Trajectory planning method and device, electronic equipment and storage medium

Abstract

The application discloses a trajectory planning method and device, electronic equipment and a storage medium, wherein the method comprises acquiring scene information of a target vehicle to generate a space-time corridor, wherein the space-time corridor comprises a plurality of steps corresponding to a plurality of time slices in an S-L-T three-dimensional space, and the step is a drivable range of the target vehicle in the corresponding time slice; a trajectory cost function is constructed based on state variables and control variables of the target vehicle, the state variables include position, speed and acceleration, and the control variables include impact degree; under the constraint of the space-time corridor, a driving trajectory of the target vehicle in the plurality of steps is planned based on the trajectory cost function, the driving trajectory comprises a plurality of spline curve segments corresponding to the plurality of steps respectively, and the constraint of the space-time corridor comprises an equality constraint on the state variables of the target vehicle and an inequality constraint on the state variables and the control variables of the target vehicle. In this way, the automatic planning of the driving trajectory of the vehicle in the automatic driving can be realized, and the availability of the automatic driving function of the vehicle is ensured.

Description

Technical Field

[0001] This application belongs to the field of autonomous driving technology, specifically relating to a trajectory planning method and apparatus, electronic device and storage medium. Background Technology

[0002] In recent years, autonomous driving technology has developed rapidly. Its goal is usually to control the vehicle to travel autonomously along the road, to reach the destination as quickly as possible while ensuring the safety of the vehicle itself, and to ensure that it does not pose a direct or indirect threat to the safety of other road users.

[0003] To achieve the above goals, autonomous driving software requires several key systems, one of which is the trajectory planning system. The goal of trajectory planning is to plan a trajectory that meets the requirements of vehicle dynamics. This trajectory needs to be able to avoid surrounding obstacles (vehicles, pedestrians, static obstacles, etc.) and meet the instructions of the decision-making layer (keeping in lane, changing lanes, pulling over). Stable and reliable trajectory planning is one of the foundations for ensuring the availability of autonomous driving functions.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this application is to provide a trajectory planning method for solving the problem of how to achieve trajectory planning for autonomous driving.

[0006] To achieve the above objectives, this application provides a trajectory planning method, the method comprising:

[0007] Scene information of the target vehicle is acquired to generate a spatiotemporal corridor, wherein the spatiotemporal corridor includes multiple steps corresponding to multiple time slices in the SLT three-dimensional space, and the steps represent the drivable range of the target vehicle within the corresponding time slice.

[0008] A trajectory cost function is constructed based on the state variables and control variables of the target vehicle, wherein the state variables include at least one of position, velocity, and acceleration, and the control variables include impact intensity.

[0009] Under the constraints of the spatiotemporal corridor, the driving trajectory of the target vehicle in multiple steps is planned based on the trajectory cost function. The driving trajectory includes multiple spline curve segments corresponding to the multiple steps respectively. The constraints of the spatiotemporal corridor include equality constraints on the target vehicle's state variables and inequality constraints on the target vehicle's state variables and control variables.

[0010] In one embodiment, a trajectory cost function is constructed based on the state variables and control variables of the target vehicle, specifically including:

[0011] A state cost subfunction is constructed based on the state variables of the target vehicle, wherein the state cost subfunction includes state component weights, and the state component weights correspond to the position, velocity, and acceleration;

[0012] A control cost subfunction is constructed based on the control quantity of the target vehicle, wherein the control cost subfunction includes control component weights, and the control component weights correspond to the impact degree;

[0013] Based on the state cost subfunction and the control cost subfunction, the trajectory cost function is constructed.

[0014] In one embodiment, the method further includes:

[0015] Based on the continuity of the spline curve segmented state quantities in adjacent steps of the spatiotemporal corridor, an equality constraint is constructed on the state quantities of the target vehicle.

[0016] And / or, the spline curve is segmented into cubic spline curves.

[0017] In one embodiment, the inequality constraints on the target vehicle state variables and control variables include joint constraints;

[0018] The method further includes:

[0019] Based on the spatiotemporal corridor and Bézier curve, an initial joint constraint is constructed on the target vehicle state variables and control variables, wherein the initial joint constraint is a constraint on the control points of the Bézier curve;

[0020] Determine the coefficient mapping between spline curves and Bézier curves;

[0021] Based on the initial joint constraints and coefficient mapping, joint constraints on the target vehicle state variables and control variables are constructed.

[0022] In one embodiment, the method specifically includes:

[0023] The spatiotemporal corridor is projected onto the ST plane and / or the LT plane, wherein the ST plane is a plane in which a drivable longitudinal trajectory is associated with time, and the LT plane is a plane in which a drivable lateral trajectory is associated with time;

[0024] Determine the intersection boundaries of adjacent steps in the spatiotemporal corridor on the ST plane and / or LT plane;

[0025] The starting and / or ending control points of the Bézier curve in the corresponding step are constrained by the intersection boundary.

[0026] In one embodiment, the method further includes:

[0027] Based on the spatiotemporal corridor and the Bézier curve, initial constraints on the control quantity of the target vehicle are constructed, wherein the initial constraints are inequality constraints on the control points of the Bézier curve;

[0028] Determine the coefficient mapping between spline curves and Bézier curves;

[0029] Based on the initial constraints and coefficient mapping, constraints on the target vehicle control quantity are constructed.

[0030] In one embodiment, under the constraints of the spatiotemporal corridor, the trajectory of the target vehicle within multiple steps is planned based on the trajectory cost function, specifically including:

[0031] Under the constraints of the spatiotemporal corridor, the longitudinal and lateral trajectories of the target vehicle in relation to time are calculated based on the trajectory cost function;

[0032] The longitudinal and lateral trajectories are combined to obtain the driving trajectory of the target vehicle.

