates that the puddle almost completely covers the drivable area on the right, and there is no enough space for the vehicle to pass. The system then prohibits the vehicle from going right.

[0118] 3. Invalidity determination for boundary crossing and offset:

[0119] like (That is, the entire puddle is to the left of the centerline), the vehicle can maintain its centerline position or make minor adjustments; there is no need to activate high-priority lateral avoidance. If (That is, the entire puddle is to the right of the center line and is more than one vehicle width away from the center line), indicating that the current lane center trajectory itself is safe. Under the above circumstances, the system terminates the strong lateral avoidance decision to prevent the vehicle from swaying unnecessarily; if there is no space to maneuver on either side, the system automatically downgrades to the "pure longitudinal speed reduction and splash prevention strategy" in Example 1.

[0120] Furthermore, when dealing with soft obstacles such as puddles, the solution in this application, in order to limit the severity of the vehicle's avoidance maneuvers, such as... Figure 5 As shown, a piecewise linear function is introduced to strongly constrain the maximum lateral offset (L_limit) of the avoidance trajectory:

[0121] (1) Close-range segment (distance) , It is the longitudinal distance from the front of the vehicle to the nearest edge of the polygon in the puddle. (Near distance threshold):

[0122] ;

[0123] This indicates the offset, meaning that when the distance to the puddle is relatively close (e.g., within 20 meters), the maximum permissible lateral offset is strictly locked at the base threshold. Within a 2.5-meter radius, avoid sudden steering wheel movements.

[0124] (2) Transition extension section (distance) ):

[0125] As distance It increases linearly with the increase of, from Gradually relax to .in, Allow for a widening factor for long distances (e.g., 4 times).

[0126] (3) Long-distance segment (distance) ):

[0127] ;

[0128] That is, when the distance from the water pit is relatively far (such as more than 80 meters), the maximum lateral offset is relaxed to 10.0 meters.

[0129] This segmented truncation model ensures that when a vehicle encounters a sudden, close-range puddle, the lateral avoidance range is strictly limited, prioritizing preventing loss of control. When a puddle is detected at a distance, the planning algorithm is allowed to find a wider and gentler detour route, balancing vehicle safety and passenger comfort.

[0130] In summary, the solution proposed in this application identifies puddles and surrounding traffic participants; assesses splash risk; and applies enhanced deceleration to reduce longitudinal speed when splash risk exists. The aforementioned lateral detour decision-making process includes calculating the intersection of the puddle polygon with the drivable road boundary, removing invalid puddle areas outside the boundary, and then calculating the lateral blocking range (BLR). Lateral detours utilize a repulsive field, and the lateral offset caused by the repulsive force is forcibly truncated and limited by a piecewise linear function based on the longitudinal distance S.

[0131] This application also provides a vehicle driving control device. It should be noted that the vehicle driving control device of this application can be used to execute the vehicle driving control method provided in this application. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0132] The following describes the vehicle driving control device provided in the embodiments of this application.

[0133] Figure 6 This is a structural block diagram of a vehicle driving control device according to an embodiment of this application. Figure 6As shown, the device includes:

[0134] The first acquisition unit 10 is used to acquire road information of the road on which the vehicle is currently traveling. The road information includes whether there is a water accumulation area in front of the vehicle and whether there are traffic participants in the area where the vehicle is located. The traffic participants include at least one of the following: pedestrians and non-motorized vehicles.

[0135] The control unit 20 is configured to control the vehicle to perform an avoidance operation when there is a water accumulation area in front of the vehicle and traffic participants are present in the area where the vehicle is located, wherein the avoidance operation includes at least one of the following: controlling the vehicle to decelerate or controlling the vehicle to bypass the water accumulation area.

[0136] In this embodiment, when a puddle is detected in front of the vehicle and there are pedestrians or non-motorized vehicles within the lateral influence range of the puddle, an avoidance operation is actively initiated. Deceleration can reduce the kinetic energy of the vehicle when passing through the puddle and the speed at which the tires throw water, thereby significantly reducing the height and range of water splash. Bypassing the puddle allows the vehicle body to avoid the water accumulation area, further reducing the splash direction, thereby preventing water from splashing onto surrounding pedestrians or non-motorized vehicles.

[0137] In the specific implementation process, the first acquisition unit includes an acquisition module, a calling module, a training module, and a processing module. The acquisition module is used to acquire an image of the road on which the vehicle is currently traveling, thus obtaining a driving image. The calling module is used to call a recognition model, wherein the recognition model is one of a CNN model, an RNN model, or a Transformer model. The training module is used to form a training set by combining historical driving images and the corresponding water accumulation labels of the historical driving images, and to train the recognition model using the training set to obtain a water accumulation area recognition model. The water accumulation labels are the presence or absence of water accumulation areas and water accumulation area selection boxes in the historical driving images of the training set. The processing module is used to input the driving image into the water accumulation area recognition model to obtain the recognition result corresponding to the driving image. The recognition result is whether there is a water accumulation area in front of the vehicle.

