Automatic driving track generation method and device, electronic equipment and read storage medium
By generating autonomous driving trajectories using a spatiotemporal semantic corridor framework, the safety and real-time performance issues of trajectory planning in complex urban environments are resolved. This achieves unified expression of semantic elements and efficient trajectory generation, ensuring the safety and real-time performance of the trajectories.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-03-13
AI Technical Summary
In complex urban environments, the safety and real-time performance of autonomous driving trajectory planning are difficult to guarantee. In particular, in scenarios with diverse semantic elements, traditional methods cannot uniformly describe and process multiple semantic elements, resulting in poor trajectory stability and versatility, and difficulty in balancing real-time performance and optimization.
The Spatio-temporal Semantic Corridor (SSC) framework is adopted to generate the drivable area of the vehicle by unifying the representation of semantic elements. An initial cube is generated using a three-dimensional configuration space, and the safety and real-time performance of trajectory planning are ensured through expansion strategies and constraint associations.
It achieves safety and real-time performance in trajectory planning in complex urban environments. By abstracting semantic elements into spatiotemporal semantic corridors, it simplifies and unifies the processing of various semantic elements, ensuring that the generated trajectory conforms to the dynamic and static constraints of the actual driving environment and avoids collision and line crossing issues.
Smart Images
Figure CN121650701A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of autonomous driving technology, and in particular to an autonomous driving trajectory generation method, device, electronic device, and read storage medium. Background Technology
[0002] In the field of autonomous driving, trajectory planning in complex urban environments is a core challenge.
[0003] Complex urban environments are filled with diverse semantic elements, including dynamic obstacles (such as pedestrians and vehicles), static obstacles (such as buildings and curbs), traffic lights, speed limits, etc.
[0004] Trajectory planning mainly includes: (1) Global path planning: Based on high-precision maps, plan the optimal route from the starting point to the end point; (2) Local behavior decision-making: Determine strategies such as overtaking, following, and lane changing, which comply with traffic rules and driving etiquette; (3) Motion trajectory generation: Generate smooth and dynamically feasible specific paths.
[0005] In related technologies, the safety and real-time performance of trajectory planning for autonomous driving cannot be guaranteed. Summary of the Invention
[0006] This application provides an autonomous driving trajectory generation method, device, electronic device, and read storage medium to ensure the safety and real-time performance of trajectory planning by unifying the representation of semantic elements and globally optimizing the trajectory.
[0007] The embodiments of this application adopt the following technical solutions:
[0008] In a first aspect, embodiments of this application provide an autonomous driving trajectory generation method, wherein the trajectory generation method includes:
[0009] In response to the vehicle's coarse trajectory state sequence information, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space, the preset three-dimensional configuration space including at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t;
[0010] Based on S, L, and t in the preset three-dimensional configuration space, an initial cube is generated in the vertical, horizontal, and time dimensions; and
[0011] The initial cube is used to generate the drivable area of the vehicle by applying an expansion strategy and constraint association.
[0012] In some embodiments, generating the drivable area of the vehicle by associating the initial cube with expansion strategies and constraints includes:
[0013] Based on the initial cube, the dynamic association speed limit, time constraint, and obstacle occupancy area in the expansion strategy are used to ensure that the generated spatiotemporal corridor conforms to the dynamic or static semantic constraints in the actual driving environment during the expansion process.
[0014] After the initial cube expands, a drivable area is formed that adheres to speed limits while avoiding collisions with obstacles.
[0015] In some embodiments, the dynamic or static semantic constraints include static obstacles and dynamic obstacles. Generating the drivable area of the vehicle by associating the initial cube with the constraints according to an expansion strategy includes:
[0016] The static obstacle is represented as an impassable area in the preset three-dimensional configuration space throughout the entire time period. ,
[0017] Wherein, the s max s min These represent the maximum and minimum longitudinal displacements, respectively. max l min These represent the maximum and minimum lateral displacements, respectively. Indicates the entire time period;
[0018] The dynamic obstacle is rendered as a dynamically occupied area in the preset three-dimensional configuration space according to the predicted trajectory;
[0019] Speed limit Wherein v min v max Related to the vehicle's speed limit; where ds / dt represents the vehicle's speed limit;
[0020] Time restrictions are imposed by limiting the range of time the vehicle can travel. , wherein t min t max Related to the vehicle's travel time, t represents the vehicle's travel time.
