Path planning method and device, equipment and storage medium

Abstract

The invention relates to a path planning method, device and equipment and a storage medium, and the method comprises the steps: obtaining a plurality of longitudinal nodes of a vehicle between a current position and a target position in the longitudinal direction, the environment information of a current longitudinal node in the plurality of longitudinal nodes, and the state information of a previous longitudinal node of the current longitudinal node; the time intervals between two adjacent longitudinal nodes are the same, and the current longitudinal node is any node except the current position in the plurality of longitudinal nodes; based on the state information of the previous longitudinal node and the environment information of the current longitudinal node, determining a path point corresponding to the vehicle at the current longitudinal node; and in response to the current longitudinal node being the target position, generating a driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node. The time dimension does not need to be considered when path planning is carried out, dimension reduction is carried out on the three-dimensional path searching problem, the calculation complexity is reduced, and the path planning efficiency is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of path planning, and relate to but are not limited to a path planning method, apparatus, device and storage medium. Background Art

[0002] Autonomous driving is one of the most popular research directions in the field of artificial intelligence. Autonomous driving technology can be divided into the following levels: environmental perception, decision-making and planning, and low-level control. Currently, environmental perception is responsible for detecting various moving and stationary obstacles and collecting various information on the road. After a decision action is given, the planning algorithm plans the spatiotemporal trajectory for each set of decision actions, and finally obtains the trajectory for vehicle control.

[0003] However, when the relevant technology performs three-dimensional path planning in time and space, the three-dimensional calculation complexity is high, which brings challenges to the real-time performance of the system. Therefore, how to efficiently complete the three-dimensional path search in time and space is an urgent problem to be solved. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a path planning method, device, equipment and storage medium to solve the problems of high complexity and low real-time performance of path planning when performing three-dimensional path planning in related technologies.

[0005] In order to achieve the above object, an embodiment of the present application provides a path planning method, the method comprising:

[0006] Acquire multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, environmental information of the current longitudinal node among the multiple longitudinal nodes and state information of the previous longitudinal node before the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time intervals between two adjacent longitudinal nodes are the same, and the current longitudinal node is any node among the multiple longitudinal nodes except the current position; determine the path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environmental information of the current longitudinal node; in response to the current longitudinal node being the target position, generate the driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node.

[0007] According to the above technical means, through multiple longitudinal nodes in the longitudinal direction between the current position and the target position, it is only necessary to calculate the lateral position of each longitudinal node as the path point of each longitudinal node, so that there is no need to consider the time dimension when performing path planning. The three-dimensional path search problem is reduced in dimension, and only the path points of each longitudinal node in the longitudinal direction are considered, which reduces the calculation complexity and the calculation overhead of the algorithm, thereby improving the running speed of path planning and the efficiency of path planning, which helps to improve the real-time performance of the driving path in a dynamic traffic environment.

[0008] In some embodiments, each longitudinal node corresponds to a driving moment of the vehicle, and the environmental information includes at least the posture of dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; determining the path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environmental information of the current longitudinal node includes: determining at least one passable space and a lateral safety boundary of each passable space at the driving moment corresponding to the current longitudinal node based on the posture of dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; determining the travel cost of each passable space based on the state information of the previous longitudinal node and the lateral safety boundary of each passable space; determining the passable space whose travel cost meets the target condition as the target passable space of the current longitudinal node; determining the path point of the vehicle corresponding to the current longitudinal node based on the target passable space and the state information of the previous longitudinal node.

[0009] According to the above technical means, the expandable rectangle algorithm is used to calculate the vehicle's passable space at the current longitudinal node, which can effectively expand the obstacle area. The expandable rectangle algorithm has a small amount of calculation and can complete path planning faster, which helps to improve real-time performance in a dynamic traffic environment. Based on the greedy search, the travel cost of each passable space can be quickly determined, and then the path point corresponding to each longitudinal node can be obtained, thereby improving the efficiency of path planning.

[0010] In some embodiments, the determining of at least one passable space and the lateral safety boundary of each passable space at the time of driving corresponding to the current longitudinal node based on the posture of the dynamic and static obstacles at the time of driving corresponding to the current longitudinal node includes: determining a plurality of lateral gaps in the lateral direction between two adjacent dynamic and static obstacles at the time of driving corresponding to the current longitudinal node based on the posture of the dynamic and static obstacles at the time of driving corresponding to the current longitudinal node; comparing the plurality of lateral gaps with the vehicle width to obtain a comparison result; determining the lateral gap whose comparison result satisfies the passing condition as the at least one passable space, and determining the lateral position of each passable space as the lateral safety boundary of each passable space based on the longitudinal reference line; wherein the longitudinal reference line is the center line of the lane where the target position is located, and the lateral direction is perpendicular to the longitudinal direction.

[0011] According to the above technical means, the vehicle's passable space at the current longitudinal node is calculated through the expansion rectangle algorithm, which can effectively expand the obstacle area. The expansion rectangle algorithm has a small amount of calculation and can complete path planning faster, which helps to improve real-time performance in a dynamic traffic environment.

[0012] In some embodiments, the status information includes at least a lateral reachable range and a lateral safety boundary of the longitudinal node based on the longitudinal baseline; the determination of the travel cost of each passable space based on the status information of the previous longitudinal node and the lateral safety boundary of each passable space includes: determining the overlapping length between the lateral reachable range of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node as the first parameter of each passable space; determining the overlapping length between the lateral safety boundary of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node as the second parameter of each passable space; determining the distance between the lateral safety boundary of each passable space of the current longitudinal node and the longitudinal baseline as the third parameter of each passable space; and performing weighted calculations on the first parameter, the second parameter and the third parameter of each passable space based on a preset weight coefficient to obtain the travel cost of each passable space.

[0013] According to the above technical means, when the travel cost of each passable space is obtained through greedy search, since greedy search only considers the local optimal solution, unnecessary calculation amount is reduced, the travel cost of each passable space can be obtained quickly, and the path search efficiency is improved.

[0014] In some embodiments, determining the path point of the vehicle corresponding to the current longitudinal node based on the target passage space and the status information of the previous longitudinal node includes: updating the lateral reachable range of the current longitudinal node based on the target passage space of the current longitudinal node and the lateral reachable range of the previous longitudinal node; determining the path point of the vehicle corresponding to the current longitudinal node based on the target passage space and the lateral reachable range of each longitudinal node.

[0015] According to the above technical means, the path point corresponding to the current longitudinal node of the vehicle is determined based on the target passage space and the lateral reachable range of the current longitudinal node, which can avoid collision with obstacles and ensure the feasibility and safety of the path point; at the same time, by comprehensively considering the lateral accessibility, it can adapt to different road conditions and environmental changes, and a smoother path can be generated subsequently, which improves the path planning performance. The generated path can travel more safely and efficiently in a complex environment.

[0016] In some embodiments, the method of determining the path point of the vehicle corresponding to the current longitudinal node based on the target passage space and lateral accessible range of each longitudinal node includes: in response to the existence of an overlapping area between the target passage space and the lateral accessible range of any longitudinal node, determining the overlapping area as the path point corresponding to the arbitrary longitudinal node; in response to the absence of an overlapping area between the target passage space and the lateral accessible range of any longitudinal node, determining the point in the lateral accessible range closest to the target passage space as the path point corresponding to the arbitrary longitudinal node.

[0017] In some embodiments, the driving path of the vehicle from the current position to the target position is generated based on the path points corresponding to each longitudinal node before the current longitudinal node, including: generating an initial path based on the path points corresponding to each longitudinal node; smoothing the initial path based on the target clearance space and lateral reachable range of each longitudinal node to obtain the driving path of the vehicle from the current position to the target position.

[0018] According to the above technical means, through smoothing processing, turning points and unnecessary turns in the path can be reduced, making the path smoother and more natural, thereby reducing driving difficulty and improving driving safety; the smoothed path has better continuity, which helps the vehicle to be more stable when executing the path and reduces vibration and impact caused by sharp turns.

[0019] In some embodiments, the method also includes: obtaining longitudinal decision information and lateral decision information of the vehicle; determining a longitudinal baseline based on the lateral decision information; determining a target position of the vehicle based on the longitudinal decision information; and generating a plurality of longitudinal nodes between the current position and the target position in the direction along the longitudinal baseline based on the current position and the target position of the vehicle.

