A coverage path planning method based on 3D terrain geometric features

Abstract

The application belongs to the field of mobile robots and 3D coverage path planning, and particularly relates to a coverage path planning method based on 3D terrain geometric features. The method comprises the following steps: S1, dividing a working area according to the geometric features of a 3D terrain to generate a 3D passable map; S2, determining the starting point, growth direction and stopping growth condition of a seed path and a companion path in the 3D passable map, and autonomously growing the two types of paths; S3, judging whether the coverage rate of the paths is greater than a preset coverage rate threshold, if not, returning to step S2 to re-determine the starting point and growth direction of the seed path and autonomously grow the two types of paths, and if yes, entering step S4; S4, obtaining all path endpoints and generating the best connection of the path endpoints based on the minimum travel cost. The application improves the adaptability and working efficiency of 3D terrain full-coverage path planning, and the method is more convenient and practical to implement.

Description

Technical Field

[0001] This application belongs to the field of mobile robots and 3D coverage path planning, and specifically relates to a coverage path planning method based on 3D terrain geometric features. Background Technology

[0002] Coverage path planning (CPP) is the process by which mobile robots find the optimal continuous path to fill a given work area, ensuring complete coverage of the workspace. It has significant applications in agriculture, mining, and manufacturing. A key aspect of CPP is the assumption that the workspace lies on a plane, and most often, the path is assumed to be zigzag. Its primary goal is to find an effective continuous coverage path even when the workspace and obstacles are irregularly shaped.

[0003] Most existing research focuses on 2D map overlay, which can solve the overlay problem of irregular 2D maps. However, 2D overlay path planning is not suitable for uneven terrain; in contrast, overlay path planning on uneven terrain has not been fully solved. Most current 3D overlay path planning research inherits the zigzag path pattern of 2D models, only changing the cost function without breaking the traditional straight-line path setup. Furthermore, path planning for hazard avoidance on 3D surfaces relies on 2D terrain accessibility analysis. Because it requires planning overlay paths within individual sub-segments, it introduces additional complex computational work and leads to unreasonable path breaks at sub-segment boundaries.

[0004] Therefore, it is desirable to have a technical solution to overcome or at least mitigate one of the aforementioned defects of the prior art. Summary of the Invention

[0005] The purpose of this application is to provide a coverage path planning method based on 3D terrain geometry features to solve the problems of poor adaptability and large computational load in current 3D coverage path planning.

[0006] The technical solution of this application is:

[0007] A coverage path planning method based on 3D terrain geometry features includes:

[0008] S1. Divide the work area according to the geometric features of the 3D terrain and generate a 3D passable map;

[0009] S2. In the 3D traversable map, determine the starting point, growth direction, and stopping conditions of the seed path and the accompanying path, and carry out autonomous growth of the two types of paths.

[0010] S3. Determine whether the coverage rate of the path is greater than the preset coverage rate threshold. If not, return to step S2 to redetermine the starting point and growth direction of the seed path and perform autonomous growth of the two types of paths. If yes, proceed to step S4.

[0011] S4. Obtain all path endpoints and generate the optimal connection of path endpoints based on the minimum travel cost.

[0012] In at least one embodiment of this application, step S1, which involves dividing the work area according to the geometric features of the 3D terrain to generate a 3D traversable map, includes:

[0013] S1.1 Construct an xy plane in the working area and rasterize a 3D terrain map on the xy plane to obtain a raster map;

[0014] S1.2 Divide the grid map into multiple grid blocks, each grid block containing N×N grid cells;

[0015] S1.3 Classify each of the grid blocks to obtain a 3D passable map.

[0016] In at least one embodiment of this application, step S1.3, classifying the individual grid blocks, includes:

[0017] If the area of ​​a grid cell that is impassable is greater than the area of ​​a grid cell that is passable, then the grid cell is defined as an impassable area; otherwise, it is defined as a passable area.

[0018] If there are more sloping grids than flat grids in a passable area, then the passable area is defined as a sloping area; otherwise, it is defined as a flat area.

[0019] In at least one embodiment of this application, step S2, which involves determining the starting point, growth direction, and stopping conditions of the seed path and accompanying path in the 3D traversable map, and performing autonomous growth of the two types of paths, includes:

[0020] S2.1 In the 3D traversable map, determine the starting point, growth direction, and stopping conditions of the seed path, and carry out the autonomous growth of the seed path;

[0021] S2.2 In the 3D traversable map, determine the starting point, growth direction, and stopping conditions of the accompanying path, and allow the accompanying path to grow autonomously on both sides of the seed path.