[0021] In some embodiments, generating an initial cube in the longitudinal, lateral, and temporal dimensions based on S, L, and t in the preset three-dimensional configuration space includes:
[0022] Select at least two consecutive trajectory points from the coarse trajectory state sequence information. , Generate the initial cube;
[0023] The longitudinal range of the initial cube Depend on and The longitudinal position determines;
[0024] The lateral range of the initial cube Determined by lane width and lateral safety distance;
[0025] The time range of the initial cube The time step is a fixed value.
[0026] In some embodiments, generating a drivable area for the vehicle by expanding the initial cube according to an expansion strategy and constraint association further includes:
[0027] The initial cube is gradually expanded along the three directions S, L, and t, and continuous monitoring is required during the expansion process.
[0028] Whether it intersects with an obstacle, wherein the obstacle includes the perceived location information of the obstacle;
[0029] Whether the semantic boundary has been crossed, the semantic boundary includes speed limit and dynamic obstacle area;
[0030] Additionally, the expansion direction stopping condition is set to stop when the cube boundary first touches the obstacle boundary or the velocity limit boundary.
[0031] In some embodiments, projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space in response to the vehicle's coarse trajectory state sequence information includes:
[0032] Obtain the coarse trajectory state sequence information obtained through three-dimensional dynamic programming. Each trajectory point therein includes at least the vehicle's position, velocity, acceleration, and heading angle at discrete time intervals;
[0033] Ensure that the generated spatiotemporal corridor includes the vehicle's final target location, based on the vehicle's target state.
[0034] In some embodiments, the step of projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space in response to the vehicle's coarse trajectory state sequence information further includes:
[0035] The acquired coarse trajectory state sequence information of the vehicle is projected into the SLt space, where S represents the longitudinal displacement along the longitudinal coordinate of the reference lane, L represents the lateral displacement relative to the lane, and t represents the time of the lateral coordinate.
[0036] Secondly, embodiments of this application also provide an autonomous driving trajectory generation device, wherein the trajectory generation device includes:
[0037] The response module is used to respond to the coarse trajectory state sequence information of the vehicle and project the coarse trajectory state sequence information onto a preset three-dimensional configuration space. The preset three-dimensional configuration space includes at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t.
[0038] An initial cube generation module is used to generate an initial cube in the longitudinal, lateral, and temporal dimensions based on S, L, and t in the preset three-dimensional configuration space; and
[0039] The expansion and constraint module is used to generate the drivable area of the vehicle by associating the initial cube with expansion strategies and constraints.
[0040] Thirdly, embodiments of this application also provide an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the above-described method.
[0041] Fourthly, embodiments of this application also provide a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the above-described method.
[0042] The at least one technical solution adopted in this application embodiment can achieve the following beneficial effects: In response to the coarse trajectory state sequence information of the vehicle, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space. Specifically, the preset three-dimensional configuration space includes at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and time t. Based on the S, L, and t in the preset three-dimensional configuration space, an initial cube is generated in the longitudinal, lateral, and time ranges. Finally, the initial cube is used to generate the drivable area of the vehicle according to the expansion strategy and constraint association.
[0043] The drivable area obtained by the above method takes into account static information such as map features, avoiding problems such as crossing lines and hitting curbs, and ensuring the effectiveness of the drivable area. In addition, by abstracting semantic elements in complex urban environments (such as dynamic obstacles, static obstacles, speed limits, etc.) into semantic boundaries in the spatiotemporal semantic corridor (SSC), a unified expression of semantic elements is achieved, overcoming the problem of inconsistent handling of multiple semantic elements in traditional trajectory planning methods. Attached Figure Description
[0044] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0045] Figure 1 This is a flowchart illustrating the autonomous driving trajectory generation method in the embodiments of this application;
[0046] Figure 2 This is a schematic diagram illustrating the expansion of the spatiotemporal semantic corridor in the autonomous driving trajectory generation method in this application embodiment;
[0047] Figure 3 This is a schematic diagram of the autonomous driving trajectory generation device in the embodiments of this application;
[0048] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation
[0049] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0050] Autonomous driving trajectory generation in complex urban environments still faces the following challenges:
[0051] (1) Diversity and non-uniformity of semantic elements: Specifically, the physical meanings and constraint properties of semantic elements are different. For example, an obstacle may be impassable, while speed limits and traffic lights are dynamic constraints. However, this diversity makes it difficult to describe the planning framework uniformly, which in turn affects the stability and universality of the trajectory.