[0020] In some embodiments, the method further includes: in response to the current longitudinal node not reaching the target position, determining a path point corresponding to the next longitudinal node of the vehicle based on the state information of the current longitudinal node and the environmental information, until the next longitudinal node is the target position.

[0021] The embodiment of the present application further provides a path planning device, which includes: an acquisition module, used to acquire multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, environmental information of the current longitudinal node among the multiple longitudinal nodes, and state information of the previous longitudinal node of the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time interval between two adjacent longitudinal nodes is the same, and the current longitudinal node is any node among the multiple longitudinal nodes except the current position; a determination module, used to determine the path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environmental information; and a generation module, used to generate the driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node in response to the current longitudinal node being the target position.

[0022] According to the above technical means, through multiple longitudinal nodes in the longitudinal direction between the current position and the target position, it is only necessary to calculate the lateral position of each longitudinal node as the path point of each longitudinal node, so that there is no need to consider the time dimension when performing path planning. The three-dimensional path search problem is reduced in dimension, and only the path points of each longitudinal node in the longitudinal direction are considered, which reduces the calculation complexity and the calculation overhead of the algorithm, thereby improving the running speed of path planning and the efficiency of path planning, which helps to improve the real-time performance of the driving path in a dynamic traffic environment.

[0023] An embodiment of the present application further provides a path planning device, which includes a memory and a processor, wherein the memory stores a computer program that can be executed on the processor, and the processor implements the above-mentioned path planning method when executing the program.

[0024] An embodiment of the present application further provides a computer-readable storage medium having executable instructions stored thereon, which is used to cause a processor to execute the executable instructions to implement the above-mentioned path planning method.

[0025] Beneficial effects of this application:

[0026] (1) When performing path planning, there is no need to consider the time dimension. The three-dimensional path search problem is reduced in dimension, and only the path points of each longitudinal node in the longitudinal direction are considered, which reduces the computational complexity and the computational overhead of the algorithm, thereby increasing the running speed of path planning and improving the efficiency of path planning. This helps to improve the real-time performance of the driving path in a dynamic traffic environment.

[0027] (2) When calculating the path points of each longitudinal node, only the local optimal solution is considered, which reduces unnecessary calculations, can quickly obtain the travel cost of each traversable space, and improves the path search efficiency;

[0028] (3) Through smoothing, turning points and unnecessary turns in the path can be reduced, making the path smoother and more natural, thereby reducing driving difficulty and improving driving safety; the smoothed path has better continuity, which helps the vehicle to be more stable when executing the path and reduces the vibration and impact caused by sharp turns. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is an optional architecture diagram of a path planning system provided in an embodiment of the present application;

[0030] Figure 2 is a schematic diagram of the structure of a path planning device provided in an embodiment of the present application;

[0031] Figure 3 This is an optional flowchart of the path planning method provided in the embodiment of the present application;

[0032] Figure 4 This is an optional flowchart of the path planning method provided in the embodiment of the present application;

[0033] Figure 5 This is an optional flowchart of the path planning method provided in the embodiment of the present application;

[0034] Figure 6 is a reference schematic diagram of an expanded rectangle provided in an embodiment of the present application;

[0035] Figure 7 This is a flowchart of the lateral path greedy search provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0037] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the technical field of the embodiments of the present application. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0038] In related technologies, after a vehicle determines a decision action (such as following the vehicle in the left lane), it usually performs short-term, real-time path planning based on the current state of the vehicle and surrounding environment information.

[0039] The relevant technology will generate a frenet coordinate system based on the vehicle's current environment, perform three-dimensional space-time rasterization, search for the initial path through a heuristic search algorithm for path planning (such as the A* algorithm), and finally smooth the path to avoid collisions.

[0040] However, the problem with the methods of related technologies is that in the time-space coordinate system, the Euclidean distance transform (ESDF) distance field is usually used for collision avoidance processing, but the distance meanings of the time dimension and the space dimension are inconsistent, and it is difficult to ensure the safety of the trajectory when the nonlinear path is smooth; the A* algorithm is used for path search in high-dimensional space, which has a high computational complexity and will bring challenges to the real-time performance of the path planning system.

[0041] Based on the problems existing in the related technology, an embodiment of the present application provides a path planning method, which obtains multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, the environmental information of the current longitudinal node among the multiple longitudinal nodes and the status information of the previous longitudinal node of the current longitudinal node, determines the path point corresponding to the current longitudinal node of the vehicle based on the environmental information of the current longitudinal node and the status information of the previous longitudinal node, and when the current longitudinal node is the target position, generates a driving path from the current position to the target position of the vehicle based on the path points corresponding to the longitudinal nodes before the current longitudinal node.

[0042] In this way, the embodiment of the present application only needs to calculate the lateral position of each longitudinal node as the path point of each longitudinal node through multiple longitudinal nodes in the longitudinal direction between the current position and the target position, so that there is no need to consider the time dimension when performing path planning. The three-dimensional path search problem is reduced in dimension, and only the path points of each longitudinal node in the longitudinal direction are considered, which reduces the computational complexity and the computational overhead of the algorithm, thereby improving the running speed of path planning and improving the efficiency of path planning.

[0043] The present application embodiment provides a path planning system, see Figure 1 , Figure 1 It is an optional architecture diagram of the path planning system provided by the embodiment of the present application. The path planning system 10 includes at least a vehicle 100, a network 200 and a server 300, wherein the server 300 can constitute the path planning device of the embodiment of the present application. The vehicle 100 is connected to the server 300 through the network 200, and the network 200 can be a wide area network or a local area network, or a combination of the two. When performing path planning, the server 300 obtains multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, the environmental information of the current longitudinal node in the multiple longitudinal nodes and the state information of the previous longitudinal node of the current longitudinal node, based on the environmental information of the current longitudinal node and the state information of the previous longitudinal node, the path point corresponding to the vehicle at the current longitudinal node is determined, in response to the current longitudinal node being the target position, based on the path points corresponding to the longitudinal nodes before the current longitudinal node, the driving path of the vehicle from the current position to the target position is generated, and the server 300 can send the driving path of the vehicle 100 to the vehicle 100 based on the network 200, and the display interface of the vehicle 100 can display the driving path.

[0044] The following describes an exemplary application of the path planning method of an embodiment of the present application. The path planning device provided in the embodiment of the present application can be located on a vehicle, or it can be implemented as a cloud server. In one implementation, the server can be an independent physical server, or it can be a server cluster or distributed system composed of multiple physical servers, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (CDN, Content Delivery Network), and big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiment of the present application. Below, an exemplary application when the path planning device is a server will be described.

[0045] It should be noted here that cloud technology refers to a hosting technology that unifies hardware, software, network and other resources in a wide area network or local area network to achieve data computing, storage, processing and sharing. Cloud technology is a general term for network technology, information technology, integration technology, management platform technology, application technology, etc. based on the cloud computing business model. It can form a resource pool that can be used on demand and is flexible and convenient. Cloud computing technology will become an important support. The backend services of the technical network system require a large amount of computing and storage resources, such as video websites, picture websites and more portal websites. With the high development and application of the Internet industry, in the future, each item may have its own identification mark, and all need to be transmitted to the backend system for logical processing. Data of different levels will be processed separately. All kinds of industry data require strong system backing support, which can only be achieved through cloud computing.

[0046] Figure 2 is a schematic diagram of the structure of a path planning device provided in an embodiment of the present application, Figure 2 The path planning device shown includes: at least one processor 210, a memory 250, at least one network interface 220 and a user interface 230. The various components in the path planning device are coupled together through a bus system 240. It is understood that the bus system 240 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 240 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 240 is not described in detail. Figure 2 Various buses are labeled as bus system 240 .

[0047] The processor 210 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0048] The user interface 230 includes one or more output devices 231 that enable presentation of media content, and one or more input devices 232 .