[0033] This application also provides a trajectory planning device, comprising:

[0034] The acquisition module is used to acquire scene information of the target vehicle to generate a spatiotemporal corridor, wherein the spatiotemporal corridor includes multiple steps corresponding to multiple time slices in the SLT three-dimensional space, and the steps are the drivable range of the target vehicle in the corresponding time slice.

[0035] A construction module is used to construct a trajectory cost function based on the state variables and control variables of the target vehicle, wherein the state variables include at least one of position, velocity, and acceleration, and the control variables include impact degree;

[0036] The planning module is used to plan the driving trajectory of the target vehicle within multiple steps based on the trajectory cost function under the constraints of the spatiotemporal corridor. The driving trajectory includes multiple spline curve segments corresponding to the multiple steps respectively. The constraints of the spatiotemporal corridor include equality constraints on the target vehicle's state variables and inequality constraints on the target vehicle's state variables and control variables.

[0037] This application also provides an electronic device, including:

[0038] At least one processor; and

[0039] A memory that stores instructions that, when executed by the at least one processor, cause the at least one processor to perform the trajectory planning method as described above.

[0040] This application also provides a machine-readable storage medium storing executable instructions that, when executed, cause the machine to perform the trajectory planning method described above.

[0041] Compared with existing technologies, the trajectory planning method of this application can generate a spatiotemporal corridor based on the scene information of the target vehicle, and use the spatiotemporal corridor to construct equality constraints on the state variables of the target vehicle, as well as inequality constraints on the state variables and control variables of the target vehicle. In this way, combined with the trajectory cost function constructed based on the state variables and control variables of the target vehicle, the trajectory planning of the target vehicle can be transformed into an optimal control problem for solution, ensuring the availability of the vehicle's autonomous driving function.

[0042] On the other hand, by utilizing the differential flatness of cubic spline curves, the driving trajectory in each step of the spatiotemporal corridor can be directly planned using cubic spline curves. While ensuring the smoothness of the driving trajectory, the constraints required for optimal control solution are reduced, thereby improving the efficiency of vehicle trajectory planning.

[0043] On the other hand, by utilizing the properties of Bézier curves, such as convex hull, endpoints, and the fact that the derivative is still a Bézier curve, the spline curves and Bézier curves used for trajectory planning are mapped by coefficients. In this way, the spatiotemporal corridor can impose hard constraints on the driving trajectory within each step of the spatiotemporal corridor from multiple aspects such as position, velocity, acceleration, and impact, thereby ensuring that the planned driving trajectory strictly conforms to the constraint expectation. Attached Figure Description

[0044] Figure 1 This is an application scenario diagram of a trajectory planning method according to an embodiment of this application;

[0045] Figure 2 This is a flowchart of a trajectory planning method according to an embodiment of this application;

[0046] Figure 3 This is a scene diagram of SLT three-dimensional space in a trajectory planning method according to an embodiment of this application;

[0047] Figure 4 This is a schematic diagram of a spatiotemporal corridor generated in SLT three-dimensional space in a trajectory planning method according to an embodiment of this application;

[0048] Figure 5 This is a schematic diagram of the spatiotemporal corridor projected onto the SL plane in a trajectory planning method according to an embodiment of this application;

[0049] Figure 6 This is a schematic diagram of a trajectory planning method according to an embodiment of the present application, in which a spatiotemporal corridor is used to constrain the driving trajectory and project it onto the ST plane;

[0050] Figure 7This is a schematic diagram of a trajectory planning method according to an embodiment of the present application, in which a spatiotemporal corridor is used to constrain the driving trajectory and project it onto the LT plane;

[0051] Figure 8 The trajectory of the target vehicle in the spacetime corridor steps is obtained in the SLT three-dimensional space according to the trajectory planning method of an embodiment of this application.

[0052] Figure 9 A block diagram of a trajectory planning device according to an embodiment of this application;

[0053] Figure 10 This is a hardware structure diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0054] The present application will now be described in detail with reference to the embodiments shown in the accompanying drawings. However, these embodiments do not limit the present application, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the protection scope of the present application.

[0055] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0056] Before introducing the embodiments of this application, the basic technologies and some technical terms involved in the embodiments of this application will be explained illustratively:

[0057] Autonomous driving refers to the ability to guide and make decisions regarding vehicle operation without requiring a driver to perform physical driving maneuvers, thus enabling the vehicle to drive safely. Autonomous driving technology typically includes high-precision mapping, environmental perception, behavioral decision-making, path planning, and motion control.

[0058] Autonomous driving systems: Systems that enable different levels of autonomous driving functions in vehicles, such as driver assistance systems (L2), high-speed autonomous driving systems requiring human supervision (L3), and highly / fully autonomous driving systems (L4 / L5).

[0059] Intelligent Traffic System (ITS), also known as Intelligent Transportation System, effectively integrates advanced science and technology (including information technology, computer technology, data communication technology, sensor technology, electronic control technology, automatic control theory, operational research, artificial intelligence, etc.) into transportation, service control and vehicle manufacturing, and strengthens the connection among vehicles, roads and users, thereby forming a comprehensive transportation system that ensures safety, improves efficiency, improves environment and saves energy.