[0138] This solution employs a multi-task-based deep learning model to model the local details of the water accumulation area, thereby enabling more accurate identification of the water accumulation area and more accurately determining whether there are puddles in front of the vehicle.

[0139] In some embodiments, the control unit includes a first control module and a second control module. The first control module is used to control the vehicle to decelerate based on the area of ​​the water accumulation area, wherein the area of ​​the water accumulation area and the speed of the vehicle are negatively correlated. The second control module is used to control the vehicle to bypass the water accumulation area based on the area of ​​the water accumulation area, wherein the area of ​​the water accumulation area and the lateral avoidance deviation of the vehicle are positively correlated.

[0140] In this solution, the deceleration intensity and lateral avoidance range are adaptively adjusted according to the size of the water accumulation area, so as to achieve more precise control of vehicle movements and optimize ride comfort.

[0141] In the specific implementation process, the first control module includes a first acquisition submodule, a second acquisition submodule, a third acquisition submodule, and a first control submodule. The first acquisition submodule is used to acquire a safe speed and a safe distance, wherein the safe speed is a preset upper limit for safely passing through the waterlogged area, and the safe distance is a preset minimum braking distance. The second acquisition submodule is used to acquire the speed difference between the vehicle and the traffic participants to obtain a relative speed difference. The third acquisition submodule is used to acquire the distance between the vehicle and the waterlogged area to obtain a longitudinal distance. The first control submodule is used to calculate the vehicle speed based on the safe speed, the safe distance, the relative speed difference, and the longitudinal distance, and control the vehicle to decelerate based on the vehicle speed, wherein the vehicle speed is the speed required to suppress water splashing.

[0142] This solution utilizes the spatial-velocity coupling relationship between traffic participants and the flooded area to achieve dynamic and adaptive longitudinal deceleration control, ensuring that unnecessary deceleration is minimized while maintaining safety and improving passenger comfort.

[0143] In some embodiments, the second control module includes an extraction submodule, a planning submodule, and a second control submodule. The extraction submodule is used to extract the polygonal outline of the waterlogged area to obtain the outline of the waterlogged area. The planning submodule is used to designate the area enclosed by the outline of the waterlogged area as a no-driving area and replan the driving path of the vehicle to obtain the planned driving path. The second control submodule is used to control the vehicle to drive according to the planned driving path so that the vehicle bypasses the waterlogged area.

[0144] This solution uses precise polygonal contour modeling and guidance to achieve safe and comfortable detours around waterlogged areas, avoiding path abrupt changes, boundary crossings, or excessive deviations caused by rigid avoidance.

[0145] In the specific implementation process, the above-mentioned device further includes a second acquisition unit, a third acquisition unit, a trimming unit, and a planning unit. The second acquisition unit is used to extract the polygonal outline of the water accumulation area and obtain the drivable boundary of the road after obtaining the outline of the water accumulation area. The third acquisition unit is used to obtain the distance between the leftmost edge point of the outline of the water accumulation area and the drivable boundary to obtain the edge spacing, or to obtain the distance between the rightmost edge point of the outline of the water accumulation area and the drivable boundary to obtain the edge spacing. The trimming unit is used to trim the outline of the water accumulation area in a direction from the outside to the inside when the edge spacing is less than the width of the vehicle to obtain the trimmed outline of the water accumulation area. The planning unit is used to take the area surrounded by the trimmed outline of the water accumulation area as the prohibited driving area and replan the driving path of the vehicle to obtain the planned driving path.

[0146] This solution effectively prevents vehicles from overshooting the boundary when avoiding floodwaters by using a dynamic trimming mechanism based on drivable boundaries, significantly improving the safety and reliability of vehicles in narrow or congested roads.

[0147] In some embodiments, the planning submodule is used to obtain the minimum lateral displacement of the vehicle's driving path relative to the centerline, thus obtaining the lateral offset; the planning submodule is used to replan the vehicle's driving path using the lateral offset as a constraint, thus obtaining the planned driving path.

[0148] This solution uses a dynamic lateral offset truncation mechanism based on longitudinal distance to achieve "loose at a distance and tight at a close distance" control of avoidance actions. This improves ride comfort and path smoothness while ensuring safety, and effectively prevents loss of control or passenger motion sickness caused by sudden close-range avoidance.

[0149] The aforementioned vehicle driving control device includes a processor and a memory. The first acquisition unit and control unit, etc., are all stored as program units in the memory, and the processor executes the program units stored in the memory to achieve the corresponding functions. All of the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0150] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters can address the problem in existing technologies where vehicles driving through puddles easily splash water onto pedestrians or non-motorized vehicles.

[0151] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0152] This invention provides a computer-readable storage medium including a stored program, wherein the program, when running, controls the device containing the computer-readable storage medium to execute the vehicle driving control method.

[0153] This invention provides a processor for running a program, wherein the program executes the vehicle driving control method.