[0052] (2) Difficulty in ensuring global trajectory safety: Most traditional trajectory generation methods are based on sampling points or discretized trajectory checks, which only verify the constraints on discrete points. This approach may miss inter-point constraints and cannot guarantee the safety of the trajectory globally, especially in complex dynamic scenarios.
[0053] (3) It is difficult to balance real-time performance and optimization: In dynamic environments, trajectory planning needs to be completed quickly to adapt to environmental changes, but traditional optimization methods are difficult to meet actual needs in terms of computational complexity and adaptability.
[0054] To address the aforementioned issues, this invention proposes an autonomous driving trajectory generation method that can generate trajectories based on a spatio-temporal semantic corridor (SSC). Furthermore, by unifying the representation of semantic elements and globally optimizing trajectories, the method ensures the safety and real-time performance of trajectory planning.
[0055] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.
[0056] This application provides an autonomous driving trajectory generation method, such as... Figure 1 The diagram shows a flowchart of an autonomous driving trajectory generation method in an embodiment of this application. The method includes at least the following steps S110 to S130:
[0057] Step S110: In response to the coarse trajectory state sequence information of the vehicle, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space, the preset three-dimensional configuration space including at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t.
[0058] For autonomous vehicles, a coarse trajectory state sequence is first generated based on upstream dynamic programming. It can be understood that each trajectory point in the coarse trajectory state sequence contains key information such as the vehicle's position, velocity, acceleration, and heading angle at specific discrete time intervals.
[0059] Then, the coarse trajectory state sequence information needs to be projected into a preset three-dimensional configuration space, that is, the obtained state sequence needs to be projected into the SLT space. The SLT space includes ①S longitudinal displacement along the reference lane; ②L lateral displacement relative to the lane; ③t time.
[0060] Step S120: Generate an initial cube in the longitudinal, transverse, and time ranges according to S, L, and t in the preset three-dimensional configuration space.
[0061] An initial cube is generated in the longitudinal, lateral, and time domains based on the longitudinal displacement along the reference lane in SLT space; the lateral displacement L relative to the lane; and time t.
[0062] It is understandable that the initial cube includes the definition of multiple dimensions.
[0063] Step S130: Generate the drivable area of the vehicle by expanding the initial cube according to the expansion strategy and constraint association.
[0064] Based on the initial cube, the initial cube is gradually expanded along the three directions of SLt according to the expansion strategy. During the expansion process, it is also necessary to pay attention to the constraint association. Only when the constraint association is satisfied can the drivable area of the generated vehicle meet the requirements.
[0065] By using the above method, dynamic and static semantic elements are abstractly represented by interconnected collision-free cubes in the spatiotemporal domain (obtained by expanding the initial cube), which simplifies and generalizes the trajectory planning problem.
[0066] The above method generates the drivable area of the vehicle by expanding the initial cube according to the expansion strategy and constraint association. The expansion method achieves high efficiency through multi-directional dynamic expansion and interactive calculation of semantic boundaries, which is suitable for the real-time trajectory generation requirements of autonomous driving systems.
[0067] The above method employs semantic-driven approaches to dynamically associate various constraints, such as speed limits, time constraints, and obstacle occupancy areas, during the expansion of the spatiotemporal semantic corridor. This dynamic association ensures that the generated spatiotemporal corridor strictly conforms to the dynamic semantic constraints of the actual driving environment.
[0068] Unlike the rasterization method used in related technologies, which mainly discretizes the spatiotemporal environment into a raster and generates trajectory boundaries by point-by-point detection, this method suffers from high computational complexity and insufficient real-time performance.
[0069] Unlike related technologies that use convex polygons to construct boundaries using geometric convex hulls, this approach is suitable for simple scenarios but suffers from limited accuracy in non-convex spaces.
