Trajectory prediction method, device and equipment for automatic driving of vehicle and storage medium

Abstract

The invention discloses a trajectory prediction method and device for vehicle automatic driving, equipment and a storage medium. According to the method provided by the invention, multiple constraint conditions are constructed according to the path information and the driving indexes; and according to multiple constraint conditions, constructing and solving a vehicle control model for automatic driving of the vehicle, and generating a path tracking optimal control quantity. And finally, on the basis of the optimal control quantity and the vehicle control model, predicting a driving track of the vehicle for automatic driving in the target path. On the basis of control and dynamic characteristics of the vehicle and in cooperation with path planning of the vehicle, exchange and information transmission with other data can be carried out more conveniently, so that high-precision prediction in a complex scene is realized, the accuracy and robustness of a trajectory prediction result are effectively improved, and the prediction efficiency is improved. Therefore, accurate data are provided for collision risk assessment in the automatic driving process of the vehicle, and the safety and reliability of automatic driving of the vehicle are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent driving, in particular to a trajectory prediction method and device for vehicle automatic driving, equipment and storage medium. BACKGROUND

[0002] With the popularity of vehicles in life, automatic driving has gradually entered people's daily life. The key to vehicle automatic driving is the accurate prediction of the automatic driving trajectory at the future time to evaluate the collision risk. Therefore, the trajectory prediction technology in the field of automatic driving is crucial for realizing safe, efficient and natural driving behavior.

[0003] Currently, for trajectory prediction of vehicle automatic driving, a deep learning method based on convolutional neural network is usually used. However, the deep learning method usually needs a large amount of labeled data for training. For trajectory prediction, a large-scale trajectory dataset is needed, and accurate labeling of the trajectory is needed, which increases the difficulty and cost of development and implementation, and the computing resources and energy consumption are high. In addition, the training of the deep learning model is based on the existing data, and their generalization ability on unseen data is limited, which leads to poor model performance and cannot accurately predict the trajectory. SUMMARY

[0004] Therefore, the present application provides a trajectory prediction method, device and equipment for vehicle automatic driving and storage medium, which mainly aims to solve the technical problems of large amount of calculation and low accuracy of prediction results in the prior art by using deep learning for trajectory prediction.

[0005] According to the first aspect of the present application, a trajectory prediction method for vehicle automatic driving is provided, which is applied to a vehicle with automatic driving function. The method comprises:

[0006] Obtaining a plurality of vehicle body parameters, path information of a target path, and a plurality of driving index parameters of a plurality of driving indexes of the vehicle for automatic driving in the target path;

[0007] According to the plurality of vehicle body parameters, the path information and the plurality of driving indexes, a plurality of constraint conditions for the vehicle to automatically drive on the target path are constructed;

[0008] According to the plurality of constraint conditions, a vehicle control model is constructed;

[0009] Solving the vehicle control model to obtain an optimal control amount of the vehicle for automatic driving in the target path;

[0010] Based on the optimal control amount and the vehicle control model, a predicted trajectory of the vehicle for automatic driving in the target path is generated.

[0011] Optionally, the step of constructing the plurality of constraint conditions for the automatic driving of the vehicle on the target path according to the plurality of vehicle body parameters, the path information and the plurality of driving indexes comprises:

[0012] constructing the inflation circle constraint of the vehicle according to the plurality of vehicle body parameters and the path information;

[0013] constructing the front wheel steering angle constraint and the front wheel steering angle rate constraint according to the plurality of driving indexes and the plurality of vehicle body parameters;

[0014] generating the system state vector according to the longitudinal coordinate, the lateral coordinate, the yaw angle, the front wheel steering angle and the front wheel steering angle rate of the midpoint of the rear axle of the vehicle on the target path;

[0015] generating the system control vector according to the derivative of the front wheel steering angle rate;

[0016] constructing the vehicle kinematics equation according to the system state vector and the system control vector;

[0017] discretizing the vehicle kinematics equation to generate the vehicle kinematics equation constraint.

[0018] Optionally, the plurality of vehicle body parameters comprises the midpoint position of the front axle and the midpoint position of the rear axle, and the path information comprises the obstacle information, and the step of constructing the inflation circle constraint of the vehicle according to the plurality of vehicle body parameters and the obstacle information comprises:

[0019] enveloping the outer contour of the vehicle according to the midpoint position of the front axle and the midpoint position of the rear axle respectively to generate a first envelope circle and a second envelope circle;

[0020] generating a first inflation circle and a second inflation circle according to the first envelope circle, the second envelope circle and the path information;

[0021] constructing the inflation circle constraint according to the first inflation circle and the second inflation circle.

[0022] Optionally, the step of generating the first inflation circle and the second inflation circle according to the first envelope circle, the second envelope circle and the path information comprises:

[0023] obtaining a first center of the first envelope circle and a second center of the second envelope circle;

[0024] forming a perpendicular line through the first center in the lateral direction of the vehicle;

[0025] generating the first inflation circle according to the straight line and the path information;

[0026] determining a center distance according to the first center and the second center;

[0027] generating the second inflation circle according to the first center, the center distance and the path information.

[0028] Optionally, multiple vehicle body parameters, including wheelbase, are used. The steps for constructing front wheel steering angle constraints and front wheel steering angle change rate constraints based on multiple driving indicators and multiple vehicle body parameters specifically include:

[0029] Based on multiple driving indicators and wheelbase, a first target mapping relationship is constructed, which represents the correspondence between front wheel steering angle and lateral acceleration.

[0030] Based on the first target mapping relationship, the front wheel steering angle change rate, and the lateral acceleration, a second target mapping relationship is constructed. The second target mapping relationship characterizes the correspondence between the front wheel steering angle change rate and the lateral acceleration.

[0031] Based on the mapping relationship between the maximum lateral acceleration and the first target, a front wheel steering angle constraint is constructed;

[0032] Based on the mapping relationship between the maximum lateral acceleration and the first target, a constraint on the rate of change of front wheel steering angle is constructed.