[0049] The memory 250 may be removable, non-removable or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drive, optical disk drive, etc. The memory 250 may optionally include one or more storage devices physically away from the processor 210. The memory 250 includes a volatile memory or a non-volatile memory, and may also include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 250 described in the embodiment of the present application is intended to include any suitable type of memory. In some embodiments, the memory 250 can store data to support various operations, and examples of these data include programs, modules, and data structures or subsets or supersets thereof, as exemplarily described below.

[0050] The operating system 251 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic businesses and process hardware-based tasks.

[0051] The network communication module 252 is used to reach other computing devices via one or more (wired or wireless) network interfaces 220. Exemplary network interfaces 220 include: Bluetooth, Wireless Fidelity (Wi-Fi), and Universal Serial Bus (USB).

[0052] The input processing module 253 is used to detect one or more user inputs or interactions from one of the one or more input devices 232 and translate the detected inputs or interactions.

[0053] In some embodiments, the device provided in the embodiments of the present application can be implemented in software. Figure 2 A path planning device 254 stored in the memory 250 is shown. The path planning device 254 may be a path planning device in a path planning device, which may be software in the form of a program or a plug-in, and includes the following software modules: an acquisition module 2541, a determination module 2542, and a generation module 2543. These modules are logical, and therefore may be arbitrarily combined or further split according to the functions implemented. The functions of each module will be described below.

[0054] In other embodiments, the device provided in the embodiments of the present application can be implemented in hardware. As an example, the device provided in the embodiments of the present application can be a processor in the form of a hardware decoding processor, which is programmed to execute the path planning method provided in the embodiments of the present application. For example, the processor in the form of a hardware decoding processor can adopt one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs) or other electronic components.

[0055] The path planning method provided in each embodiment of the present application can be executed by a path planning device, wherein the path planning device can be any vehicle-mounted device with a path planning processing function, or it can also be a server, that is, the path planning method in each embodiment of the present application can be executed by a vehicle-mounted device, or it can be executed by a server, or it can also be executed by interaction between the vehicle-mounted device and the server.

[0056] See also Figure 3 , Figure 3 This is an optional flow chart of the path planning method provided in the embodiment of the present application. Figure 3 The steps shown are explained, and it should be noted that Figure 3 The path planning method in the embodiment of the present application is described by taking the server as the execution subject as an example. The path planning method provided in the embodiment of the present application can be implemented by steps S301 to S303:

[0057] Step S301, obtaining multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, environmental information of the current longitudinal node among the multiple longitudinal nodes and state information of the previous longitudinal node of the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time intervals between two adjacent longitudinal nodes are the same, and the current longitudinal node is any node among the multiple longitudinal nodes except the current position.

[0058] In the embodiment of the present application, the target position of the vehicle may refer to the position to which the vehicle is going. The target position is determined according to the driving decision of the vehicle. The driving decision includes lateral decision and longitudinal decision. The lateral decision may be left lane change, right lane change, lane keeping, etc. The longitudinal decision may be target following, target courtesy, constant speed driving, etc. Here, if the current decision action of the vehicle is to change lanes left and catch up with a vehicle in the left lane, then the position behind the vehicle in the lane where the vehicle is located is the target position.

[0059] Here, after determining the target position, the center line of the lane where the target position is located can be used as the longitudinal reference line, the current position of the vehicle can be used as the origin, and a coordinate system can be established (which can be a Frenet coordinate system, with the longitudinal reference line as the S axis and the axis perpendicular to the longitudinal reference line as the L axis). An intelligent driver (IDM, Intelligent Driver Model) or a similar deduction model is used to calculate multiple longitudinal nodes of the vehicle between the current position and the target position under the longitudinal decision label, that is, the st discrete point trajectory of the vehicle. Here, the s value in the st trajectory is the longitudinal axis coordinate of the vehicle in the Frenet coordinate system at each time t, and the value range of t can be [0:0.1:8.0] seconds. The time interval between two adjacent longitudinal nodes is the same, that is, time t starts from 0 seconds, increases to 8.0 seconds at intervals of 0.1 seconds, and the values ​​of t include 0 seconds, 0.1 seconds, 0.2 seconds, ... 7.9 seconds and 8.0 seconds. These time points constitute a discrete time series, and the end point of the longitudinal node is the farthest longitudinal distance that the vehicle can reach within 8 seconds at the current speed.

[0060] Here, each of the multiple vertical nodes corresponds to a time.

[0061] In the embodiment of the present application, the environmental information of the current longitudinal node may refer to static obstacles and dynamic obstacles around the longitudinal node at the current moment. Static obstacles refer to objects that do not move on the road, such as stationary vehicles, stone pillars in the middle of the road surface, guardrails, curbs, construction signs, fences, crash barrels, etc. Dynamic obstacles refer to moving objects, such as vehicles and pedestrians traveling at the current moment. Based on the environmental information, the traversable space of the vehicle at the current longitudinal node can be calculated.

[0062] In some embodiments, the server may obtain environmental information based on the vehicle's camera, lidar, millimeter-wave radar, etc., or may obtain municipal information to obtain the vehicle's environmental information at the current longitudinal node.

[0063] In an embodiment of the present application, the state information of the previous longitudinal node of the current longitudinal node may refer to the lateral reachable range and lateral safety boundary of the previous longitudinal node based on the longitudinal baseline. Here, the lateral reachable range reachable_lat_range refers to the set of all possible states that can be reached in the lateral direction starting from the position of the vehicle at the previous longitudinal node, that is, the maximum range [reachable_lat_left, reachable_lat_right] that the vehicle can reach in the lateral direction from the previous longitudinal node to the current longitudinal node. Here, the lateral reachable range may be the lateral reachable range on one side of the vehicle, because under the lateral decision, it can be known whether the vehicle will travel left or right. The lateral safety boundary bound_range refers to the range interval [left_bound, right_bound] on the horizontal axis coordinate of the frenet coordinate system. The lateral safety boundary expresses that the vehicle has no risk of collision in the lateral direction within this interval.

[0064] The path planning method provided in the embodiment of the present application can be used to perform path planning during driving. Therefore, after the server obtains this information and performs path planning, the path planning is sent to the vehicle, and the vehicle executes driving decisions based on the path during driving.

[0065] Step S302: determining a path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environment information of the current longitudinal node.

[0066] In an embodiment of the present application, the state information includes at least the lateral reachable range and lateral safety boundary of the longitudinal node based on the longitudinal baseline. Based on the environmental information of the current longitudinal node, the passable space of the vehicle at the current longitudinal node, that is, the space that can be passed without collision, can be determined. Here, the current longitudinal node can have multiple passable spaces, and based on the lateral reachable range and lateral safety boundary of the previous node, a final passable space can be determined from the multiple passable spaces. Based on the passable space and the lateral reachable range of the longitudinal node, the path point of the longitudinal node can be determined, that is, the position of the vehicle when it travels to the longitudinal node.

[0067] Here, if at the current longitudinal node, the lateral reachable range of the vehicle cannot reach the passable space of the vehicle, then the point in the lateral reachable range of the vehicle closest to the passable space can be determined as the path point; if the lateral reachable range of the vehicle is within the passable space of the vehicle, any point in the lateral reachable range can be selected as the path point, and the path point can be selected based on the vehicle's speed, acceleration, curvature, etc.

[0068] Step S303: In response to the current longitudinal node being the target position, generating a driving path of the vehicle from the current position to the target position based on path points corresponding to each longitudinal node before the current longitudinal node.

[0069] In an embodiment of the present application, if the current longitudinal node is the target position, that is, the current longitudinal node is the last longitudinal node in the st discrete point trajectory, then the vehicle's driving path from the current position to the target position can be generated based on the path points corresponding to the longitudinal nodes before the current longitudinal node.

[0070] Here, when generating the driving path, the path may be smoothed based on a path smoothing algorithm, and finally a smooth driving path is obtained.

[0071] In some embodiments, the path planning method provided in the embodiments of the present application may further include step S1:

[0072] Step S1: In response to the current longitudinal node not reaching the target position, based on the state information of the current longitudinal node and the environmental information, determine the path point of the vehicle corresponding to the next longitudinal node until the next longitudinal node is the target position.