[0070] Unlike related technologies that use the latent field method, this method generates boundaries based on the potential energy field, which is suitable for dynamic obstacle avoidance, but is prone to getting stuck in local optima and has difficulty handling complex scenarios.
[0071] In one embodiment of this application, generating a drivable area for a vehicle by expanding the initial cube according to an expansion strategy and constraint association includes: dynamically associating speed limits, time constraints, and obstacle occupancy areas according to the expansion strategy of the initial cube; dynamically associating these constraints during the expansion of the spatiotemporal semantic corridor to ensure that the generated spatiotemporal corridor conforms to the dynamic or static semantic constraints in the actual driving environment; and forming a drivable area that complies with speed limits while avoiding collisions with obstacles after the expansion of the initial cube is completed.
[0072] The Spatio-Temporal Semantic Corridor (SSC) is a framework for trajectory planning in autonomous driving. It generates safe and dynamically feasible paths by unifying the modeling of semantic elements (such as obstacles and constraints) in the environment. Semantic elements are categorized into obstacles (e.g., vehicles, pedestrians) and constraints (e.g., speed limits, traffic lights, lane boundaries). Obstacles define inaccessible areas, while constraints are uniformly represented by semantic boundaries. A three-dimensional configuration space is constructed based on the Frenet coordinate system (vertical, horizontal, and temporal), mapping dynamic obstacles and static constraints to the spatio-temporal domain. Generation steps include: Seed generation: initializing the reference trajectory; Cube expansion: expanding the safe area; Constraint association: attaching semantic constraints to the cube; Cube relaxation: optimizing trajectory smoothness; Trajectory planning: employing piecewise Bézier curves combined with quadratic programming to ensure the trajectory satisfies all constraints.
[0073] Based on the initial cube's expansion strategy, which dynamically correlates various factors such as speed limits, time constraints, and obstacle occupancy areas, dynamic correlation is used during the expansion of the spatiotemporal semantic corridor to ensure that the generated spatiotemporal corridor conforms to the dynamic or static semantic constraints of the actual driving environment. Furthermore, after the initial cube's expansion is complete, a drivable area is formed that adheres to speed limits while avoiding collisions with obstacles, such as... Figure 2 As shown, multiple constraints, such as speed limits, time constraints, and obstacle occupancy areas, are dynamically associated during the expansion of the spatiotemporal semantic corridor. This dynamic association ensures that the generated spatiotemporal corridor strictly conforms to the dynamic semantic constraints of the actual driving environment.
[0074] In one embodiment of this application, the dynamic or static semantic constraints include static obstacles and dynamic obstacles. Generating a drivable area for the vehicle by expanding the initial cube according to an expansion strategy and constraint association includes: representing the static obstacles as impassable areas in the preset three-dimensional configuration space throughout the entire time period. , wherein s max s min These represent the maximum and minimum longitudinal displacements, respectively. max l min These represent the maximum and minimum lateral displacements, respectively. Represents the entire time period; the dynamic obstacles are rendered as dynamically occupied areas in the preset three-dimensional configuration space according to the predicted trajectory; speed limit. Wherein v min v max Related to the vehicle's speed limit; where ds / dt represents the vehicle's speed limit; time limit is defined by the vehicle's travel time range. , wherein t min t max Related to the vehicle's travel time, t represents the vehicle's travel time.
[0075] The dynamic or static semantic constraints include static obstacles and dynamic obstacles, and the constraint associations for generating the initial cube include, but are not limited to:
[0076] (1) Obstacle semantic elements: Static obstacles are represented as impassable regions in the SLT space throughout the entire time period: , representing the maximum and minimum values in the vertical and horizontal directions, with a time range covering all time periods.
[0077] (2) Dynamic obstacles will be rendered as dynamically occupied areas in SLT space according to the predicted trajectory. This step mainly fine-tunes the initial trajectory so that the loss of control of the body sensation is controlled within a reasonable range.
[0078] Regarding the situation in (2) above, the focus is actually on: speed limits, the impact of obstacles, and whether traffic regulations are followed. Among these, the semantic elements of dynamic constraints include:
[0079] The speed limit is defined as: (Mainly refers to longitudinal speed, the perceived speed is greater than that of the obstacle).