[0033] Optionally, the multiple driving indicators include: yaw angle, understeer coefficient, longitudinal speed, turning radius, lateral acceleration, and lateral jerk. The step of constructing a first target mapping relationship based on these multiple driving indicators and wheelbase specifically includes:

[0034] A third target mapping relationship is constructed based on yaw angle, front wheel steering angle, understeer coefficient, longitudinal velocity, turning radius, and lateral acceleration;

[0035] Based on wheelbase, turning radius, and front wheel steering angle, construct a fourth target mapping relationship;

[0036] Based on the understeer coefficient and wheelbase, construct the fifth target mapping relationship;

[0037] Based on the third, fourth, and fifth target mapping relationships, the first target mapping relationship is generated.

[0038] Optionally, the steps of solving the vehicle control model to obtain the optimal control variables for the vehicle specifically include:

[0039] The vehicle control model is abstracted.

[0040] The augmented Lagrangian method is used to solve the abstracted vehicle control model to obtain the optimal control quantity.

[0041] According to a second aspect of this application, a trajectory prediction device for autonomous driving of a vehicle is provided, applied to a vehicle with autonomous driving capabilities, the device comprising:

[0042] The acquisition module is used to acquire multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of the vehicle for automatic driving within the target path.

[0043] The first construction module is used to construct multiple constraints for the vehicle to drive automatically on the target path based on multiple vehicle body parameters, path information and multiple driving indicators.

[0044] The second construction module is used to build a vehicle control model based on multiple constraints.

[0045] The first generation module is used to solve the vehicle control model and obtain the optimal control quantity for the vehicle to drive automatically within the target path.

[0046] The second generation module is used to generate a predicted trajectory for the vehicle to drive automatically within the target path based on the optimal control quantity and the vehicle control model.

[0047] According to a third aspect of this application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods in the first aspect.

[0048] According to a fourth aspect of this application, a readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0049] By employing the aforementioned technical solution, this application provides a trajectory prediction method, device, electronic device, and readable storage medium for autonomous driving of vehicles. Based on path information and driving indicators, multiple constraints are constructed. According to these constraints, a vehicle control model for autonomous driving is constructed and solved, generating optimal control quantities for path tracking. Finally, based on the optimal control quantities and the vehicle control model, the driving trajectory of the vehicle in autonomous driving on the target path is predicted. This method can more clearly explain the relationship between the prediction results and vehicle characteristics, driving environment, and control parameters, providing more intuitive and interpretable trajectory prediction results. Furthermore, this method, based on the vehicle's control and dynamic characteristics and in conjunction with the vehicle's path planning, can more easily exchange and transmit information with other data to achieve high-precision prediction in complex scenarios. This effectively improves the accuracy and robustness of trajectory prediction results, thereby providing accurate data for collision risk assessment during autonomous driving and enhancing the safety and reliability of autonomous driving.

[0050] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0051] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0052] Figure 1 A schematic flowchart of a trajectory prediction method for autonomous driving of a vehicle provided in an embodiment of this application is shown;

[0053] Figure 2 This paper illustrates a schematic flowchart of another trajectory prediction method for autonomous driving of vehicles provided in an embodiment of this application.

[0054] Figure 3 This illustration shows a structural diagram of the vehicle envelope circle and expansion circle provided in an embodiment of this application;

[0055] Figure 4 A schematic diagram of the vehicle control system provided in an embodiment of this application is shown;

[0056] Figure 5 A schematic diagram of the structure of a trajectory prediction device for autonomous driving of a vehicle provided in an embodiment of this application is shown. Detailed Implementation

[0057] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0058] This application provides a trajectory prediction method for autonomous driving of vehicles, such as... Figure 1 As shown, the method includes:

[0059] S101. Obtain multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of the vehicle for automatic driving within the target path.

[0060] The trajectory prediction method for autonomous driving provided in this application is applicable to vehicles with autonomous driving functions. The vehicle body parameters describe the vehicle's appearance and dimensions, including the front axle position, rear axle position, front overhang position, rear overhang position, and wheelbase.

[0061] The target trajectory can be a pre-planned path on the map or the trajectory of the road ahead as perceived in real time. The path information of the target path includes: geometric features such as the shape of the road, markings, road edges, and lane lines; obstacle information: including the location and type of obstacles on both sides of the road (such as trash cans, trees, etc.).

[0062] Several vehicle driving parameters include: yaw angle, understeer coefficient, longitudinal speed, turning radius, lateral acceleration, and lateral jerk. Considering the vehicle's handling capabilities, driving parameters are set for the vehicle during autonomous driving to generate smooth and safe driving maneuvers.

[0063] S102. Based on multiple vehicle body parameters, path information, and multiple driving indicators, construct multiple constraints for the vehicle to drive automatically on the target path.

[0064] In this step, in order to ensure that the vehicle meets a series of safety, compliance and performance requirements during autonomous driving, various safety constraints need to be set based on multiple vehicle body parameters, target path path information and multiple driving indicators to limit the vehicle, such as expansion circle constraints, front wheel steering angle constraints, front wheel steering angle change rate constraints and kinematic equation constraints, which help to optimize the driving strategy and route planning of autonomous vehicles.

[0065] S103. Construct a vehicle control model based on multiple constraints.

[0066] S104. Solve the vehicle control model to obtain the optimal control variables for the vehicle to drive automatically within the target path.

[0067] In steps S103 and S104, considering multiple generated constraints, a vehicle control model based on safe passage is constructed and solved to obtain the optimal control quantities for the vehicle to perform autonomous driving in the first future time period, including reasonable acceleration, deceleration, and smooth turning. By predicting the optimal control quantities, it can be ensured that the vehicle always stays within a safe range during autonomous driving. Following the optimal control quantities minimizes potential collision risks and enhances the safety of autonomous driving.

[0068] S105. Based on the optimal control quantity and vehicle control model, predict the trajectory of the vehicle for automatic driving within the target path.