[0073] In some embodiments, if the current longitudinal node is not the target position, it indicates that the path point of the next longitudinal node needs to be calculated. At this time, the path point corresponding to the next longitudinal node of the vehicle can be determined based on the state information and environmental information of the current longitudinal node, and then it can be determined whether the next longitudinal node has reached the target position. If it has not reached, the path point of the next longitudinal node will continue to be calculated until the next longitudinal node is the target position. At this time, the driving path of the vehicle from the current position to the target position can be generated based on the path points corresponding to the longitudinal nodes before the current longitudinal node.

[0074] The embodiment of the present application uses multiple longitudinal nodes in the longitudinal direction between the current position and the target position, and only needs to calculate the lateral position of each longitudinal node as the path point of each longitudinal node, so that there is no need to consider the time dimension when performing path planning. The three-dimensional path search problem is reduced in dimension, and only the path points of each longitudinal node in the longitudinal direction are considered, which reduces the calculation complexity and the calculation overhead of the algorithm, thereby improving the running speed of path planning and the efficiency of path planning, which helps to improve the real-time performance of the driving path in a dynamic traffic environment.

[0075] In some embodiments, each longitudinal node corresponds to a driving moment of the vehicle, and the environmental information includes at least the posture of dynamic and static obstacles at the driving moment corresponding to the current longitudinal node. The dynamic and static obstacles are static obstacles and dynamic obstacles around the longitudinal node at the current moment. The posture can be the position, direction, driving path, boundary and other information of the dynamic obstacles and the position, size, boundary and other information of the static obstacles.

[0076] Figure 4 is an optional flow chart of the path planning method provided in the embodiment of the present application, see Figure 4 , step S302 in the path planning method provided in the embodiment of the present application can be implemented through steps S401 to S404:

[0077] Step S401: based on the postures of the dynamic and static obstacles at the time of driving corresponding to the current longitudinal node, determine at least one passable space and a lateral safety boundary of each passable space at the time of driving corresponding to the current longitudinal node.

[0078] In an embodiment of the present application, all dynamic and static obstacles on the current longitudinal node can be distributed as discrete points in the frenet coordinate system, sorted according to the L-axis direction, and the gaps between all obstacles in the L direction are calculated. The gaps are sorted, and the gaps with a width greater than the vehicle width are screened out, and the gaps are used as the passable space of the current longitudinal node. Here, there can be multiple passable spaces. If the width of all gaps is less than the vehicle width, it means that there is no passable space at the current longitudinal node. According to the current behavior decision, speed planning cannot be performed, and a collision-free trajectory cannot be searched. At this time, the trajectory planning can be exited, and new data can be obtained for trajectory planning when trajectory planning is required later.

[0079] At a longitudinal node, the passable space may be divided into multiple intervals due to the existence of obstacles. The generation calculation of the passable intervals is determined by rectangular collision, and each passable space corresponds to an expansion rectangle.

[0080] Here, the lateral safety boundary bound_range of each passable space may refer to a range interval [left_bound, right_bound] on the frenet horizontal axis (ie, L axis) coordinate of each passable space, within which there is no risk of collision for the vehicle.

[0081] Step S402: Determine the travel cost of each passable space based on the state information of the previous longitudinal node and the transverse safety boundary of each passable space.

[0082] In the embodiment of the present application, the passage cost, ie, cost, of each passable space may be calculated through a greedy search rule based on the state information of the previous longitudinal node and the transverse safety boundary of each passable space.

[0083] In some embodiments, the cost of calculating the expanded rectangle corresponding to each search node can be calculated based on the degree of overlap of the left and right safety boundaries bound_range of the previous and next longitudinal nodes, the overlap rate of the lateral reachable range reachable_lat_range of the previous longitudinal node and the bound_range of the current longitudinal node, and the lateral distance of bound_range from the longitudinal baseline.

[0084] Step S403: Determine the passable space whose pass cost meets the target condition as the target passable space of the current longitudinal node.

[0085] In the embodiment of the present application, the traversable space with the minimum traversable cost may be determined as the target traversable space of the current longitudinal node to reduce the risk of collision.

[0086] Step S404: determining a path point corresponding to the current longitudinal node of the vehicle based on the target passage space and the state information of the previous longitudinal node.

[0087] In the embodiment of the present application, after the target clearance of the current longitudinal node is determined, the lateral reachable range reachable_lat_range of the previous longitudinal node and the target clearance space bound_range of the previous longitudinal node can be used to update the lateral reachable range reachable_lat_range of the current longitudinal node. Then, based on the lateral reachable range and the target clearance space of the current longitudinal node, the path point corresponding to the vehicle at the current longitudinal node is determined.

[0088] In an embodiment of the present application, the vehicle's passable space at the current longitudinal node is calculated using an expansion rectangle algorithm, which can effectively expand the obstacle area. The expansion rectangle algorithm has a small amount of calculation and can complete path planning more quickly, which helps to improve real-time performance in a dynamic traffic environment. Based on the greedy search, the travel cost of each passable space can be quickly determined, and then the path point corresponding to each longitudinal node can be obtained, thereby improving the efficiency of path planning.

[0089] In some embodiments, step 401 may be implemented through steps S4011 to S4013:

[0090] Step S4011: at the driving moment corresponding to the current longitudinal node, based on the postures of the dynamic and static obstacles at the driving moment corresponding to the current longitudinal node, determine a plurality of lateral gaps between two adjacent dynamic and static obstacles in the lateral direction.

[0091] In the embodiment of the present application, all the dynamic and static obstacles at the driving moment corresponding to the current longitudinal node can be projected as discrete points in the Frenet coordinate system at that moment, and the lateral gap between two adjacent obstacles in the L direction can be calculated.

[0092] Here, this can be achieved through the expansion rectangle algorithm. The lower longitudinal limit of the expansion rectangle is the current longitudinal node s minus the distance from the center point of the vehicle's rear axle to the rear bumper; the upper longitudinal limit of the expansion rectangle is the s value of the next longitudinal node plus the distance from the center point of the vehicle's rear axle to the front bumper; the upper and lower lateral limits of the expansion rectangle are based on the posture of the dynamic and static obstacles at the current longitudinal node corresponding to the driving moment.

[0093] Step S4012: Compare the multiple lateral gaps with the vehicle width to obtain a comparison result.

[0094] The comparison result includes each lateral gap being greater than or less than or equal to the vehicle width.

[0095] Step S4013, determine the lateral gap that meets the traffic conditions in the comparison result as the at least one passable space, and determine the lateral position of each passable space as the lateral safety boundary of each passable space based on the longitudinal reference line; wherein the longitudinal reference line is the center line of the lane where the target position is located, and the lateral direction is perpendicular to the longitudinal direction.

[0096] In the embodiment of the present application, a lateral gap that is larger than the width of the vehicle may be determined as a passable space for the vehicle, and there may be at least one passable space.

[0097] In some embodiments, the longitudinal reference line is the center line of the lane where the target position is located, that is, the s-axis of the Frenet coordinates. After the target passable space is determined, the longitudinal reference line can be used as a reference to determine the lateral position of each passable space as the lateral safety boundary of each passable space, that is, the coordinates of each passable space of the longitudinal node [left_bound, right_bound].

[0098] In an embodiment of the present application, the vehicle's passable space at the current longitudinal node is calculated using an expansion rectangle algorithm, which can effectively expand the obstacle area. The expansion rectangle algorithm has a small amount of calculation and can complete path planning faster, which helps to improve real-time performance in a dynamic traffic environment.

[0099] In some embodiments, the state information includes at least the lateral reachable range and lateral safety boundary of the longitudinal node based on the longitudinal reference line, and the passage cost of each traversable space can be obtained by greedy search. Step 402 can be implemented by steps S4021 to S4024:

[0100] Step S4021: Determine the overlapping length between the lateral reachable range of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node as the first parameter of each passable space.

[0101] In the embodiment of the present application, the first parameter Dreach of each traversable space represents the overlapping length between the lateral reachable range of the previous longitudinal node and the lateral safety boundary of each traversable space of the current longitudinal node.

[0102] Step S4022: Determine the overlapping length between the transverse safety boundary bound_range of the previous longitudinal node and the transverse safety boundary of each passable space of the current longitudinal node as the second parameter of each passable space.