[0080] For the situation in (2) above, the time limit is expressed by the range of the vehicle's travel time: This maps the physical meaning of obstacles and dynamic constraints to the expansion range of the spacetime corridor, ensuring that the expanded cube strictly follows the semantic boundaries.
[0081] In one embodiment of this application, generating an initial cube in the longitudinal, lateral, and temporal ranges based on S, L, and t in the preset three-dimensional configuration space includes: selecting at least two consecutive trajectory points from the coarse trajectory state sequence information. , Generate an initial cube; the longitudinal range of the initial cube. Depend on and The longitudinal position is determined by the initial cube's lateral range. Determined by lane width and lateral safety distance; the time range of the initial cube The time step is a fixed value.
[0082] For generating the initial cube, first select at least two consecutive trajectory points from the initial point set sequence. , , used to generate the initial cube, such as Figure 2As shown in the dark-colored small cuboid, the three dimensions of the initial cube are defined as: vertical range : Depend on and Its vertical position determines its horizontal range. Consider lane width and lateral safety distance. Time range. The time step is a fixed 1 second (empirical value / fixed value).
[0083] In one embodiment of this application, the initial cube is used to generate a drivable area for the vehicle according to an expansion strategy and constraint association. The method further includes: gradually expanding the initial cube along the three directions S, L, and t, and continuously monitoring during the expansion process: whether it intersects with an obstacle, the obstacle including the perceived obstacle position information; whether it crosses a semantic boundary, the semantic boundary including a speed limit and a dynamic obstacle area; and setting the expansion direction stopping condition to stop when the cube boundary first touches the obstacle boundary or the speed limit boundary.
[0084] The specific expansion strategy (which needs to consider passenger comfort, whether it meets planning requirements, and whether it complies with traffic regulations) includes: the initial cube gradually expands along the three directions of SLt. During the expansion process, it is necessary to: ① detect whether it intersects with obstacles (obstacles such as...) Figure 2 (As shown by the medium gray quadrilateral). ② Detect whether it crosses semantic boundaries (such as speed limits or dynamic obstacle regions, e.g.) Figure 2 (As shown by the horizontal lines in the middle). ③ The expansion stops immediately when the cube boundary first touches the obstacle boundary or the speed limit boundary.
[0085] In one embodiment of this application, the step of projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space in response to the vehicle's coarse trajectory state sequence information includes: acquiring the coarse trajectory state sequence information obtained through three-dimensional dynamic programming. Each trajectory point here includes at least the vehicle's position, velocity, acceleration, and heading angle at discrete time intervals; the generated spatiotemporal corridor is ensured to include the vehicle's final target position according to the vehicle's target state.
[0086] Receive upstream decision coarse trajectory (coarse-filtered trajectory) and initial trajectory point generation: obtain the coarse trajectory state sequence calculated by upstream dynamic programming, such as 3D DP search. Each trajectory point contains the vehicle's position, velocity, acceleration, and heading angle at a specific discrete time interval (1 second in this case). The target state is also introduced. This ensures that the generated spatiotemporal corridor contains the final target location. Only when the target state is clearly defined can coarse trajectory state sequence information be projected.
[0087] In one embodiment of this application, the step of projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space in response to the vehicle's coarse trajectory state sequence information further includes: projecting the acquired vehicle's coarse trajectory state sequence information into an SLt space, where S represents the longitudinal displacement along the longitudinal coordinate of the reference lane, L represents the lateral displacement relative to the lane, and t represents the time of the lateral coordinate.
[0088] like Figure 2 As shown in the diagram, the light-colored cuboid represents the drivable area formed after the expansion on the ST plane, which adheres to speed limits and avoids collisions with obstacles. The calculation method for the drivable area considers static information such as map features to avoid issues like crossing lines and hitting curbs, ensuring the effectiveness of the drivable area. By abstracting semantic elements in complex urban environments into semantic boundaries in the Spatiotemporal Semantic Corridor (SSC), a unified expression of semantic elements is achieved, overcoming the inconsistency in handling multiple semantic elements in traditional trajectory planning methods. Specific implementation schemes include:
[0089] Step S1: In response to the coarse trajectory state sequence information of the vehicle, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space. The preset three-dimensional configuration space includes at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t.