[0022] In at least one embodiment of this application, step S2.1, which involves determining the starting point, growth direction, and stopping conditions of the seed path in the 3D traversable map, and performing autonomous growth of the seed path, includes:

[0023] Determine the starting point of the seed path:

[0024] Choose the center point of the slope region with the maximum slope as the starting point of seed path 0;

[0025] Determine the growth direction of the seed path:

[0026] Obtain the first cost function of the growth direction of the seed path, and select the direction with the minimum first cost function as the growth direction of the seed path. The first cost function is:

[0027]

[0028] Among them, J s1 (k) is the first cost function, k = 1, 2, 3, 4... are the indices of the robot's direction of travel, and Δω k Let λ be the turning angle at the current point. s1 Here, α is the weighting coefficient, α is the slope angle, and p is the path index. The maximum allowable tilt angle is set to a custom value.

[0029] Determine the conditions for stopping growth along the seed path:

[0030] When a seed path grows into the covered area, the seed path stops growing;

[0031] The coverage area is:

[0032] Ω=S t ∪S up

[0033] Where Ω represents the coverage area, and S t For the explored area, S up This is an impassable area.

[0034] In at least one embodiment of this application, when the robot travels on a flat area, the first cost function is:

[0035] J s1 (k)=λ s1 *tan(Δω k ).

[0036] In at least one embodiment of this application, step S2.2, which involves determining the starting point, growth direction, and stopping conditions of the accompanying path in the 3D traversable map, and performing autonomous growth of the accompanying path on both sides of the seed path, includes:

[0037] Determine the starting point of the accompanying path:

[0038] The starting point of the accompanying path is on the perpendicular line to the tangent point of the seed path growth direction, and the distance from the tangent point is the coverage radius of the robot's travel.

[0039] Determine the growth direction of the accompanying path:

[0040] The initial growth direction of the accompanying path is parallel to the tangent of the seed path;

[0041] Obtain the second cost function of the growth direction of the accompanying path, and select the direction with the minimum second cost function as the growth direction of the accompanying path. The second cost function is:

[0042]

[0043] Among them, J s2 (k) is the second cost function, d pl β is the coverage radius of the robot's movement. s1 Here, d(k) is the weighting coefficient, and d(k) is the distance between the next expected point of the accompanying path and the next expected point of its seed path.

[0044] Determine the growth cessation conditions of the accompanying path:

[0045] When the accompanying path grows into the coverage area, the accompanying path stops growing;

[0046] The coverage area is:

[0047] Ω=S t ∪S up

[0048] Where Ω represents the coverage area, and S t For the explored area, S up This is an impassable area.

[0049] In at least one embodiment of this application, when the robot travels on a flat area, the second cost function is:

[0050]

[0051] In at least one embodiment of this application, step S3, which involves redetermining the starting point and growth direction of the seed path and performing autonomous growth of the two types of paths, includes:

[0052] The starting point and initial growth direction of the seed path are redefined, and the autonomous growth of the seed path and its accompanying paths is carried out, specifically as follows:

[0053] A neighborhood is defined centered on the grid cell containing the starting and ending points of the seed path that stopped growing in the previous cycle. Each neighborhood is a square region comprising multiple grid cells, and the side length of each neighborhood is l. se=2d pl ;

[0054] Calculate the neighborhood of the starting point respectively and the endpoint neighborhood The middle part belongs to the area yet to be explored. The grid area will be the starting neighborhood. and the endpoint neighborhood The middle part belongs to the area yet to be explored. The neighborhood with the larger grid area is used as the starting neighborhood of the seed path for this loop, and the endpoints within the starting neighborhood are denoted as Q. n ;

[0055] The initial neighborhood belongs to the area to be explored. The grid cell is selected as the candidate grid cell containing the starting point of the seed path for this loop, with the candidate grid cell as the center and a side length of l. ss Determine candidate square regions and classify them as areas to be explored. The region with the largest grid area is taken as the starting region, and the center point Q of the starting region is set as the region to which the starting region belongs. ns As the starting point of the seed path for this iteration, d pl <l ss <2d pl ;

[0056] The initial growth direction of the seed path in this cycle is direction.