[0103] Here, the second parameter Dspace of each passable space represents the overlapping length between the transverse safety boundary of the previous longitudinal node and the transverse safety boundary of each passable space of the current longitudinal node.

[0104] Step S4023: Determine the distance between the transverse safety boundary of each passable space of the current longitudinal node and the longitudinal reference line as the third parameter of each passable space.

[0105] Here, the third parameter Dtarget of each passable space represents the distance between the transverse safety boundary of each passable space at the current longitudinal node and the longitudinal reference line. If the longitudinal reference line is within the transverse safety boundary of the passable space, the third parameter Dtarget is 0.

[0106] Step S4024: Based on the preset weight coefficient, weighted calculation is performed on the first parameter, the second parameter and the third parameter of each passable space to obtain the passage cost of each passable space.

[0107] Here, the preset weight coefficient can be obtained based on experience. If there are multiple passable spaces at the current vertical node being searched, the passable space is selected by the following formula (1):

[0108] cost=W1*D reach +W2*D space +W3*D target (1);

[0109] When the embodiment of the present application obtains the travel cost of each passable space through greedy search, since the greedy search only considers the local optimal solution, unnecessary calculation amount is reduced, the travel cost of each passable space can be obtained quickly, and the path search efficiency is improved.

[0110] In some embodiments, step 404 may be implemented through steps S4041 to S4042:

[0111] Step S4041: Update the lateral reachable range of the current longitudinal node based on the target passage space of the current longitudinal node and the lateral reachable range of the previous longitudinal node.

[0112] In the embodiment of the present application, after the target passage space of each longitudinal node is insufficient, the lateral reachable range reachable_lat_range of the current longitudinal node is updated based on the lateral reachable range reachable_lat_range of the previous longitudinal node and the lateral safety boundary of the current longitudinal node.

[0113] For example, if the lateral reachable range D_prev of the previous longitudinal node is less than or equal to the lateral safety boundary B_curr of the current longitudinal node, the lateral reachable range D_curr of the current longitudinal node can be updated to D_prev. If the lateral reachable range D_prev of the previous longitudinal node is greater than the lateral safety boundary B_curr of the current longitudinal node, the lateral reachable range D_curr of the current longitudinal node should be updated to B_curr, because the safety boundary of the current longitudinal node limits its reachable range.

[0114] Step S4042: Based on the target passage space and lateral reachable range of each longitudinal node, determine the path point corresponding to the vehicle at the current longitudinal node.

[0115] In some embodiments, constraint conditions may be generated through the target passage space and the lateral reachable range of the current longitudinal node to obtain a path point on each longitudinal node that avoids collision.

[0116] In some embodiments, in response to the existence of an overlapping area between the target passage space and the lateral reachable range of any longitudinal node, the overlapping area is determined as a path point corresponding to the arbitrary longitudinal node. In this case, when generating a path, a point in the overlapping area can be selected as the path point of the longitudinal node based on the principle of path smoothness.

[0117] In response to the absence of an overlapping area between the target passage space and the transverse reachable range of any longitudinal node, the point in the transverse reachable range closest to the target passage space is determined as the path point corresponding to the arbitrary longitudinal node. That is, if the transverse reachable range cannot reach the target passage space at the arbitrary longitudinal node, the closest point in the transverse reachable range is used as the path point of the longitudinal node.

[0118] The embodiment of the present application determines the path point of the vehicle corresponding to the current longitudinal node based on the target passage space and the lateral reachable range of the current longitudinal node, which can avoid collision with obstacles and ensure the feasibility and safety of the path point; at the same time, by comprehensively considering the lateral accessibility, it can adapt to different road conditions and environmental changes, and can subsequently generate a smoother path, thereby improving the path planning performance. The generated path can travel more safely and efficiently in a complex environment.

[0119] In some embodiments, step S303 may also be implemented through steps S3031 to S3032:

[0120] Step S3031: Generate an initial path based on the path points corresponding to each longitudinal node.

[0121] In an embodiment of the present application, after obtaining the path points corresponding to each longitudinal node, the path points can be connected in sequence to generate an initial path.

[0122] Step S3032: Based on the target travel space and lateral reachable range of each longitudinal node, the initial path is smoothed to obtain a travel path of the vehicle from the current position to the target position.

[0123] In an embodiment of the present application, in order to make the path more streamlined, the initial path needs to be smoothed. The lateral reachable range of the target travel space of the vehicle at each longitudinal node can be used as a constraint condition. The initial path is smoothed by a spatiotemporal trajectory smoothing algorithm such as an iterative linear quadratic regulator (iLQR) to obtain a driving path.

[0124] In the embodiment of the present application, through smoothing processing, turning points and unnecessary turns in the path can be reduced, making the path smoother and more natural, thereby reducing driving difficulty and improving driving safety; the smoothed path has better continuity, which helps the vehicle to be more stable when executing the path and reduces vibration and impact caused by sharp turns.

[0125] In some embodiments, Figure 5 is an optional flow chart of the path planning method provided in the embodiment of the present application, see Figure 5 The path planning method provided in the embodiment of the present application may further include steps S501 to S504:

[0126] Step S501: Acquire longitudinal decision information and lateral decision information of the vehicle.

[0127] In the embodiment of the present application, the lateral decision information may be left lane change, right lane change, lane keeping, etc., and the longitudinal decision information may be target following, target courtesy, constant speed driving, etc. Here, if the current decision action of the vehicle is to change lanes left and catch up with a vehicle in the left lane, then the position of the vehicle is the target position.

[0128] Step S502: Determine a longitudinal reference line based on the transverse decision information.

[0129] Here, the lateral information of the target position can be obtained based on the lateral decision information, and the center line of the lane where the target position is located is used as the longitudinal reference line.

[0130] Step S503: Determine the target position of the vehicle based on the longitudinal decision information.

[0131] Step S504: Based on the current position and the target position of the vehicle, generate a plurality of longitudinal nodes between the current position and the target position in a direction along the longitudinal reference line.

[0132] In an embodiment of the present application, the center line of the lane where the target position is located can be used as the longitudinal reference line, the current position of the vehicle can be used as the origin, and a coordinate system can be established (which can be a Frenet coordinate system, with the longitudinal reference line as the S axis and the axis perpendicular to the longitudinal reference line as the L axis). The IDM model or a similar deduction model can be used to calculate multiple longitudinal nodes of the vehicle between the current position and the target position under the longitudinal decision label, that is, the st discrete point trajectory of the vehicle. Here, the s value in the st trajectory is the longitudinal axis coordinate of the vehicle in the Frenet coordinate system, and the value range of t can be [0:0.1:8.0] seconds, that is, the time t starts from 0 seconds, increases to 8.0 seconds at intervals of 0.1 seconds, and the values ​​of t include 0 seconds, 0.1 seconds, 0.2 seconds, ... 7.9 seconds and 8.0 seconds. These time points constitute a discrete time series, and the end point of the longitudinal node is the farthest longitudinal distance that the vehicle can reach within 8 seconds at the current speed.

[0133] The embodiment of the present application uses multiple longitudinal nodes in the vertical direction between the current position and the target position, and only needs to calculate the lateral position of each longitudinal node as the path point of each longitudinal node, so that there is no need to consider the time dimension when performing path planning. The three-dimensional path search problem is reduced in dimensionality, and only the path point of each longitudinal node in the vertical direction is considered, thereby reducing the calculation complexity.

[0134] The embodiment of the present application further provides an application of a path planning method in an actual scenario.

[0135] Based on the problems existing in the related art, the embodiment of the present application proposes an efficient spatiotemporal path search algorithm, which completes the path search under the given lateral and longitudinal behavior decisions, and the path can satisfy the collision safety constraints.

[0136] The path planning method provided by the present application can be deployed and run under the algorithm structure of decision-making and planning coupling. The algorithm structure includes a decision module and an algorithm module. The decision module samples the action space of the vehicle, and the planning algorithm module performs spatiotemporal trajectory planning for each group of decision actions. Finally, the decision module evaluates all planned trajectories (for example, calculates the cost of each group of trajectories), and selects and sorts them to select the best trajectory for vehicle control. This method is used for the part of path planning after a given decision action.