[0090] By receiving the coarse trajectory from the upstream decision-making module and simultaneously determining the initial trajectory generation method, a preset three-dimensional configuration space can be established using the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and time t, thus serving as the basis for the expansion of the spatiotemporal corridor. A semantically driven method is employed to dynamically associate various constraints such as speed limits, time constraints, and obstacle occupancy areas during the expansion of the spatiotemporal semantic corridor. Step S2 involves generating an initial cube based on the S, L, and t in the preset three-dimensional configuration space, across the longitudinal, lateral, and time ranges. Then, through dynamic association, it is ensured that the generated spatiotemporal corridor strictly conforms to the dynamic semantic constraints of the actual driving environment.
[0091] Initial cube generation: Select two consecutive trajectory points from the initial point set (sequence) to generate an initial cube, such as... Figure 2 The dark cuboid in the image is the initial cube shown.
[0092] Step S3: Generate the drivable area of the vehicle by expanding the initial cube according to the expansion strategy and constraint association.
[0093] The expansion process follows a pre-defined expansion strategy, and the constraint associations include obstacle semantic elements and dynamic constraint semantic elements. Specifically, the expansion method achieves high efficiency through multi-directional dynamic expansion and interactive computation of semantic boundaries, making it suitable for the real-time trajectory generation requirements of autonomous driving systems. By using interconnected collision-free cubes in the spatiotemporal domain to abstractly represent dynamic and static semantic elements, the trajectory planning problem is simplified and generalized.
[0094] This application also provides an autonomous driving trajectory generation device 300, such as... Figure 3 As shown, a schematic diagram of the structure of an autonomous driving trajectory generation device in an embodiment of this application is provided. The autonomous driving trajectory generation device 300 includes at least: a response module 310, an initial cube generation module 320, and an expansion and constraint module 330, wherein:
[0095] In one embodiment of this application, the response module 310 is specifically used to: in response to the coarse trajectory state sequence information of the vehicle, project the coarse trajectory state sequence information onto a preset three-dimensional configuration space, wherein the preset three-dimensional configuration space includes at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t.
[0096] For autonomous vehicles, a coarse trajectory state sequence is first generated based on upstream dynamic programming. Each trajectory point in the coarse trajectory state sequence contains key information such as the vehicle's position, velocity, acceleration, and heading angle at specific discrete time intervals.
[0097] Then, the coarse trajectory state sequence information needs to be projected into a preset three-dimensional configuration space, that is, the obtained state sequence needs to be projected into the SLT space. The SLT space includes ①S longitudinal displacement along the reference lane; ②L lateral displacement relative to the lane; ③t time.
[0098] In one embodiment of this application, the initial cube generation module 320 is specifically used to: generate an initial cube in the longitudinal range, the transverse range, and the time range according to the S, the L, and the t in the preset three-dimensional configuration space.
[0099] An initial cube is generated in the longitudinal, lateral, and time domains based on the longitudinal displacement along the reference lane in SLT space; the lateral displacement L relative to the lane; and time t.
[0100] It is understandable that the initial cube includes the definition of multiple dimensions.
[0101] In one embodiment of this application, the expansion and constraint module 330 is specifically used to: generate a drivable area for the vehicle by expanding the initial cube according to the expansion strategy and constraint association.
[0102] Based on the initial cube, the initial cube is gradually expanded along the three directions of SLt according to the expansion strategy. During the expansion process, it is also necessary to pay attention to the constraint association. Only when the constraint association is satisfied can the drivable area of the generated vehicle meet the requirements.
[0103] In one embodiment of this application, the expansion and constraint module 330 is further configured to:
[0104] Based on the initial cube, the dynamic association speed limit, time constraint, and obstacle occupancy area in the expansion strategy are used to ensure that the generated spatiotemporal corridor conforms to the dynamic or static semantic constraints in the actual driving environment during the expansion process.
[0105] After the initial cube expands, a drivable area is formed that adheres to speed limits while avoiding collisions with obstacles.
[0106] In one embodiment of this application, the initial cube generation module 320 is further configured to:
[0107] The static obstacle is represented as an impassable area in the preset three-dimensional configuration space throughout the entire time period. ,
[0108] Wherein s max s min These represent the maximum and minimum longitudinal displacements, respectively. max l min These represent the maximum and minimum lateral displacements, respectively. Indicates the entire time period;
[0109] The dynamic obstacle is rendered as a dynamically occupied area in the preset three-dimensional configuration space according to the predicted trajectory;
[0110] Speed limit Wherein v min v max Related to; where ds / dt represents the vehicle's speed limit;
[0111] Time restrictions are imposed by limiting the range of time the vehicle can travel. , wherein t min t max In relation to this, t represents the vehicle's travel time.