[0060] Intelligent Vehicle Infrastructure Cooperative Systems (IVICS), referred to as vehicle-infrastructure cooperative system, is a development direction of intelligent transportation systems. The vehicle-infrastructure cooperative system adopts advanced technologies such as wireless communication and the new generation of Internet to realize all-round dynamic real-time information interaction between vehicles and vehicles, and between vehicles and roads. On the basis of full-space-time dynamic traffic information collection and fusion, it carries out active vehicle safety control and road cooperative management, fully realizes effective coordination among human, vehicles and roads, ensures traffic safety and improves traffic efficiency, thereby forming a safe, efficient and environmentally friendly road traffic system.

[0061] Trajectory planning: Given the initial state (including starting position, velocity and acceleration) of the vehicle, the target state (including target position, velocity and acceleration), the position of obstacles, and constraint conditions such as dynamics and comfort, a smooth trajectory is calculated so that the vehicle can reach the target state along the trajectory. Trajectory planning usually includes two parts: path planning and velocity planning. Path planning is responsible for calculating a smooth path from the starting position to the target position, and velocity planning calculates the velocity of each path point based on this path to form a velocity curve.

[0062] Frenet coordinate system: also called road coordinate system, which takes the starting position of the vehicle as the origin, has mutually perpendicular coordinate axes, and is divided into S-axis direction (that is, the tangent direction along the road reference line, called the longitudinal direction) and L-axis direction (that is, the current normal direction of the reference line, called the lateral direction), and the coordinate is represented as (S, L).

[0063] The trajectory planning method provided in the embodiments of the present application can be applied to autonomous vehicles, including L2, L3, L4 and higher level autonomous driving systems.

[0064] Reference Figure 1Taking an application scenario of the trajectory planning method provided by the embodiments of the present application as an example. A user can drive a vehicle manually, or can perform automatic driving by means of the intelligent driving system of the vehicle. Whether in the process of manual driving or automatic driving, a terminal can collect scene information based on sensors, lidar, cameras, millimeter-wave radar, navigation systems, positioning systems, high-precision maps, etc., and provide decision basis information for vehicle control. Wherein, the terminal may be the vehicle driven by the user, or an intelligent on-board device / module on the vehicle, or a desktop computer, a notebook computer, a smart phone and a tablet computer configured on the vehicle when the user drives the vehicle, as well as a portable wearable device carried by the user, etc.

[0065] Reference Figure 2 , an embodiment of the trajectory planning method of the present application is introduced. In this embodiment, the method comprises:

[0066] S11, acquiring scene information of a target vehicle to generate a space-time corridor.

[0067] In automatic driving of vehicles, the scene information may include positioning information, map information, environment information, end point information, vehicle-related information, and the like. Wherein, the vehicle-related information may be related information of the target vehicle and vehicles adjacent thereto, that is, the vehicle-related information may include the speed and acceleration of the target vehicle, and the speed and acceleration of vehicles adjacent to the target vehicle, and the vehicle-related information may further include the positional relationship between the target vehicle and vehicles adjacent thereto.

[0068] Different types of scene information can be obtained respectively through one or more on-vehicle devices. For example, coordinates of the target vehicle in a lane coordinate system can be obtained through a global navigation satellite system (GNSS); relative speed, relative distance and the like between the target vehicle and a preceding vehicle can be obtained through an ultrasonic radar, a camera, or a manner of ultrasonic radar fused with a camera.

[0069] The definition of the space-time corridor can rely on a configured S-L-T three-dimensional space, wherein S and L can refer to the Frenet coordinate system described above, which are respectively called a lateral direction and a longitudinal direction; and T represents a time direction.

[0070] With reference to Figure 3Based on the scene information obtained above, a scenario is illustrated, showing static and dynamic obstacles in the environment surrounding the target vehicle. The x and y axes represent the lateral (S) and longitudinal (L) axes, respectively. Static obstacles, whose positions do not change over time, are represented in SLT 3D space as spatial volumes extending upwards along the T-axis. Dynamic obstacles (such as vehicles), predicted to travel at a velocity v along the x-axis in the future, are represented in SLT 3D space as spatial volumes extending along the T-axis while being diagonally stretched towards the x-axis, with their projection slope on the ST plane representing the velocity v.

[0071] Coordination Figure 4 The combination of the spatial volumes corresponding to the aforementioned static and dynamic obstacles constitutes the spatiotemporal obstacle region for the target vehicle within the SLT three-dimensional space. The spatiotemporal corridor corresponds to this obstacle region, providing the target vehicle with a drivable range within the SLT three-dimensional space by avoiding the space occupied by the obstacle region. It can be seen that the spatiotemporal corridor can be considered as the "solution set" of all drivable trajectories of the target vehicle within the SLT three-dimensional space.

[0072] Coordination Figure 5 The projection of the spacetime corridor onto the SL plane can be seen as a "top view" of the spacetime corridor in the SLT 3D space, where each step occupies the time of a corresponding time slice. It can be understood that the overlap of the projections on the SL plane does not represent the overlap of the steps in the SLT 3D space. In fact, in the SLT 3D space, the steps are continuously stacked on the time axis T.

[0073] Similarly, in conjunction with the participants Figure 6 and Figure 7 The ST and LT planes retain the time axis, representing the longitudinal and lateral trajectories of the drivable path, respectively. Adjacent steps in the spacetime corridor intersect at the projection boundaries of the ST and LT planes.

[0074] The optimal generation of spatiotemporal corridors can rely on the optimization solution of the motion planning layer of autonomous vehicles. Typical generation processes include seed generation, cube expansion and constraint association, and cube relaxation.

[0075] Specifically:

[0076] ① Seed Generation: The seeds for the spatiotemporal corridor are generated by projecting the forward simulation state of the behavior planner onto the SLT 3D space. Since the forward simulation state is discretized, the feasibility of the spatiotemporal corridor generation process depends on the complexity of the environment and the seed resolution. To ensure the success of the spatiotemporal corridor generation process, it is required that the initial cube constructed from consecutive seeds be collision-free. The motivation for generating spatiotemporal corridors around seeds is to fully model the topologically equivalent free space while preserving the same high-level behavior. Here, dilated seeds can be obtained by sampling in both the ST and LT topological spaces (planes).