[0137] The path planning provided in the embodiment of the present application is mainly divided into topology node generation and path search.

[0138] Topological node generation can mean that given a decision action, according to the lateral decision information (left lane change, right lane change, lane keeping), the corresponding lane reference line is obtained, and the frenet coordinate system is generated based on the lane reference line, and the frenet coordinates (L, S) are generated according to the ego positioning of the ego vehicle. According to the longitudinal decision type, the time curve of the ego vehicle position is deduced. For example: the lateral decision is to change lanes left, and the longitudinal decision is to follow the target vehicle agent1 in the left lane. Speed ​​deduction can use the IDM model or other driver models such as the Gipps model to deduce the rough longitudinal discrete trajectory that the ego vehicle can execute under the current longitudinal decision, and represent the longitudinal trajectory as a sequence of st discrete point pairs, t∈[0:0.1:8.0], S=[s1, s2, …, sn].

[0139] After a decision action is given, the longitudinal direction of the path search is fixed along the increasing direction of the frenet longitudinal axis coordinate s. The lateral search direction is a greedy search for the target offset Ltar given by the lateral decision. For example, when changing lanes left, the lateral coordinate L corresponding to the left lane line is usually given as the target offset.

[0140] Topological node generation includes steps S1 to S4, and path search includes steps S5 to S8:

[0141] Step S1, obtaining the state information of the dynamic targets participating in the interaction (including the positioning coordinates, orientation and vehicle posture of the dynamic targets), and the lateral and longitudinal decision information of the vehicle, wherein lateral decisions include left lane change, right lane change, and lane keeping, and longitudinal decisions include target following, target courtesy, and constant speed.

[0142] Step S2: According to the lateral decision type, select the target lane reference line, and generate the corresponding frenet coordinate system based on the reference line, and calculate the SL coordinate of the vehicle in the frenet coordinate system and the SL coordinate of the target with longitudinal decision on the reference line.

[0143] Here, the target lane reference line refers to the lane to which the vehicle is changing, and the center line of that lane is the target lane reference line. Generating the corresponding frenet coordinate system based on the reference line means that the frenet coordinate system is established with the reference line as the S axis of the coordinate system, i.e. the longitudinal line, and the center of the rear axle of the vehicle as the origin of the coordinate system, and the L axis perpendicular to the S axis, i.e. the transverse line, is obtained.

[0144] In the embodiment of the present application, this block can be to obtain a positioning of the vehicle, and perform coordinate conversion to obtain the SL coordinates of the vehicle in the frenet coordinate system. The target SL coordinates with longitudinal decisions on the reference line means that if a certain vehicle is to be followed longitudinally, the target SL coordinates are the coordinates of the vehicle.

[0145] Step S3: According to the longitudinal decision type, the IDM model or other deduction models with the same function are used to calculate the st discrete point trajectory corresponding to the vehicle under the longitudinal decision label.

[0146] The value range of t in the st trajectory is [0:0.1:8.0] seconds, which means that the time t starts from 0 seconds and increases to 8.0 seconds at intervals of 0.1 seconds. The values ​​of t include 0 seconds, 0.1 seconds, 0.2 seconds, ..., 7.9 seconds and 8.0 seconds. The s value is the vertical axis coordinate defined by the ego in the frenet coordinate system in step S2, which can be the farthest distance that can be reached in the vertical direction within the time t.

[0147] Step S4: Perform dynamic target projection.

[0148] Here, dynamic target projection can refer to projecting the dynamic target with the vertical decision label on the reference line onto the time layer according to the discrete time value. That is to say, according to the discrete time, the position of the dynamic target at each moment is obtained, and dynamic avoidance is performed on the same time layer.

[0149] In the embodiment of the present application, the subsequent path search module, when searching at each time node on st, only needs to consider collision avoidance search for the dynamic target posture on the t time layer.

[0150] Step S5: Obtain input environment information, dynamic target information and longitudinal st discrete point trajectory.

[0151] In the embodiment of the present application, environmental information refers to static obstacles on the road where the vehicle is located. All environmental information can be obtained at one time and can be obtained through the vehicle's sensors. Environmental information can include the traversable space between static obstacles, semantic obstacle space information, lane guide lines, lane right of way information, etc. Dynamic target information can refer to information about dynamic targets containing longitudinal decision labels, such as position, posture, forward speed, etc. The longitudinal st discrete point trajectory refers to the st discrete point trajectory obtained in step S3.

[0152] In some embodiments, the attribute information of each topological node includes: 1. Frenet vertical axis coordinate value s, s∈S=[s1,s2,…,sn]; 2. Discrete time value t, used to match the dynamic target predicted trajectory corresponding to time t, and project the posture of the dynamic target at time t to calculate the collision avoidance space; 3. Left and right safety boundaries bound_range (i.e., passable lateral space), defined as the range interval [left_bound,right_bound] on the Frenet horizontal axis coordinate, expressing that the ego has no collision risk within this range. At a longitudinal s value, the passable space may be divided into multiple intervals due to the existence of obstacles, and the generation calculation of the passable interval adopts the rectangular collision method. When there are multiple passable intervals, greedy search will be used to select in the path search link; 4. The lateral reachable range reachable_lat_range that the current node ego can reach, defined as the range interval [reachable_lat_left,reachable_lat_right] on the Frenet horizontal axis coordinate. After each step of greedy search is completed, the reachable_lat_right of the current node can be updated according to the reachable_lat_right of the previous topological node and the bound_range of the current node.

[0153] Step S6: Match the lane guide line according to the vehicle positioning information and initialize the search starting point.

[0154] Here, the lane guide line refers to the S axis of the coordinate axis established in the previous step, and the search starting point refers to the current vehicle position.

[0155] In some embodiments, the attributes of each search node in the st discrete point trajectory include the s value of the current node, the t value of the current node, the lateral reachable range of the vehicle in the frenet coordinate system (i.e., the maximum boundary that can be moved to in the L direction of the coordinate system) and the left and right safety boundaries (i.e., the maximum boundary that will not hit an obstacle in the L direction of the coordinate system).

[0156] Step S7: Use the expansion rectangle expand_rectangle to calculate the left and right safety boundaries bound_range of the current node.

[0157] Figure 6 is a reference schematic diagram of an expansion rectangle provided in an embodiment of the present application, such as Figure 6 As shown, the lower limit of s of the expansion rectangle 601 is the s value of the current search node 602 minus the distance from the center point of the rear axle of the ego vehicle to the rear bumper; the upper limit of s of the expansion rectangle is the s value of the next search node 603 plus the distance from the center point of the rear axle of the ego vehicle to the front bumper.

[0158] Here, the upper and lower bounds of the expansion rectangle in the L direction can be calculated based on the postures of the static obstacles and dynamic targets in the environment at the current node t, as follows: Figure 6 As shown, all dynamic obstacles 604 and static obstacles 605 are sorted in the form of discrete point distribution according to the L-axis direction in the frenet coordinate system, the gap between two adjacent obstacles in the L direction is calculated, the gaps are sorted, and the gaps with a width greater than the ego car width car_width are selected, and the gaps are used as the left and right safety boundaries bound_range of the current node.

[0159] Here, on a search node, there may be multiple gaps whose width is greater than the width of the ego car, that is, on this node, there are multiple passable spaces, and each passable space corresponds to an expansion rectangle expand_rectangle.

[0160] Step S8: Based on the greedy search algorithm framework, determine the left and right safety boundaries corresponding to the current s value, and obtain the L value corresponding to the current S value.

[0161] First, the path search direction is always searched step by step along the increasing direction of s, and the single-step search step length is △s, which is the s position corresponding to each discrete moment t of the longitudinal trajectory.

[0162] Secondly, each search node finally determines an expand_rectangle. When calculating the expansion rectangle corresponding to each search node, the cost considered includes the overlap degree of the left and right safety boundaries bound_range of the front and rear nodes, the overlap rate of the lateral reachable range reachable_lat_range of the previous node and the bound_range of the current node, and the lateral distance of bound_range from the target lane.