[0069] In this step, the vehicle control model incorporates the vehicle's dynamic characteristics and can predict the vehicle's autonomous driving state based on the input optimal control variables. In a simulation environment, using the vehicle control model, based on vehicle dynamics and constraints, the model predicts the vehicle's autonomous driving trajectory along the target path. This facilitates subsequent assessment of the vehicle's collision risk during autonomous driving based on the predicted trajectory, effectively improving the traffic safety of autonomous driving.

[0070] The trajectory prediction method for autonomous driving provided in this application constructs multiple constraints based on path information and driving indicators. Based on these constraints, a vehicle control model for autonomous driving is constructed and solved, generating optimal control quantities for path tracking. Finally, based on the optimal control quantities and the vehicle control model, the driving trajectory of the vehicle on the target path is predicted. This method can more clearly explain the relationship between the prediction results and vehicle characteristics, driving environment, and control parameters, providing more intuitive and interpretable trajectory prediction results. Furthermore, this method, based on the vehicle's control and dynamic characteristics and in conjunction with the vehicle's path planning, can more easily exchange and transmit information with other data to achieve high-precision prediction in complex scenarios. This effectively improves the accuracy and robustness of trajectory prediction results, thereby providing accurate data for collision risk assessment during autonomous driving and enhancing the safety and reliability of autonomous driving.

[0071] Furthermore, such as Figure 2 As shown, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, this application provides another trajectory prediction method for autonomous driving of vehicles, which includes:

[0072] S201. Obtain multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of multiple driving indicators for the vehicle to drive automatically within the target path.

[0073] This step and Figure 1 The steps S101 shown are the same and will not be repeated here.

[0074] S202. Based on multiple vehicle body parameters and path information, construct the vehicle's expansion circle constraint.

[0075] In this step, the expanded circle constraint is a motion constraint used to describe the vehicle's lateral movement and collision avoidance capabilities during autonomous driving. Using the vehicle's body parameters and the target path information, the expanded circle constraint is constructed to generate control commands, limiting the driving step size within the expanded circle. This ensures that the generated trajectory can travel within the expanded circle, guaranteeing that the vehicle's trajectory will not collide with roadside obstacles during autonomous driving, thus ensuring the vehicle's safe distance and collision avoidance capabilities.

[0076] In this embodiment of the application, optionally, the multiple vehicle body parameters include the midpoint position of the front axle and the midpoint position of the rear axle, and the path information includes obstacle information. In step S202, the expansion circle constraint of the vehicle is constructed based on the multiple vehicle body parameters and the path information, specifically including: enveloping the outer contour of the vehicle according to the midpoint position of the front axle and the midpoint position of the rear axle respectively to generate a first envelope circle and a second envelope circle; generating a first expansion circle and a second expansion circle according to the first envelope circle, the second envelope circle and the path information; and constructing the expansion circle constraint according to the first expansion circle and the second expansion circle.

[0077] In this embodiment, the vehicle's envelope circle refers to a circular region defined to represent the maximum spatial range occupied by the vehicle during movement. This circular region can be used to represent the vehicle's physical dimensions and dynamic characteristics. During path planning and obstacle avoidance, the vehicle's envelope circle can be used to calculate the required safe distances, ensuring that the vehicle has sufficient space to safely pass through or bypass obstacles during autonomous driving. Specifically, a first envelope circle is formed by drawing a circle with the midpoint of the vehicle's front axle as the center point and the radius from the midpoint of the front axle to the left / right front vertex of the vehicle. Simultaneously, a second envelope circle is formed by drawing a circle with the midpoint of the vehicle's rear axle as the center point and the radius from the midpoint of the rear axle to the left / right rear vertex of the vehicle.

[0078] Furthermore, the expansion circle is a further extension of the vehicle's envelope circle, used to represent the safe distance required for the vehicle's movement. It adds an extra distance to the vehicle's envelope circle to account for the safe separation between the vehicle and surrounding obstacles. Therefore, after constructing the vehicle's envelope circle, an expansion circle is constructed based on the vehicle's envelope circle and the distance between the vehicle and obstacles on the target path to ensure the vehicle can safely avoid obstacles. Subsequently, based on the first and second expansion circles, expansion circle constraints are formed for the vehicle during autonomous driving, helping to determine the safe distance between the vehicle and obstacles and preventing the path from intersecting with obstacles.

[0079] In this embodiment of the application, optionally, generating a first expansion circle and a second expansion circle based on a first envelope circle, a second envelope circle, and path information specifically includes: obtaining a first center of the first envelope circle and a second center of the second envelope circle; forming a straight line along the lateral direction of the autonomous vehicle, passing through the first center; generating a first expansion circle based on the straight line and path information; determining the center-to-center distance based on the first center and the second center; and generating a second expansion circle based on the first center, the center-to-center distance, and path information.

[0080] In this embodiment, the first center of the first envelope circle is obtained, and a straight line is formed along the lateral direction of the vehicle and passing through the first center. Based on the path information of the target path, the passable area and obstacle area in the target path are identified. The passable area is the area where the vehicle can move freely, and the obstacle area is the area that needs to be avoided. Starting from the initial position of the vehicle on the target path, an initial expansion circle is generated, with any point on this straight line as the center. This initial expansion circle is placed in the passable area. The radius of the initial expansion circle is continuously expanded until it contacts the nearest obstacle, ensuring a safe distance between the expansion circle and the obstacle area. Subsequently, while maintaining a safe distance between the expansion circle and the obstacle area, the position of the center of the expansion circle is adjusted to maximize its radius and achieve the maximum expansion effect. Through multiple iterations of the above steps, by continuously adjusting the position and size of the expansion circle, a first expansion circle with a larger radius is found on the straight line passing through the first envelope circle, maximizing the radius of the first expansion circle and effectively isolating the passable area and the obstacle area.