[0057] In at least one embodiment of this application, step S4, which involves obtaining all path endpoints and generating the optimal connection of the path endpoints based on the minimum travel cost, includes:

[0058] The two endpoints of a single path are defined as follows:

[0059] Constructing the travel cost equation:

[0060]

[0061]

[0062] F = min(d) Q [1,p],d Q [2,p])

[0063] Where, d Q [1, p] are the endpoints of the connecting path P. The travel cost of path P-1, d Q [2,p] are the endpoints of the connecting path P. The travel cost of path P-1 Let p be the distance between the two endpoints of path P, p be the index of path P, p ≥ 2, and F be the fitness.

[0064] Solving the travel cost equation yields the endpoint connection sequence with the minimum travel cost, i.e., the optimal connection of the path endpoints.

[0065] The invention has at least the following beneficial technical effects:

[0066] The coverage path planning method based on 3D terrain geometry features proposed in this application can provide self-growing terrain-adaptive coverage paths with high energy efficiency and short path length. It is robust to various terrain conditions, improves the adaptability and efficiency of 3D terrain full coverage path planning, and makes the method implementation process more convenient and practical. Attached Figure Description

[0067] Figure 1 This is a flowchart of a coverage path planning method based on 3D terrain geometry features according to one embodiment of this application;

[0068] Figure 2 This is a schematic diagram of a 3D passable map according to one embodiment of this application;

[0069] Figure 3 This is a schematic diagram of the autonomous growth of a path in one embodiment of this application;

[0070] Figure 4 This is a schematic diagram illustrating a novel path starting point determination method according to one embodiment of this application;

[0071] Figure 5 This is a schematic diagram of a genetic node according to one embodiment of this application;

[0072] Figure 6 This is a schematic diagram of the connection between a genetic node and all possible endpoint sequences in one embodiment of this application;

[0073] Figure 7 This is a comparison diagram of the coverage path planning based on 3D terrain geometry features and the traditional zigzag path planning results of one embodiment of this application. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0075] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application 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 the scope of protection of this application.

[0076] The following is in conjunction with the appendix Figures 1 to 7 This application will be described in further detail.

[0077] This application provides a coverage path planning method based on 3D terrain geometry features, including the following steps:

[0078] S1. Divide the work area according to the geometric features of the 3D terrain and generate a 3D passable map;

[0079] S2. In the 3D walkable map, determine the starting point, growth direction and stopping conditions of the seed path and the accompanying path, and carry out autonomous growth of the two types of paths.

[0080] S3. Determine whether the path coverage is greater than the preset coverage threshold. If not, return to step S2 to redetermine the starting point and growth direction of the seed path and perform autonomous growth of the two types of paths. If yes, proceed to step S4.

[0081] S4. Obtain all path endpoints and generate the optimal connection of path endpoints based on the minimum travel cost.

[0082] The coverage path planning method based on 3D terrain geometry features of this application first divides the working area based on the traditional 2D passable map and combines the geometric features of 3D terrain to generate a 3D passable map.

[0083] In a preferred embodiment of this application, the specific process for generating the 3D accessible map is as follows:

[0084] S1.1 Construct an xy plane in the working area and rasterize a 3D terrain map on the xy plane to obtain a raster map;

[0085] S1.2 Divide the raster map into multiple raster blocks, each raster block containing N×N raster cells, where N is greater than 1;

[0086] S1.3 Classify each grid block to obtain a 3D passable map.

[0087] In this embodiment, classifying each grid block includes: if the area of ​​impassable grid cells in a grid block is greater than the area of ​​passable grid cells, then the grid block is defined as an impassable region; otherwise, the grid block is defined as a passable region. If there are more sloping grid cells than flat grid cells in a passable region, then the passable region is defined as a ramp region; otherwise, the passable region is defined as a flat region. Based on the passability of each grid block, the grid blocks are divided into impassable and passable regions. If more than half of the area of ​​a grid block consists of impassable grid cells, then the grid block is defined as an impassable region; otherwise, it is a passable region. The passable regions are further divided into flat and ramp regions, ultimately generating a 3D passable map, such as... Figure 2 As shown.

