Vehicle control methods, devices and vehicles

Abstract

This application relates to a vehicle control method, device, and vehicle. The method includes: generating a speed base curve corresponding to the current parking path based on historical parking memory data; determining environmental risk coefficients corresponding to each path mileage on the current parking path using a pre-established environmental risk weighting function based on road environment information on the current parking path; obtaining a driving style coefficient, wherein the driving style coefficient represents the driver's preference for adjusting vehicle speed; and correcting the speed base curve based on the environmental risk coefficient and the driving style coefficient to determine a target speed curve, thereby controlling the vehicle to perform memory parking according to the target speed curve. This application can improve the safety of the vehicle and the user experience during memory parking in different scenarios.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and specifically to a vehicle control method, apparatus, vehicle, and computer-readable storage medium. Background Technology

[0002] With the rapid development of autonomous driving technology, memory parking, with its advantages of automatically reproducing frequently used parking routes and eliminating the need for manual operation, has become a key technology for improving vehicle intelligence and user convenience. The control of longitudinal speed during memory parking is a core element determining the user experience, directly impacting safety, comfort, and user acceptance. Therefore, improving vehicle safety and user experience during memory parking is a pressing issue that needs to be addressed. Summary of the Invention

[0003] This application provides a vehicle control method, apparatus, vehicle, and computer-readable storage medium that can improve vehicle safety and user experience during memory parking.

[0004] This application provides a vehicle control method, the method including: Based on historical parking data, a speed baseline curve is generated for the current parking path; the speed baseline curve represents the relationship between path mileage and vehicle speed. Using a pre-established environmental risk weighting function, the environmental risk coefficients corresponding to each path mileage on the current parking path are determined based on the road environment information on the current parking path; whereby the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficients; Obtain the driving style coefficient; where the driving style coefficient represents the driver's preference for adjusting the vehicle speed. The speed base curve is corrected based on the environmental risk coefficient and driving style coefficient to determine the target speed curve, so as to control the vehicle to perform memory parking according to the target speed curve.

[0005] Based on the aforementioned technical means, the speed baseline curve generated by learning from the driver's historical data makes the speed rhythm, cornering habits, and straight-line acceleration of the automatic parking system highly similar to the driver's own, greatly enhancing the user's sense of trust. Furthermore, by correcting the speed baseline curve through a continuously changing environmental risk weighting function and driving style coefficient, the system can proactively adapt to the needs of users with different driving preferences while meeting the safety requirements of autonomous driving. This can improve the safety of the vehicle in different scenarios during the memory parking process and enhance the user's experience.

[0006] In some embodiments, the speed base curve is corrected based on environmental risk coefficients and driving style coefficients to determine the target speed curve, including: The speed base curve is corrected based on the environmental risk coefficient and driving style coefficient to determine the first speed curve; The speed base curve is corrected based on the driving style coefficient to determine the second speed curve; The first and second velocity curves are weighted and fused to determine the target velocity curve. Based on the aforementioned technical means, by generating a first speed curve that takes into account both environmental risks and driving style, and a second speed curve that only reflects driving preferences, and dynamically weighting and fusing them, it is possible to prioritize safe deceleration in high-risk areas and fully restore the user's habitual speed rhythm in low-risk areas. This satisfies the safety requirements of advanced autonomous driving while preserving a natural, smooth, and personalized parking control feel to the greatest extent, effectively enhancing the user's trust in the memory parking system and improving the comfort experience. In some embodiments, road environment information includes one or more of road width information, road slope information, road curvature information, and intersection occlusion information; the environmental risk weighting function includes one or more of the width sub-function, slope sub-function, curvature sub-function, and intersection occlusion sub-function.

[0007] Based on the aforementioned technical means, by modeling key environmental factors such as road width, slope, curvature, and intersection obstruction as independent sub-functions and weighting and integrating them into a unified environmental risk coefficient, it is possible to conduct a refined and continuous quantitative assessment of risk sources at different locations on the parking path. This makes speed correction more consistent with actual roads, thereby enabling targeted speed adjustment in memory parking, which can improve the adaptability to complex environments and the accuracy of safety control. In some embodiments, based on road environment information along the current parking path, and using a pre-established environmental risk weighting function, the environmental risk coefficient corresponding to each path mileage along the current parking path is determined, including: Input the road width information into the width sub-function to obtain the first weight value corresponding to the mileage of each path on the current parking path; The road slope information is input into the slope sub-function to obtain the second weight value corresponding to the mileage of each path on the current parking path; The road curvature information is input into the curvature sub-function to obtain the third weight value corresponding to the mileage of each path on the current parking path. Input the intersection occlusion information into the intersection occlusion sub-function to obtain the fourth weight value corresponding to the mileage of each path on the current parking path; The first, second, third, and fourth weight values ​​are combined to obtain the environmental risk coefficient corresponding to each path mileage.

[0008] Based on the aforementioned technical means, by inputting road width, slope, curvature, and intersection obstruction information into corresponding independent sub-functions, the weight values ​​of different risk dimensions under each path mileage are obtained. Then, multi-dimensional fusion is performed to form a unified environmental risk coefficient, which enables precise decoupling and quantitative assessment of the risk sources at each location on the parking path. This avoids the dominance of a single factor or the distortion caused by risk superposition, thereby improving the safety adaptability and speed control rationality of memory parking in scenarios such as narrow roads, steep slopes, sharp bends, and blind spots. In some embodiments, the above method further includes: Determine the status information of target obstacles within the preset detection range; the status information includes the position, speed, and direction of movement of the target obstacles; Based on the state information, determine the relative distance and relative speed between the target obstacle and the vehicle; The relative distance is compared with the preset safety distance to obtain the distance comparison result; Based on the distance comparison results, relative distance, relative speed, and the vehicle's current speed, determine the vehicle's safe target speed.

[0009] Based on the aforementioned technical means, by detecting the position, speed, and direction of the target obstacle in real time, and calculating the relative distance and speed between it and the vehicle, the safe target speed of the vehicle can be dynamically and accurately determined. This enables the memory parking system to have the ability to actively and smoothly adjust speed when encountering dynamic obstacles (such as pedestrians or other vehicles), thus avoiding the risk of sudden braking or collision, ensuring the continuity and comfort of the parking process, and significantly improving driving safety in complex dynamic environments. In some embodiments, the preset safe distance includes a first safe distance and a second safe distance; determining the vehicle's safe target speed based on distance comparison results, relative distance, relative speed, and the vehicle's current speed includes: If the distance comparison result indicates that the relative distance is less than or equal to the first safe distance and greater than the second safe distance, the first safe target speed of the vehicle is determined based on the relative distance and the relative speed; wherein, the first safe target speed is positively correlated with the relative distance, the first safe target speed is negatively correlated with the absolute value of the relative speed, and the first safe distance is greater than the second safe distance; The vehicle's current speed is reduced to a first safe target speed based on the first rate of change of speed.

[0010] Based on the aforementioned technical means, when the relative distance between the vehicle and the obstacle is in the transition range between the first safe distance and the second safe distance, by constructing a first safe target speed that is positively correlated with the relative distance and negatively correlated with the absolute value of the relative speed, and by using a first speed change rate to smoothly reduce speed, it is possible to avoid sudden braking or premature deceleration while ensuring safety. This makes the speed adjustment process more in line with human driving's anticipation and gradual braking habits, thereby taking into account safety, comfort, and parking continuity during dynamic obstacle avoidance. In some embodiments, the preset safe distance includes a second safe distance; determining the vehicle's safe target speed based on distance comparison results, relative distance, relative speed, and the vehicle's current speed includes: If the distance comparison result indicates that the relative distance is less than or equal to the second safe distance, the second safe target speed is determined based on the vehicle's current speed and the relative distance. The vehicle's current speed is reduced to a second safe target speed based on a second speed change rate; wherein the second speed change rate is less than or equal to a preset change rate threshold, and the first speed change rate is less than the second speed change rate.

[0011] Based on the aforementioned technical means, when the relative distance between the vehicle and the target obstacle (such as the vehicle in front or the obstacle) is less than or equal to the second safe distance (i.e., entering a more urgent safe zone), the second safe target speed can be calculated in real time based on the current driving speed and relative distance. By setting a relatively large second speed change rate (but still not exceeding the preset change rate threshold), the vehicle speed can be smoothly reduced to the target speed. This ensures driving safety and avoids collisions while preventing driving discomfort or the risk of rear-end collisions caused by excessive deceleration. Under the premise of meeting the safety braking requirements, the smoothness, ride comfort, and responsiveness of the vehicle driving are maximized. In some embodiments, the above method further includes: The target speed curve is smoothed based on a preset smoothing algorithm to obtain a smoothed target speed curve, so as to control the vehicle to perform memory parking according to the smoothed target speed curve. The smoothed target velocity curve satisfies one or more of the following constraints: On the target velocity curve, the absolute difference between the velocity values ​​corresponding to any two adjacent sampling times is less than or equal to a first preset threshold; on the target velocity curve, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a second preset threshold; on the target velocity curve, the absolute difference between the jerk values ​​corresponding to any two adjacent sampling times is less than or equal to a third preset threshold; and the absolute value of the jerk is less than or equal to a preset jerk threshold.

[0012] Based on the above technical means, a preset smoothing algorithm is used for global smoothing, which ensures the absolute continuity of the final speed curve in the three dimensions of speed, acceleration, and jerk. This eliminates speed jumps and acceleration abrupt changes at the junctions of different road sections and during normal acceleration and deceleration, achieving a shock-free and jerky ride quality and effectively avoiding motion sickness.

[0013] This application provides a vehicle control device, the device comprising: The generation unit is used to generate the speed base curve corresponding to the current parking path based on historical parking data; wherein, the speed base curve represents the correspondence between path mileage and vehicle speed; The determination unit is used to determine the environmental risk coefficient corresponding to each path mileage on the current parking path based on the road environment information on the current parking path using a pre-established environmental risk weighting function; wherein, the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; The acquisition unit is used to acquire the driving style coefficient; wherein, the driving style coefficient represents the driver's preference for adjusting the vehicle speed; The correction unit is used to correct the speed base curve based on the environmental risk coefficient and driving style coefficient, determine the target speed curve, and control the vehicle to perform memory parking according to the target speed curve.

[0014] This application provides a vehicle including a processor and a memory. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the steps in any of the above methods.

[0015] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in any of the above methods.

[0016] This application provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps of any of the above methods.

[0017] The beneficial effects of this application are: (1) The speed base curve generated by learning from the driver's historical data makes the speed rhythm, cornering habits, and straight-line acceleration of the automatic parking highly similar to the owner, which greatly enhances the user's sense of trust. Furthermore, by modifying the speed base curve through the continuously changing environmental risk weighting function and driving style coefficient, it can actively adapt to the needs of users with different driving preferences while meeting the safety requirements of autonomous driving, thereby improving the safety of the vehicle in different scenarios and the user's experience during the memory parking process.