[0077] ② Cube dilation with semantic boundaries: The spatiotemporal corridor is generated through iterative seeding. Seeds already included in the last dilated cube are skipped because they are topologically equivalent. The initial cube is generated based on two consecutive seeds, which are treated as two cube vertices. A key feature of cube dilation is the consideration of semantic boundaries. The goal of the cube dilation process is to generate cubes that match the semantic boundaries so that constraints can be easily associated.

[0078] Specifically, when the initial cube intersects a semantic boundary, the expansion direction opposite to the entry direction is disabled, so that the expanded cube almost matches the semantic boundary. For an expansion step, expansion alternates between SLT directions, and expansion terminates if the step collides with an obstacle or intersects a semantic boundary.

[0079] ③ Cube Relaxation: After the cube expansion process, the expanded cube almost matches the semantic boundary. However, some constraints, such as lane change duration constraints, are soft constraints and should allow for additional optimization space. Therefore, a cube relaxation process can be used to relax the cube boundaries. The maximum allowable margin for relaxation is systematically determined by the constraints applied to the two consecutive cubes. For example, longitudinally, the margin can be determined based on the speed constraint through the speed matching distance. For laterally (i.e., in the lane change case), the margin can be calculated through the allowable fluctuation of the lane change duration.

[0080] Coordination Figure 4 The spatiotemporal corridor generated by the above method can include multiple steps (i.e., the cubes mentioned above) corresponding to multiple time slices. Each step represents the drivable range of the target vehicle within the corresponding time slice, and each step occupies a time t. k .

[0081] S12. Construct a trajectory cost function based on the state variables and control variables of the target vehicle.

[0082] In various embodiments of this application, it is desirable to use spline curves to plan the travel trajectory of each step in a spacetime corridor. The spline curve can be a common quintic spline curve, cubic spline curve, etc. In particular, in this embodiment, it is desirable to use cubic spline curves to plan the travel trajectory of each step in a spacetime corridor. Correspondingly, the travel trajectory of each step in the spacetime corridor can be referred to as a "spline curve segment".

[0083] Taking the ST plane as an example, the cubic spline curve segmentation in a step of the spacetime corridor can be represented as:

[0084]

[0085] The domain of the cubic spline curve segment extends from t. i To t i+1 That is, the length h of a step in the spacetime corridor along the T direction, which shows that h = t i+1 -t i (i.e., △t), where j represents the sequence number of the steps in the spacetime corridor.

[0086] Let χ(t) = f j (t), χ i and χ i+1 These represent the ordinates of the cubic spline curve segments at their starting and ending positions within the corresponding steps (lateral coordinates in the ST plane and longitudinal coordinates in the LT plane). We can obtain:

[0087]

[0088]

[0089]

[0090]

[0091] in, Representing χ i In t i The first, second, and third derivatives at that point, that is, Corresponding to the target vehicle at t i Velocity, acceleration, and impact intensity at the point.

[0092] Similarly, we can also obtain:

[0093]

[0094]

[0095]

[0096]

[0097] in, Representing χ i+1 In t i+1 The first, second, and third derivatives at that point, that is, Corresponding to the target vehicle at t i+1 Velocity, acceleration, and impact intensity at the point.

[0098] In various embodiments of this application, the state variables may include at least one of the target vehicle's position, velocity, and acceleration, and the control variables may include the target vehicle's impact degree. Here, the trajectory planning method of this embodiment will be described using the example of simultaneously utilizing the position, velocity, and acceleration in the state variables and the impact degree in the control variables to construct a trajectory cost function.

[0099] Specifically, setting state variables Control quantity We can obtain:

[0100]

[0101] Combining equations (2) and (3), equation (4) can be written in matrix form to obtain:

[0102]

[0103] After obtaining the relationship model of the spline curve piecewise in a step of the corresponding spatiotemporal corridor with respect to equation (5), the current state quantity of the target vehicle is... Given the given information, let's define the entire spacetime corridor as follows:

[0104]

[0105]

[0106] in, and These are the state variables and control variables of the target vehicle within the entire spatiotemporal corridor.

[0107] Next, state cost sub-functions can be constructed based on the state variables of the target vehicle, and control cost sub-functions can be constructed based on the control variables of the target vehicle.

[0108] State variables, such as position, velocity, and acceleration, are called "state components." The state cost subfunction can then include state component weights, which correspond to position, velocity, and acceleration. The state cost subfunction can be expressed as:

[0109]

[0110] in, represent The transpose of the matrix is ​​given, where Q represents the state component weights. It can be understood that the state component weights are in matrix form, and the weights of the corresponding state components in the control cost subfunction can be adjusted individually.

[0111] Similarly, if the impact included in the control quantity is referred to as the "control component," then the control cost sub-function includes control component weights, which correspond to the impact. The control cost sub-function can be expressed as:

[0112]

[0113] in, represent The transpose of the matrix is ​​given, where R represents the control component weights. It can be understood that the control component weights are in matrix form, and the weights of the control components in the control cost subfunction can be adjusted.

[0114] Finally, the trajectory cost function is constructed based on the state cost subfunction and the control cost subfunction. Exemplarily, the trajectory cost function can be determined by directly summing or by weighted summing of the state cost subfunction and the control cost subfunction. In one embodiment, considering the goal of minimizing acceleration and impact in the planned driving trajectory, the trajectory cost function can be expressed as:

[0115]

[0116] Here, w1 and w2 are the weights of the corresponding state cost subfunction and control cost subfunction, respectively. It is understood that w1 and / or w2 in the trajectory cost function of equation (9) can also be omitted.