[0081] Furthermore, after constructing the first expansion circle according to the above steps, obtain the second center of the second envelope circle, connect the first center and the second center, take the center-to-center distance between the first center and the second center as the radius, take the first center as the center, and find the second expansion circle such that the radius of the second expansion circle is maximized and can isolate the passable area and the obstacle area.

[0082] In practical application scenarios, such as Figure 3The diagram illustrates the structure of the vehicle's envelope circle and expansion circle. Circles O1 and O2 enclose the vehicle's outer contour, transforming the vehicle collision detection problem into a collision detection problem involving the first envelope circle O1 and the second envelope circle O2. Subsequently, the first envelope circle O1 and the second envelope circle O2 are expanded. A point P is selected near the front of the vehicle, at the midpoint of the front axle, and a Cartesian coordinate system is established with P as the origin. A straight line is constructed with the direction as the y-axis of the coordinate system and passing through the point of the first envelope circle O1. On this line, a first expansion circle A1 is found, maximizing its radius and effectively isolating the passable area from the obstacle area. A second expansion circle A2 is then found on the first envelope circle O1 with radius |O1O2| and center at point O1, maximizing its radius and effectively isolating the passable area from the obstacle area. By constructing a first expansion circle A1 and a second expansion circle A2 that can isolate the passable area and the obstacle area, the collision detection problem of the first envelope circle O1 and the second envelope circle O2 is transformed into a problem where the envelope circle is constrained within the expansion circle. Specifically, when the distance between the centers of the two envelope circles |O1O2| is greater than the distance between the centers of the two expansion circles |A1A2|, the vehicle cannot fully utilize the lateral distance for driving close to the edge in a straight-moving scenario, compressing the feasible solution space and easily leading to solution failure. When the distance between the centers of the two envelope circles |O1O2| is less than the distance between the centers of the two expansion circles |A1A2|, the vehicle cannot fully utilize the lateral distance for driving close to the edge in a straight-moving scenario, compressing the feasible solution space and easily leading to solution failure. When the distance between the centers of the two envelope circles |O1O2| is equal to the distance between the centers of the two expansion circles |A1A2|, the vehicle can fully utilize the lateral distance for driving close to the edge in a straight-moving scenario, maximizing the utilization of the feasible solution space. Therefore, when expanding the first envelope circle O1 and the second envelope circle O2 respectively, it is necessary to ensure that the distance between the centers of the two expansion circles, |A1A2|, is equal to the distance between the centers of the two envelope circles, |O1O2|. The first envelope circle is constrained within the first expansion circle, the second envelope circle is constrained within the second expansion circle, and the vehicle's outer contour is constrained within the envelope circles. That is, the vehicle's outer contour is always constrained within the two expansion circles. Therefore, during autonomous driving, it is necessary to select an appropriate step size to ensure that the envelope circles are always constrained within the expansion circles, thus ensuring that the vehicle's outer contour is always constrained within the expansion circles.

[0083] The above method generates an expansion strategy based on the boundary of the obstacle area of ​​the target path. It searches for the expansion circle based on the envelope circle, which can ensure the existence of the expansion circle and achieve the maximum expansion effect while ensuring the isolation of the passable area and the obstacle area.

[0084] S203. Based on multiple driving indicators and multiple vehicle body parameters, construct front wheel steering angle constraints and front wheel steering angle change rate constraints.

[0085] In this step, the front wheel steering angle constraint refers to the maximum steering angle limit of the vehicle's front wheels during steering, typically determined based on the vehicle's maximum steering capability and handling limitations. This constraint ensures that the vehicle does not exceed its controllable range during steering by limiting the maximum value of the front wheel steering angle. The front wheel steering angle constraint is constructed using multiple vehicle driving indicators and body parameters to ensure that the vehicle's front wheel steering angle remains within an acceptable range.

[0086] Furthermore, the front wheel steering angle change rate constraint refers to the constraint on the rate of change of the front wheel steering angle during vehicle steering. An excessively large front wheel steering angle change rate may generate excessive steering torque, leading to instability or loss of control during steering. Therefore, by using multiple driving indicators and vehicle parameters, a front wheel steering angle change rate constraint is formed to plan the vehicle's steering process, generate a smooth steering trajectory, and ensure that the rate of change of the vehicle's steering angle does not exceed the limit during steering.

[0087] In this embodiment of the application, optionally, the multiple vehicle body parameters include wheelbase. In step S203, front wheel steering angle constraints and front wheel steering angle change rate constraints are constructed based on multiple driving indicators and multiple vehicle body parameters. Specifically, this includes: constructing a first target mapping relationship based on multiple driving indicators and wheelbase, the first target mapping relationship representing the correspondence between front wheel steering angle and lateral acceleration; constructing a second target mapping relationship based on the first target mapping relationship, front wheel steering angle change rate, and lateral acceleration, the second target mapping relationship representing the correspondence between front wheel steering angle change rate and lateral acceleration; constructing a front wheel steering angle constraint based on the maximum value of lateral acceleration and the first target mapping relationship; and constructing a front wheel steering angle change rate constraint based on the maximum value of lateral acceleration and the first target mapping relationship.

[0088] In this embodiment, lateral acceleration is represented by multiple vehicle driving indicators and the wheelbase between the two front wheels / two rear wheels to form a correspondence between lateral acceleration and front wheel steering angle, ensuring stability and handling during cornering in autonomous driving.

[0089] Furthermore, to precisely control the vehicle's dynamic characteristics during cornering in autonomous driving and achieve a higher level of handling performance and safety, a second target mapping relationship is constructed between the front wheel steering angle change rate and lateral acceleration, based on the mapping relationship between the front wheel steering angle and lateral acceleration. This allows for precise control of the vehicle's cornering speed and lateral acceleration during cornering, enabling the vehicle to follow the planned path and improve its stability during cornering.

[0090] Furthermore, given the maximum lateral acceleration and maximum lateral jerk of the vehicle during autonomous driving, the front wheel steering angle constraint of the vehicle is obtained based on the established first target mapping relationship; and the front wheel steering angle change rate constraint of the vehicle is obtained based on the established second target mapping relationship.