[0018] (2) By generating a first speed curve that takes into account both environmental risks and driving style, and a second speed curve that only reflects driving preferences, and dynamically weighting and fusing them, it can prioritize safe deceleration in high-risk areas and fully restore the user's habitual speed rhythm in low-risk areas. Thus, while meeting the safety requirements of advanced autonomous driving, it can retain the natural, smooth and personalized parking control feel to the greatest extent, effectively improving the user's trust in the memory parking system and the comfort experience.

[0019] (3) By modeling key environmental factors such as road width, slope, curvature, and intersection obstruction as independent sub-functions and weighting and integrating them into a unified environmental risk coefficient, a refined and continuous quantitative assessment of risk sources at different locations on the parking path can be achieved. This makes speed correction more consistent with actual roads, thereby enabling targeted speed adjustment in memory parking and improving adaptability to complex environments and the accuracy of safety control. Furthermore, by inputting road width, slope, curvature, and intersection obstruction information into the corresponding independent sub-functions, the weight values ​​of different risk dimensions under each path mileage are obtained. Then, multi-dimensional fusion is performed to form a unified environmental risk coefficient, enabling a refined decoupling and quantitative assessment of risk sources at each location on the parking path. This avoids distortion caused by a single factor or superimposed risks, thereby improving the safety adaptability and speed control rationality of memory parking in scenarios such as narrow roads, steep slopes, sharp bends, and blind spots. (4) By detecting the position, speed and direction of the target obstacle in real time, and calculating the relative distance and speed between it and the vehicle, the safe target speed of the vehicle can be determined dynamically and accurately. This enables the memory parking system to have active and smooth speed adjustment capability when encountering dynamic obstacles (such as pedestrians and other vehicles), which avoids the risk of sudden braking or collision, and ensures the continuity and comfort of the parking process, significantly improving driving safety and system robustness in complex dynamic environments. (5) When the relative distance between the vehicle and the obstacle is in the transition range between the first safe distance and the second safe distance, by constructing a first safe target speed that is positively correlated with the relative distance and negatively correlated with the absolute value of the relative speed, and by using the first speed change rate to smoothly reduce speed, it is possible to avoid sudden braking or premature deceleration under the premise of ensuring safety, so that the speed adjustment process is more in line with the human driving prediction and gradual braking habits, thereby taking into account safety, comfort and parking continuity in the dynamic obstacle avoidance process. (6) When the relative distance between the vehicle and the target obstacle (such as the vehicle in front or the obstacle) is less than or equal to the second safe distance (i.e., entering a more urgent safe zone), the second safe target speed can be calculated in real time based on the current driving speed and relative distance. The vehicle speed can be smoothly reduced to the target speed by setting a relatively large second speed change rate (but still not exceeding the preset change rate threshold). This ensures driving safety and avoids collisions while avoiding driving discomfort or rear-end collisions caused by excessive deceleration. Under the premise of meeting the safety braking requirements, the smoothness, ride comfort and following ability of the vehicle are maximized.

[0020] (7) A preset smoothing algorithm is used for global smoothing, which ensures the absolute continuity of the final speed curve in the three dimensions of speed, acceleration and jerk. This eliminates speed jumps and acceleration changes at the junctions of different road sections and during normal acceleration and deceleration, achieving a ride quality without shock or jerking, and effectively avoiding motion sickness. Attached Figure Description

[0021] Figure 1 A schematic flowchart of a vehicle control method provided in an embodiment of this application; Figure 2 A schematic diagram of smooth speed baseline curves for different driving styles in various scenarios provided in the embodiments of this application; Figure 3 This is a schematic diagram illustrating the continuous change of the environmental risk weighted field provided in the embodiments of this application; Figure 4 This is a schematic diagram of the overall architecture of the memory parking speed control system provided in the embodiments of this application; Figure 5 This is a schematic diagram of the composition structure of a vehicle control device provided in an embodiment of this application; Figure 6 This is a schematic diagram of the hardware entity of a vehicle provided in an embodiment of this application.

[0022] It should be noted that the terms "first" and "second" mentioned above are only used to distinguish between different options and do not represent the degree of superiority or inferiority of the options or their priority in the implementation process. Detailed Implementation

[0023] The embodiments of this application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be understood that the preferred embodiments are only for illustrating this application and are not intended to limit the scope of protection of this application.

[0024] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0025] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0026] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0027] In this embodiment, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist simultaneously, and object B exists alone.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0029] Currently, with the rapid development of autonomous driving technology, memory parking, as one of the core functions of L2+ level autonomous driving, has become a key technology for improving vehicle intelligence and user convenience due to its advantages of automatically reproducing frequently used parking routes without requiring manual operation. The control of longitudinal speed in memory parking is a crucial aspect determining the user experience, directly impacting safety, comfort, and user acceptance. However, current speed control solutions applied to memory parking generally suffer from one or more technical defects, resulting in a poor user experience and limited applicability.

[0030] Some solutions employ fixed speed points, resulting in rigid and impersonal control. This approach completely ignores the driver's personal habits, leading to mechanical and abrupt vehicle behavior, severe speed jumps during environmental transitions, and poor driving comfort. While it can set safe speed limits for different scenarios (such as "narrow roads" and "intersections"), the target speed undergoes abrupt changes at the boundaries of different environmental labels. This discontinuity in speed commands causes the vehicle to accelerate and decelerate abruptly, resulting in strong jerking shocks, manifested as a noticeable "nodding" or "lifting" sensation. This is far removed from the flexible and adaptive driving style of human drivers, resulting in a poor driving experience.

[0031] Other solutions attempt to enhance the human-like experience by faithfully recording and replicating the driver's speed during the learning phase. However, during demonstrations, drivers may maintain high speeds in high-risk scenarios such as narrow passages, sharp bends, slopes, and intersections with obstructed visibility due to negligence, habit, or over-familiarity with the road conditions. If the system blindly replicates these behaviors, it will result in extremely high risks of scrapes, collisions, or even loss of control, essentially sacrificing safety for experience and failing to meet the needs of diverse users.

[0032] Next, embodiments of this application will introduce several related technologies of the memory parking control method.

[0033] In related technology 1, a method and device for adaptive speed adjustment of memory parking function is provided. It mainly focuses on the priority strategy of speed adjustment, such as brake / accelerator / target detection / scene / manual. However, the strategy is rigid, the actual experience is not adequately considered, there is no human-like baseline, no personalized style, scene weighting and smoothing are insufficient, and the speed will inevitably jump, the impact is large and the comfort is poor when scene is switched.

[0034] Related technology 2 provides a parking method based on millimeter-wave radar and fuzzy control, which focuses on making real-time obstacle avoidance decisions through sensor information. However, it does not conduct a forward-looking and refined environmental risk assessment and label segmentation of the known memory path, nor does it address how to smoothly connect the target speeds of different road segments from a planning perspective. Therefore, it cannot fundamentally solve the speed jump and impact problem caused by environmental switching.

[0035] Related technology 3 provides a general vehicle obstacle avoidance control method, the core of which is to establish a mapping relationship between the real-time distance to obstacles and the safe vehicle speed, which is a typical reactive control method. This method does not utilize the valuable prior information of "global path known" in memory parking, and does not involve optimizing and smoothing the global speed curve of a fixed path. Therefore, it is not suitable for solving speed smoothness planning in memory parking scenarios.

[0036] Related Technology 4 provides a parking control method, device, electronic equipment, and vehicle, focusing on anthropomorphic speed control strategies that attempt to improve the experience by mimicking driver behavior. However, this solution lacks a quantitative assessment of environmental safety risks and a rigid constraint mechanism integrated into the decision-making process. In dangerous scenarios, the system may still reproduce unsafe high-speed driver behavior, making safety risks uncontrollable.

[0037] Related technology five provides an automatic parking control method and vehicle based on user profiles, which optimizes the parking experience from the perspective of building user driving profiles, representing a higher level of personalization strategy. However, this patent does not address the most fundamental technical contradiction in memory parking speed control—that is, how to achieve a smooth and imperceptible transition of target speed between road sections with different safety levels from the underlying control algorithm level. Therefore, it fails to solve the fundamental pain point of acceleration and deceleration shock.

[0038] In summary, controlling the longitudinal speed during memory parking is a core element determining the user experience, directly impacting safety, comfort, and user acceptance. Therefore, improving vehicle safety and user experience during memory parking is a pressing issue that needs to be addressed.

[0039] Based on this, embodiments of this application provide a vehicle control method, the method comprising: generating a speed base curve corresponding to the current parking path based on historical parking data; wherein the speed base curve represents the correspondence between path mileage and vehicle speed; using a pre-established environmental risk weighting function, determining the environmental risk coefficient corresponding to each path mileage on the current parking path based on road environment information on the current parking path; wherein the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; obtaining a driving style coefficient; wherein the driving style coefficient represents the driver's preference for adjusting vehicle speed; and correcting the speed base curve based on the environmental risk coefficient and the driving style coefficient to determine a target speed curve, so as to control the vehicle to perform memory parking according to the target speed curve. In this way, the speed baseline curve generated based on the driver's historical data makes the speed rhythm, cornering habits, and straight-line acceleration of the automatic parking highly similar to the owner's, greatly enhancing the user's sense of trust. Furthermore, by correcting the speed baseline curve through a continuously changing environmental risk weighting function and driving style coefficient, it can proactively adapt to the needs of users with different driving preferences while meeting the safety requirements of autonomous driving, thereby improving the safety of the vehicle in different scenarios during the memory parking process and the user's experience.

[0040] The technical solutions in the embodiments of this application will now be clearly and completely described with reference to the accompanying drawings.

[0041] It should be noted that the vehicle control method provided in the embodiments of this application can be executed by the electronic control unit in the vehicle.

[0042] Figure 1 This is a flowchart illustrating a vehicle control method provided in an embodiment of this application, as shown below. Figure 1 As shown, it may include S101 to S104, wherein: S101, based on historical parking data, generates the speed baseline curve corresponding to the current parking path.

[0043] Here, historical parking data refers to the complete driving trajectory and environmental data collected and stored in real time by the vehicle's underlying system when the user manually drives the vehicle to complete parking route learning (teaching). This data may include, but is not limited to, vehicle motion state data and environmental perception feature data. Vehicle motion state data may include the mileage coordinate sequence of the historical path, the vehicle speed at the corresponding historical location, historical acceleration / deceleration changes, the speed pattern during successful parking, steering wheel angle, gear status, and brake / accelerator pedal opening. Environmental perception feature data may include parking lot environmental elements (such as lane lines, parking spaces, walls, pillars, etc.) collected by sensors such as cameras, ultrasonic radar, and millimeter-wave radar, used for subsequent positioning matching and obstacle avoidance.

[0044] The speed baseline curve represents the relationship between the path mileage (the cumulative driving distance calculated from the parking starting point) and the vehicle speed. In other words, the speed baseline curve can be used to indicate the target baseline driving speed that the vehicle should match when it reaches each specific mileage position (e.g., 50 meters or 100 meters from the starting point) on the memorized route.