[0117] S13. Under the constraints of the spatiotemporal corridor, plan the driving trajectory of the target vehicle within multiple steps based on the trajectory cost function.

[0118] In this embodiment, the constraints of the spatiotemporal corridor include equality constraints on the target vehicle state variables and inequality constraints on the target vehicle state variables and control variables.

[0119] ① Equality constraints

[0120] In this embodiment, it is desired that the state variables at the spline curve segment connection points of adjacent steps in the spatiotemporal corridor are the same, in order to ensure the smoothness and comfort of the planned driving trajectory. That is, it is desired that the spline curve segment state variables of adjacent steps in the spatiotemporal corridor have continuity.

[0121] Based on equation (4) above, the relationship between the starting and ending positions of the spline curve segments in each step of the spacetime corridor has been given. For two adjacent steps, the ending position of the spline curve segment in the earlier step corresponds to the starting position of the spline curve segment in the later step. Therefore, the continuity of the state quantities of the spline curve segments in each step can be defined by the form of equation (4) itself.

[0122] ② Inequality constraints

[0123] In this embodiment, the planned driving trajectory is a segment of spline curve within the steps of the spacetime corridor. Due to the differential flatness of such polynomial curves, curves of a fixed order can be directly determined by the state of the two endpoints of the curve. Hard constraints cannot be used to constrain the curve coefficients, that is, the path points of the driving trajectory and their derivatives (corresponding to the speed, acceleration, impact, etc. of the target vehicle) cannot be explicitly constrained.

[0124] Bézier curves exhibit the convex hull property, where the convex hull is the smallest convex polygon containing all control points of the Bézier curve. Extending any side of a convex polygon will extend all other sides to that side; therefore, the Bézier curve will always lie within this smallest convex polygon containing all control points. Bézier curves also possess the endpoint property, meaning they only pass through two control points (the starting and ending control points); all other control points are approximations and generally not traversed. Furthermore, the derivative of a Bézier curve is also a Bézier curve.

[0125] Based on the above characteristics of Bézier curves, this application proposes to transform the constraint problem of spline curves into a constraint problem of Bézier curves by using the coefficient mapping between spline curves and Bézier curves, thereby adding hard constraints to the path points and derivatives of the travel trajectory in each step of the spacetime corridor.

[0126] Specifically, based on the spatiotemporal corridor and Bézier curves, initial joint constraints on the target vehicle's state variables and control variables can be constructed, and the coefficient mapping between spline curves and Bézier curves can be determined; then, based on the initial joint constraints and coefficient mapping, joint constraints on the target vehicle's state variables and control variables can be constructed.

[0127] The initial joint constraints are constraints on the control points of the Bézier curve. These control points mathematically correspond to the coefficients of the Bézier curve, and therefore can be converted to each other through the coefficient mapping between spline curves and Bézier curves. The following will use the mapping between cubic spline curves and third-order Bézier curves as an example for specific explanation.

[0128] First, the cubic spline curve f(t) and the third-order Bézier curve B(t) in each step of the spacetime corridor can be represented piecewise as follows:

[0129]

[0130]

[0131] Where t is the range of each time slice, C is the control point of each Bézier curve segment, and b is the Bernstein basis function.

[0132] In this embodiment, in order to correspond to the cubic spline curve segment, the Bézier curve segment should be a third-order Bézier curve, that is, each Bézier curve segment in equation (10) has four control points C (i = 0 to 3).

[0133] The cubic spline curve f(t) of equation (10) and the Bézier curve B(t) of equation (11) can be written in matrix form:

[0134]

[0135]

[0136] Next, the order is:

[0137]

[0138]

[0139]

[0140] Based on this, assuming that the segments of the Bézier curve and the spline curve are the same, then we can have:

[0141]

[0142] It can be seen that equation (14) can be understood as the coefficient mapping between spline curves and Bézier curves.

[0143] Based on equation (14), the spline curve can be converted into the form of Bézier curve control points. The derivation process is as follows.

[0144] Equation (2) can be written in matrix form:

[0145]

[0146] Combining equations (14) and (15), we can obtain:

[0147]

[0148] make:

[0149]

[0150] Since the length h (i.e. Δt) of each step in the spacetime corridor in the T direction may not be the same, the mapping relationship between the Bézier curve segment and the spline curve segment in each step may not be the same, and the mapping needs to be considered segment by segment.

[0151] In this embodiment, since the Bézier curve has constraints on all control points, and the (third-order) Bézier curve includes the state and control variables of the target vehicle mentioned in this embodiment, an initial joint constraint can be determined based on the derivative formula of the Bézier curve.

[0152] The formula for the (k) derivative of an (n)-order Bézier curve is expressed as:

[0153]

[0154] For ease of representation, we convert Δt in equation (11) to the form h, and obtain:

[0155]

[0156] By combining equations (18) and (19) and performing substitutions, we can obtain:

[0157]

[0158] Equation (20) determines the forms of the first (k=1), second (k=2), and third (k=3) derivatives of the Bézier curve, thereby constraining the position, velocity, acceleration, and impact of the target vehicle's trajectory at each control point. Thus, the initial joint constraints on the target vehicle can be written as:

[0159]

[0160] Since the spline curves and Bézier curves defined in this application are two-dimensional curves projected onto the ST and LT planes from the travel trajectory in the spatiotemporal corridor, the spatiotemporal corridor constraints are also applied using the projection of the spatiotemporal corridor onto the ST and LT planes, respectively. In the ST plane, p j_min v min a min j min p j_max v max a max j max These are the constraints imposed by each step of the spacetime corridor on the target vehicle's minimum lateral position, minimum velocity, minimum acceleration, minimum impact, maximum position, maximum velocity, maximum acceleration, and maximum impact; similarly, in the LT plane, p j_min v min amax j min p j_max v max a max j max These are the constraints imposed by each step of the spacetime corridor on the target vehicle's longitudinal position, minimum speed, minimum acceleration, minimum impact, maximum position, maximum speed, maximum acceleration, and maximum impact.

