A path planning method, device, equipment and storage medium

Abstract

The present invention provides a path planning method, apparatus, device and storage medium, which obtain a first reference path obtained by performing path planning with the center of the rear axle of the target vehicle as the center; convert the first reference path into a second reference path with the center of the front axle as the center; construct a plurality of road grid lines with the second reference path as the reference line, and the longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line; take the current planned position on the second reference path as the initial state, traverse the grid positions corresponding to each road network line, and obtain the path expressions of a plurality of candidate paths through polynomial fitting; calculate the loss values of the loss functions of each candidate path respectively by using the path expressions of each candidate path based on the positions of obstacles around the target vehicle; and screen the target planned path from each candidate path based on the loss values. It avoids the vehicle head from exceeding the lane and improves the accuracy of path planning.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving, and particularly to a path planning method, device, equipment and storage medium. Background Art

[0002] With the continuous acceleration of the research, application and commercialization process of autonomous driving technology, the demand for driving safety technology is also increasing. More closed-scene driverless systems require the removal of safety officers to achieve true driverless driving. In a driverless system, trajectory planning is divided into path planning and speed planning. Among them, path planning mainly plans a reasonable driving path for the vehicle according to the information of obstacles, and then provides the result of path planning for speed planning to use. Finally, the results of path planning and speed planning are merged and sent to the control module to calculate the steering wheel control amount, brake and throttle control amounts, etc. to control the vehicle to drive.

[0003] The existing path planning uses the rear axle of the vehicle for path planning, regards the vehicle as a particle, only considers simple scenarios, converts the reference path and obstacles to the frenet coordinate system for design, and then converts to the Cartesian coordinate system. Due to the distortion of obstacles during the conversion process and the failure to consider the differences in the paths of the front and rear axles of the vehicle, the accuracy of the path planning result is poor, and the vehicle head is likely to exceed the lane. Especially for vehicle models with a long body such as a semi-trailer truck, the existing path planning method is difficult to apply. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a path planning method, device, equipment and storage medium to overcome the problem that the path planning method based on the rear axle of the vehicle in the prior art is prone to the phenomenon that the vehicle head exceeds the lane and is difficult to apply in vehicle models with a long body.

[0005] According to a first aspect, an embodiment of the present invention provides a path planning method, including:

[0006] Obtaining a first reference path obtained by performing path planning with the rear axle of the target vehicle as the center;

[0007] Converting the first reference path into a second reference path with the front axle as the center;

[0008] Constructing a plurality of road grid lines with the second reference path as the reference line, the longitudinal intervals of the corresponding position points of each road grid line are the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line;

[0009] Taking the current planning position on the second reference path as the initial state, traversing the grid positions corresponding to each road network line, and obtaining the path expressions of a plurality of candidate paths through polynomial fitting;

[0010] Based on the positions of the obstacles around the target vehicle, calculate the loss values of the loss functions of each candidate path respectively using the path expressions of each candidate path;

[0011] Based on the loss values of the loss functions of each candidate path, screen the target planning path from each candidate path.

[0012] Optionally, convert the first reference path into a second reference path centered on the front axle through the following formula:

[0013]

[0014]

[0015] where x f , y f represent the position coordinates of the front axle on the second reference path, L w represents the distance between the front and rear axles, x r , y r represent the position coordinates of the rear axle on the first reference path, represents the heading angle of the first reference path.

[0016] Optionally, based on the positions of the obstacles around the target vehicle, calculate the loss values of the loss functions of each candidate path respectively using the path expressions of each candidate path, including:

[0017] Based on the positions of the obstacles around the target vehicle, calculate the distance cost between the current candidate path and all obstacles;

[0018] Based on the path expression of the current candidate path, calculate the centripetal acceleration, first derivative, second derivative of the current candidate path and its deviation from the reference path;

[0019] Calculate the loss value of the loss function corresponding to the current candidate path based on the distance cost, centripetal acceleration, first derivative, second derivative and deviation.