[0045] In some embodiments, when a vehicle is parked, memory parking data of the vehicle can be recorded and stored. In the next memory parking process, machine learning algorithms (such as Gaussian process regression and spline fitting) are used to analyze, denoise and generalize the historical memory parking data to generate an anthropomorphic speed base curve with continuously changing smoke path mileage, which can reflect the driver's natural driving rhythm.

[0046] S102, using a pre-established environmental risk weighting function, determine the environmental risk coefficient corresponding to each path mileage on the current parking path based on the road environment information on the current parking path.

[0047] The environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; that is, the environmental risk weighting function is used to indicate the different environmental risk coefficients that different road environments on the parking path may correspond to.

[0048] Here, road environment information refers to external road feature data that can directly affect driving safety and passability, obtained by the vehicle on the current parking path through sensors (such as cameras and lidar) or high-precision maps.

[0049] In some embodiments, road environment information may include one or more of road width information, road slope information, road curvature information, and intersection occlusion information; the environmental risk weighting function may include one or more of the width sub-function, slope sub-function, curvature sub-function, and intersection occlusion sub-function.

[0050] Road width information refers to the lateral width of the passable area (such as lanes, passageways, and parking space entrances) at a certain mileage point along the parking path. The narrower the width, the smaller the distance between the vehicle and obstacles on either side (walls, pillars, other vehicles), resulting in a higher collision risk and a generally larger environmental risk coefficient. Correspondingly, the width sub-function is related to road width information and can be used to map "road width information" to "risk coefficient," typically showing an inverse proportional or piecewise functional relationship.

[0051] Road gradient information refers to the longitudinal inclination at a certain mileage point on the parking path, usually expressed as a percentage (%) or angle (°). A steeper gradient requires greater driving or braking force to maintain the desired speed, and the vehicle is more prone to rolling back when starting / stopping on an incline, thus increasing the risk factor. Correspondingly, the gradient sub-function is related to road gradient information and can be used to assess the impact of gradient on vehicle control stability and perception accuracy.

[0052] Road curvature information refers to the degree of curvature at a certain mileage point on the parking path, usually expressed as a curvature value (1 / radius, unit: meters). - ¹) indicates that the greater the curvature (i.e., the smaller the turning radius), the larger the steering angle required by the vehicle, and the more prone it is to sideslip or exceed the driving boundary at high speeds. Therefore, the risk factor increases with the increase of curvature. Correspondingly, the curvature sub-function is related to the road curvature information and can be used to quantify the centrifugal force risk and handling difficulty brought about by turning.

[0053] Intersection occlusion information refers to the presence of areas near a certain mileage point on the parking path (such as T-junctions, garage exits, and blind spots between pillars) where pedestrians, non-motorized vehicles, or other vehicles may suddenly appear and obstruct the view. The more severe the occlusion (e.g., walls, pillars, or large parked vehicles obstructing the view), the weaker the driver's or sensors' ability to detect obstacles in advance, and the greater the risk factor. Correspondingly, the intersection occlusion subfunction is related to intersection occlusion information and can be used to assess the risk of sudden collisions.

[0054] In some embodiments, road environment information (one or more of road width information, road slope information, road curvature information, and intersection occlusion information) can be acquired on the current parking path using a camera or lidar. Furthermore, a pre-established environmental risk weighting function and road environment information can be used to determine the environmental risk coefficient corresponding to each road environment information, and the environmental risk coefficients corresponding to each road environment information can be weighted and fused to obtain the environmental risk coefficient corresponding to each path mileage on the current parking path.

[0055] In other embodiments, road environment information (one or more of road width information, road slope information, road curvature information, and intersection occlusion information) can be obtained on the current parking path using a camera or lidar. Furthermore, a pre-established environmental risk weighting function and road environment information can be used to determine the environmental risk coefficient corresponding to each road environment information, and the environmental risk coefficient with the smallest value can be determined as the target environmental risk coefficient from the environmental risk coefficients corresponding to each road environment information. Thus, the target environmental risk coefficients corresponding to each path mileage on the current parking path can be obtained.

[0056] S103, obtain driving style coefficient; S104, based on the environmental risk coefficient and driving style coefficient, corrects the speed base curve to determine the target speed curve, so as to control the vehicle to perform memory parking according to the target speed curve.

[0057] Here, the driving style coefficient refers to a numerical parameter used to quantify whether a driver is aggressive or conservative in their approach to vehicle speed. In other words, the driving style coefficient characterizes a driver's preference for adjusting vehicle speed. For example, a larger coefficient indicates a greater pursuit of efficiency and speed, while a smaller coefficient indicates a greater emphasis on safety and stability.

[0058] In one possible implementation, to meet the personalized needs of different users, the system can pre-set multiple style coefficients, which users can select on the vehicle's infotainment interface. For example, three driving style coefficients can be set: Conservative mode: K_style=0.8. In all scenarios, the final speed will be further reduced based on the base curve and environmental influences, prioritizing safety and smoothness. Standard mode: K_style=1.0 (default). Faithfully reflects the combined effect of the "base curve" and "environment weighting," balancing safety and efficiency. Aggressive mode: K_style=1.2. While meeting hard safety constraints, it maximizes driving efficiency, bringing the speed closer to or reaching the upper limit allowed by the environment.

[0059] In another possible implementation, the driving style coefficient can be determined by analyzing the driver's actual driving behavior data during the historical parking instruction phase (i.e., the first time manually recording the route) or during historical driving. For example, the average vehicle speed, absolute acceleration (reflecting the degree of acceleration and deceleration), and brake pedal frequency and depth can be collected and statistically analyzed in real time during the instruction or historical driving process. These characteristics are then compared and analyzed with the legal speed limit or theoretical safe speed of the current road segment to automatically quantify the driver's unique driving habits and generate a personalized style coefficient, achieving personalized adaptive parking without manual user intervention.

[0060] Next, this application embodiment will detail several ways to determine the target speed curve by modifying the speed base curve based on the environmental risk coefficient and driving style coefficient.

[0061] In one possible implementation, after determining the base speed curve, environmental risk coefficient, and driving style coefficient, the base speed curve can be reduced based on the environmental risk coefficient at each location on the current path: maintaining the original speed for low-risk sections, appropriately reducing speed for medium-risk sections, and reducing speed to near a stop for extremely high-risk sections. Then, the results are adjusted overall based on the driving style coefficient: a conservative style further reduces the overall speed, while an aggressive style appropriately increases the speed within safe limits. Finally, the entire curve is smoothed, limiting the rate of speed change between adjacent locations (avoiding sudden acceleration and deceleration), resulting in the final target speed curve.

[0062] In another possible implementation, after determining the speed base curve, environmental risk coefficient, and driving style coefficient, two curves can be pre-established for the current parking path—one is the "most conservative lower limit curve" (the safe speed assuming the worst environment and the most conservative driver), and the other is the "most aggressive upper limit curve" (the permissible speed assuming the best environment and the most aggressive driver). Then, an "aggression weight" between 0 and 1 is calculated based on the actual environmental risk coefficient and driving style coefficient: the higher the environmental risk, the closer the weight is to 0; the more aggressive the driving style, the closer the weight is to 1. Finally, at each path location, the corresponding median value between the lower limit curve and the upper limit curve is taken according to this weight, thereby generating a target speed curve that adapts to the current environment and personalized preferences.

[0063] In this embodiment, the speed baseline curve generated based on the driver's historical data makes the speed rhythm, cornering habits, and straight-line acceleration of the automatic parking highly similar to the owner's, greatly enhancing the user's sense of trust. Furthermore, by correcting the speed baseline curve through a continuously changing environmental risk weighting function and driving style coefficient, the system can proactively adapt to the needs of users with different driving preferences while meeting the safety requirements of autonomous driving, thereby improving the safety of the vehicle in different scenarios during the memory parking process and enhancing the user's experience.

[0064] In some embodiments, the step S104 above, "correcting the speed base curve based on the environmental risk coefficient and driving style coefficient to determine the target speed curve," may include the following steps: S1041, based on the correction of the speed base curve according to the environmental risk coefficient and driving style coefficient, the first speed curve is determined.

[0065] Understandably, both external environmental factors and driver preferences can be considered simultaneously to make a correction to the speed base curve.

[0066] In some embodiments, for each mileage position on the current parking path, an environmental risk coefficient (reflecting safety risks caused by road width, slope, curvature, obstruction, etc.) can be obtained, along with the currently selected driving style coefficient (reflecting the driver's preference for speed; less than 1 indicates conservative, and greater than 1 indicates aggressive). Further, the product of the base speed value of the speed baseline curve at the path mileage position and the environmental risk coefficient can be calculated, and then scaled according to the driving style coefficient.

[0067] For example, the formula for calculating the first velocity curve can be expressed as: V_env(s)=V_human(s)·W_env(s)·K_style (1) Where V_env(s) represents the first velocity curve; V_human(s) represents the velocity base curve; W_env(s) represents the environmental risk coefficient; and K_styl represents the driving style coefficient.

[0068] Thus, after the combined correction of the two coefficients mentioned above, a first speed curve is obtained that simultaneously reflects environmental adaptability and personalized preferences. This curve can ensure automatic deceleration in dangerous road sections while also satisfying the driver's preference for overall parking rhythm.

[0069] S1042, the speed base curve is corrected based on the driving style coefficient to determine the second speed curve.

[0070] It is understandable that the speed base curve is independently modified based solely on the driver's individual preferences.

[0071] In some embodiments, the product between the driving style coefficient and the speed base curve is calculated, that is, each speed value on the entire speed base curve is uniformly multiplied by the driving style coefficient, thereby obtaining the second speed curve.

[0072] For example, when the driving style coefficient is less than 1, the speed decreases proportionally at all points along the speed curve, reflecting the conservative driver's desire to complete parking more slowly and safely. When the driving style coefficient is greater than 1, the speed increases proportionally at all points along the speed curve, reflecting the aggressive driver's desire to complete parking faster while maintaining safety. The resulting second speed curve reflects the driver's subjective speed preference, without including adaptive adjustments to road environment risks. This curve can serve as a "personalized baseline" in the subsequent fusion process.

[0073] For example, the formula for calculating the second velocity curve can be expressed as: V1(s)=V_human(s)·K_style (2) Where V1(s) represents the second velocity curve.

[0074] S1043, perform weighted fusion of the first velocity curve and the second velocity curve to determine the target velocity curve.

[0075] In some embodiments, a dynamic safety factor can be determined by a real-time risk assessment model, which can characterize the correspondence between path mileage and dynamic safety factor; that is, the dynamic safety factor can be different at different parking mileage locations.

[0076] After determining the dynamic safety factor, the first speed curve and the second speed curve can be weighted and fused based on the dynamic safety factor. During the fusion, the first speed curve value at each mileage position is multiplied by its corresponding weight, and the second speed curve value is multiplied by its corresponding weight to obtain the target speed value at that position, thus obtaining the target speed curve.