[0161] When constraining the position of the target vehicle at the control point using the spatiotemporal corridor, the starting and / or ending control points of the Bézier curve segments in each step of the spatiotemporal corridor can be constrained as shown below.

[0162] Continue to cooperate with the participants Figure 6 and Figure 7 The spacetime corridor can be projected onto the ST plane and / or LT plane, and the intersection boundary of each adjacent step in the spacetime corridor on the ST plane and / or LT plane can be determined. Then, the starting control point and / or ending control point of the Bézier curve segment in the corresponding step can be constrained by the intersection boundary.

[0163] The endpoint property of Bézier curves can be used to achieve this constraint. Taking the ST plane as an example, the position of the termination control point of the Bézier curve segment in the first block is constrained in the interval [0, 100], while the position of the starting control point of the Bézier curve segment in the second block is constrained in the interval [50, 150]. Obviously, the intersection of these two intervals (the intersection boundary) can jointly constrain the aforementioned termination control point and starting control point, that is, both are located on this intersection boundary.

[0164] Furthermore, equation (21) can be written in the following form:

[0165]

[0166] in, and φ represents the minimum and maximum constraint quantities of the initial joint constraints, respectively. j Representative and The coefficient matrix of the multiplication.

[0167] Furthermore, the state and control variables of the travel trajectory in each step of the spacetime corridor can be written in the following matrix form:

[0168]

[0169] By substituting equation (23) with equations (16) and (17), we can obtain:

[0170]

[0171] Furthermore, H in equation (24) can be... M Transpose the terms to the right side of the equation and write them as piecewise control points of the Bézier curve. The relationship between the state variables and control variables of the driving trajectory is as follows:

[0172]

[0173] Finally, based on equations (25) and (22), the hard constraint on the travel trajectory (i.e., spline curve) in the spatiotemporal corridor can be obtained as follows:

[0174]

[0175] Among them, b unequ_min and Corresponding to b unequ_max and correspond.

[0176] Based on the joint constraints on the target vehicle state variables and control variables constructed above, this embodiment can further impose separate constraints on the target vehicle control variables to ensure the comfort of the planned driving trajectory of the target vehicle.

[0177] Similarly, initial constraints on the target vehicle control quantity can be constructed based on the spatiotemporal corridor and Bézier curves. Then, by combining the coefficient mapping of spline curves and Bézier curves, constraints on the target vehicle control quantity can be constructed. The initial constraints here are also inequality constraints on the control points of the Bézier curves. The derivation process of converting the initial constraints into constraints on the target vehicle control quantity can refer to the derivation of the joint constraints mentioned above, and will not be repeated here.

[0178] The inequality constraints on the target vehicle control quantity take the following form:

[0179] box unequ_min ≤U≤box unequ_max (27)

[0180] In this embodiment, under the equality constraints of Equation (4) and the inequality constraints of Equations (26) and (27), the trajectory of the target vehicle in multiple steps of the spatiotemporal corridor can be planned with the goal of minimizing the trajectory cost function of Equation (9).

[0181] Coordination Figure 8 As described above, in various embodiments of this application, the spatiotemporal corridor can be projected onto the ST plane and the LT plane to constrain the target vehicle's driving trajectory (spline curve). Therefore, when planning the driving trajectory, the solution obtained is the lateral trajectory on the ST plane and the longitudinal trajectory on the LT plane, and then the longitudinal trajectory and the lateral trajectory can be merged to obtain the target vehicle's driving trajectory in the spatiotemporal corridor (SLT space).

[0182] Reference Figure 9 , an embodiment of the trajectory planning apparatus of the present application is described. In this embodiment, the trajectory planning apparatus comprises an acquisition module 21, a construction module 22, and a planning module 23.

[0183] The acquisition module is configured to acquire scene information of a target vehicle to generate a spatiotemporal corridor, wherein the spatiotemporal corridor comprises a plurality of steps corresponding to a plurality of time slices in an S-L-T three-dimensional space, and the steps are drivable ranges of the target vehicle within the corresponding time slices; the construction module is configured to construct a trajectory cost function based on state quantities and control quantities of the target vehicle, wherein the state quantities comprise at least one of position, speed and acceleration, and the control quantities comprise jerk; the planning module is configured to plan a driving trajectory of the target vehicle in the plurality of steps based on the trajectory cost function under the constraint of the spatiotemporal corridor, wherein the driving trajectory comprises a plurality of spline curve segments respectively corresponding to the plurality of steps, and the constraints of the spatiotemporal corridor comprise equality constraints on state quantities of the target vehicle and inequality constraints on state quantities and control quantities of the target vehicle.

[0184] In one embodiment, the construction module 22 is specifically configured to construct a state cost sub-function based on the state quantities of the target vehicle, wherein the state cost sub-function comprises state component weights, and the state component weights correspond to the position, speed and acceleration; construct a control cost sub-function based on the control quantities of the target vehicle, wherein the control cost sub-function comprises control component weights, and the control component weights correspond to the jerk; construct the trajectory cost function based on the state cost sub-function and the control cost sub-function.