[0091] By using the above method, based on the established mapping relationship, appropriate front wheel steering angle constraints and front wheel steering angle change rate constraints are determined, and upper and lower limits of the steering angle are set to ensure that the lateral acceleration is within a safe range and to avoid loss of vehicle control or instability.

[0092] In this embodiment of the application, optionally, the multiple driving indicators include: yaw angle, understeer coefficient, longitudinal speed, turning radius, lateral acceleration, and lateral jerk. Based on the multiple driving indicators and wheelbase, a first target mapping relationship is constructed, specifically including: constructing a third target mapping relationship based on yaw angle, front wheel angle, understeer coefficient, longitudinal speed, turning radius, and lateral acceleration; constructing a fourth target mapping relationship based on wheelbase, turning radius, and front wheel angle; constructing a fifth target mapping relationship based on understeer coefficient and wheelbase; and generating the first target mapping relationship based on the third, fourth, and fifth target mapping relationships.

[0093] In this embodiment, the vehicle yaw angle is the angle of rotation of the vehicle around its vertical axis, representing the vehicle's lateral attitude; the front wheel steering angle is the steering angle of the vehicle's front wheels; the understeer coefficient is the nonlinear relationship between the front wheel steering angle and the yaw angle when the vehicle is moving laterally; the longitudinal velocity is the vehicle's speed in the longitudinal direction; and the turning radius is the radius of curvature of the turning path the vehicle is traveling on. Substituting the yaw angle, understeer coefficient, longitudinal velocity, turning radius, and lateral acceleration into the formula to represent lateral acceleration forms the third objective mapping relationship.

[0094] The third objective mapping relationship is:

[0095] a y =rθ 2 =r(kvtanδ) 2 ;

[0096] Among them, a y θ is the lateral acceleration; r is the turning radius; θ is the yaw angle; k is the understeer coefficient; v is the longitudinal velocity.

[0097] Furthermore, the vehicle wheelbase, front wheel steering angle, and understeer coefficient are imported into the formula to establish a fourth mapping relationship and a fifth target mapping relationship;

[0098] The fourth target mapping relationship is:

[0099]

[0100] Where L is the wheelbase of the vehicle's front / rear axle;

[0101] The fifth objective mapping relationship is as follows:

[0102]

[0103] Furthermore, substituting the fourth and fifth objective mapping relationships into the third objective mapping relationship, we obtain:

[0104]

[0105] Therefore, the first target mapping relationship between the vehicle's front wheel steering angle and lateral acceleration can be obtained as follows:

[0106]

[0107] Furthermore, assuming the vehicle's longitudinal velocity remains stable, the second objective mapping relationship between the front wheel steering angle change rate and the lateral acceleration can be obtained as follows:

[0108]

[0109] Where γ is the rate of change of the front wheel steering angle; Jy is the lateral acceleration.

[0110] In summary, given the maximum lateral acceleration and lateral jerk of the vehicle during autonomous driving, the front wheel steering angle constraint and the front wheel steering angle change rate constraint can be obtained from the first target mapping relationship and the second target mapping relationship.

[0111] S204. Generate the system state vector based on the longitudinal coordinates, lateral coordinates, yaw angle, front wheel angle, and front wheel angle change rate of the vehicle's rear axle midpoint on the target path.

[0112] In this step, the system state vector is a vector representing the vehicle's state, containing relevant information describing the vehicle's motion and environmental perception. It provides crucial input for subsequent autonomous driving calculations, used to generate appropriate control commands. Specifically, the system state vector is defined based on the longitudinal and lateral coordinates of the vehicle's rear axle midpoint on the target path, the vehicle's yaw angle, the vehicle's front wheel steering angle, and the rate of change of the front wheel steering angle.

[0113] S205. Generate the system control vector based on the derivative of the rate of change of the front wheel steering angle.

[0114] In this step, the derivative of the rate of change of the vehicle's front wheel steering angle is typically expressed as the steering angular velocity or steering rate. By adjusting and controlling this variable, precise steering behavior can be achieved, ensuring that the vehicle can accurately travel along a specified path. Therefore, based on the derivative of the rate of change of the vehicle's front wheel steering angle, a system control vector is defined to achieve precise vehicle handling and trajectory tracking.

[0115] S206. Construct the vehicle kinematic equations based on the system state vector and system control vector.

[0116] In this step, kinematic equations describe the motion and behavior of the vehicle within specified time intervals during autonomous driving. By iteratively applying the kinematic equations, information such as the vehicle's position, heading, and speed can be tracked and predicted to achieve precise control and adaptive driving for autonomous driving. Specifically, after generating the vehicle's system state vector and system control vector, the system state vector and system control vector are combined based on the vehicle's kinematic properties to establish kinematic equations that describe the vehicle's motion behavior.

[0117] In practical applications, based on the longitudinal coordinate x, lateral coordinate y, yaw angle θ, front wheel steering angle δ, and front wheel steering angle change rate γ of the vehicle's rear axle midpoint, the system state vector is defined as z = [x yθδγ]. T Based on the derivative of the rate of change of the front wheel rotation angle γ, the system control vector is defined. The vehicle's kinematic equations can then be expressed as:

[0118]

[0119] S207. Discretize the vehicle kinematic equations to generate vehicle kinematic equation constraints.

[0120] In this step, discretization transforms the continuous-time kinematic equations into approximate solutions at discrete time points. In autonomous driving, the purpose of discretization is to utilize discrete-time states and inputs in control algorithms and trajectory planning to better achieve vehicle control and path tracking. Therefore, the vehicle's kinematic equations are approximated, and the vehicle state at discrete time points is calculated iteratively. Subsequently, based on requirements and application scenarios, appropriate constraints are introduced to limit the vehicle's motion. By considering these constraints in the iterative calculations, it can be ensured that the vehicle's motion behavior during autonomous driving meets specific performance indicators and safety requirements, providing a stable, safe, efficient, and expected autonomous driving experience.