[0077] For example, the formula for calculating the target velocity curve can be expressed as: V_ final (s)=α·V_env(s)+(1-α)·V1(s) (3) Where V_final(s) represents the target velocity curve; α represents the dynamic safety factor, which is output by the real-time risk assessment module and is α∈[0,1].

[0078] Understandably, in high-risk scenarios (such as entering extremely narrow blind spot intersections): α→1. In this case, V_final(s)≈V_env(s), the system is dominated by the safety constraint curve, exhibiting extreme caution. In relaxed scenarios (wide straight roads): α→0. In this case, V_final(s)≈V_human(s)·K_style, the system is dominated by the personalized anthropomorphic curve, resulting in efficient and natural driving.

[0079] In other embodiments, the dynamic safety factor can be customized by the user. For example, if the user prioritizes safety, the value of α can be set to be close to 1, that is, the first speed curve (which takes into account both environmental risks and driving style) can be given a higher weight; if the user wants to retain personalized speed preference characteristics, the value of α can be set to be close to 0.

[0080] For example, Figure 2 This is a schematic diagram of the smooth speed baseline curves for different driving styles in various scenarios provided in the embodiments of this application.

[0081] like Figure 2 As shown, the speed control performance of the vehicle under different driving styles in various scenarios is compared. The three curves visually demonstrate the impact of driving style on speed and how the system generates appropriate speed curves based on environmental risks and user needs under different driving style preferences, achieving safe, human-like, smooth, and personalized speed control. The conservative mode curve represents the vehicle's final speed (V_final(s)) in various scenarios when the user selects a conservative driving style. The overall speed is relatively low to ensure safety. The standard mode curve represents the vehicle's final speed (V_final(s)) in various scenarios when the user selects a standard driving style. The speed falls between conservative and aggressive styles, balancing safety and efficiency. The aggressive mode curve represents the vehicle's final speed (V_final(s)) in various scenarios when the user selects an aggressive driving style. The overall speed is relatively high, pursuing higher traffic efficiency within safe limits.

[0082] In this embodiment, by generating a first speed curve that takes into account both environmental risks and driving style, and a second speed curve that only reflects driving preferences, and dynamically weighting and fusing them, safe deceleration can be prioritized in high-risk areas, while fully restoring the user's habitual speed rhythm in low-risk areas. This satisfies the safety requirements of advanced autonomous driving while preserving a natural, smooth, and personalized parking control feel to the greatest extent, effectively enhancing the user's trust in the memory parking system and improving the comfort experience. In some embodiments, the step S102 above, "using a pre-established environmental risk weighting function to determine the environmental risk coefficient corresponding to each path mileage on the current parking path based on the road environment information on the current parking path," may further include the following steps: S1021, input the road width information into the width sub-function to obtain the first weight value corresponding to the mileage of each path on the current parking path.

[0083] In some embodiments, the width sub-function can be a function determined based on the negative correlation between road width and safety risk, or it can be a neural network model trained based on a large amount of data labeled with road width and corresponding risk level. This application does not limit this.

[0084] In some embodiments, during the parking process, after determining the road width information of the current path, the road width information can be directly input into the width sub-function for mapping calculation, thereby obtaining the first weight value corresponding to the mileage position of the path.

[0085] It is understandable that the narrower the channel, the smaller the weight value (which can approach 0.3), and the stronger the suppression of the base curve velocity; the wider the channel, the weight approaches 1.0, and there is almost no suppression.

[0086] S1022, input the road slope information into the slope sub-function to obtain the second weight value corresponding to the mileage of each path on the current parking path.

[0087] In some embodiments, the slope sub-function can be a function determined based on the positive correlation between the absolute value of the slope and the safety risk. The slope sub-function can also be a neural network model trained based on the difficulty of controlling the vehicle under different slopes in historical parking data. This application does not limit this.

[0088] In some embodiments, during the parking process, after determining the road slope information of the current path, the road slope information can be directly input into the slope sub-function for mapping calculation, thereby obtaining the second weight value corresponding to the mileage position of the path. The second weight value is used to quantify the environmental risk coefficient of the position caused by road slope factors (insufficient power or risk of rolling backward when going uphill, and risk of loss of vehicle speed when going downhill).

[0089] It is understandable that the steeper the slope, the smaller the weight value, and the stronger the suppression of the base curve velocity; the gentler the slope, the closer the weight value is to 1, and there is almost no suppression.

[0090] S1023, input the road curvature information into the curvature sub-function to obtain the third weight value corresponding to each path mileage on the current parking path.

[0091] In some embodiments, the curvature sub-function can be a function determined based on the positive correlation between curvature magnitude and safety risk, or it can be a neural network model trained based on vehicle tracking deviation and collision probability data under different curvature paths. This application does not limit this.

[0092] In some embodiments, during the parking process, after determining the road curvature information of the current path, the road curvature information can be directly input into the curvature sub-function for mapping calculation, thereby obtaining the third weight value corresponding to the mileage position of the path. The third weight value is used to quantify the risk coefficient of the position caused by road curvature factors (delayed steering response at curves, increased risk of vehicle body sweep).

[0093] Understandably, the smaller the curvature, such as less than the straight-line threshold (close to 0), the smaller the weight value, indicating that there is no additional risk in straight-line driving. The larger the curvature, the larger the weight value, indicating that the vehicle cannot pass normally or is extremely prone to collision.

[0094] S1024, input the intersection occlusion information into the intersection occlusion sub-function to obtain the fourth weight value corresponding to the mileage of each path on the current parking path.

[0095] Here, intersection occlusion information can include occlusion rate, occlusion type, occlusion distance, etc.

[0096] In some embodiments, the intersection occlusion sub-function can be a function determined based on the positive correlation between the severity of occlusion and safety risk; the intersection occlusion sub-function can also be a neural network model trained based on accident rates or sensor perception attenuation data under different occlusion scenarios, and this application embodiment does not limit this.

[0097] In some embodiments, during the parking process, after determining the intersection obstruction information of the current path, the intersection obstruction information can be directly input into the intersection obstruction sub-function for mapping calculation, thereby obtaining the fourth weight value corresponding to the mileage position of the path. The fourth weight value can be used to quantify the environmental risk coefficient of the position caused by the intersection obstruction factor (blind spot causing inability to detect pedestrians or non-motorized vehicles crossing in time).

[0098] It is understandable that when the degree of occlusion is no occlusion, it means that the field of vision is good and the weight value is close to 1; when the degree of occlusion is a complete blind spot or partial occlusion, it means that it is impossible to predict suddenly appearing obstacles or pedestrians, and the weight value can be set to a value much less than 1 (such as 0.3~0.5).

[0099] S1025, the first weight value, the second weight value, the third weight value and the fourth weight value are integrated to obtain the environmental risk coefficient corresponding to each path mileage.

[0100] In some embodiments, after determining the first weight value, second weight value, third weight value and fourth weight value corresponding to each path mileage, the product of the first weight value, second weight value, third weight value and fourth weight value can be calculated to obtain the environmental risk coefficient corresponding to each path mileage.

[0101] For example, the environmental risk coefficient corresponding to each path mileage can be expressed as: W_env(s)=W_w(s)·W_s(s)·W_c(s)·W_x(s) (4) Where W_env(s) represents the environmental risk weighting function; W_w(s) represents the width sub-function; W_s(s) represents the slope sub-function; W_c(s) represents the curvature sub-function; and W_x(s) represents the intersection obstruction sub-function.

[0102] For example, Figure 3 This is a schematic diagram illustrating the continuous change of the environmental risk weighted field provided in the embodiments of this application.

[0103] like Figure 3 As shown, the graph visually illustrates the changes in environmental risk during vehicle travel. The continuous curve demonstrates how the environmental risk weighting value W_env(s) changes continuously at different path locations due to variations in environmental factors (such as lane width, slope, curvature, and intersection visibility). The coordinate axes in the graph are: the horizontal axis represents path mileage s (meters), indicating the vehicle's path location; the vertical axis represents the environmental risk weighting value W_env(s), ranging from 0 to 1, reflecting the degree of environmental risk at different path locations. The curve is a continuous, smooth fluctuation between 0 and 1, representing the change of the environmental risk weighting function W_env(s) with path mileage s.

[0104] In this embodiment, by inputting road width, slope, curvature, and intersection obstruction information into corresponding independent sub-functions, the weight values ​​of different risk dimensions under each path mileage are obtained. Then, multi-dimensional fusion is performed to form a unified environmental risk coefficient, thereby achieving fine decoupling and quantitative assessment of the risk sources at each location on the parking path. This avoids distortion caused by a single factor or superimposed risks, thereby improving the safety adaptability and speed control rationality of memory parking in scenarios such as narrow roads, steep slopes, sharp bends, and blind spots.

[0105] In some embodiments, during the process of controlling the vehicle to perform memory parking according to a target speed curve, the above method further includes: S201, determine the status information of the target obstacle within the preset detection range of the vehicle.

[0106] Understandably, during the parking memory process, the vehicle can use an onboard environmental perception system (such as one or more of millimeter-wave radar, lidar, ultrasonic sensors, and surround-view cameras) to detect dynamic obstacles in real time within a preset detection range (e.g., a fan-shaped or rectangular area with a radius of 10 to 30 meters centered on the vehicle), and identify dynamic obstacles that pose a potential interference risk to the current parking path as target obstacles.

[0107] Here, state information refers to a set of parameters used to describe the current motion state of the target obstacle; for example, state information may include the target obstacle's position information, speed, and direction of motion.

[0108] S202, based on state information, determines the relative distance and relative speed between the target obstacle and the vehicle.

[0109] In some embodiments, the Euclidean distance between the target obstacle and the vehicle's current position can be calculated based on the target obstacle's position information and the vehicle's current position information, and this distance can be used as the relative distance. Further, based on the target obstacle's speed and direction of movement, as well as the vehicle's current speed and direction of movement, the component of the target obstacle's speed along the line connecting the vehicle and the obstacle can be subtracted from the component of the vehicle's speed along that line to obtain the relative speed along the line. The relative speed is positive when the target obstacle is moving towards the vehicle (indicating it is approaching), and negative when it is moving away from the vehicle (indicating it is moving away).

[0110] S203, compare the relative distance with the preset safety distance to obtain the distance comparison result; Here, the distance comparison result refers to the relationship between the relative distance and the preset safety distance, which is used to classify different risk levels; the preset safety distance refers to the pre-set distance threshold used to judge the degree of danger of the target obstacle and trigger different avoidance strategies.

[0111] For example, the preset safety distance may include a first safety distance and a second safety distance. The first safety distance may be used to mark the boundary of the "warning zone" (e.g., 5 meters), and the second safety distance may be used to mark the boundary of the "emergency zone" (e.g., 2 meters).

