Intelligent vehicle local path planning method based on frenet coordinate system lateral and longitudinal decoupling

Abstract

The application discloses a kind of intelligent vehicle local path planning method, device and storage medium, wherein method includes: according to positioning information and high-precision map information, the reference driving path required for intelligent car local path planning is obtained;According to the planning starting point of current planning frame is solved according to tracking effect;Smooth the reference driving path, obtain path frenet coordinate system according to the reference driving path after smoothing;Static obstacles and oncoming vehicles are projected to path frenet coordinate system, SL graph is generated, and path planning trajectory is obtained according to SL graph;Dynamic obstacle short-time prediction trajectory is projected to speed frenet coordinate system, ST graph is generated, and speed planning trajectory is obtained according to ST graph;Path planning trajectory and speed planning trajectory are coupled on natural coordinate system, and restore into high-dimensional local planning trajectory.The application uses SL graph and ST graph to process path and speed respectively, improves efficiency, and can be widely applied to trajectory planning technical field in unmanned driving.

Description

Technical Field

[0001] This invention relates to the field of trajectory planning technology in autonomous driving, and in particular to a method, device and storage medium for local path planning of intelligent vehicles. Background Technology

[0002] Autonomous vehicles can be categorized based on their driving scenarios into two types: those with clear road structures and known lane shapes, such as highways and urban arterial roads, and those with unclear road structures and unknown lane shapes, such as underground parking garages and rural roads. Structured scenarios are characterized by fixed lane shapes, highly dynamic environments, high vehicle speeds, and short response times for the autonomous driving system. Since vehicles primarily encounter structured road scenarios in urban areas, local path planning on structured roads is a crucial aspect of autonomous driving technology research.

[0003] Currently, the mainstream approach for this type of scenario is to perform local trajectory planning based on sampling, including two main categories of methods: random sampling and deterministic sampling.

[0004] Random sampling methods mainly include Probabilistic Roadmap (PRM) and Rapidly-exploring Random Tree (RRT). These methods explore the environment step-by-step through random sampling, quickly obtaining local trajectories while avoiding getting trapped in local optima. However, due to the randomness of the exploration process, the planning results of this method are difficult to predict, resulting in a certain degree of uncontrollability in practical industrial applications.

[0005] Deterministic sampling methods generate a certain number of sampled trajectories to traverse all possible vehicle motions, thereby obtaining a relatively optimal solution that satisfies the constraints. This method typically calculates a large number of sampled trajectories by ignoring constraints or using parametric trajectory models, resulting in relatively good planning results. However, it also wastes a significant amount of computational resources and time. Summary of the Invention

[0006] In order to at least partially solve one of the technical problems existing in the prior art, the present invention aims to provide a method, device and storage medium for local path planning of intelligent vehicles.

[0007] The technical solution adopted in this invention is:

[0008] A local path planning method for intelligent vehicles includes the following steps:

[0009] Obtain the location information of the intelligent vehicle, and based on the location information and high-precision map information, obtain the reference driving path required for the local path planning of the intelligent vehicle;

[0010] Obtain the tracking effect of the intelligent vehicle in the previous frame, and solve the planning starting point of the current planning frame based on the tracking effect;

[0011] Starting from the planning point, the reference driving path is smoothed, and the Frenet coordinate system of the path is obtained based on the smoothed reference driving path.

[0012] Project static obstacles and oncoming vehicles onto the path Frenet coordinate system to generate an SL map, and obtain the path planning trajectory based on the SL map.

[0013] Based on the path planning trajectory, the velocity Frenet coordinate system is obtained. The short-time predicted trajectory of the dynamic obstacle is projected onto the velocity Frenet coordinate system to generate the ST diagram. The velocity planning trajectory is then obtained from the ST diagram.

[0014] The path planning trajectory and velocity planning trajectory are coupled in the natural coordinate system to restore the high-dimensional local planning trajectory.

[0015] Furthermore, the step of obtaining the intelligent vehicle's location information and obtaining a reference driving path required for local path planning of the intelligent vehicle based on the location information and high-precision map information includes:

[0016] Before each local path planning begins, the vehicle extracts the road segment that it needs to travel within a preset time period from the global path by combining the location information and high-precision map information, and converts the road segment into a reference driving path for the intelligent vehicle's local path planning.