[0121] In practical applications, by using the fourth-order Runge-Kutta integral method to discretize the kinematic equations, we can obtain:

[0122]

[0123] Where h is the computational step size; K1, K2, K3, and K4 are the coefficients of the fourth-order Runge-Kutta integral method, which can be expressed as:

[0124]

[0125]

[0126]

[0127]

[0128] S208. Construct a vehicle control model based on multiple constraints.

[0129] S209. Abstract the vehicle control model.

[0130] S210. Using the augmented Lagrangian method, solve the abstracted vehicle control model to obtain the optimal control quantity.

[0131] In steps S208 to S210, a vehicle control model based on a safe passage corridor is established according to the expansion circle constraint, front wheel steering angle constraint, front wheel steering angle change rate constraint, and kinematic equation constraint:

[0132]

[0133]

[0134] Where E represents the set of equality constraints; I represents the set of inequality constraints;

[0135] JN is represented as:

[0136] J N =W N，y (y N -y N，ref ) 2 +W N，θ (θ N -θ N，ref ) 2 +W N，δ (δ N ) 2 ;

[0137] J k Represented as:

[0138] J k =W k，y (y k -y k，ref ) 2 +W k，θ (θk -θ k，ref ) 2 +W k，δ (δ k ) 2 +W k，γ (γ k ) 2 ;

[0139] Among them, W N，y W N，θ W N，δ W k，y W k，θ W k，δ W k，γ y represents the weighting coefficient; k，ref and θ k，ref The reference lateral coordinates and reference yaw angles are used for the midpoint of the vehicle's rear axle.

[0140] Furthermore, the constrained optimization problem described by the vehicle control model is abstracted as follows:

[0141] minf(x)

[0142]

[0143] Furthermore, the simplified constrained optimization problem is solved using the augmented Lagrangian method to obtain the optimal control quantity for path tracking in autonomous vehicle driving based on a safe passage corridor. The specific solution steps are as follows:

[0144] S1: Maintain λ t and μ t Without changing the constraints, solve the unconstrained optimization problem described by the following formula to obtain x. t+1 .

[0145]

[0146] S2: Update the augmented Lagrange multipliers as follows:

[0147]

[0148] S3: Update the penalty parameter as follows:

[0149]

[0150] S4: Update the augmented Lagrangian function using the following formula. If the termination condition is not met, jump to S1 to continue execution.

[0151]

[0152] The projection function ∏ onto the positive semi-axis is... k (λ t+1 +μt+1 c(x) can be expressed as:

[0153]

[0154] The projection function expression on the positive half-axis described in the above formula indicates that the condition c(x) is satisfied. t+1 ) < 0 and λ t+1 The inequality constraint that equals 0 is in the initial state x of the next iteration. t+1 If the function is inactive, the corresponding penalty parameter is set to zero. Furthermore, comparing the above formulas, we can see that the augmented Lagrangian function for the next iteration includes both active inequality constraints and all equality constraints.

[0155] S211. Based on the optimal control quantity and vehicle control model, predict the trajectory of the vehicle for automatic driving within the target path.

[0156] This step and Figure 1 The method for step S105 shown is the same, and will not be repeated here.

[0157] In one embodiment of this application, the execution entity of the vehicle autonomous driving trajectory prediction method of this application embodiment is a vehicle control system capable of controlling vehicle autonomous driving based on a safe passage corridor strategy, specifically located on the server side. For example... Figure 4The diagram shows the structure of a vehicle control system, which includes a safe passage corridor constraint generation module, a lateral acceleration constraint generation module, a lateral jerk constraint generation module, a path tracking control problem construction module for autonomous driving, and a path tracking control problem solving module. The safe passage corridor constraint generation module expands the vehicle's outer contour envelope circle to isolate the passable area and obstacle area. Connecting the passable areas yields the safe passage corridor for autonomous driving, simplifying the data description of the vehicle collision detection problem. The lateral acceleration constraint generation module and the lateral jerk constraint generation module establish linear mapping relationships between front wheel angle and lateral acceleration, and between the front wheel angle change rate and lateral jerk, respectively, thus transforming the lateral acceleration and jerk constraint problems into front wheel angle and front wheel angle change rate constraint problems. The system includes modules for constructing and solving the path tracking control problem during autonomous driving. These modules comprehensively consider constraints such as the expansion circle constraint, front wheel steering angle constraint, and front wheel steering angle change rate constraint to construct the path tracking control problem for autonomous driving. This results in a vehicle control model, and the optimal control quantity for path tracking is solved to generate a predicted trajectory. This trajectory accurately reflects the vehicle's predicted path during autonomous driving, improving the accuracy of collision risk assessment. Furthermore, the output optimal control quantity drives the vehicle away from obstacles during autonomous driving, enhancing safety and efficiency in confined spaces.

[0158] Furthermore, as Figure 1 In a specific implementation of the method, this application provides a trajectory prediction device 300 for autonomous driving of vehicles, applicable to vehicles with autonomous driving functions, such as... Figure 5 As shown, the device includes: an acquisition module 301, a first construction module 302, a second construction module 303, a first generation module 304, and a second generation module 305.

[0159] The acquisition module 301 is used to acquire multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of multiple driving indicators for the vehicle to drive automatically within the target path.

[0160] The first construction module 302 is used to construct multiple constraints for the vehicle to drive automatically on the target path based on multiple vehicle body parameters, path information and multiple driving indicators.

[0161] The second construction module 303 is used to construct a vehicle control model based on multiple constraints.

[0162] The first generation module 304 is used to solve the vehicle control model to obtain the optimal control quantity for the vehicle to drive automatically within the target path.

[0163] The second generation module 305 is used to generate a predicted trajectory for the vehicle to drive automatically within the target path based on the optimal control quantity and the vehicle control model.

[0164] In specific application scenarios, in order to construct multiple constraints for the vehicle to drive automatically on the target path, the first construction module 302 includes: a first construction unit, a second construction unit, a first generation unit, a second generation unit, a third construction unit, and a third generation unit.