[0112] It should be noted that the first safe distance is greater than the second safe distance. The preset safe distance can be customized according to the user's driving habits, but this application embodiment does not limit this.

[0113] For example, the distance comparison results may include the following situations: the relative distance is less than or equal to the first safe distance and greater than the second safe distance (i.e., the vehicle has entered the warning zone but has not yet entered the emergency zone); the relative distance is less than or equal to the second safe distance (i.e., the vehicle has entered the emergency zone); the relative distance is greater than or equal to the first safe distance (i.e., the vehicle is in the safe zone and no active deceleration intervention is required).

[0114] S204. Based on the distance comparison results, relative distance, relative speed, and the vehicle's current speed, determine the vehicle's safe target speed.

[0115] In some embodiments, after determining the distance comparison result, relative distance, relative speed, and the vehicle's current speed, the safe target speed of the vehicle can be determined using the distance comparison result, relative distance, relative speed, and the vehicle's current speed.

[0116] Next, embodiments of this application will introduce several ways to determine the safe target speed of a vehicle based on distance comparison results, relative distance, relative speed, and the vehicle's current driving speed.

[0117] In one possible implementation, the above-mentioned S204 "determining the vehicle's safe target speed based on the distance comparison result, relative distance, relative speed, and the vehicle's current speed" may include the following steps: S2041, if the distance comparison result indicates that the relative distance is less than or equal to the first safe distance and greater than the second safe distance, the first safe target speed of the vehicle is determined based on the relative distance and the relative speed.

[0118] Among them, the speed of the first safe target is positively correlated with the relative distance, that is, the closer the relative distance, the lower the permissible speed of the safe target; the speed of the first safe target is negatively correlated with the absolute value of the relative speed, that is, the faster the obstacle approaches, the lower the permissible speed of the safe target.

[0119] In some embodiments, a mapping relationship between "expected stopping distance" and "expected stopping time" can be pre-established, the expected safe distance to be maintained can be calculated based on the relative speed, and the maximum allowable safe speed can be derived by combining the relative distance.

[0120] For example, a linear adjustment method can be used, where the first safe target speed can be obtained by subtracting a speed correction amount from the vehicle's current speed. This speed correction amount is determined by both the relative distance deviation and the relative speed, where the closer the relative distance or the faster the speed approaches the obstacle, the greater the speed correction amount.

[0121] Another example is the use of a lookup table method, where safe speed values ​​for different combinations of relative distance and relative speed are pre-defined to form a two-dimensional mapping table. In actual operation, the first safe target speed is obtained by looking up the table and interpolating.

[0122] S2042, based on a first rate of change of speed, reduces the vehicle's current speed to a first safe target speed.

[0123] Here, the first rate of change of speed refers to the amount of speed reduction per unit time during deceleration (i.e., deceleration), which is used to control the smoothness of deceleration.

[0124] Understandably, the first speed change rate can be determined based on a preset comfort deceleration threshold, such as between 0.8 m / s² and 1.2 m / s², to ensure that the deceleration process does not cause discomfort to passengers; or, the first speed change rate can be determined in real time based on the current road surface adhesion coefficient, and a smaller deceleration can be used when the road surface is slippery to avoid wheel lock-up or skidding.

[0125] In some embodiments, the vehicle's longitudinal control system can use the first rate of change of speed as the maximum permissible deceleration constraint, and generate a smooth deceleration trajectory in real time, starting from the current driving speed and targeting the first safe target speed. For example, the electronic control unit can use a closed-loop control algorithm (such as PID control or model predictive control) to dynamically adjust the braking pressure according to the deviation between the current vehicle speed and the target speed at the corresponding moment on the trajectory, so that the actual vehicle speed continuously decreases along the trajectory until the first safe target speed is reached.

[0126] Understandably, during the entire deceleration process, the vehicle's electronic control unit can continuously monitor the relative distance and relative speed. If the obstacle's condition changes, the first safe target speed is recalculated and the deceleration trajectory is updated. If the relative distance increases or the obstacle moves away, the deceleration process is exited early, and the original speed baseline curve is restored. This ensures the comfort and smoothness of the deceleration process and avoids the discomfort caused by sudden braking.

[0127] In this embodiment of the application, when the relative distance between the vehicle and the obstacle is in the transition range between the first safe distance and the second safe distance, by constructing a first safe target speed that is positively correlated with the relative distance and negatively correlated with the absolute value of the relative speed, and by using a first speed change rate to smoothly reduce speed, it is possible to avoid sudden braking or premature deceleration while ensuring safety. This makes the speed adjustment process more in line with the anticipation and gradual braking habits of human drivers, thereby taking into account safety, comfort and parking continuity during dynamic obstacle avoidance.

[0128] In another possible implementation, S204 above may also include the following steps: If the distance comparison result indicates that the relative distance is less than or equal to the second safe distance, the second safe target speed is determined based on the vehicle's current speed and the relative distance.

[0129] Among them, the speed of the second safety target is positively correlated with the relative distance.

[0130] In some embodiments, the vehicle's electronic control unit can monitor the relative distance between the vehicle and a target obstacle (such as a vehicle in front or an obstacle) in real time. When it is determined that the relative distance is less than or equal to a preset second safe distance threshold, it means that the current distance is in a dangerous state, and active intervention to slow down is necessary. The electronic control unit can read the vehicle's current speed and the measured relative distance, and use these two data as the main inputs to determine the "second safe target speed" through a preset mapping relationship (or fuzzy rule) between the current speed and relative distance and the second safe target speed. Understandably, the second safe target speed is usually a low instantaneous speed value that can guarantee no collision within the current distance. The higher the current speed and the closer the distance, the lower the calculated second safe target speed will be, and it may even be zero.

[0131] The vehicle's current speed is reduced to a second safe target speed based on the second rate of change of speed.

[0132] Here, the second speed change rate refers to the amount of speed reduction per unit time during emergency deceleration of the vehicle; for example, the second speed change rate can be set according to user-defined settings, wherein the second speed change rate is less than or equal to a preset change rate threshold, and the first speed change rate is less than the second speed change rate.

[0133] In some embodiments, after determining the second safe target speed, a second speed change rate can be invoked to issue precise control commands to the vehicle's drive or braking system. The vehicle's control unit (such as a VCU or brake controller) can monitor the difference between the current driving speed and the second safe target speed in real time, and smoothly apply braking torque or recover driving torque according to the second speed change rate until the current driving speed precisely converges and stabilizes at the second safe target speed, thereby eliminating the risk of collision while avoiding severe vehicle shaking or rear-end collisions caused by sudden braking.

[0134] In this embodiment, when the relative distance between the vehicle and the target obstacle (such as the vehicle in front or the obstacle) is less than or equal to the second safe distance (i.e., entering a more urgent safe zone), the second safe target speed can be calculated in real time based on the current driving speed and the relative distance. By setting a relatively large second speed change rate (but still not exceeding the preset change rate threshold), the vehicle speed is smoothly reduced to the target speed. This ensures driving safety and avoids collisions while avoiding driving discomfort or the risk of rear-end collisions caused by excessive deceleration. Under the premise of meeting the safety braking requirements, the smoothness of vehicle driving, ride comfort and following performance are maximized.

[0135] It is understandable that after generating the target speed curve, there may be speed jumps, acceleration abrupt changes, or discontinuities in acceleration (i.e., the rate of change of acceleration) at certain local locations. If this curve is directly sent to the vehicle's longitudinal control system for execution, it may cause the vehicle to experience sudden acceleration, sudden deceleration, or jerking, affecting the ride comfort and control smoothness during parking. It may even cause overshoot in the actuators (such as the braking system or motor) due to sudden changes in commands.

[0136] In some embodiments, after determining the target velocity curve, the above method may further include: The target speed curve is smoothed based on a preset smoothing algorithm to obtain a smoothed target speed curve, so as to control the vehicle to perform memory parking according to the smoothed target speed curve.

[0137] Here, the pre-defined smoothing algorithm filters or reprograms the original velocity sequence to remove high-frequency noise and localized sharp fluctuations, making the velocity curve smoother, more continuous, and physically easier to track without significantly deviating from its overall shape. For example, the pre-defined smoothing algorithm may include, but is not limited to, fifth-order polynomial smoothing, moving average filtering, Gaussian filtering, spline smoothing, low-pass filtering based on finite impulse response, or trajectory optimization methods based on model predictive control.

[0138] In some embodiments, a preset smoothing algorithm is used to perform global smoothing on the target speed curve, making the speed curve smoother, more continuous and physically easier to track without significant deviation in its overall shape. This results in a smoothed target speed curve. Furthermore, the vehicle's control system can use the smoothed target speed curve as a basis to generate corresponding throttle, braking and regenerative braking commands, thereby completing a smooth and comfortable memory parking.

[0139] The smoothed target velocity curve satisfies one or more of the following constraints: On the target velocity curve, the absolute difference between the velocity values ​​corresponding to any two adjacent sampling times is less than or equal to a first preset threshold; on the target velocity curve, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a second preset threshold; on the target velocity curve, the absolute difference between the jerk values ​​corresponding to any two adjacent sampling times is less than or equal to a third preset threshold; and the absolute value of the jerk is less than or equal to a preset jerk threshold.

[0140] Here, the sampling time refers to the point in time obtained by the control system discretizing the target velocity curve at fixed time intervals (e.g., 0.02 seconds or 0.05 seconds) when performing velocity tracking. The velocity difference between two adjacent sampling times reflects how fast the velocity changes with time, that is, an approximation of instantaneous acceleration.

[0141] In this embodiment, the absolute difference between the speed values ​​corresponding to any two adjacent sampling times is less than or equal to a first preset threshold. This can be understood as the speed curve not exhibiting excessive speed jumps between any two consecutive sampling points. The first preset threshold (e.g., 0.1 m / s to 0.2 m / s) is the maximum allowable single-step speed change. Meeting this condition indicates that the speed curve is continuous in the time dimension and will not produce abrupt changes exceeding the response capability of the actuator, thereby avoiding sudden acceleration or braking pitch in the vehicle.

[0142] Here, the acceleration value can be obtained by the difference between velocity and time. The acceleration difference between two adjacent sampling times reflects the rate of change of acceleration, that is, an approximation of jerk.

[0143] In this embodiment, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a second preset threshold. This can be understood as requiring the acceleration curve itself to be continuous, without any abrupt changes in acceleration. The second preset threshold (e.g., 0.2 m / s² to 0.5 m / s²) limits the maximum allowable variation in acceleration values ​​between adjacent sampling points. Meeting this condition can further eliminate "inflection points" in the speed curve, making the changes in the vehicle's driving or braking force smoother, avoiding the back-and-forth swaying of the occupants' bodies caused by sudden acceleration changes, thereby improving ride comfort during parking.

[0144] Here, jerk (jerk value) is the derivative of acceleration with respect to time. The difference in jerk between two adjacent sampling times reflects the rate of change of jerk, that is, the second-order change of jerk.