[0017] Furthermore, the step of obtaining the tracking effect of the intelligent vehicle in the previous frame and solving for the planning starting point of the current planning frame based on the tracking effect includes:

[0018] Obtain the matching point of the intelligent vehicle's current position on the planned trajectory in the previous frame, and calculate the corresponding lateral and longitudinal errors based on the matching point;

[0019] Real-time error is calculated based on lateral and longitudinal errors to characterize the tracking performance.

[0020] The planning starting point of the current planning frame is determined based on the real-time error.

[0021] Furthermore, the smoothing of the reference driving path starting from the planned starting point includes:

[0022] A k-point discrete smoothing algorithm is used to smooth the reference driving path; the smoothing requirements are as follows:

[0023] 1) The connection curves of discrete points are required to be smooth, which is reflected in the cost of continuous smoothing;

[0024] 2) It requires uniform spacing between discrete points, which manifests as a relatively compact cost;

[0025] 3) It is required that the smoothed discrete points do not deviate from the original discrete points, which is represented by the original similarity cost.

[0026] Further, the step of projecting static obstacles and oncoming vehicles onto the path Frenet coordinate system to generate an SL map, and obtaining the path planning trajectory based on the SL map, includes:

[0027] Project static obstacles and oncoming vehicles onto the path Frenet coordinate system to generate an SL map;

[0028] In the SL diagram, uniform sampling of ▽s is performed along the S dimension, resulting in a total of n discrete points;

[0029] Suppose that among n discrete points, the coordinate value of each discrete point in the L direction is x. i The first, second, and third derivatives of the corresponding S-dimensional region are x′, x″, and x″′, respectively.

[0030] An optimization problem is constructed based on the costs of proximity to the reference centerline, lateral velocity, lateral acceleration, and lateral jerk, and the path planning trajectory is obtained by solving the optimization problem.

[0031] Furthermore, the expression for the optimization problem is:

[0032]

[0033] In the formula, ω x ω x′ ω x″ ω x″′ These are the weight parameters for each cost.

[0034] Further, the step of obtaining the velocity Frenet coordinate system based on the path planning trajectory, projecting the short-time predicted trajectory of the dynamic obstacle onto the velocity Frenet coordinate system to generate an ST map, and obtaining the velocity planning trajectory based on the ST map includes:

[0035] The generated path planning trajectory is converted into a velocity Frenet coordinate system for velocity planning reference;

[0036] The short-time predicted trajectory of dynamic obstacles is projected onto the velocity Frenet coordinate system to generate an ST diagram.

[0037] Discrete sampling is performed on the ST diagram at intervals of Δt to generate n discrete states S1, S2, S3...S n-1 ;

[0038] Based on the discrete state, a dynamic programming algorithm is used to generate the optimal velocity planning curve;

[0039] Within the convex space surrounding the velocity planning curve, quadratic programming optimization is used to obtain the final velocity planning trajectory.

[0040] Furthermore, suppose that the number of discrete points obtained by dynamic programming is n, and the ordinate s of each point is s. i Considering the cost of offsetting the reference position, the cost of the reference velocity, the cost of the reference acceleration, and the cost of the reference jerk, the expression for the optimization objective is as follows:

[0041]

[0042] In the formula, ω s ω v ω a ω j This represents the corresponding cost weight.

[0043] Another technical solution adopted in this invention is:

[0044] A local path planning device for intelligent vehicles, comprising:

[0045] At least one processor;

[0046] At least one memory for storing at least one program;

[0047] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.

[0048] Another technical solution adopted in this invention is:

[0049] A computer-readable storage medium storing a processor-executable program, which, when executed by a processor, performs the method described above.

[0050] The beneficial effects of this invention are: This invention performs high-dimensional decoupling based on the Frenet coordinate system, projects obstacles, and uses SL and ST diagrams to process the path and velocity separately, avoiding the problem of slow solution speed caused by directly solving in high dimensions. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following description is provided with accompanying drawings of the relevant technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1This is a flowchart of a local path planning method for intelligent vehicles oriented towards structured roads, as described in an embodiment of the present invention.