[0163] In some embodiments, if there are multiple traversable spaces on the topological node currently being searched, the traversable space is selected by the following formula (2):

[0164] cost=W1*D reach +W2*D space +W3*D target (2);

[0165] Among them, D reach Indicates the overlapping length of the current bound_range and the lateral reachable_lat_range of the previous topological node; D space Indicates the overlapping length of the current bound_range and the bound_range of the previous topological node; D target Indicates the distance from the current passage space bound_range to the lateral decision target offset Ltar. When bound_range contains Ltar, D target =0; Ltar is the target offset given by the lateral decision; for example, when changing lanes left, the lateral coordinate L corresponding to the left lane line is usually given as the target offset.

[0166] Here, the node with the smallest cost can be selected as the final left and right safety boundaries of the search node to determine the L value of the node (ie, the path point).

[0167] Again, after the single-step search is completed, the lateral reachable range of the current node, reachable_lat_range, is updated based on the lateral reachable range of the previous step node, reachable_lat_range, and the current node, bound_range.

[0168] Finally, determine whether the updated node has reached the end of S. If not, repeat the above steps. If it has reached it, generate a path by determining the L value corresponding to each S value.

[0169] Here, if the width of all the expansion rectangles expand_rectangle of the current node is smaller than the car model car_width, that is, the speed planning is carried out according to the current behavior decision, and a collision-free trajectory cannot be searched in the time and space, then the path search program is exited.

[0170] Figure 7 is a flowchart of a greedy search of a horizontal path provided in an embodiment of the present application, such as Figure 7 As shown, the lateral path greedy search can be implemented through steps S701 to S708:

[0171] Step S701: Initialize the current node, lateral safety margin and lateral reachable distance of the vehicle.

[0172] In an embodiment of the present application, the current node of the vehicle can be the position of the current node of the vehicle in the frenet coordinate system. The frenet coordinate system is based on the vehicle selecting the target lane reference line according to the lateral decision type, and generating the corresponding frenet coordinate system based on the reference line, and at the same time calculating the SL coordinate of the vehicle in the frenet coordinate system.

[0173] Here, the lateral safety boundary is the maximum boundary within which the vehicle will not collide with an obstacle at the current node. The lateral safety boundary of the current node can be calculated by expanding the rectangle; the lateral reachable distance is the maximum boundary to which the vehicle can move around the current node.

[0174] Step S702: Calculate the lateral safety margin at the next node.

[0175] In an embodiment of the present application, the lateral safety boundary at the next node can be obtained by expanding the rectangle according to the positions of the dynamic and static targets at the next node at the corresponding time of the next node.

[0176] Step S703: Determine whether the width of the lateral safety margin at the next node is smaller than the vehicle width.

[0177] In an embodiment of the present application, if the width of the lateral safety boundary at the next node is smaller than the width of the vehicle, it means that the current vehicle cannot pass through the dynamic and static obstacles at the next moment, that is, the speed planning is performed according to the current behavior decision, and a collision-free trajectory cannot be searched in time and space. At this time, path planning is no longer performed and the path search program is exited.

[0178] In some embodiments, if at least one of the widths of the lateral safety margins at the next node is greater than the vehicle width, step S704 is executed.

[0179] Step S704: Calculate the cost of each horizontal safety boundary and sort them.

[0180] The cost of each lateral safety boundary at the next node is calculated according to formula (2) in the above embodiment and sorted.

[0181] Step S705: Select the next node with the minimum cost as the current node of the vehicle.

[0182] In this application, the next node with the lowest cost can be used as the current node of the vehicle.

[0183] Step S706: Update the horizontal reachable range of the next node.

[0184] After the next node is used as the current node of the vehicle, the lateral reachable range of the current node can be updated based on the lateral reachable range of the previous node and the lateral safety margin of the current node.

[0185] Step S707: Determine whether the next node reaches the end of the path.

[0186] Determine whether the vehicle has reached the end of the path when it reaches the next node. If so, execute step S708; if not, execute step S702.

[0187] Step S708: Generate an initial path based on all nodes.

[0188] Based on all planned nodes, an initial path is generated to smooth out a collision-free trajectory.

[0189] The embodiment of the present application reduces the time-space multi-dimensional path search problem (multi-dimensional includes longitudinal, lateral and time) to two dimensions (only longitudinal and lateral) through dynamic target projection, thereby improving the search efficiency. When searching for a path, the search is performed based on the greedy search algorithm architecture, and the search is performed along the s-axis direction (i.e. longitudinal) of the Frenet coordinate, which can further compress the problem search dimension, and the time complexity can reach O(n). During the search process, the lateral safety boundary near the vehicle state point at each moment is calculated, and the lateral reachable range at the corresponding moment can be directly used as a constraint condition of the time-space joint smoothing algorithm, which can ensure trajectory smoothness and strictly guarantee collision avoidance.

[0190] Based on the above embodiments, the present application provides a path planning device, such as Figure 2 As shown, the path planning device 254 includes an acquisition module 2541 , a determination module 2542 and a generation module 2543 .

[0191] Among them, the acquisition module 2541 is used to obtain multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, the environmental information of the current longitudinal node among the multiple longitudinal nodes and the state information of the previous longitudinal node of the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time interval between two adjacent longitudinal nodes is the same, and the current longitudinal node is any node among the multiple longitudinal nodes except the current position; the determination module 2542 is used to determine the path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environmental information; the generation module 2543 is used to generate the driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node in response to the current longitudinal node being the target position.

[0192] In some embodiments, each longitudinal node corresponds to a driving moment of the vehicle, and the environmental information includes at least the posture of dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; the determination module 2542 is also used to determine at least one passable space and the lateral safety boundary of each passable space at the driving moment corresponding to the current longitudinal node based on the posture of dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; determine the passage cost of each passable space based on the state information of the previous longitudinal node and the lateral safety boundary of each passable space; determine the passable space whose passage cost meets the target condition as the target passage space of the current longitudinal node; determine the path point of the vehicle corresponding to the current longitudinal node based on the target passage space and the state information of the previous longitudinal node.

[0193] In some embodiments, the determination module 2542 is also used to determine, at the driving moment corresponding to the current longitudinal node, a plurality of lateral gaps between two adjacent dynamic and static obstacles in the lateral direction based on the postures of the dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; compare the plurality of lateral gaps with the vehicle width to obtain a comparison result; determine the lateral gap that meets the traffic conditions in the comparison result as the at least one passable space, and determine the lateral position of each passable space as the lateral safety boundary of each passable space based on the longitudinal reference line; wherein the longitudinal reference line is the center line of the lane where the target position is located, and the lateral direction is perpendicular to the longitudinal direction.

[0194] In some embodiments, the status information includes at least the lateral reachable range and lateral safety boundary of the longitudinal node based on the longitudinal baseline; the determination module 2542 is also used to determine the overlapping length between the lateral reachable range of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node as the first parameter of each passable space; the overlapping length between the lateral safety boundary of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node is determined as the second parameter of each passable space; the distance between the lateral safety boundary of each passable space of the current longitudinal node and the longitudinal baseline is determined as the third parameter of each passable space; based on a preset weight coefficient, the first parameter, the second parameter and the third parameter of each passable space are weightedly calculated respectively to obtain the passage cost of each passable space.

[0195] In some embodiments, the determination module 2542 is also used to update the lateral reachable range of the current longitudinal node based on the target pass space of the current longitudinal node and the lateral reachable range of the previous longitudinal node; and determine the path point of the vehicle corresponding to the current longitudinal node based on the target pass space and lateral reachable range of each longitudinal node.

[0196] In some embodiments, the determination module 2542 is also used to, in response to the existence of an overlapping area between the target passage space and the lateral accessible range of any longitudinal node, determine the overlapping area as the path point corresponding to the arbitrary longitudinal node; in response to the absence of an overlapping area between the target passage space and the lateral accessible range of any longitudinal node, determine the point in the lateral accessible range closest to the target passage space as the path point corresponding to the arbitrary longitudinal node.

[0197] In some embodiments, the generation module 2543 is also used to generate an initial path based on the path points corresponding to each longitudinal node; based on the target passage space and lateral reachable range of each longitudinal node, the initial path is smoothed to obtain the driving path of the vehicle from the current position to the target position.