[0145] In this embodiment, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a third preset threshold. This can be understood as a higher-order requirement for the smoothness of the velocity curve, ensuring that the acceleration itself is also continuously changing, rather than a step change. The third preset threshold (e.g., 0.5 m / s³ to 1.0 m / s³) limits the upper limit of the variation range of adjacent points of the acceleration value. Meeting this condition allows for extremely subtle changes in the vehicle's power output, eliminating any impact that may be perceived by passengers, and is particularly suitable for memory parking scenarios in high-end passenger vehicles where smoothness is a high priority.

[0146] In this embodiment, the absolute value of the jerk is less than or equal to a preset jerk threshold. This can be understood as meaning that, at any sampling moment, the absolute value of the jerk (i.e., whether it is a forward or reverse impact) cannot exceed a preset physical upper limit. The preset jerk threshold is typically set based on human comfort studies and the capabilities of the vehicle's actuators, with a typical value of 2 to 5 m / s³. Jerk exceeding this threshold will significantly cause discomfort to occupants (such as head shaking, motion sickness), and may even impact the vehicle's transmission system. This constraint ensures, in absolute terms, that the entire speed curve will not produce an impact exceeding the limits of human tolerance at any point, and is an important baseline constraint to ensure the safety and comfort of the parking process.

[0147] It should be noted that the above four constraints can be used individually or in combination, and the comparison of the embodiments in this application is not limited.

[0148] In this embodiment, a preset smoothing algorithm is used for global smoothing, which ensures the absolute continuity of the final speed curve in the three dimensions of speed, acceleration, and jerk. This eliminates speed jumps and acceleration abrupt changes at the junctions of different road sections and during normal acceleration and deceleration. The memory parking system can provide users with different driving style preferences with a personalized, smooth and comfortable automatic parking experience while ensuring safe obstacle avoidance.

[0149] The following describes the application of the vehicle control method provided in the embodiments of this application in a real-world scenario.

[0150] Based on the aforementioned technical issues, the embodiments of this application achieve a unified and optimal balance of safety, human-likeness, smoothness, and personalization in memory parking speed control, thereby solving the following core technical problems: 1) Solve the problem of stiff, unnatural vehicle behavior that does not conform to real human driving habits caused by fixed speed control.

[0151] 2) To address the issue of insufficient safety and high collision risk when simply replicating the driver's historical speed in high-risk scenarios such as obstructed intersections, sharp bends, and slopes.

[0152] 3) To address the issue of sudden changes in speed commands and acceleration at the boundary of different scene switching in the threshold control scheme based on environmental labels, which leads to strong vehicle impact and poor driving comfort.

[0153] 4) To solve the problem that existing solutions cannot adapt to the personalized needs of users with different driving styles such as conservative, standard, and aggressive, and achieve "one-size-fits-all" control.

[0154] To achieve the above objectives, this application proposes a "four-layer integrated" speed curve generation and control system. The core idea is to use an anthropomorphic speed curve generated from the driver's historical data as the "base curve," dynamically weight and constrain it through a continuously changing environmental risk function to ensure safety, introduce a configurable driving style coefficient to achieve "personalized" scaling, and finally use a high-order polynomial for global smoothing to ensure the absolute continuity of motion.

[0155] 1) Construct a continuous anthropomorphic velocity basis curve V_human(s).

[0156] Instead of simply recording discrete speed points from the driver's multiple successful parking demonstrations, machine learning algorithms (such as Gaussian process regression and spline fitting) are used to learn and generalize, generating a human-like speed baseline curve that continuously varies along the path mileage s. V_human(s) = f(path point sequence, historical velocity sequence, acceleration / deceleration operation mode) This curve reflects a particular driver's natural driving rhythm, speed preference (such as how fast they like to drive on straightaways and how much they like to slow down on curves), and acceleration and deceleration habits on this path, which is the basis for the anthropomorphism of the system.

[0157] 2) Design the continuous weighted function for environmental risk, W_env(s).

[0158] The environmental weighting function is not a simple piecewise constant, but a function that varies continuously with the path position s, obtained by multiplying the weighted sub-functions of the four core environmental factors: W_env(s)=W_w(s)·W_s(s)·W_c(s)·W_x(s) (4) Where: W_w(s): channel width weighting. The narrower the channel, the smaller the weight value (approaching 0.3), and the stronger the suppression of the base curve velocity; the wider the channel, the weight approaches 1.0, and there is almost no suppression.

[0159] W_s(s): Gradient-weighted. The weight is reduced for inclines and declines, especially steep inclines, to reduce vehicle speed and ensure safety and comfort.

[0160] W_c(s): Curvature weighting. The greater the curvature of the road (the smaller the radius of curvature), the smaller the weight value, forcing vehicles to pass through curves more smoothly.

[0161] W_x(s): Crossroads visibility weighting. This is key to safety constraints. For "blind spot intersections" where perception is completely or partially obstructed, the weight is set to a value much less than 1 (e.g., 0.3~0.5); for "unobstructed intersections" with open vision, the weight can be close to 1.0.

[0162] This continuous function ensures that safety constraints transition smoothly, rather than abruptly, when a vehicle enters / exits different risk areas.

[0163] 3) Introduce a three-level configurable driving style coefficient K_style.

[0164] To meet the personalized needs of different users, the system has three preset style coefficients, which users can select on the vehicle's infotainment interface: Conservative mode: K_style=0.8. In all scenarios, the final speed will be further reduced based on the base curve and environmental influences, prioritizing safety and stability.

[0165] Standard mode: K_style=1.0 (default). Faithfully reflects the combined effect of the "base curve" and "environment weighting", balancing safety and efficiency.

[0166] Aggressive mode: K_style=1.2. While meeting hard safety constraints, maximize driving efficiency, bringing the speed closer to or reaching the environmentally permissible upper limit.

[0167] 4) Generate the environmental constraint velocity curve V_env(s).

[0168] Combining the above three elements generates a preliminary speed curve that incorporates environmental safety constraints and personalized style: V_env(s)=V_human(s)·W_env(s)·K_style (1) This curve already possesses the characteristics of "personalization" and "environmental safety adaptation," but in regions where the risk weight W_env(s) changes drastically, it may still not be smooth.

[0169] 5) Dynamic weighted fusion and global smoothing are used to generate the final velocity curve V_final(s).

[0170] To ensure optimal smoothness globally and to flexibly handle extreme scenarios, a dynamic safety factor α is introduced for secondary fusion and smoothing: ① Dynamic fusion: Calculate a dynamic weighted fusion curve.

[0171] V_final(s)=α·V_env(s)+(1-α)·V_human(s)·K_style (5) The dynamic safety factor α∈[0, 1] is output by the real-time risk assessment module. In high-risk scenarios (such as entering extremely narrow blind spot intersections): α→1. At this time, V_final(s)≈V_env(s), the system is dominated by the safety constraint curve and is extremely cautious.

[0172] In a relaxed scenario (wide, straight road): α→0. At this point, V_final(s)≈V_human(s)·K_style, the system is dominated by the personalized anthropomorphic curve, resulting in efficient and natural driving.

[0173] This mechanism achieves a second dynamic trade-off between security strategies and anthropomorphic strategies based on the aforementioned integration, resulting in more refined and intelligent control.

[0174] ② Fifth-order polynomial global smoothing: Perform global smoothing optimization on the V_final(s) curve based on the fifth-order polynomial.

[0175] It should be noted that a fifth-order polynomial was chosen because it can simultaneously satisfy the following three key kinematic constraints: Position (velocity) continuity: ensures that the velocity curve itself does not jump.

[0176] The first derivative of velocity (acceleration) is continuous: This ensures that the acceleration changes smoothly without abrupt changes.

[0177] The second derivative of velocity (jerk) is continuous and bounded: this ensures that the rate of change of acceleration is gentle, which is the core of eliminating the sense of impact and achieving a "silky smooth" experience. The smoothing process is strictly limited by |Jerk| ≤ J_max (J_max is usually set to 1.0-1.5 m / s³).

[0178] The smoother uses V_final(s) as the reference path to generate a completely new final execution speed curve V_final(s) that fully satisfies the above continuity constraints.

[0179] 6) Trajectory execution and dynamic safety monitoring.

[0180] The vehicle's longitudinal controller tracks the V_final(s) curve. Simultaneously, a separate safety monitoring module uses ultrasonic radar and surround-view cameras to detect dynamic obstacles (such as pedestrians and suddenly appearing vehicles) along the path in real time.

[0181] When a potential risk is detected, instead of using the traditional AEB emergency braking, a smooth deceleration command is triggered: within a set buffer time (such as 0.5 seconds), the smoother is required to linearly or quadraticly polynomially reduce the current target speed to a new safe speed, achieving "predictive and gentle deceleration".

[0182] If an obstacle rapidly approaches to the minimum safe distance, initiate smooth emergency braking to bring the vehicle to a stop as quickly as possible while ensuring that the acceleration does not exceed the limit.

[0183] Compared with the prior art, the beneficial effects of this application include at least the following: 1) Achieved truly personalized driving: By introducing three configurable style coefficients K_style (conservative (0.8), standard (1.0), and aggressive (1.2), the system can proactively adapt to the deep needs of users with different driving preferences, covering the entire user group from novices to experienced drivers, and completely changing the "one-size-fits-all" control mode.

[0184] 2) It retains a highly human-like driving experience: The continuous base curve V_human(s) generated based on real driver historical data makes the speed rhythm, cornering habits, straight-line acceleration and other details during automatic parking highly similar to the owner. The vehicle behavior "understands you" and is natural and smooth, which greatly enhances the user's sense of trust and intimacy.

[0185] 3) A global proactive safety system has been constructed: Through the continuously changing environmental risk weighting function W_env(s) and dynamic safety coefficient α, the system can conduct a forward-looking risk assessment of the entire path and automatically and smoothly apply speed constraints in high-risk areas (narrow roads, blind spots, steep slopes), eliminating safety hazards from the planning source and meeting the safety requirements of high-level autonomous driving.

[0186] 4) Achieved ultimate smoothness: Global smoothing is performed using a fifth-order polynomial, mathematically guaranteeing the absolute continuity of the final velocity curve V_final(s) in the three dimensions of velocity, acceleration, and jerk. This completely eliminates speed jumps and acceleration abrupt changes at the junctions of different road sections and during normal acceleration and deceleration, achieving a shock-free, jerky, and "experienced driver"-like ride quality, effectively avoiding motion sickness.

[0187] 5) The system achieves intelligent adjustment that adapts to different scenarios: It can automatically identify and distinguish dozens of subdivided scenario combinations such as wide straight roads, unobstructed intersections, obstructed intersections, slopes, small curves, and sharp curves, and make precise and continuous speed adjustments based on preset parameter tables, demonstrating a high degree of intelligence.