[0185] In one embodiment, the planning module 23 is further configured to construct an equality constraint on the state quantities of the target vehicle based on continuity of state quantities of the spline curve segments in adjacent steps of the spatiotemporal corridor.

[0186] In one embodiment, the spline curve segments are cubic spline curves.

[0187] In one embodiment, the inequality constraints on the state quantities and control quantities of the target vehicle comprise joint constraints;

[0188] the planning module 23 is further configured to construct an initial joint constraint on the state quantities and control quantities of the target vehicle based on the spatiotemporal corridor and a Bezier curve, wherein the initial joint constraint is a constraint on control points of the Bezier curve; determine a coefficient mapping between a spline curve and the Bezier curve; construct a joint constraint on the state quantities and control quantities of the target vehicle based on the initial joint constraint and the coefficient mapping.

[0189] In one embodiment, the planning module 23 is specifically used to project the spatiotemporal corridor onto the ST plane and / or the LT plane, wherein the ST plane is a plane in which the drivable longitudinal trajectory is associated with time, and the LT plane is a plane in which the drivable lateral trajectory is associated with time; determine the intersection boundary of each adjacent step in the spatiotemporal corridor on the ST plane and / or the LT plane; and constrain the start control point and / or end control point of the Bézier curve in the corresponding step with the intersection boundary.

[0190] In one embodiment, the planning module 23 is further configured to construct initial constraints on the target vehicle control quantity based on the spatiotemporal corridor and the Bézier curve, wherein the initial constraints are inequality constraints on the control points of the Bézier curve; determine the coefficient mapping of the spline curve and the Bézier curve; and construct constraints on the target vehicle control quantity based on the initial constraints and the coefficient mapping.

[0191] In one embodiment, the planning module 23 is specifically used to calculate the longitudinal and lateral trajectories of the target vehicle in relation to time based on the trajectory cost function under the constraints of the spatiotemporal corridor; and to merge the longitudinal and lateral trajectories to obtain the driving trajectory of the target vehicle.

[0192] As referred above Figures 1 to 8 The trajectory planning method according to embodiments of this specification has been described. The details mentioned in the above description of the method embodiments also apply to the trajectory planning device of the embodiments of this specification. The above trajectory planning device can be implemented in hardware, software, or a combination of hardware and software.

[0193] Figure 10 A hardware structure diagram of an electronic device according to an embodiment of this specification is shown. Figure 10 As shown, the electronic device 30 may include at least one processor 31, a memory 32 (e.g., non-volatile memory), a memory 33, and a communication interface 34, and the at least one processor 31, memory 32, memory 33, and communication interface 34 are connected together via an internal bus 35. The at least one processor 31 executes at least one computer-readable instruction stored or encoded in the memory 32.

[0194] It should be understood that the computer-executable instructions stored in memory 32, when executed, cause at least one processor 31 to perform the above-described combinations in the various embodiments of this specification. Figures 1 to 8 The description includes various operations and functions.

[0195] In the embodiments of this specification, electronic device 30 may include, but is not limited to: personal computer, server computer, workstation, desktop computer, laptop computer, notebook computer, mobile electronic device, smartphone, tablet computer, cellular phone, personal digital assistant (PDA), handheld device, messaging device, wearable electronic device, consumer electronic device, etc.

[0196] According to one embodiment, a program product, such as a machine-readable medium, is provided. The machine-readable medium may have instructions (i.e., the elements implemented in software as described above), which, when executed by a machine, cause the machine to perform the above-described combinations of the various embodiments of this specification. Figures 1-8 The various operations and functions described. Specifically, a system or apparatus equipped with a readable storage medium storing software program code that implements the functions of any of the embodiments described above, and enabling the computer or processor of the system or apparatus to read and execute the instructions stored in the readable storage medium.

[0197] In this case, the program code read from the readable medium itself can perform the functions of any of the above embodiments, and therefore the machine-readable code and the readable storage medium storing the machine-readable code constitute a part of this specification.

[0198] Examples of readable storage media include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD-RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer or the cloud via a communication network.

[0199] Those skilled in the art will understand that the various embodiments disclosed above can be modified and varied without departing from the spirit of the invention. Therefore, the scope of protection of this specification should be defined by the appended claims.

[0200] It should be noted that not all steps and units in the above process and system structure diagrams are mandatory; some steps or units can be omitted according to actual needs. The execution order of each step is not fixed and can be determined as needed. The device structure described in the above embodiments can be a physical structure or a logical structure. That is, some units may be implemented by the same physical client, or some units may be implemented by multiple physical clients, or they may be jointly implemented by certain components in multiple independent devices.

[0201] In the above embodiments, the hardware units or modules can be implemented mechanically or electrically. For example, a hardware unit, module, or processor may include permanent dedicated circuitry or logic (such as a dedicated processor, FPGA, or ASIC) to perform the corresponding operation. The hardware unit or processor may also include programmable logic or circuitry (such as a general-purpose processor or other programmable processor), which can be temporarily configured by software to perform the corresponding operation. The specific implementation method (mechanical, dedicated permanent circuitry, or temporarily configured circuitry) can be determined based on cost and time considerations.

[0202] The specific embodiments described above with reference to the accompanying drawings are exemplary embodiments, but do not represent all embodiments that can be implemented or fall within the scope of the claims. The term "exemplary" as used throughout this specification means "serving as an example, instance, or illustration" and does not imply that it is "preferred" or "advantageous" compared to other embodiments. Specific details are included to provide an understanding of the described techniques. However, these techniques can be practiced without these specific details. In some instances, well-known structures and apparatuses are shown in block diagram form to avoid obscuring the concepts of the described embodiments.