[0053] Figure 2 This is a schematic diagram of the current frame planning starting point in an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of the k-point discrete smoothing algorithm in an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of the local path environment of the vehicle in an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of path planning sampling optimization in an embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of the dynamic programming result in the speed planning in an embodiment of the present invention;

[0058] Figure 7 This is a schematic diagram of the formation of convex space in the velocity planning of this invention embodiment;

[0059] Figure 8 This is a schematic diagram of the quadratic programming result in the convex space during velocity programming in an embodiment of the present invention. Detailed Implementation

[0060] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0061] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0062] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0063] In the description of this invention, unless otherwise explicitly defined, terms such as "setting," "installing," and "connecting" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0064] To save computational resources and time, the Frenet coordinate system can be used for decoupling, and a hierarchical optimization approach can be adopted to decompose high-dimensional variables into low-dimensional variables for separate optimization. However, the path planning and velocity planning in this method are optimized completely independently, and the calculation results may not be optimal. Therefore, a further method is proposed that uses the Frenet coordinate system for high-dimensional decoupling, projects obstacles, and then uses SL and ST diagrams to sample and optimize the path and velocity separately, avoiding the problem of slow solution speed caused by directly solving in high dimensions. In addition, the method of decision-making before optimization is abandoned in the path planning layer, achieving optimality in path planning. At the same time, this invention makes reasonable inferences and predictions about the planning starting point of each frame, which not only provides good initial values ​​for the optimization of the current frame, but also ensures the temporal consistency of the planning system. In complex road conditions of structured roads, this invention has the advantages of high real-time computational efficiency and strong adaptability.

[0065] like Figure 1 As shown, this embodiment provides a local path planning method for intelligent vehicles on structured roads, including the following steps:

[0066] S1. Obtain the location information of the intelligent vehicle, and obtain the reference driving path required for local path planning of the intelligent vehicle based on the location information and high-precision map information.

[0067] Determine the attributes of the planning period and identify the local reference path required for the current planning period.

[0068] At the beginning of each planning cycle, the first step is to determine whether the current planning cycle is an initial planning, replanning, or normal planning. Let the local reference path length required for each planning frame be d. total The effective reference path length of the previous frame is l valid .

[0069] For initial planning and replanning, since there is no reference value from the previous frame or the reference value from the previous frame is invalid, it is necessary to re-match a length of d based on the high-precision map.total Unoptimized local reference path.

[0070] For normal planning, it is only necessary to have a valid reference path length of l in the previous frame. valid Then match d total -l valid Unoptimized local reference path of length.

[0071] S2. Obtain the tracking effect of the intelligent vehicle in the previous frame, and solve the planning starting point of the current planning frame based on the tracking effect.

[0072] As an optional implementation method, a strategy combining prediction and matching was adopted to determine the starting point of the autonomous vehicle planning for the current planning cycle.

[0073] Find the matching point between the current position of the intelligent vehicle and the planned trajectory in the previous frame, and calculate the corresponding horizontal and vertical errors. Based on the real-time errors, determine the planning starting point of the current planning frame.

[0074] like Figure 2 Let the current time be t, and the current position of the vehicle be point A (A x A y Each frame has a planning period of T seconds, and the lateral error of the vehicle following the planning result of the previous frame at the current moment is E. d The longitudinal error is E s The vehicle's current heading speed is ν, and its yaw angle is... The acceleration is a.

[0075] If the horizontal and vertical errors at the current moment are too large (the horizontal error threshold is E), dth The lateral error threshold is E sth If the position of the vehicle is C, which is T seconds later, then the position is determined by dynamic extrapolation. x C y Otherwise, use the matching result from the previous frame T seconds later as the planning starting point C, and the horizontal coordinate value of its matching point is C. xmatch The vertical coordinate value is C ymatch .

[0076] Within each planning period of T seconds, the vehicle continues to move under the control of the control module. Let's say the vehicle reaches point C after T seconds; its position can be mathematically expressed as follows:

[0077]

[0078]

[0079] S3. Starting from the planning starting point, smooth the reference driving path and obtain the path Frenet coordinate system based on the smoothed reference driving path.

[0080] Starting from the planning origin, the k-point discrete smoothing algorithm is used to smooth the reference path generated in step S1, and the smoothed reference path is converted into the Frenet coordinate system referenced in the local path planning.