[0154] This invention provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps. The device described herein may be a server, PC, PAD, mobile phone, etc.

[0155] This application also provides a computer program product that, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps.

[0156] This application also provides a vehicle including one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include methods for performing any of the above-described vehicle driving control methods.

[0157] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0158] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0160] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0162] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0163] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0164] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0165] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0166] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0167] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A vehicle driving control method, characterized in that, include: Obtain road information of the current road the vehicle is traveling on, wherein the road information includes whether there is a water accumulation area ahead of the vehicle and whether there are traffic participants in the area where the vehicle is located, wherein the traffic participants include at least one of the following: pedestrians and non-motorized vehicles; When there is a waterlogged area in front of the vehicle and the traffic participants are present in the area where the vehicle is located, the vehicle is controlled to perform an avoidance operation, wherein the avoidance operation includes at least one of the following: controlling the vehicle to decelerate, or controlling the vehicle to bypass the waterlogged area.

2. The method according to claim 1, characterized in that, Obtain road information for the current route the vehicle is traveling on, including: Obtain an image of the road the vehicle is currently traveling on to obtain a driving image; Invoke the recognition model, wherein the recognition model is one of the CNN model, RNN model, and Transformer model; A training set is formed by combining historical driving images and the corresponding water accumulation labels of the historical driving images. The recognition model is trained using the training set to obtain a water accumulation area recognition model. The water accumulation label is the presence or absence of a water accumulation area and a selection box of the water accumulation area in the historical driving images of the training set. The driving image is input into the water accumulation area recognition model to obtain the recognition result corresponding to the driving image, wherein the recognition result is whether there is a water accumulation area in front of the vehicle.

3. The method according to claim 1, characterized in that, Controlling the vehicle to perform an avoidance maneuver includes at least one of the following: The vehicle speed is controlled to decelerate based on the area of ​​the water accumulation area, wherein the size of the water accumulation area and the speed of the vehicle are negatively correlated. Based on the area of ​​the waterlogged area, the vehicle is controlled to bypass the waterlogged area, wherein the size of the waterlogged area and the lateral avoidance offset of the vehicle are positively correlated.

4. The method according to claim 3, characterized in that, Controlling the vehicle to decelerate includes: Obtain a safe speed and a safe distance, wherein the safe speed is a preset upper limit for safely passing through the waterlogged area, and the safe distance is a preset minimum braking distance; The speed difference between the vehicle and the traffic participant is obtained to obtain the relative speed difference; The distance between the vehicle and the flooded area is obtained to determine the longitudinal distance; The vehicle speed is calculated based on the safe speed, the safe distance, the relative speed difference, and the longitudinal distance. The vehicle speed is then controlled to decelerate, wherein the vehicle speed is the speed required to suppress water splashing.

5. The method according to claim 3, characterized in that, Controlling the vehicle to bypass the flooded area includes: Extract the polygonal outline of the water accumulation area to obtain the outline of the water accumulation area; The area enclosed by the outline of the waterlogged area is designated as a no-driving zone, and the driving path of the vehicle is replanned to obtain the planned driving path. The vehicle is controlled to travel along the planned route so that it bypasses the flooded area.

6. The method according to claim 5, characterized in that, After extracting the polygonal outline of the water accumulation area to obtain the outline of the water accumulation area, the method further includes: Obtain the drivable boundaries of the road; The distance between the leftmost edge point of the water accumulation area outline and the drivable boundary is obtained to obtain the edge spacing; or the distance between the rightmost edge point of the water accumulation area outline and the drivable boundary is obtained to obtain the edge spacing. If the edge spacing is less than the width of the vehicle, the outline of the water accumulation area is cut out from the outside to the inside to obtain the cut outline of the water accumulation area. The area enclosed by the cut-out outline of the waterlogged area is designated as the no-driving zone, and the vehicle's driving path is replanned to obtain the planned driving path.

7. The method according to claim 5, characterized in that, The vehicle's travel path is replanned to obtain the planned travel path, including: The minimum lateral displacement of the vehicle's travel path relative to the centerline is obtained to determine the lateral offset. Using the lateral offset as a constraint, the vehicle's driving path is replanned to obtain the planned driving path.

8. A vehicle driving control device, characterized in that, include: The first acquisition unit is used to acquire road information of the road on which the vehicle is currently traveling, wherein the road information includes whether there is a water accumulation area in front of the vehicle and whether there are traffic participants in the area where the vehicle is located, wherein the traffic participants include at least one of the following: pedestrians and non-motorized vehicles; A control unit is configured to control the vehicle to perform an avoidance maneuver when there is a water accumulation area in front of the vehicle and the traffic participant is present in the area where the vehicle is located, wherein the avoidance maneuver includes at least one of the following: controlling the vehicle to decelerate, or controlling the vehicle to bypass the water accumulation area.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the vehicle driving control method according to any one of claims 1 to 7.

10. A vehicle, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing the vehicle driving control method according to any one of claims 1 to 7.

Images