[0088] The coverage path planning method based on 3D terrain geometry features in this application, secondly, requires the autonomous growth of two types of paths in the 3D traversable map, including:

[0089] S2.1 In the 3D walkable map, determine the starting point, growth direction and stopping conditions of the seed path, and carry out the autonomous growth of the seed path;

[0090] S2.2 In the 3D walkable map, determine the starting point, growth direction and stopping conditions of the accompanying path, and allow the accompanying path to grow autonomously on both sides of the seed path.

[0091] In a preferred embodiment of this application, the autonomous growth process of the seed path is as follows:

[0092] Determine the starting point of the seed path:

[0093] Choose the center point of the slope region with the maximum slope as the starting point of seed path 0;

[0094] Determine the growth direction of the seed path:

[0095] Obtain the first cost function of the growth direction of the seed path, and select the direction with the minimum first cost function as the growth direction of the seed path. The first cost function is:

[0096]

[0097] Among them, J s1 (k) is the first cost function, k = 1, 2, 3, 4... are the indices of the robot's direction of travel, and Δω k Let λ be the turning angle at the current point. s1 Here, α is the weighting coefficient, α is the slope angle, and p is the path index. The maximum allowable tilt angle is set to a custom value.

[0098] In this embodiment, the seed path grows from the starting point, and the growth direction of each point is determined by introducing two constraints: climbing cost and turning cost. Climbing cost is the energy the robot needs to travel along a slope in a sloping area (see the first term on the right-hand side of the first cost function). Turning cost is the energy required for the robot to turn (see the second term on the right-hand side of the first cost function). These two terms are defined as the cost function of the seed path's growth direction, and the direction with the minimum cost function is selected as the growth direction. It is understood that when the robot travels in a flat area, the influence of the terrain slope is negligible, the first term on the right-hand side of the equation becomes invalid, and the first cost function simplifies to:

[0099] J s1 (k)=λ s1 *tan(Δω k )

[0100] Determine the conditions for stopping growth along the seed path:

[0101] When a seed path grows into the covered area, the seed path stops growing;

[0102] The coverage area is:

[0103] Ω=S t ∪S up

[0104] Where Ω represents the coverage area, and S t For the explored area, S up This is an impassable area.

[0105] In a preferred embodiment of this application, the process of autonomous growth of the accompanying path is as follows:

[0106] As soon as a seed path begins to grow, two accompanying paths will grow on either side of it. The starting point of each accompanying path is determined as follows:

[0107] The starting point of the accompanying path lies on the perpendicular line to the tangent point of the seed path's growth direction, and the distance from the tangent point is the coverage radius d of the robot's travel. pl ;

[0108] Determine the growth direction of the accompanying path:

[0109] The initial growth direction of the accompanying path is parallel to the tangent of the seed path;

[0110] Obtain the second cost function of the growth direction of the accompanying path, and select the direction with the minimum second cost function as the growth direction of the accompanying path. The second cost function is:

[0111]

[0112] In this embodiment, the accompanying path growth strategy evaluates the climbing cost (first term on the right-hand side of the equation), turning cost (second term on the right-hand side of the equation), and deviation from the seed path shape (third term on the right-hand side of the equation) as cost functions, selecting the direction with the minimum cost function as the growth direction. It is understood that when the robot travels on a flat area, the influence of terrain slope can be ignored, the first term on the right-hand side of the equation can be eliminated, and the second cost function simplifies to:

[0113]

[0114] Among them, J s2 (k) is the second cost function, d pl β is the coverage radius of the robot's movement. s1 Here, d(k) is the weighting coefficient, and d(k) is the distance between the next expected point of the accompanying path and the next expected point of its seed path.

[0115] Determine the growth cessation conditions of the accompanying path:

[0116] The growth cessation condition of the accompanying path is the same as that of the seed path; when the accompanying path grows into the coverage area, the accompanying path stops growing.

[0117] The coverage area is:

[0118] Ω=S t ∪S up

[0119] Where Ω represents the coverage area, and S t For the explored area, S up This is an impassable area.

[0120] The area traversed by the path is called the explored area, denoted by S. t Impassable areas are denoted as S. upDefine Ω as the coverage area, which includes the explored area and the impassable area. The remaining area is the area to be explored, denoted as Ω. When a seed path grows into the covered area, that seed path stops growing; when an accompanying path grows into the covered area, that accompanying path stops growing.