[0203] The foregoing description of this disclosure is provided to enable any person skilled in the art to implement or use this disclosure. Various modifications to this disclosure will be apparent to those skilled in the art, and the general principles applicable herein can be applied to other variations without departing from the scope of this disclosure. Therefore, this disclosure is not limited to the examples and designs described herein, but is consistent with the widest scope of the principles and novel features disclosed herein.

Claims

1. A trajectory planning method, characterized in that, The method includes: Scene information of the target vehicle is acquired to generate a spatiotemporal corridor, wherein the spatiotemporal corridor includes multiple steps corresponding to multiple time slices in the SLT three-dimensional space, and the steps represent the drivable range of the target vehicle within the corresponding time slice. A trajectory cost function is constructed based on the state variables and control variables of the target vehicle, wherein the state variables include at least one of position, velocity, and acceleration, and the control variables include impact intensity. Under the constraints of the spatiotemporal corridor, the driving trajectory of the target vehicle within multiple steps is planned based on the trajectory cost function. The driving trajectory includes multiple spline curve segments corresponding to the multiple steps. The constraints of the spatiotemporal corridor include equality constraints on the target vehicle's state variables and inequality constraints on the target vehicle's state variables and control variables. The equality constraints are constructed based on the continuity of the spline curve segment state variables in adjacent steps of the spatiotemporal corridor. The inequality constraints include joint constraints. Constructing the joint constraints involves building initial joint constraints on the target vehicle's state variables and control variables based on the spatiotemporal corridor and Bézier curves, where the initial joint constraints are constraints on the control points of the Bézier curves; determining the coefficient mapping between the spline curves and the Bézier curves; and constructing joint constraints on the target vehicle's state variables and control variables based on the initial joint constraints and the coefficient mapping.

2. The trajectory planning method according to claim 1, characterized in that, The trajectory cost function is constructed based on the state variables and control variables of the target vehicle, specifically including: A state cost subfunction is constructed based on the state variables of the target vehicle, wherein the state cost subfunction includes state component weights, and the state component weights correspond to the position, velocity, and acceleration; A control cost subfunction is constructed based on the control quantity of the target vehicle, wherein the control cost subfunction includes control component weights, and the control component weights correspond to the impact degree; Based on the state cost subfunction and the control cost subfunction, the trajectory cost function is constructed.

3. The trajectory planning method according to claim 1, characterized in that, The spline curve is segmented into cubic spline curves.

4. The trajectory planning method according to claim 1, characterized in that, The method specifically includes: The spatiotemporal corridor is projected onto the ST plane and / or the LT plane, wherein the ST plane is a plane in which a drivable longitudinal trajectory is associated with time, and the LT plane is a plane in which a drivable lateral trajectory is associated with time; Determine the intersection boundaries of adjacent steps in the spatiotemporal corridor on the ST plane and / or LT plane; The starting and / or ending control points of the Bézier curve in the corresponding step are constrained by the intersection boundary.

5. The trajectory planning method according to claim 1, characterized in that, The method further includes: Based on the spatiotemporal corridor and the Bézier curve, initial constraints on the control quantity of the target vehicle are constructed, wherein the initial constraints are inequality constraints on the control points of the Bézier curve; Determine the coefficient mapping between spline curves and Bézier curves; Based on the initial constraints and coefficient mapping, constraints on the target vehicle control quantity are constructed.

6. The trajectory planning method according to claim 1, characterized in that, Under the constraints of the spatiotemporal corridor, the trajectory of the target vehicle within multiple steps is planned based on the trajectory cost function, specifically including: Under the constraints of the spatiotemporal corridor, the longitudinal and lateral trajectories of the target vehicle in relation to time are calculated based on the trajectory cost function; The longitudinal and lateral trajectories are combined to obtain the driving trajectory of the target vehicle.

7. A trajectory planning device, characterized in that, include: The acquisition module is used to acquire scene information of the target vehicle to generate a spatiotemporal corridor, wherein the spatiotemporal corridor includes multiple steps corresponding to multiple time slices in the SLT three-dimensional space, and the steps are the drivable range of the target vehicle in the corresponding time slice. A construction module is used to construct a trajectory cost function based on the state variables and control variables of the target vehicle, wherein the state variables include at least one of position, velocity, and acceleration, and the control variables include impact degree; The planning module is used to plan the driving trajectory of the target vehicle within multiple steps based on the trajectory cost function under the constraints of the spatiotemporal corridor. The driving trajectory includes multiple spline curve segments corresponding to the multiple steps. The constraints of the spatiotemporal corridor include equality constraints on the target vehicle's state variables and inequality constraints on the target vehicle's state variables and control variables. The equality constraints are constructed based on the continuity of the spline curve segment state variables in adjacent steps of the spatiotemporal corridor. The inequality constraints include joint constraints. Constructing the joint constraints involves constructing initial joint constraints on the target vehicle's state variables and control variables based on the spatiotemporal corridor and Bézier curves, wherein the initial joint constraints are constraints on the control points of the Bézier curves; determining the coefficient mapping between the spline curves and the Bézier curves; and constructing joint constraints on the target vehicle's state variables and control variables based on the initial joint constraints and the coefficient mapping.

8. An electronic device, comprising: At least one processor; as well as A memory that stores instructions that, when executed by the at least one processor, cause the at least one processor to perform the trajectory planning method as described in any one of claims 1 to 6.

9. A machine-readable storage medium storing executable instructions that, when executed, cause the machine to perform the trajectory planning method as described in any one of claims 1 to 6.

Images