[0020] Optionally, based on the loss values of the loss functions of each candidate path, screen the target planning path from each candidate path, including:

[0021] Sort each candidate path in ascending order of the loss value of the loss function;

[0022] Based on the sorting result, screen the target planning path that meets the preset requirements.

[0023] Optionally, the method further includes:

[0024] Convert the target planning path into a third planning path centered on the rear axle;

[0025] Perform safety detection on the third planned path, where the safety detection includes: collision detection and constraint detection;

[0026] When the third planned path fails the safety detection, delete the target planned path from the candidate paths, and return to the step of screening the target planned path from the candidate paths based on the loss values of the loss functions of the candidate paths.

[0027] Optionally, when the third planned paths corresponding to all candidate paths do not pass the safety test, control the target vehicle to stop.

[0028] Optionally, when the third planned path passes the safety detection, perform speed planning on the target vehicle based on the third planned path.

[0029] According to a second aspect, an embodiment of the present invention provides a path planning device, including:

[0030] An acquisition module, configured to acquire a first reference path obtained by performing path planning with the rear axle of the target vehicle as the center;

[0031] A first processing module, configured to convert the first reference path into a second reference path with the front axle as the center;

[0032] A second processing module, configured to construct a plurality of road grid lines with the second reference path as the reference line, where the longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line;

[0033] A third processing module, configured to use the current planned position on the second reference path as the initial state, traverse the grid positions corresponding to each road network line, and obtain the path expressions of a plurality of candidate paths through polynomial fitting;

[0034] A fourth processing module, configured to calculate the loss values of the loss functions of the candidate paths respectively by using the path expressions of the candidate paths based on the positions of the obstacles around the target vehicle;

[0035] A fifth processing module, configured to screen the target planned path from the candidate paths based on the loss values of the loss functions of the candidate paths.

[0036] According to a third aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method of the first aspect of the present invention and any of its optional methods is implemented.

[0037] According to a fourth aspect, an embodiment of the present invention provides a path planning device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the method according to the first aspect of the present invention and any of its optional implementations.

[0038] The technical solution of the present invention has the following advantages:

[0039] The path planning method provided by the embodiment of the present invention includes: obtaining a first reference path obtained by performing path planning with the rear axle of the target vehicle as the center; converting the first reference path into a second reference path with the front axle as the center; constructing a plurality of road grid lines with the second reference path as the reference line, where the longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line; taking the current planned position on the second reference path as the initial state, traversing the grid positions corresponding to each road network line, and obtaining the path expressions of a plurality of candidate paths through polynomial fitting; calculating the loss values of the loss functions of each candidate path respectively based on the positions of the obstacles around the target vehicle by using the path expressions of each candidate path; and screening the target planned path from each candidate path based on the loss values of the loss functions of each candidate path. Therefore, by converting the reference path with the rear axle as the center to the front axle as the reference line, constructing the road grid by defining the positional relationship between the road grid and the reference line, obtaining several candidate paths by using the road grid through polynomial fitting, and using the loss values of the loss functions of each path as the screening basis to obtain the target planned path, the problem of the vehicle head exceeding the lane is effectively avoided by comprehensively constructing candidate paths with the front and rear axles, and the entire path planning process is completed in the Cartesian coordinate system without coordinate conversion, avoiding the problem of obstacle distortion during coordinate conversion and improving the accuracy of path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0041] Figure 1 It is a flowchart of the path planning method in the embodiment of the present invention;

[0042] Figure 2 It is a comparison schematic diagram of the front axle path and the rear axle path in the embodiment of the present invention;

[0043] Figure 3 It is a schematic diagram of the road grid lines in the embodiment of the present invention;

[0044] Figure 4 Schematic diagram of the specific working process of path planning in an embodiment of the present invention;

[0045] Figure 5 Schematic diagram of the structure of a path planning device in an embodiment of the present invention;

[0046] Figure 6 Schematic diagram of the structure of a path planning device in an embodiment of the present invention. Detailed implementation manners

[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] In addition, the technical features involved in different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0049] In the existing path planning, the rear axle of the vehicle is used for path planning, and the vehicle is regarded as a particle. Only simple scenarios are considered, and the reference path and obstacles are both converted to the frenet coordinate system for design and then converted to the Cartesian coordinate system. Due to the distortion of obstacles during the conversion process and the difference in the paths of the front and rear axles of the vehicle not being considered, the accuracy of the path planning result is poor, and the vehicle head is likely to exceed the lane. Especially for vehicle models with a long body such as a semi-trailer truck, the existing path planning method is difficult to apply.