[0121] The 3D terrain geometry-based coverage path planning method of this application, after the autonomous growth of seed paths and accompanying paths, requires further growth of the coverage path if the path coverage is not greater than a preset coverage threshold. Coverage is the percentage of the covered area that occupies the entire 3D walkable map. When the coverage area of ​​the seed path and accompanying path does not reach the set value β, a new seed path's starting point and initial growth direction need to be set, and the autonomous growth of the seed path and accompanying path is performed. Steps S2-S3 are repeated until the set value β is reached, at which point the coverage path growth is complete. Figure 4 As shown.

[0122] In a preferred embodiment of this application, the starting point and initial growth direction of the new seed path are determined through the following steps:

[0123] A neighborhood is defined centered on the grid cell containing the starting and ending points of the seed path that stopped growing in the previous loop. Each neighborhood is a square region containing multiple grid cells, and its side length is twice the robot's coverage radius. se =2d pl ;

[0124] Calculate the neighborhood of the starting point respectively and the endpoint neighborhood The middle part belongs to the area yet to be explored. The grid area will be the starting neighborhood. and the endpoint neighborhood The middle part belongs to the area yet to be explored. The neighborhood with the larger grid area is used as the starting neighborhood of the seed path for this loop, and the endpoints within this starting neighborhood are denoted as Q. n That is, the starting or ending point of the seed path that stopped growing in the previous cycle;

[0125] The initial neighborhood belongs to the area to be explored. The grid cell that serves as the starting point of the seed path for this loop is the candidate grid cell, with the candidate grid cell as the center and a side length of l. ss Identify candidate square regions and include them in the area to be explored. The region with the largest grid area is taken as the starting region, and the center point Q of the starting region is set as the starting region. ns As the starting point of the seed path for this iteration, d pl <l ss <2dpl In this embodiment, a sliding window is designed to calculate the value of a candidate square region with a side length of l. ss Within the square, the area to be explored The sliding window that occupies the largest area is the region to which the selected starting point belongs;

[0126] The initial growth direction of the seed path in this cycle is Direction. The initial growth direction of the new seed path and the tangent to the corresponding endpoint of the old seed path. They are in the same direction.

[0127] The coverage path planning method based on 3D terrain geometry features in this application finally connects discrete path endpoints based on minimum travel cost to generate the optimal connection of path endpoints.

[0128] Specifically, it includes:

[0129] Define the two endpoints of a single path as p is the index of path P, and the two endpoints of path P are represented as follows: and like Figure 5 , Figure 6 As shown;

[0130] Constructing the travel cost equation:

[0131]

[0132]

[0133] F = min(d) Q [1,p],d Q [2,p])

[0134] Where, d Q [1, p] are the endpoints of the connecting path P. The travel cost of path P-1, d Q [2,p] are the endpoints of the connecting path P. The travel cost of path P-1 Let p be the distance between the two endpoints of path P, p be the index of path P, p ≥ 2, and F be the fitness.

[0135] Solving the travel cost equation yields the sequence of endpoint connections with the minimum travel cost, i.e., the optimal connection of the path endpoints.

[0136] The coverage path planning method based on 3D terrain geometry features in this application transforms the calculation of the minimum connection cost between all discrete paths into a dynamic programming problem. The corresponding state transition equations are shown in the first and second equations of the travel cost equation. The path connection state d between the two endpoints of path P... Q [1,p],d Q [2, p] represent connections respectively. or The travel cost of path P-1 is calculated. The endpoint connection sequence with minimum travel cost is obtained by calculating the fitness F using the state transition equation.

[0137] This application presents a coverage path planning method based on 3D terrain geometry features. The path grows autonomously according to the geometry of the work area's terrain surface. The growth rule utilizes 3D terrain accessibility analysis, automatically avoiding entry into dangerous areas. Furthermore, an improved genetic algorithm connects irregularly distributed paths in an optimal sequence, ultimately completing the coverage path planning. In one embodiment of this application, as... Figure 7 As shown, the left column displays the coverage path planning results based on 3D terrain geometry features of this application, and the right column displays the traditional zigzag path planning results.

[0138] The 3D terrain geometry-based coverage path planning method proposed in this application can autonomously generate 3D coverage paths parallel to the contour, adapting to all terrain conditions. The generated coverage paths enable robots to avoid all dangerous areas. The calculation of the optimal path connection reduces the amount of computation, and for typical cases, driving energy consumption is reduced by 30%-50%, the path is shortened by 8%, and the total path cost is significantly reduced.