[0081] Smoothing evaluation criteria fall into three main categories: one is to make the connection curves of discrete points as smooth as possible; another is to make the intervals between discrete points as uniform as possible; and the third is to make the smoothed discrete points as close as possible to the original discrete points. These correspond to smoothing cost, compaction cost, and geometric similarity cost, respectively.

[0082] like Figure 3 As shown, let P i Let be one of n points to be optimized, and its corresponding coordinates be (x, y, y). i ,y i );p ir For a series of optimized points, the corresponding coordinates are (x... ir ,y ir Let X be the set of points to be optimized, and buff be the threshold of the sampling constraint space. Then, the specific mathematical problem of the k-point discrete smoothing algorithm is expressed as:

[0083]

[0084] st-buff≤x ir -x i ≤buff

[0085] Let X = [x1 … x n-1 ] T Expanding the smoothing cost, compaction cost, and geometric similarity cost, let ω smooth ω length ω ref The weights of each cost are represented by the following binary programming form:

[0086] cost smooth =ω smooth ·X T A1 T A1X

[0087] cost length =ω length ·X T A2 T A2X

[0088] cost ref =ω ref ·(X T A3 T A3X+h T X)

[0089] Where X is a 2n*1 column vector, A1 is a (2n-4)*2n matrix, A2 is a 2n*(2n-2) matrix, A3 is a 2n*2n matrix, and h is a 2n*1 column vector.

[0090] Combining the three optimization costs and simplifying them, the mathematical expression for the total cost of the quadratic programming is:

[0091] cost total =X T (ω smooth A1 T A1+ω length A2 T A2+ω ref A3 T A3)X+ω ref h T X

[0092] =X T HX+f T X

[0093] in:

[0094] H=(ω smooth A1 T A1+ω length A2 T A2+ω ref A3 T A3)

[0095] f = ω ref h T

[0096] The optimized result is as follows Figure 3 As shown, it is then used as the Frenet coordinate system required for path planning.

[0097] S4. Project static obstacles and oncoming vehicles onto the path Frenet coordinate system to generate an SL diagram, and obtain the path planning trajectory based on the SL diagram.

[0098] The static obstacles and oncoming vehicles are projected into the Frenet coordinate system generated in step S3 to generate the SL diagram. Then, the drivable range is constrained on the SL diagram to perform trajectory optimization and generate a path planning trajectory that satisfies vehicle dynamics.

[0099] Starting from the path planning origin, project obstacles into the Frenet coordinate system generated in step S3. Local path conditions are as follows: Figure 4 As shown, the SL diagram generated after projection is as follows: Figure 5 .exist Figure 5 In the process, uniform sampling of S is performed with a value of ▽s to obtain a series of discrete points X. iAssume a total of n points are sampled, and the L-direction coordinate value of each point is x. i Its corresponding first, second, and third derivatives are x′, x″, and x″′, respectively.

[0100] Based on the costs of proximity to the reference centerline, lateral velocity, lateral acceleration, and lateral jerk, the following optimization problem is constructed, and its mathematical expression is as follows:

[0101]

[0102] Because x′″ i ·Δs=x″ i+1 -x″ i ,have Substituting it in, we get:

[0103]

[0104] After simplification and rewriting in the form of a quadratic programming problem, we have:

[0105]

[0106] st l≤AX≤u where:

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118] Finally, the optimization mathematical model constructed above is solved using optimization tools to obtain the final path planning trajectory, as shown below. Figure 4 In the context of cycling motion planning or Figure 5 The optimized path planning points.

[0119] S5. Obtain the velocity Frenet coordinate system based on the path planning trajectory, project the short-time predicted trajectory of the dynamic obstacle onto the velocity Frenet coordinate system to generate the ST diagram, and obtain the velocity planning trajectory based on the ST diagram.

[0120] The path planning trajectory generated in step S4 is transformed into a Frenet coordinate system for velocity planning reference, and the short-time predicted trajectory of dynamic obstacles is projected onto this Frenet coordinate system to generate an ST diagram. Dynamic programming is used on the ST diagram to obtain a rough trajectory with obstacle avoidance function. Then, a convex space is created around the rough trajectory, and finally, a quadratic programming algorithm is used to optimize and generate a velocity planning trajectory that satisfies vehicle dynamics.