[0050] Based on the above problems, an embodiment of the present invention provides a path planning method, as Figure 1 shown, and the path planning method specifically includes the following steps:

[0051] Step S101: Obtain a first reference path obtained by performing path planning with the rear axle of the target vehicle as the center.

[0052] Specifically, the first reference path can be obtained according to the method of performing path planning based on the rear axle of the vehicle in the prior art, and the detailed implementation process will not be elaborated here.

[0053] Step S102: Convert the first reference path into a second reference path with the front axle as the center.

[0054] Exemplarily, as Figure 2 shown, when the vehicle is represented by a rectangle, during its driving process, it can be seen that the paths of the front axle and the rear axle of the vehicle are different. Since the vehicle steering is at the front wheels, the heading angle of the vehicle is the heading angle on the rear axle path.

[0055] Specifically, the above step S102 converts the first reference path into a second reference path centered on the front axle through the following formula:

[0056]

[0057]

[0058] where x f , y f represent the front axle position coordinates on the second reference path, L w represents the distance between the front and rear axles, x r , y r represent the rear axle position coordinates on the first reference path, represents the heading angle of the first reference path.

[0059] In practical applications, in order to simplify the calculation and further improve the efficiency of path planning, before performing the above step S102, the first reference path can be downsampled and sparsely processed to obtain a series of discrete coordinate points, and then these discrete coordinate points are processed according to the above formula (1) to obtain the second reference path. The specific sampling interval can be flexibly set according to the actual path planning accuracy and computing power, and the present invention is not limited thereto. In addition, after obtaining the second reference path composed of discrete coordinate points, the second reference path can be smoothed, such as: spline smoothing, or quadratic programming smoothing, or discrete point smoothing, or spiral curve smoothing, etc., so that the second reference path is closer to the actual path of vehicle driving and excludes the interference of discrete points, improving the accuracy of the converted path.

[0060] Step S103: Construct a plurality of road grid lines with the second reference path as the reference line, and the longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line. Exemplarily, constructing a plurality of road grid lines with the second reference path as the reference line is as Figure 3 shown.

[0061] Step S104: Taking the current planned position on the second reference path as the initial state, traverse the grid positions corresponding to each road network line, and obtain the path expressions of a plurality of candidate paths through polynomial fitting.

[0062] Specifically, taking the current planned position of the vehicle (x0, y0, x0′, y0′, x0″, y0″, s0) as the initial state, traverse the positions (x k , y k , x k ′, y k ′, x k″,y k ″,s k ), the path expression is obtained by fitting a fifth-order polynomial as shown in formulas (2) and (3):

[0063]

[0064]

[0065] Among them, x0, y0, x0′, y0′, x0″, y0″, s0 represent the position coordinates of the current planned position, the first-order derivative of the path, the second-order derivative of the path, and the path moving distance, respectively. k ,y k , x k ′,y k ′,x k ″,y k ″,s k , respectively represent the position coordinates of the current planned position corresponding to the road grid, the first-order derivative of the path, the second-order derivative of the path and the path moving distance on the road grid line, a0 to a5 and b0 to b5 are fifth-order polynomial parameters. The specific solution process of the above fifth-order polynomial is a prior art and will not be repeated here. By fitting with the above 5ci polynomial, expressions of multiple candidate paths can be obtained, thereby providing multiple driving plans for vehicle path planning. The richness of the path plan can further improve the accuracy of the path planning results, which is conducive to ensuring the performance of the final screening of the driving path.