[0188] For example, Figure 4 This is a schematic diagram of the overall architecture of the memory parking speed control system provided in the embodiments of this application; as shown below. Figure 4As shown in the figure, this diagram illustrates the overall architecture of the memory parking speed control system and the relationships between its various modules, as well as the interaction between the system and the external environment. It presents the entire process from acquiring historical path and real-time environmental information, to learning anthropomorphic speed curves, assessing environmental risks, fusing speed curves, smoothing the process, and finally to vehicle execution, demonstrating the logic of collaborative work among the various parts of the system.

[0189] In some embodiments, the overall architecture of the memory parking speed control system may include a data acquisition module 401, a path memory and mileage analysis module 402, an anthropomorphic speed base curve learning module 403, an environmental perception and weighted field construction module 404, a driving style configuration module 405, a speed curve dynamic fusion module 406, a global smoothing module 407, a dynamic obstacle smooth deceleration module 408, and a vehicle longitudinal execution module 409.

[0190] Among them, the data acquisition module 401 is used to acquire historical path data and store the path information of the driver's previous successful parking, for the anthropomorphic speed base curve learning module to learn.

[0191] The path memory and mileage parsing module 402 is used to store, manage, and retrieve the memory parking paths that the user has successfully learned, and to map the discrete point sequence of the path to the continuous mileage s.

[0192] The anthropomorphic speed base curve learning module 403 is used to interact with the path memory module. It uses machine learning algorithms to analyze, denoise, and generalize historical driving speed data to generate a continuous and smooth anthropomorphic speed base curve V_human(s).

[0193] The environmental perception and weighted field construction module 404 is used to connect ultrasonic radar, cameras, millimeter-wave radar, IMU, wheel speed sensors, etc., to perceive the vehicle's surrounding environment in real time. This module includes an environmental label classifier and a weighted function calculator, which can output W_env(s) that changes continuously along the path.

[0194] The driving style configuration module 405 provides a human-machine interface (HMI) for users to select "conservative", "standard" or "aggressive" driving styles and passes the corresponding K_style coefficients to downstream modules.

[0195] The velocity curve dynamic fusion module 406 is used to receive V_human(s), W_env(s), and K_style, and determine the dynamic safety factor α by integrating the real-time risk assessment results, and calculate V_env(s) and V_final(s) in sequence.

[0196] The global smoothing module 407 is used to receive the V_final(s) curve, call the fifth-order polynomial smoothing algorithm, and solve for the globally optimal final execution speed curve V_final(s) under the constraints that the velocity, acceleration, and jerk are continuous and bounded.

[0197] The vehicle longitudinal execution module 409 is used to receive the V_final(s) curve and accurately track the target speed through the vehicle's drive-by-wire (throttle) and brake-by-wire (brake) systems using model predictive control or PID control algorithms.

[0198] The dynamic obstacle smooth deceleration module 408 operates as an independent high-priority safety layer. It processes environmental perception information in real time, detects dynamic obstacles, assesses risk levels, and sends smooth deceleration commands to the smooth controller to manage deceleration and stopping processes, ensuring safety and comfort in dynamic scenarios.

[0199] Next, this application will provide a detailed explanation of the solution through specific road scenarios.

[0200] In one implementation (Example 1), a wide standard straight road scenario (standard driving style).

[0201] Scenario description: The vehicle is traveling in a straight line on the ground or main road of the underground parking garage with a width greater than 3.5 meters, with a wide field of vision and no interference from non-motorized vehicles or pedestrians.

[0202] Workflow: 1) The environmental perception module determines the label as "wide straight road" and queries the core parameter table: environmental weighted W_env(s) ≈ 0.8~1.0, dynamic safety coefficient α is 0.1~0.3.

[0203] 2) The learning value of the anthropomorphic velocity base curve V_human(s) for this road section is 20km / h.

[0204] 3) The user selects "Standard Mode", K_style=1.0.

[0205] 4) Calculate V_env(s) = 20 • 0.9 • 1.0 = 18 km / h (take W_env(s) = 0.8).

[0206] 5) Calculate V_final(s) = 0.2•18 + 0.8•20 = 19.6 km / h. (Take α = 0.2).

[0207] 6) The smoother performs fifth-order polynomial smoothing on V_final(s) and the speeds of the preceding and following road segments, and outputs V_final(s).

[0208] 7) Implementation Results: Vehicles pass through smoothly and continuously at speeds ranging from 12 to 20 km / h (see Table 1). Due to the relaxed environment and small α value, the final speed is almost entirely determined by the driver's historical habits, resulting in a highly human-like experience, high traffic efficiency, and no sudden changes in speed or acceleration throughout the journey.

[0209] In one implementation (Example 2), an unobstructed intersection scenario (aggressive driving style).

[0210] Scenario description: The vehicle needs to pass through a crossroads with good visibility, where oncoming vehicles from all directions can be detected from a distance.

[0211] Workflow: 1) The environment label is "unobstructed intersection". From the table: W_env)(s)≈0.7~0.9, α is 0.5~0.7.

[0212] 2) The learning value of V_human(s) is 15km / h.

[0213] 3) The user selects "aggressive mode", K_style=1.2.

[0214] 4) Calculate V_env(s) = 15 • 0.8 • 1.2 = 14.4 km / h. (Take W_env(s) = 0.8) 5) Calculate V_final(s) = 0.6•14.4 + 0.4•(15•1.2) = 15.84 km / h. (Take α = 0.6) 6) The smoother generates the final curve.

[0215] 7) Implementation Results: Vehicles pass through the intersection at a smooth speed of 13-18 km / h. Although the intersection weight W_env reduces the base speed, the combined effect of the aggressive style coefficient K_style=1.2 and the dynamic fusion mechanism significantly improves the final speed, satisfying the pursuit of traffic efficiency by aggressive users while remaining within safety constraints.

[0216] In one implementation (Example 3), a right-angle blind spot intersection scenario with visual obstruction (conservative driving style). Scenario description: At a T-shaped intersection in an underground parking garage, a wall on the left side of the vehicle completely obstructs the view of oncoming vehicles from the side, posing an extremely high risk.

[0217] Workflow: 1) The environment label is "obstructed blind spot intersection". From the table: W_env(s)≈0.3~0.5, α is 0.7~0.9.

[0218] 2) The learning value of V_human(s) is 8km / h.

[0219] 3) The user selects "conservative mode", K_style=0.8.

[0220] 4) Calculate V_env(s) = 8 • 0.4 • 0.8 = 2.56 km / h. (For example, take W_env(s) = 0.4) 5) Calculate V_final(s) = 0.8•2.56 + 0.2•(8•0.7) = 1.79 + 1.12 = 3.168 km / h. (For example, take α = 0.8) 6) The smoother generates the final curve.

[0221] 7) Implementation Results: Vehicles smoothly approach and pass through intersections at speeds of 4-7 km / h. In this high-risk scenario, the system exhibits high caution, but by optimizing the weighted parameters, the speed is significantly improved compared to earlier solutions. Vehicles begin very gentle linear deceleration in advance, ensuring extremely high safety, and the deceleration process is smooth, without any tension caused by sudden braking.

[0222] In one implementation (Example 4), an uphill / downhill scenario (standard driving style).

[0223] Scene description: The vehicle enters the underground parking garage ramp that connects different floors and has a slope of approximately 8%.

[0224] Workflow: 1) The environment label is "steep slope". Look up the table: W_env(s)≈0.5~0.7, α is 0.5~0.7.

[0225] 2) The learning value of V_human(s) is 12km / h.

[0226] 3) K_style=1.0.

[0227] 4) Calculate V_env(s) = 12•0.6•1.0 = 7.2 km / h.

[0228] 5) Calculate V_final(s) = 0.6•7.2 + 0.4•12 = 9.12 km / h.

[0229] 6) The smoother generates the final curve.

[0230] 7) Implementation Results: Vehicles ascend or descend slopes at a smooth speed of 8-12 km / h. The slope weight W_env(s) prevents vehicles from rushing uphill or losing speed on slopes. Smooth control ensures the continuity of power output and avoids the "nose-up" or "nodding" phenomenon caused by sudden speed changes at the start and end of the slope, resulting in a very smooth ride.

[0231] In one implementation (Example 5), dynamic pedestrians appear in the path.

[0232] Scenario description: The vehicle is traveling at a speed of 10km / h along the planned path when the dynamic monitoring module suddenly detects a pedestrian crossing the path 5 meters ahead.

[0233] Workflow: 1) The safety monitoring module calculates the time to collision (TTC), determines it to be a "warning level" risk, and immediately sends an instruction to the smoother: "Smoothly reduce the target speed to 3km / h within 0.5 seconds."

[0234] 2) The smoother does not interrupt the current planning, but performs a quadratic polynomial replanning on the subsequent part of V_final(s) online to generate a smooth deceleration curve with limited acceleration from the current speed to 3 km / h.

[0235] 3) The vehicle begins to decelerate gently and linearly.

[0236] 4) Implementation effect: Passengers experience a smooth and predictable deceleration, similar to a skilled driver's anticipatory release of the accelerator and light application of the brakes, completely lacking the uncomfortable sudden braking impact and panic of traditional AEB. If a pedestrian crosses quickly, the vehicle can continue to accelerate smoothly; if the pedestrian stops, the vehicle will come to a smooth stop.

[0237] In one implementation (Example 6), a narrow passage scenario (standard driving style).

[0238] Scenario description: The vehicle needs to pass through a narrow passage with a width of approximately 2.6 meters.

[0239] System workflow: 1) The environment label is "narrow channel". Look up the table: W_env(s)≈0.4~0.6, α is 0.6~0.8.

[0240] 2) The learning value of V_human(s) is 9km / h.

[0241] 3) The user selects "Standard Mode", K_style=1.0.

[0242] 4) Calculate V_env(s) = 9 • 0.5 • 1.0 = 4.5 km / h. (For example, take W_env(s) = 0.5) 5) Calculate V_final(s) = 0.7•4.5 + 0.3•9 = 3.15 + 2.7 = 5.85 km / h. (For example, take α = 0.7) 6) The smoother generates the final curve.

[0243] 7) Implementation Results: Vehicles pass through narrow passages at a smooth speed of 5-8 km / h. Environmental weighting effectively reduces vehicle speed to ensure a safety margin, while dynamic fusion and smooth planning ensure the continuity of speed changes. Vehicles drive stably without the tension of being too close to the edge, and sudden speed changes within the passage are also avoided.

[0244] For example, Table 1 is a table of expected speed ranges and core parameters for various scenarios provided in the embodiments of this application.

[0245] Table 1

[0246] It should be noted that the above speed is the approximate range of the final execution speed V_final(s) after smoothing. The actual value is dynamically calculated from the base curve, specific weights, and style coefficients.