[0121] Using dynamic programming and quadratic programming methods, the velocity dimension is sampled and optimized based on the path planning trajectory generated in step S4 to generate a velocity planning trajectory that satisfies vehicle dynamics. Specifically, the velocity planning trajectory mainly consists of three steps:

[0122] Step 1: Use the path plan generated in step S4 as the Frenet coordinate system for velocity planning, and then project the dynamic obstacles onto this coordinate system to form the corresponding ST diagram, such as... Figure 6 Then, discrete sampling is performed on the ST diagram. Let there be n discrete sampling point states, denoted as S1, S2, S3...S... n-1 .

[0123] Step 2: Using dynamic programming, at n sampling points S1, S2, S3...S... n-1 In this process, find the optimal combination of states. See also Figure 7 Then, a convex space is formed around the optimal combination of discrete points according to the constraint relationship.

[0124] Step 3: See Figure 8 Within the coarse discrete trajectory generated in the first step and the convex space generated in the second step, an optimization method is used to solve for the optimal velocity trajectory. Specifically, let the ordinates of the n discrete points obtained by dynamic programming be s. i Considering the costs of deviation from the centerline, deviation from the reference velocity, acceleration, and jerk, the mathematical problem can be constructed as follows:

[0125]

[0126] Will Substituting into the above equation and rearranging, we get:

[0127]

[0128] After simplification and rewriting in the form of a quadratic programming problem, we have:

[0129]

[0130] st l≤AX≤u

[0131] Similar to path planning, where:

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138] Finally, the optimization problem constructed above is solved using optimization tools to obtain the final velocity planning trajectory, such as... Figure 8 As shown.

[0139] S6. Couple the path planning trajectory and velocity planning trajectory in the natural coordinate system to restore the high-dimensional local planning trajectory.

[0140] The path planning trajectory generated in step S4 and the velocity planning trajectory generated in step S5 are coupled in the natural coordinate system to restore a high-dimensional local planning trajectory. Finally, the high-dimensional planning trajectory that satisfies vehicle driving safety, vehicle dynamics, and human comfort is sent to the control module for execution.

[0141] As can be seen from the above, this embodiment provides a local path planning method for intelligent vehicles on structured roads, which solves the time consistency problem and the problem of slow solution speed for high-dimensional optimization. Furthermore, it utilizes high-precision maps for matching, which effectively improves the positioning accuracy of the system and thus ensures the reliability of local paths for vehicles without vehicles.

[0142] In summary, the method of this embodiment has the following advantages and beneficial effects compared with the prior art:

[0143] (1) The embodiments of the present invention adopt a strategy that combines prediction and matching, which can accurately and dynamically predict the position of the vehicle after one planning cycle. This provides a good initial value for the subsequent optimization algorithm and improves the robustness and time consistency of the system.

[0144] (2) In this embodiment of the invention, a precise local reference path is obtained through a high-precision map, and then the local reference path is smoothed in advance to ensure the accuracy of the transformed Frenet coordinate system, effectively reduce the error of obstacle projection and collision detection in the later stage, and improve the accuracy of the system.

[0145] (3) This embodiment of the invention employs a Frenet coordinate system decoupling method, solving path planning and velocity planning separately. This avoids the problem of sacrificing solution speed for higher-dimensional optimization, and the separate optimization method improves the real-time performance and response speed of the system.

[0146] (4) In the speed planning process, the embodiment of the present invention uses a method of point sampling combined with dynamic planning based on the ST diagram generated by real-time obstacle projection to open up the convex space required for secondary planning, thus ensuring that the necessary conditions in the optimization algorithm are met.

[0147] (5) In this embodiment of the invention, the obtained path planning and speed planning are coupled to restore high-dimensional information in the natural coordinate system, and then sent to the control module for execution, thus ensuring decoupling between modules.

[0148] This embodiment also provides an intelligent vehicle local path planning device, including:

[0149] At least one processor;

[0150] At least one memory for storing at least one program;

[0151] When the at least one program is executed by the at least one processor, the at least one processor implements Figure 1 The method shown.

[0152] This embodiment of the intelligent vehicle local path planning device can execute the intelligent vehicle local path planning method provided in the method embodiment of the present invention, and can execute any combination of implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0153] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform... Figure 1 The method shown.