[0066] Step S105: Based on the positions of obstacles around the target vehicle, the loss values ​​of the loss functions of the candidate paths are calculated using the path expressions of the candidate paths.

[0067] Specifically, the obstacle information can be obtained through the perception module such as radar set on the vehicle to obtain the obstacle heading. Position(x obs ,y obs ) and obstacle size, etc. In addition, the vehicle's own positioning module, such as the GPS positioning system, can also be used to obtain the vehicle's current heading θ ego , the current position of the vehicle (x ego ,y eg0 ) etc. Since the planned path of the vehicle is directly related to the obstacle situation, in order to achieve obstacle-avoiding driving, the position of the obstacle has a direct impact on the evaluation of the planned path. Therefore, by using the position information of the obstacle to calculate the loss function of each candidate path, it can ensure that the evaluation result is more in line with the requirements of actual vehicle driving and further improve the performance of the final planned path.

[0068] Step S106: Screen the target planned path from each candidate path based on the loss values of the loss functions of each candidate path.

[0069] Specifically, in one embodiment, the above step S106 is specifically implemented by sorting each candidate path in ascending order of the loss value of the loss function; screening the target planned path that meets the preset requirements based on the sorting result. Among them, the smaller the loss value, the better the comprehensive performance of the candidate path. One or more candidate paths can be selected and output according to the sorting of the loss values according to the actual path planning requirements. For example, the candidate path with the smallest loss value can be selected as the optimal path for subsequent speed planning, etc., or the three candidate paths with the smallest loss values can be selected and provided to the user for reference at the same time. The specific selection method can be flexibly set according to actual needs, and the present invention is not limited thereto.

[0070] By performing the above steps, the path planning method provided by the embodiment of the present invention converts the reference path centered on the rear axle to the front axle as the reference line, constructs the road grid by defining the positional relationship between the road grid and the reference line, and uses the road grid to obtain several candidate paths by polynomial fitting. Taking the loss values of the loss functions of each path as the screening basis, the target planned path is obtained. By comprehensively constructing candidate paths for the front and rear axles, the problem that the vehicle head exceeds the lane is effectively avoided, and the entire path planning process is completed in the Cartesian coordinate system without coordinate conversion, avoiding the problem of obstacle distortion during coordinate conversion and improving the accuracy of path planning.

[0071] Specifically, in one embodiment, the above step S105 specifically includes the following steps:

[0072] Step S501: Calculate the distance cost between the current candidate path and all obstacles based on the positions of the obstacles around the target vehicle.

[0073] Exemplarily, the distance cost between the current candidate path and all obstacles can be determined by calculating the sum of the distances between each obstacle and the current candidate path, and then determining the distance cost according to the magnitude of the sum of the distances. For example, the sum of the distances is determined as the distance cost, or the distance cost is determined according to the corresponding relationship between the preset sum of the distances and the distance cost.

[0074] Step S502: Calculate the centripetal acceleration, first derivative, second derivative of the current candidate path and its deviation from the reference path based on the path expression of the current candidate path.

[0075] Specifically, after knowing the path expression of the current candidate path, the corresponding centripetal acceleration, first derivative, and second derivative values can be directly calculated according to this expression. The specific calculation process is prior art and will not be elaborated here. The shortest distance from the point on the current candidate path to the reference path is calculated by obtaining the path expression of the reference path, and this shortest distance is determined as the deviation between the current candidate path and the reference path.

[0076] Step S503: Calculate the loss value of the loss function corresponding to the current candidate path based on the distance cost, centripetal acceleration, first derivative, second derivative, and deviation.

[0077] In practical applications, to meet the requirements of different application scenarios and improve the adaptability of the target planned path, the loss value of the loss function can be calculated by setting different weight coefficients for the above parameters such as the distance cost, centripetal acceleration, first derivative, second derivative, and deviation. Exemplarily, the calculation method of the loss function cost is shown in formula (4):

[0078] J = w1J r + w2J d + w3J dd + w4J e + w5J o (4)

[0079] Where J represents the loss value of the loss function corresponding to the current candidate path, and J r 、J d 、J dd 、J e 、J o represent the centripetal acceleration, first derivative, second derivative, deviation from the reference line, and distance cost from all obstacles of the current candidate path respectively, and w1 to w5 represent the corresponding weight coefficients respectively.