[0165] The first building unit is used to construct the vehicle's expansion circle constraint based on multiple vehicle body parameters and path information;

[0166] The second building unit is used to build front wheel steering angle constraints and front wheel steering angle change rate constraints based on multiple driving indicators and multiple vehicle body parameters;

[0167] The first generation unit is used to generate a system state vector based on the longitudinal coordinates, lateral coordinates, yaw angle, front wheel steering angle, and front wheel steering angle change rate of the vehicle's rear axle midpoint on the target path.

[0168] The second generation unit is used to generate the system control vector based on the derivative of the front wheel steering angle change rate;

[0169] The third building unit is used to construct the vehicle kinematic equations based on the system state vector and the system control vector;

[0170] The third generation unit is used to discretize the vehicle kinematic equations and generate constraints for the vehicle kinematic equations.

[0171] In specific application scenarios, multiple vehicle body parameters include the midpoint position of the front axle and the midpoint position of the rear axle, and path information includes obstacle information. In order to construct the vehicle's expansion circle constraint, the first construction unit includes: a first generation sub-unit, a second generation sub-unit, and a construction sub-unit.

[0172] The first generation subunit is used to enclose the outer contour of the vehicle according to the midpoint position of the front axle and the midpoint position of the rear axle, respectively, to generate a first envelope circle and a second envelope circle.

[0173] The second generation subunit is used to generate a first expansion circle and a second expansion circle based on the first envelope circle, the second envelope circle, and path information;

[0174] Construct sub-units to build expansion circle constraints based on the first expansion circle and the second expansion circle.

[0175] In specific application scenarios, in order to generate the first expansion circle and the second expansion circle, the first building unit also includes: an acquisition sub-unit, a third generation sub-unit, a fourth generation sub-unit, a determination sub-unit, and a fifth generation sub-unit.

[0176] Obtain sub-units to obtain the first center of the first envelope circle and the second center of the second envelope circle;

[0177] The third generating sub-unit is used to form a perpendicular line along the lateral direction of the vehicle, passing through the center of the first circle.

[0178] The fourth generation sub-unit is used to generate the first expansion circle based on the straight line and path information;

[0179] Determine the sub-unit, which is used to determine the center-to-center distance based on the first and second center points;

[0180] The fifth generation sub-unit is used to generate the second expansion circle based on the first circle center, the circle center distance, and the path information.

[0181] In specific application scenarios, multiple vehicle body parameters, including wheelbase, are used to construct front wheel steering angle constraints and front wheel steering angle change rate constraints. The first construction module 302 includes: a first construction sub-unit, a second construction sub-unit, a third construction sub-unit, and a fourth construction sub-unit.

[0182] The first construction subunit is used to construct a first target mapping relationship based on multiple driving indicators and wheelbase. The first target mapping relationship represents the correspondence between the front wheel steering angle and the lateral acceleration.

[0183] The second construction subunit is used to construct a second target mapping relationship based on the first target mapping relationship, the front wheel steering angle change rate and the lateral acceleration. The second target mapping relationship characterizes the correspondence between the front wheel steering angle change rate and the lateral acceleration.

[0184] The third construction subunit is used to construct the front wheel steering angle constraint based on the mapping relationship between the maximum lateral acceleration and the first target;

[0185] The fourth sub-unit is used to construct the front wheel steering angle change rate constraint based on the mapping relationship between the maximum lateral acceleration and the first target.

[0186] In specific application scenarios, multiple driving indicators include: yaw angle, understeer coefficient, longitudinal speed, turning radius, lateral acceleration, and lateral jerk. In order to construct the first target mapping relationship, the first construction sub-unit includes: the fifth construction sub-unit, the sixth construction sub-unit, the seventh construction sub-unit, and the sixth generation sub-unit.

[0187] The fifth construction subunit is used to construct the third target mapping relationship based on the yaw angle, front wheel steering angle, understeer coefficient, longitudinal velocity, turning radius, and lateral acceleration;

[0188] The sixth construction subunit is used to construct the fourth target mapping relationship based on the wheelbase, turning radius, and front wheel steering angle;

[0189] The seventh construction sub-unit is used to construct the fifth target mapping relationship based on the understeering coefficient and wheelbase;

[0190] The sixth generation subunit is used to generate the first target mapping relationship based on the third target mapping relationship, the fourth target mapping relationship, and the fifth target mapping relationship.

[0191] In specific application scenarios, in order to obtain the optimal control quantity, the first generation module 304 includes a processing unit and a fourth generation unit.

[0192] The processing unit is used to abstract the vehicle control model;

[0193] The fourth generation unit is used to solve the abstracted vehicle control model using the augmented Lagrangian method to obtain the optimal control quantity.

[0194] The trajectory prediction device 300 for autonomous driving provided in this application constructs multiple constraints based on path information and driving indicators. Based on these constraints, a vehicle control model for autonomous driving is constructed and solved, generating optimal control quantities for path tracking. Finally, based on the optimal control quantities and the vehicle control model, the driving trajectory of the vehicle on the target path is predicted. This method can more clearly explain the relationship between the prediction results and vehicle characteristics, driving environment, and control parameters, providing more intuitive and interpretable trajectory prediction results. Furthermore, this method, based on the vehicle's control and dynamic characteristics and in conjunction with the vehicle's path planning, can more easily exchange and transmit information with other data to achieve high-precision prediction in complex scenarios. This effectively improves the accuracy and robustness of trajectory prediction results, thereby providing accurate data for collision risk assessment during autonomous driving and improving the safety and reliability of autonomous driving.

[0195] In an exemplary embodiment, this application also provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor executes the program stored in the memory to perform the trajectory prediction method for autonomous driving of a vehicle in the above embodiments.

[0196] In an exemplary embodiment, this application also provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the trajectory prediction method for autonomous driving of the vehicle.

[0197] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented in hardware or by using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile readable storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause an electronic device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0198] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application.