[0247] Based on the above embodiments, this application also provides a vehicle control device. Figure 5 This is a schematic diagram of the composition structure of a vehicle control device provided in an embodiment of this application, as shown below. Figure 5 As shown, the vehicle control device 500 includes a generation unit 501, a determination unit 502, an acquisition unit 503, and a correction unit 504, wherein: The generation unit 501 is used to generate the speed base curve corresponding to the current parking path based on historical parking data; wherein, the speed base curve represents the correspondence between path mileage and vehicle speed. The determination unit 502 is used to determine the environmental risk coefficient corresponding to each path mileage on the current parking path based on the road environment information on the current parking path using a pre-established environmental risk weighting function; wherein, the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; The acquisition unit 503 is used to acquire the driving style coefficient; wherein, the driving style coefficient represents the driver's preference for adjusting the vehicle speed; The correction unit 504 is used to correct the speed base curve based on the environmental risk coefficient and driving style coefficient, determine the target speed curve, and control the vehicle to perform memory parking according to the target speed curve.

[0248] In some embodiments of this application, the correction unit 504 is specifically used to correct the speed base curve based on the environmental risk coefficient and the driving style coefficient to determine the first speed curve; The speed base curve is corrected based on the driving style coefficient to determine the second speed curve; The first and second velocity curves are weighted and fused to determine the target velocity curve.

[0249] In some embodiments of this application, the road environment information includes one or more of the following: road width information, road slope information, road curvature information, and intersection occlusion information; the environmental risk weighting function includes one or more of the following: width subfunction, slope subfunction, curvature subfunction, and intersection occlusion subfunction.

[0250] In some embodiments of this application, the determining unit 502 is specifically used to input the road width information into the width sub-function to obtain the first weight value corresponding to each path mileage on the current parking path; The road slope information is input into the slope sub-function to obtain the second weight value corresponding to the mileage of each path on the current parking path; The road curvature information is input into the curvature sub-function to obtain the third weight value corresponding to the mileage of each path on the current parking path. Input the intersection occlusion information into the intersection occlusion sub-function to obtain the fourth weight value corresponding to the mileage of each path on the current parking path; The first, second, third, and fourth weight values ​​are combined to obtain the environmental risk coefficient corresponding to each path mileage.

[0251] In some embodiments of this application, the vehicle control device 500 further includes a comparison unit, wherein: The determining unit 502 is also used to determine the state information of the target obstacle within the preset detection range of the vehicle; the state information includes the position information, speed and direction of movement of the target obstacle; The determining unit 502 is also used to determine the relative distance and relative speed between the target obstacle and the vehicle based on the state information; The comparison unit is used to compare the relative distance with the preset safety distance to obtain the distance comparison result; The determining unit 502 is also used to determine the safe target speed of the vehicle based on the distance comparison result, relative distance, relative speed and the current driving speed of the vehicle.

[0252] In some embodiments of this application, the determining unit 502 is specifically used to determine the first safe target speed of the vehicle based on the relative distance and the relative speed when the distance comparison result indicates that the relative distance is less than or equal to the first safe distance and greater than the second safe distance; wherein the first safe target speed is positively correlated with the relative distance, the first safe target speed is negatively correlated with the absolute value of the relative speed, and the first safe distance is greater than the second safe distance; The vehicle's current speed is reduced to a first safe target speed based on the first rate of change of speed.

[0253] In some embodiments of this application, the determining unit 502 is specifically used to determine the second safe target speed based on the vehicle's current driving speed and the relative distance when the distance comparison result indicates that the relative distance is less than or equal to the second safe distance; The vehicle's current speed is reduced to a second safe target speed based on a second speed change rate; wherein the second speed change rate is less than or equal to a preset change rate threshold, and the first speed change rate is less than the second speed change rate.

[0254] In some embodiments of this application, the vehicle control device 500 further includes: a smoothing processing unit, wherein: The smoothing unit is used to smooth the target speed curve based on a preset smoothing algorithm to obtain a smoothed target speed curve, so as to control the vehicle to perform memory parking according to the smoothed target speed curve. The smoothed target velocity curve satisfies one or more of the following constraints: On the target velocity curve, the absolute difference between the velocity values ​​corresponding to any two adjacent sampling times is less than or equal to a first preset threshold; on the target velocity curve, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a second preset threshold; on the target velocity curve, the absolute difference between the jerk values ​​corresponding to any two adjacent sampling times is less than or equal to a third preset threshold; and the absolute value of the jerk is less than or equal to a preset jerk threshold. It should be noted that, in the embodiments of this application, if the above methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of software products. These software products are stored in a storage medium and include several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0255] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the above-described method.

[0256] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. The computer-readable storage medium can be transient or non-transient.

[0257] This application also provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement some or all of the steps in the above-described method. This computer program product can be implemented specifically through hardware, software, or a combination thereof.

[0258] In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0259] It should be noted that, Figure 6 This is a schematic diagram of the hardware entity of a vehicle provided in an embodiment of this application, such as... Figure 6 As shown, the hardware entity of the vehicle 600 includes: a processor 601, a communication interface 602, and a memory 603, wherein: The processor 601 typically controls the overall operation of the vehicle 600.

[0260] The communication interface 602 enables the vehicle 600 to communicate with other terminals or servers via a network.

[0261] The memory 603 is configured to store instructions and applications executable by the processor 601, and can also cache data to be processed or already processed by the processor 601 and various modules in the vehicle 600 (e.g., image data, audio data, voice communication data, and video communication data), and can be implemented using flash memory or RAM. Data can be transferred between the processor 601, the communication interface 602, and the memory 603 via the bus 604.

[0262] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0263] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above steps / processes do not imply a sequential order of execution; the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0264] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. 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.

[0265] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0266] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0267] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0268] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory, magnetic disks, or optical disks.

[0269] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0270] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A vehicle control method, characterized in that, The method includes: Based on historical parking data, a speed baseline curve corresponding to the current parking path is generated; wherein, the speed baseline curve represents the correspondence between path mileage and vehicle speed. Using a pre-established environmental risk weighting function, the environmental risk coefficient corresponding to each path mileage on the current parking path is determined based on the road environment information on the current parking path; wherein, the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; Obtain the driving style coefficient; wherein the driving style coefficient represents the driver's preference for adjusting the vehicle speed; The speed base curve is corrected based on the environmental risk coefficient and the driving style coefficient to determine the target speed curve, so as to control the vehicle to perform memory parking according to the target speed curve.

2. The method according to claim 1, characterized in that, The step of correcting the speed base curve based on the environmental risk coefficient and the driving style coefficient to determine the target speed curve includes: The speed base curve is corrected based on the environmental risk coefficient and driving style coefficient to determine the first speed curve; The speed base curve is corrected based on the driving style coefficient to determine the second speed curve; The first velocity curve and the second velocity curve are weighted and fused to determine the target velocity curve.

3. The method according to claim 1, characterized in that, The road environment information includes one or more of the following: road width information, road slope information, road curvature information, and intersection occlusion information; the environmental risk weighting function includes one or more of the following: width sub-function, slope sub-function, curvature sub-function, and intersection occlusion sub-function.

4. The method according to claim 3, characterized in that, The step of determining the environmental risk coefficient corresponding to each path mileage on the current parking path based on road environment information on the current parking path and using a pre-established environmental risk weighting function includes: The road width information is input into the width sub-function to obtain the first weight value corresponding to each path mileage on the current parking path; The road slope information is input into the slope sub-function to obtain the second weight value corresponding to the mileage of each path on the current parking path; The road curvature information is input into the curvature sub-function to obtain the third weight value corresponding to each path mileage on the current parking path; The intersection occlusion information is input into the intersection occlusion sub-function to obtain the fourth weight value corresponding to each path mileage on the current parking path; The first weight value, the second weight value, the third weight value, and the fourth weight value are fused to obtain the environmental risk coefficient corresponding to each path mileage.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: The status information of the target obstacle within the preset detection range of the vehicle is determined; the status information includes the position information, speed and direction of movement of the target obstacle; Based on the state information, the relative distance and relative speed between the target obstacle and the vehicle are determined; The relative distance is compared with the preset safety distance to obtain the distance comparison result; Based on the distance comparison results, the relative distance, the relative speed, and the vehicle's current speed, the vehicle's safe target speed is determined.

6. The method according to claim 5, characterized in that, The preset safety distance includes a first safety distance and a second safety distance; determining the vehicle's safe target speed based on the distance comparison result, the relative distance, the relative speed, and the vehicle's current speed includes: If the distance comparison result indicates that the relative distance is less than or equal to the first safe distance and greater than the second safe distance, a first safe target speed for the vehicle is determined based on the relative distance and the relative speed; wherein the first safe target speed is positively correlated with the relative distance, the first safe target speed is negatively correlated with the absolute value of the relative speed, and the first safe distance is greater than the second safe distance; The vehicle's current speed is reduced to the first safe target speed based on the first rate of change of speed.

7. The method according to claim 5, characterized in that, The preset safety distance includes a second safety distance; determining the vehicle's safe target speed based on the distance comparison result, the relative distance, the relative speed, and the vehicle's current speed includes: If the distance comparison result indicates that the relative distance is less than or equal to the second safe distance, a second safe target speed is determined based on the vehicle's current speed and the relative distance. The vehicle's current speed is reduced to the second safe target speed based on the second speed change rate; wherein the second speed change rate is less than or equal to a preset change rate threshold, and the first speed change rate is less than the second speed change rate.

8. The method according to any one of claims 1 to 4, characterized in that, The method further includes: The target speed curve is smoothed based on a preset smoothing algorithm to obtain a smoothed target speed curve, so as to control the vehicle to perform memory parking according to the smoothed target speed curve. The smoothed target velocity curve satisfies one or more of the following constraints: On the target velocity curve, the absolute difference between the velocity values ​​corresponding to any two adjacent sampling times is less than or equal to a first preset threshold; on the target velocity curve, the absolute difference between the acceleration values ​​corresponding to any two adjacent sampling times is less than or equal to a second preset threshold; on the target velocity curve, the absolute difference between the jerk values ​​corresponding to any two adjacent sampling times is less than or equal to a third preset threshold; and the absolute value of the jerk is less than or equal to a preset jerk threshold.

9. A vehicle control device, characterized in that, The device includes: The generation unit is used to generate a speed base curve corresponding to the current parking path based on historical parking data; wherein, the speed base curve represents the correspondence between path mileage and vehicle speed. The determining unit is used to determine the environmental risk coefficient corresponding to each path mileage on the current parking path based on the road environment information on the current parking path using a pre-established environmental risk weighting function; wherein, the environmental risk weighting function represents the correspondence between path mileage and environmental risk coefficient; An acquisition unit is used to acquire a driving style coefficient; wherein the driving style coefficient represents the driver's preference for adjusting the vehicle's speed. The correction unit is used to correct the speed base curve based on the environmental risk coefficient and the driving style coefficient, and determine the target speed curve so as to control the vehicle to perform memory parking according to the target speed curve.

10. A vehicle, characterized in that, It includes a processor and a memory, the memory storing a computer program that can run on the processor, the processor executing the computer program to implement the steps of the method according to any one of claims 1 to 8.

Images