[0080] The specific values of each weight coefficient can be set flexibly, and the present invention is not limited thereto.

[0081] Specifically, in one embodiment, the path planning method provided by the embodiment of the present invention further includes the following steps:

[0082] Step S107: Convert the target planned path into a third planned path centered on the rear axle.

[0083] Specifically, the target planned path can be converted into a third planned path centered on the rear axle through the following formula (5).

[0084]

[0085] Where L w represents the distance between the front and rear axles, x, y, are the position coordinates and heading angle of the target planned path, x3, y3, represent the position coordinates and heading angle of the third reference path.

[0086] Step S108: Perform safety detection on the third planned path.

[0087] Among them, the safety detection includes: collision detection and constraint detection. The purpose of collision detection is to detect whether the planned path will collide with obstacles, and the purpose of constraint detection is to detect whether the planned path will have problems such as exceeding the road boundary and the curvature being greater than the threshold, which do not meet the path planning constraint conditions.

[0088] Specifically, since the heading angle of the vehicle is the heading angle of the rear axle of the vehicle, the collision detection of obstacles is performed by using the heading angle of the third planned path calculated by the above-mentioned step S107. The specific detection processes of the specific collision detection and constraint detection are prior arts, and specific implementations can be referred to the relevant descriptions of the prior arts and will not be elaborated here.

[0089] Step S109: When the third planned path fails the safety detection, delete the target planned path from the candidate paths and return to step S106. Exemplarily, as Figure 4 shown, assuming that the planned path fails the collision detection, it means that there is a risk of collision between the planned path and obstacles. For the consideration of safe driving, it is necessary to eliminate this planned path to ensure safe driving; assuming that the planned path fails the constraint detection, such as detecting that it will exceed the road boundary, for the consideration of safe driving and meeting the corresponding road driving requirements, it is necessary to eliminate this planned path to further improve the overall performance of the final planned path and enhance the user experience.

[0090] Step S110: When the third planned paths corresponding to all candidate paths do not pass the safety test, control the target vehicle to stop.

[0091] Specifically, if all candidate paths do not meet the safety test requirements, it means that the vehicle cannot currently drive safely to the target position. Therefore, by controlling the vehicle to stop in time to avoid potential safety hazards and further ensure the safety of vehicle driving.

[0092] Step S111: When the third planned path passes the safety detection, perform speed planning for the target vehicle based on the third planned path.

[0093] Specifically, since the current speed planning is also implemented based on the path planning centered on the rear axle, in order to improve the adaptability and compatibility of the path planning, the speed planning of the vehicle is performed by using the third planned path without changing the vehicle speed planning algorithm, which improves the efficiency of the overall autonomous driving planning of the vehicle.

[0094] Thus, by converting the planned path of the rear axle to the front axle, based on the planned path of the front axle, and according to information such as obstacle information, road boundaries, and path constraints, a path planning is designed based on the front axle of the vehicle, and then the corresponding path planning result of the rear axle is obtained by conversion. This can effectively solve the problem that the front of the vehicle with a longer body exceeds the lane. Moreover, the entire path planning process directly generates path reference points in the Cartesian coordinate system, selects the optimal path, and does not need to be converted to the frenet coordinate system for path planning design. The entire path planning process is more convenient, and the path planning result is more accurate.

[0095] The embodiment of the present invention also provides a path planning device, as Figure 5 shown, the path planning device includes:

[0096] An acquisition module 101, configured to acquire a first reference path obtained by performing path planning with the rear axle of the target vehicle as the center. For detailed content, refer to the relevant description of step S101 in the above method embodiment, and details will not be elaborated here.

[0097] A first processing module 102, configured to convert the first reference path into a second reference path with the front axle as the center. For detailed content, refer to the relevant description of step S102 in the above method embodiment, and details will not be elaborated here.