[0199] Those skilled in the art will understand that the modules in the apparatus of the implementation scenario can be distributed within the apparatus of the implementation scenario as described, or they can be located in one or more apparatuses different from this implementation scenario with corresponding changes. The modules of the above-mentioned implementation scenario can be combined into one module, or they can be further divided into multiple sub-modules.

[0200] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of the implementation scenario.

[0201] The above disclosures are only a few specific implementation scenarios of this application. However, this application is not limited to these. Any variations that can be conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A trajectory prediction method for autonomous driving of vehicles, characterized in that, Applied to vehicles with autonomous driving capabilities, the method includes: The vehicle acquires multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of the vehicle for automatic driving within the target path. Based on the multiple vehicle body parameters, the path information, and the multiple driving indicators, a number of constraints are constructed for the vehicle to automatically drive on the target path. Based on the aforementioned constraints, a vehicle control model is constructed. Solve the vehicle control model to obtain the optimal control parameters for the vehicle to drive automatically within the target path; Based on the optimal control quantity and the vehicle control model, a predicted trajectory is generated for the vehicle to drive automatically within the target path.

2. The method according to claim 1, characterized in that, The step of constructing multiple constraints for the vehicle to automatically drive on the target path based on the multiple vehicle body parameters, the path information, and the multiple driving indicators specifically includes: Based on the multiple vehicle body parameters and the path information, construct the expansion circle constraint of the vehicle; Based on the multiple driving indicators and the multiple vehicle body parameters, front wheel steering angle constraints and front wheel steering angle change rate constraints are constructed. A system state vector is generated based on the longitudinal and lateral coordinates of the rear axle midpoint of the vehicle on the target path, the yaw angle, the front wheel steering angle, and the rate of change of the front wheel steering angle. The system control vector is generated based on the derivative of the rate of change of the front wheel steering angle; Based on the system state vector and the system control vector, construct the vehicle kinematic equations; Discretize the vehicle kinematic equations to generate vehicle kinematic equation constraints.

3. The method according to claim 2, characterized in that, The multiple vehicle body parameters include the midpoint positions of the front and rear axles, and the path information includes obstacle information. The step of constructing the vehicle's expansion circle constraint based on the multiple vehicle body parameters and the obstacle information specifically includes: Based on the midpoint positions of the front axle and the rear axle, the outer contour of the vehicle is enveloped to generate a first envelope circle and a second envelope circle. Based on the first envelope circle, the second envelope circle, and the path information, a first expansion circle and a second expansion circle are generated; The expansion circle constraint is constructed based on the first expansion circle and the second expansion circle.

4. The method according to claim 3, characterized in that, The step of generating the first expansion circle and the second expansion circle based on the first envelope circle, the second envelope circle, and the path information specifically includes: Obtain the first center of the first envelope circle and the second center of the second envelope circle; A perpendicular line is formed along the lateral direction of the vehicle, passing through the center of the first circle; The first expansion circle is generated based on the straight line and the path information; Determine the center-to-center distance based on the first center and the second center; The second expanded circle is generated based on the first center, the distance between the centers, and the path information.

5. The method according to claim 2, characterized in that, The multiple vehicle body parameters include wheelbase. The step of constructing front wheel steering angle constraints and front wheel steering angle change rate constraints based on the multiple driving indicators and the multiple vehicle body parameters specifically includes: Based on the multiple driving indicators and the wheelbase, a first target mapping relationship is constructed, which characterizes the correspondence between the front wheel steering angle and the lateral acceleration. Based on the first target mapping relationship, the front wheel steering angle change rate, and the lateral acceleration, a second target mapping relationship is constructed, which characterizes the correspondence between the front wheel steering angle change rate and the lateral acceleration. Based on the mapping relationship between the maximum lateral acceleration and the first target, the front wheel steering angle constraint is constructed; Based on the maximum lateral acceleration and the mapping relationship with the first target, the constraint on the front wheel steering angle change rate is constructed.

6. The method according to claim 5, characterized in that, The multiple driving indicators include: yaw angle, understeer coefficient, longitudinal speed, turning radius, lateral acceleration, and lateral jerk. The step of constructing a first target mapping relationship based on the multiple driving indicators and the wheelbase specifically includes: A third target mapping relationship is constructed based on the yaw angle, the front wheel steering angle, the understeer coefficient, the longitudinal velocity, the turning radius, and the lateral acceleration; Based on the wheelbase, the turning radius, and the front wheel steering angle, a fourth target mapping relationship is constructed; Based on the understeering coefficient and the wheelbase, a fifth target mapping relationship is constructed; The first target mapping relationship is generated based on the third target mapping relationship, the fourth target mapping relationship, and the fifth target mapping relationship.

7. The method according to claim 1, characterized in that, The step of solving the vehicle control model to obtain the optimal control quantity of the vehicle specifically includes: The vehicle control model is abstracted. The optimal control quantity is obtained by solving the abstracted vehicle control model using the augmented Lagrange method.

8. A trajectory prediction device for autonomous driving of vehicles, characterized in that, The device, applicable to vehicles with autonomous driving capabilities, includes: The acquisition module is used to acquire multiple vehicle body parameters, path information of the target path, and multiple driving indicator parameters of the vehicle for automatic driving within the target path. The first construction module is used to construct multiple constraints for the vehicle to drive automatically on the target path based on the multiple vehicle body parameters, the path information, and the multiple driving indicators. The second construction module is used to construct a vehicle control model based on the multiple constraints. The first generation module is used to solve the vehicle control model to obtain the optimal control quantity for the vehicle to drive automatically within the target path; The second generation module is used to generate a predicted trajectory for the vehicle to drive automatically within the target path based on the optimal control quantity and the vehicle control model.

9. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the steps of any one of claims 1 to 7 when executing a computer program.

10. A readable storage medium having a computer program stored thereon, characterized in that, When a computer program is executed by a processor, it implements the steps of any one of claims 1 to 7.

Images