[0198] In some embodiments, the device also includes: a first acquisition module, used to acquire longitudinal decision information and lateral decision information of the vehicle; a first determination module, used to determine a longitudinal baseline based on the lateral decision information; a second determination module, used to determine a target position of the vehicle based on the longitudinal decision information; a first generation module, used to generate a plurality of longitudinal nodes between the current position and the target position in the direction along the longitudinal baseline based on the current position and the target position of the vehicle.

[0199] In some embodiments, the device also includes: a third determination module, which is used to determine, in response to the current longitudinal node not reaching the target position, the path point of the vehicle corresponding to the next longitudinal node based on the state information of the current longitudinal node and the environmental information, until the next longitudinal node is the target position.

[0200] It should be noted that the description of the device of the embodiment of the present application is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment, so it is not repeated. For technical details not disclosed in the embodiment of the device, please refer to the description of the method embodiment of the present application for understanding.

[0201] If the technical solution of this application involves personal information, the product using the technical solution of this application has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using the technical solution of this application has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, the personal information processing rules are notified by obvious signs / information, and the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

[0202] The embodiment of the present application provides a computer program product or a computer program, which includes an executable instruction, which is a computer instruction; the executable instruction is stored in a computer-readable storage medium. When the processor of the path planning device reads the executable instruction from the computer-readable storage medium and the processor executes the executable instruction, the path planning device executes the method described in the embodiment of the present application.

[0203] An embodiment of the present application provides a storage medium storing executable instructions, wherein executable instructions are stored. When the executable instructions are executed by a processor, the processor will be caused to execute the method provided by the embodiment of the present application.

[0204] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0205] As an example, executable instructions may, but need not necessarily, correspond to a file in a file system, may be stored as part of a file storing other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions). As an example, executable instructions may be deployed to be executed on one electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed at multiple locations and interconnected by a communication network.

[0206] It should be noted here that the description of the above storage medium is similar to the description of the above method embodiment and has similar beneficial effects as the method embodiment. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the description of the method embodiment of this application for understanding.

[0207] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of hardware embodiments, software embodiments, or embodiments in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0208] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0209] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0210] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0211] In addition, all functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be separately configured as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0212] The above is only an embodiment of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent substitutions and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. A path planning method, characterized in that: The method comprises: Acquire multiple longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, the environment information of the current longitudinal node among the multiple longitudinal nodes and the state information of the previous longitudinal node of the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time intervals between two adjacent longitudinal nodes are the same, and the current longitudinal node is any node among the multiple longitudinal nodes except the current position; Determine a path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environment information of the current longitudinal node; In response to the current longitudinal node being the target position, a driving path of the vehicle from the current position to the target position is generated based on path points corresponding to each longitudinal node before the current longitudinal node.

2. The path planning method according to claim 1, characterized in that: Each longitudinal node corresponds to a driving moment of the vehicle, and the environmental information at least includes the posture of the dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; The determining, based on the state information of the previous longitudinal node and the environment information of the current longitudinal node, a path point corresponding to the current longitudinal node of the vehicle comprises: Determine at least one passable space and a lateral safety boundary of each passable space at the time of driving corresponding to the current longitudinal node based on the postures of the dynamic and static obstacles at the time of driving corresponding to the current longitudinal node; Determine the travel cost of each passable space based on the state information of the previous longitudinal node and the transverse safety boundary of each passable space; Determine the passable space whose pass cost meets the target condition as the target passable space of the current longitudinal node; Based on the target passage space and the state information of the previous longitudinal node, a path point of the vehicle corresponding to the current longitudinal node is determined.

3. The path planning method according to claim 2, characterized in that: The determining, based on the posture of the dynamic and static obstacles at the driving time corresponding to the current longitudinal node, at least one passable space at the driving time corresponding to the current longitudinal node and the lateral safety boundary of each passable space includes: At the driving moment corresponding to the current longitudinal node, determining a plurality of lateral gaps between two adjacent dynamic and static obstacles in the lateral direction based on the postures of the dynamic and static obstacles at the driving moment corresponding to the current longitudinal node; Comparing the plurality of lateral gaps with the vehicle width of the vehicle to obtain a comparison result; The lateral gap whose comparison result satisfies the pass conditions is determined as the at least one passable space, and the lateral position of each passable space is determined as the lateral safety boundary of each passable space based on the longitudinal reference line; wherein the longitudinal reference line is the center line of the lane where the target position is located, and the lateral direction is perpendicular to the longitudinal direction.

4. The path planning method according to claim 3, characterized in that: The state information at least includes a lateral reachable range and a lateral safety boundary of the longitudinal node based on the longitudinal reference line; The determining the passage cost of each passable space based on the state information of the previous longitudinal node and the transverse safety boundary of each passable space includes: Determine the overlapping length between the lateral reachable range of the previous longitudinal node and the lateral safety boundary of each passable space of the current longitudinal node as the first parameter of each passable space; Determine the overlapping length between the transverse safety boundary of the previous longitudinal node and the transverse safety boundary of each passable space of the current longitudinal node as the second parameter of each passable space; Determine the distance between the transverse safety boundary of each traversable space of the current longitudinal node and the longitudinal reference line as the third parameter of each traversable space; Based on the preset weight coefficient, the first parameter, the second parameter and the third parameter of each passable space are weightedly calculated to obtain the passage cost of each passable space.

5. The path planning method according to any one of claims 2 to 4, characterized in that: The determining, based on the target passage space and the state information of the previous longitudinal node, a path point corresponding to the current longitudinal node of the vehicle comprises: Based on the target passage space of the current longitudinal node and the transverse reachable range of the previous longitudinal node, updating the transverse reachable range of the current longitudinal node; Based on the target travel space and lateral reachable range of each longitudinal node, a path point corresponding to the current longitudinal node of the vehicle is determined.

6. The path planning method according to claim 5, characterized in that: The determining of the path point corresponding to the current longitudinal node of the vehicle based on the target passage space and the lateral reachable range of each longitudinal node includes: In response to the existence of an overlapping area between the target passage space and the lateral reachable range of any longitudinal node, determining the overlapping area as a path point corresponding to the any longitudinal node; In response to the fact that there is no overlapping area between the target passage space and the transverse reachable range of any longitudinal node, a point in the transverse reachable range closest to the target passage space is determined as a path point corresponding to the any longitudinal node.

7. The path planning method according to claim 5, characterized in that: The generating a driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node includes: Generate an initial path based on the path points corresponding to each longitudinal node; Based on the target travel space and lateral reachable range of each longitudinal node, the initial path is smoothed to obtain a travel path of the vehicle from the current position to the target position.

8. The path planning method according to any one of claims 1 to 4, characterized in that: The method further comprises: Acquiring longitudinal decision information and lateral decision information of the vehicle; Determining a longitudinal reference line based on the transverse decision information; determining a target position of the vehicle based on the longitudinal decision information; Based on the current position and the target position of the vehicle, a plurality of longitudinal nodes are generated between the current position and the target position in a direction along the longitudinal reference line.

9. The path planning method according to any one of claims 1 to 4, characterized in that: The method further comprises: In response to the current longitudinal node not reaching the target position, a path point corresponding to a next longitudinal node of the vehicle is determined based on the state information of the current longitudinal node and the environmental information until the next longitudinal node is the target position.

10. A path planning device, characterized in that: The device comprises: an acquisition module, used to acquire a plurality of longitudinal nodes of the vehicle in the longitudinal direction between the current position and the target position, environmental information of a current longitudinal node among the plurality of longitudinal nodes and state information of a previous longitudinal node of the current longitudinal node; wherein the longitudinal direction is the driving direction of the vehicle, the time intervals between two adjacent longitudinal nodes are the same, and the current longitudinal node is any node among the plurality of longitudinal nodes except the current position; A determination module, configured to determine a path point of the vehicle corresponding to the current longitudinal node based on the state information of the previous longitudinal node and the environmental information; A generation module is used to generate a driving path of the vehicle from the current position to the target position based on the path points corresponding to the longitudinal nodes before the current longitudinal node in response to the current longitudinal node being the target position.

11. A path planning device, characterized in that: The path planning device includes a processor and a memory storing instructions executable by the processor; when the instructions are executed by the processor, the method according to any one of claims 1 to 9 is implemented.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images