[0098] A second processing module 103, configured to construct a plurality of road grid lines with the second reference path as the reference line. The longitudinal distances of the corresponding position points of each road grid line are the same at the same moment, and the connection lines of the corresponding position points at the same moment are perpendicular to the reference line. For detailed content, refer to the relevant description of step S103 in the above method embodiment, and details will not be elaborated here.

[0099] A third processing module 104, configured to use the current planned position on the second reference path as the initial state, traverse the grid positions corresponding to each road network line, and obtain the path expressions of a plurality of candidate paths through polynomial fitting. For detailed content, refer to the relevant description of step S104 in the above method embodiment, and details will not be elaborated here.

[0100] A fourth processing module 105, configured to calculate the loss values of the loss functions of each candidate path respectively based on the positions of the obstacles around the target vehicle by using the path expressions of each candidate path. For detailed content, refer to the relevant description of step S105 in the above method embodiment, and details will not be elaborated here.

[0101] A fifth processing module 106, configured to screen the target planned path from each candidate path based on the loss values of the loss functions of each candidate path. For detailed content, refer to the relevant description of step S106 in the above method embodiment, and details will not be elaborated here.

[0102] The further function descriptions of the above modules are the same as those in the corresponding method embodiments described above, and will not be elaborated here.

[0103] Through the collaborative cooperation of the above-mentioned components, the path planning device provided by the embodiment of the present invention converts the reference path centered on the rear axle to the front axle as the reference line, constructs the road grid by defining the positional relationship between the road grid and the reference line, and uses the road grid to obtain several candidate paths by polynomial fitting. Taking the loss value of the loss function of each path as the screening basis, the target planned path is obtained. By comprehensively constructing candidate paths for the front and rear axles, the problem of the vehicle head exceeding the lane is effectively avoided, and the entire path planning process is completed in the Cartesian coordinate system without coordinate conversion, avoiding the problem of obstacle distortion during coordinate conversion and improving the accuracy of path planning.

[0104] The embodiment of the present invention also provides a path planning device, which can be a domain controller mounted on the vehicle end or a cloud server. When the path planning device is a domain controller mounted on the vehicle end, the domain controller collects the current driving scenario corresponding to the vehicle, performs path planning based on the current driving scenario, and then can control the vehicle to automatically drive according to the path planning result; when the path planning device is a cloud server, the cloud server can communicate with the vehicle end, obtain the current driving scenario of the vehicle through the current driving scenario collection device set on the vehicle end, then perform path planning in the cloud, and send the planned path planning result back to the vehicle end controller, so that the vehicle end controller controls the vehicle to drive according to the path planning result.

[0105] As Figure 6 shown, the path planning device may include a processor 901 and a memory 902, where the processor 901 and the memory 902 may be connected by a bus or other means, Figure 6 taking the connection by bus as an example.

[0106] The processor 901 may be a central processing unit (CPU). The processor 901 may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. chips, or a combination of the above types of chips.

[0107] The memory 902, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of the present invention. The processor 901 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions, and modules stored in the memory 902, that is, implements the above method.

[0108] The memory 902 may include a program storage area and a data storage area. Among them, the program storage area can store an operating device and application programs required for at least one function; the data storage area can store data created by the processor 901 and the like. In addition, the memory 902 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory 902 may optionally include a memory remotely disposed relative to the processor 901, and these remote memories can be connected to the processor 901 through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0109] One or more modules are stored in the memory 902 and, when executed by the processor 901, implement the above method.

[0110] For specific details of the above path planning device, reference can be made to the corresponding relevant descriptions and effects in the above method embodiments for understanding, and details are not described herein again.

[0111] Those skilled in the art can understand that to implement all or part of the processes in the above method embodiments, it can be completed by instructing relevant hardware through a computer program. The implemented program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (abbreviation: HDD), or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0112] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: still modifications or equivalent replacements can be made to the specific embodiments of the present invention, and any modifications or equivalent replacements without departing from the spirit and scope of the present invention should be covered by the scope of the claims of the present invention.