[0112] In one embodiment of this application, the initial cube generation module 320 is further configured to:
[0113] Select at least two consecutive trajectory points from the coarse trajectory state sequence information. , Generate the initial cube;
[0114] The longitudinal range of the initial cube Depend on and The longitudinal position determines;
[0115] The lateral range of the initial cube This includes lane width and lateral safety distance;
[0116] The time range of the initial cube The time step is a fixed value.
[0117] In one embodiment of this application, the initial cube generation module 320 is further configured to:
[0118] The initial cube is gradually expanded along the three directions S, L, and t, and continuous monitoring is required during the expansion process.
[0119] Whether it intersects with an obstacle, wherein the obstacle includes the perceived location information of the obstacle;
[0120] Whether the semantic boundary has been crossed, the semantic boundary includes speed limit and dynamic obstacle area;
[0121] Additionally, the expansion direction stopping condition is set to stop when the cube boundary first touches the obstacle boundary or the speed limit boundary.
[0122] In one embodiment of this application, the response module 310 is further configured to:
[0123] Obtain the coarse trajectory state sequence information obtained through three-dimensional dynamic programming. Each trajectory point therein includes at least the vehicle's position, velocity, acceleration, and heading angle at discrete time intervals;
[0124] Ensure that the generated spatiotemporal corridor includes the vehicle's final target location, based on the vehicle's target state.
[0125] In one embodiment of this application, the response module 310 is further configured to:
[0126] The acquired coarse trajectory state sequence information of the vehicle is projected into the SLt space, where S includes the longitudinal displacement along the longitudinal coordinate of the reference lane, L includes the lateral displacement relative to the lane, and t includes the time of the lateral coordinate.
[0127] It is understood that the above-mentioned autonomous driving trajectory generation device can realize each step of the autonomous driving trajectory generation method provided in the foregoing embodiments. The relevant explanations of the autonomous driving trajectory generation method are applicable to the autonomous driving trajectory generation device, and will not be repeated here.
[0128] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Please refer to it. Figure 4 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for other business operations.
[0129] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.
[0130] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.
[0131] The processor reads the corresponding computer program from non-volatile memory into main memory and then runs it, forming an autonomous driving trajectory generation device at the logical level. The processor executes the program stored in memory and specifically performs the following operations:
[0132] In response to the vehicle's coarse trajectory state sequence information, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space, the preset three-dimensional configuration space including at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t;
[0133] Based on S, L, and t in the preset three-dimensional configuration space, an initial cube is generated in the vertical, horizontal, and time dimensions; and
[0134] The initial cube is used to generate the drivable area of the vehicle by applying an expansion strategy and constraint association.
[0135] The above is as stated in this application. Figure 1 The method executed by the autonomous driving trajectory generation device disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.
[0136] The electronic device can also perform Figure 1 The method for executing the autonomous driving trajectory generation device, and the implementation of the autonomous driving trajectory generation device in Figure 1 The functions of the embodiments shown are not described in detail here.
[0137] This application also proposes a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by an electronic device including multiple applications, enable the electronic device to perform... Figure 1 The method executed by the autonomous driving trajectory generation device in the illustrated embodiment is specifically used to perform:
[0138] In response to the vehicle's coarse trajectory state sequence information, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space, the preset three-dimensional configuration space including at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t;
[0139] Based on S, L, and t in the preset three-dimensional configuration space, an initial cube is generated in the vertical, horizontal, and time dimensions; and
[0140] The initial cube is used to generate the drivable area of the vehicle by applying an expansion strategy and constraint association.
[0141] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0142] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0143] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0144] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0145] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0146] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0147] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0148] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0149] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0150] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A method for generating autonomous driving trajectories, wherein, The trajectory generation method includes: In response to the vehicle's coarse trajectory state sequence information, the coarse trajectory state sequence information is projected onto a preset three-dimensional configuration space, the preset three-dimensional configuration space including at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t; Based on S, L, and t in the preset three-dimensional configuration space, an initial cube is generated in the vertical, horizontal, and time dimensions; and The initial cube is used to generate the drivable area of the vehicle by applying an expansion strategy and constraint association.