[0154] This embodiment also provides a storage medium storing instructions or programs that can execute the intelligent vehicle local path planning method provided in the method embodiment of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the method has the corresponding functions and beneficial effects.

[0155] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0156] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0157] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0158] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0159] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0160] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0161] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0162] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0163] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A local path planning method for intelligent vehicles, characterized in that, Includes the following steps: Obtain the location information of the intelligent vehicle, and based on the location information and high-precision map information, obtain the reference driving path required for the local path planning of the intelligent vehicle; Obtain the matching point of the intelligent vehicle's current position on the planned trajectory in the previous frame, and calculate the corresponding lateral and longitudinal errors based on the matching points; calculate the real-time error based on the lateral and longitudinal errors to characterize the tracking effect; and solve for the planning starting point of the current planning frame based on the real-time error. Starting from the planning starting point, the reference driving path is smoothed, and the first Frenet coordinate system is obtained based on the smoothed reference driving path. Project static obstacles and oncoming vehicles onto the first Frenet coordinate system to generate an SL map, and obtain the path planning trajectory based on the SL map; The second Frenet coordinate system is obtained based on the path planning trajectory. The short-time predicted trajectory of the dynamic obstacle is projected onto the second Frenet coordinate system to generate the ST diagram. The velocity planning trajectory is obtained based on the ST diagram. The path planning trajectory and velocity planning trajectory are coupled in the natural coordinate system to restore the high-dimensional local planning trajectory. The step of projecting static obstacles and oncoming vehicles onto a first Frenet coordinate system to generate an SL map, and obtaining the path planning trajectory based on the SL map, includes: Project static obstacles and oncoming vehicles onto the first Frenet coordinate system to generate an SL map; In the SL diagram, it is performed along the S dimension. The uniform sampling forms a total of n discrete points; Assume that among n discrete points, the L-direction coordinate value of each discrete point is... The first, second, and third derivatives of the corresponding S-dimensional space are respectively , , ; An optimization problem is constructed based on the costs of proximity to the reference centerline, lateral velocity, lateral acceleration, and lateral jerk, and the path planning trajectory is obtained by solving the optimization problem. The expression for the optimization problem is: In the formula, , , , These are the weight parameters for each cost; The step of obtaining the second Frenet coordinate system based on the path planning trajectory, projecting the short-time predicted trajectory of the dynamic obstacle onto the second Frenet coordinate system to generate the ST map, and obtaining the velocity planning trajectory based on the ST map includes: The generated path planning trajectory is transformed into a second Frenet coordinate system for velocity planning reference; The short-time predicted trajectory of dynamic obstacles is projected onto the second Frenet coordinate system to generate the ST diagram. On the ST chart Discrete sampling is performed at intervals to generate n discrete states. , , ; Based on the discrete state, a dynamic programming algorithm is used to generate the optimal velocity planning curve; Within the convex space surrounding the velocity planning curve, quadratic programming optimization is used to obtain the final velocity planning trajectory; Suppose that the dynamic programming yields n discrete points, and the ordinate s of each point is... Considering the cost of offsetting the reference position, the cost of the reference velocity, the cost of the reference acceleration, and the cost of the reference jerk, the expression for the optimization objective is as follows: In the formula, , , , This represents the corresponding cost weight.

2. The intelligent vehicle local path planning method according to claim 1, characterized in that, The process of obtaining the location information of the intelligent vehicle and obtaining a reference driving path required for local path planning of the intelligent vehicle based on the location information and high-precision map information includes: Before each local path planning begins, the vehicle extracts the road segment that it needs to travel within a preset time period from the global path by combining the location information and high-precision map information, and converts the road segment into a reference driving path for the intelligent vehicle's local path planning.

3. The intelligent vehicle local path planning method according to claim 1, characterized in that, The smoothing of the reference driving path starting from the planned starting point includes: A k-point discrete smoothing algorithm is used to smooth the reference driving path; the smoothing requirements are as follows: 1) The connection curves of discrete points are required to be smooth, which is a continuous smoothing cost; 2) It requires uniform spacing between discrete points, which manifests as a relatively compact cost; 3) It is required that the smoothed discrete points do not deviate from the original discrete points, which is represented by the original similarity cost.

4. A local path planning device for intelligent vehicles, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method of any one of claims 1-3.

5. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-3.

Images