Claims

1. A path planning method, characterized in that, Including: Obtain a first reference path obtained by path planning centered on the rear axle of the target vehicle; Convert the first reference path into a second reference path centered on the front axle; Construct multiple road grid lines with the second reference path as the reference line. The longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line; Taking the current planned position on the second reference path as the initial state, traverse the grid positions corresponding to each road network line, and obtain the path expressions of multiple candidate paths through polynomial fitting; Based on the positions of obstacles around the target vehicle, use the path expressions of each candidate path to calculate the loss values of the loss functions of each candidate path respectively, including: Based on the positions of obstacles around the target vehicle, calculate the distance cost between the current candidate path and all obstacles, where the distance cost is determined by the sum of the distances calculated by each obstacle and the current candidate path; Based on the path expression of the current candidate path, calculate the centripetal acceleration, first derivative, second derivative of the current candidate path, and its deviation from the reference path, where the deviation is determined by the shortest distance from the points on the current candidate path to the reference path; Calculate the loss value of the loss function corresponding to the current candidate path based on the distance cost, centripetal acceleration, first derivative, second derivative, and deviation; Based on the loss values of the loss functions of each candidate path, screen the target planned path from each candidate path.

2. The method according to claim 1, characterized in that, Convert the first reference path into a second reference path centered on the front axle through the following formula: Among them, x f , y f represent the front axle position coordinates on the second reference path, L w represents the distance between the front and rear axles, x r , y r represent the rear axle position coordinates on the first reference path, represents the heading angle of the first reference path.

3. The method according to claim 1, characterized in that, Based on the loss values of the loss functions of each candidate path, screen the target planned path from each candidate path, including: Sort each candidate path in ascending order of the loss value of the loss function; Based on the sorting result, screen the target planned path that meets the preset requirements.

4. The method according to claim 1, characterized in that, The method further includes: Convert the target planned path into a third planned path centered on the rear axle; Perform safety detection on the third planned path. The safety detection includes: collision detection and constraint detection; When the third planned path fails the safety detection, delete the target planned path from the candidate paths, and return to the step of screening the target planned path from each candidate path based on the loss values of the loss functions of each candidate path.

5. The method according to claim 4, characterized in that, When the third planned paths corresponding to all candidate paths do not pass the safety test, control the target vehicle to stop.

6. The method according to claim 4, characterized in that, When the third planned path passes the safety detection, perform speed planning on the target vehicle based on the third planned path.

7. A path planning device, characterized in that, Including: An acquisition module for acquiring a first reference path obtained by path planning centered on the rear axle of the target vehicle; A first processing module for converting the first reference path into a second reference path centered on the front axle; A second processing module for constructing multiple road grid lines with the second reference path as the reference line. The longitudinal spacing of the corresponding position points of each road grid line is the same at the same moment, and the connection line of the corresponding position points at the same moment is perpendicular to the reference line; A third processing module for taking the current planned position on the second reference path as the initial state, traversing the grid positions corresponding to each road network line, and obtaining the path expressions of multiple candidate paths through polynomial fitting; A fourth processing module, configured to calculate loss values of loss functions of respective candidate paths by using path expressions of the respective candidate paths based on positions of obstacles around a target vehicle, including: Calculating a distance cost between a current candidate path and all obstacles based on positions of obstacles around the target vehicle, where the distance cost is determined by a magnitude of a sum of distances calculated between each obstacle and the current candidate path; Calculating a centripetal acceleration, a first derivative, a second derivative of the current candidate path and a deviation thereof from a reference path based on the path expression of the current candidate path, where the deviation is determined by a shortest distance from a point on the current candidate path to the reference path; Calculating a loss value of a loss function corresponding to the current candidate path by using the distance cost, the centripetal acceleration, the first derivative, the second derivative and the deviation; A fifth processing module, configured to screen a target planning path from the respective candidate paths based on the loss values of the loss functions of the respective candidate paths.

8. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1-6 is implemented.

9. A path planning device, characterized in that, Including: A memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to execute the method according to any one of claims 1-6.

Images