2. The method as described in claim 1, wherein, The step of generating the drivable area of the vehicle by inflating the initial cube according to the expansion strategy and constraint association includes: Based on the initial cube, the dynamic association speed limit, time constraint, and obstacle occupancy area in the expansion strategy are used to ensure that the generated spatiotemporal corridor conforms to the dynamic or static semantic constraints in the actual driving environment during the expansion process. After the initial cube expands, a drivable area is formed that adheres to speed limits while avoiding collisions with obstacles.
3. The method as described in claim 2, wherein, The dynamic or static semantic constraints include static and dynamic obstacles. The initial cube is expanded according to an expansion strategy and constraint associations to generate the vehicle's drivable area, including: The static obstacle is represented as an impassable area in the preset three-dimensional configuration space throughout the entire time period. , Wherein, the s max s min These represent the maximum and minimum longitudinal displacements, respectively. max l min These represent the maximum and minimum lateral displacements, respectively. Indicates the entire time period; The dynamic obstacle is rendered as a dynamically occupied area in the preset three-dimensional configuration space according to the predicted trajectory; Speed limit Wherein v min v max Related to the vehicle's speed limit; where ds / dt represents the vehicle's speed limit; Time restrictions are imposed by limiting the range of time the vehicle can travel. , wherein t min t max Related to the vehicle's travel time, t represents the vehicle's travel time.
4. The method of claim 1, wherein, The step of generating an initial cube in the longitudinal, lateral, and temporal ranges based on S, L, and t in the preset three-dimensional configuration space includes: Select at least two consecutive trajectory points from the coarse trajectory state sequence information. , Generate the initial cube; The longitudinal range of the initial cube Depend on and The longitudinal position determines; The lateral range of the initial cube Determined by lane width and lateral safety distance; The time range of the initial cube The time step is a fixed value.
5. The method of claim 4, wherein, The process of generating a drivable area for the vehicle by inflating the initial cube according to an expansion strategy and constraint association also includes: The initial cube is gradually expanded along the three directions S, L, and t, and continuous monitoring is required during the expansion process. Whether it intersects with an obstacle, wherein the obstacle includes the perceived location information of the obstacle; Whether the semantic boundary has been crossed, the semantic boundary includes speed limit and dynamic obstacle area; Additionally, the expansion direction stopping condition is set to stop when the cube boundary first touches the obstacle boundary or the velocity limit boundary.
6. The method of claim 1, wherein, The response to the vehicle's coarse trajectory state sequence information, projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space, includes: Obtain the coarse trajectory state sequence information obtained through three-dimensional dynamic programming. Each trajectory point therein includes at least the vehicle's position, velocity, acceleration, and heading angle at discrete time intervals; Ensure that the generated spatiotemporal corridor includes the vehicle's final target location, based on the vehicle's target state.
7. The method of claim 1, wherein, The method of responding to the vehicle's coarse trajectory state sequence information and projecting the coarse trajectory state sequence information onto a preset three-dimensional configuration space further includes: The acquired coarse trajectory state sequence information of the vehicle is projected into the SLt space, where S represents the longitudinal displacement along the longitudinal coordinate of the reference lane, L represents the lateral displacement relative to the lane, and t represents time.
8. An autonomous driving trajectory generation device, wherein, The trajectory generation device includes: The response module is used to respond to the coarse trajectory state sequence information of the vehicle and project the coarse trajectory state sequence information onto a preset three-dimensional configuration space. The preset three-dimensional configuration space includes at least the longitudinal displacement S along the reference lane, the lateral displacement L relative to the lane, and the time t. An initial cube generation module is used to generate an initial cube in the longitudinal, lateral, and temporal dimensions based on S, L, and t in the preset three-dimensional configuration space; and The expansion and constraint module is used to generate the drivable area of the vehicle by associating the initial cube with expansion strategies and constraints.
9. An electronic device, comprising: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method of any one of claims 1 to 7.
10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of applications, cause the electronic device to perform the method of any one of claims 1 to 7.