[0139] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A coverage path planning method based on 3D terrain geometry features, characterized in that, include: S1. Divide the work area according to the geometric features of the 3D terrain and generate a 3D passable map; The work area is divided according to the passable area and slope of the 3D terrain; S2. In the 3D walkable map, determine the starting point, growth direction, and stopping conditions of the seed path and accompanying path, and perform autonomous growth of the two types of paths; including: Step S2.1: In the 3D traversable map, determine the starting point, growth direction, and stopping conditions of the seed path, and perform autonomous growth of the seed path, including: Determine the starting point of the seed path: Choose the center point of the slope region with the maximum slope as the starting point of seed path 0; Determine the growth direction of the seed path: Obtain the first cost function of the growth direction of the seed path, and select the direction with the minimum first cost function as the growth direction of the seed path. The first cost function is: Among them, J s1 (k) is the first cost function, k = 1, 2, 3, 4... are the indices of the robot's direction of travel, and Δω k Let λ be the turning angle at the current point. s1 Here, α is the weighting coefficient, α is the slope angle, and p is the path index. The maximum allowable tilt angle is set to a custom value. Determine the conditions for stopping growth along the seed path: When a seed path grows into the covered area, the seed path stops growing; The coverage area is: Ω=S t ∪S up Where Ω represents the coverage area, and S t For the explored area, S up This area is impassable. Step S2.2: In the 3D walkable map, determine the starting point, growth direction, and stopping conditions of the accompanying path, and perform autonomous growth of the accompanying path on both sides of the seed path, including: Determine the starting point of the accompanying path: The starting point of the accompanying path is on the perpendicular line to the tangent point of the seed path growth direction, and the distance from the tangent point is the coverage radius of the robot's travel. Determine the growth direction of the accompanying path: The initial growth direction of the accompanying path is parallel to the tangent of the seed path; Obtain the second cost function of the growth direction of the accompanying path, and select the direction with the minimum second cost function as the growth direction of the accompanying path. The second cost function is: Among them, J s2 (k) is the second cost function, d pl β is the coverage radius of the robot's movement. s1 Here, d(k) is the weighting coefficient, and d(k) is the distance between the next expected point of the accompanying path and the next expected point of its seed path. Determine the growth cessation conditions of the accompanying path: When the accompanying path grows into the coverage area, the accompanying path stops growing; The coverage area is: Ω=S t ∪S up Where Ω represents the coverage area, and S t For the explored area, S up This area is impassable. S3. Determine whether the coverage rate of the path is greater than the preset coverage rate threshold. If not, return to step S2 to redetermine the starting point and growth direction of the seed path and perform autonomous growth of the two types of paths. If yes, proceed to step S4. S4. Obtain all path endpoints and generate the optimal connection of path endpoints based on the minimum travel cost.

2. The coverage path planning method based on 3D terrain geometry features according to claim 1, characterized in that, In step S1, the step of dividing the working area according to the geometric features of the 3D terrain and generating a 3D passable map includes: S1.1 Construct an xy plane in the working area and rasterize a 3D terrain map on the xy plane to obtain a raster map; S1.2 Divide the grid map into multiple grid blocks, each grid block containing N×N grid cells; S1.3 Classify each of the grid blocks to obtain a 3D passable map.

3. The coverage path planning method based on 3D terrain geometry features according to claim 2, characterized in that, In step S1.3, classifying each of the grid blocks includes: If the area of ​​a grid cell that is impassable is greater than the area of ​​a grid cell that is passable, then the grid cell is defined as an impassable area; otherwise, it is defined as a passable area. If there are more sloping grids than flat grids in a passable area, then the passable area is defined as a sloping area; otherwise, it is defined as a flat area.

4. The coverage path planning method based on 3D terrain geometry features according to claim 3, characterized in that, When the robot travels on a flat area, the first cost function is: J s1 (k)=λ s1 *tan(See k )。 5. The coverage path planning method based on 3D terrain geometry features according to claim 4, characterized in that, When the robot travels on a flat area, the second cost function is:

6. The coverage path planning method based on 3D terrain geometry features according to claim 1, characterized in that, In step S3, the process of redetermining the starting point and growth direction of the seed path and enabling autonomous growth of the two types of paths includes: The starting point and initial growth direction of the seed path are redefined, and the autonomous growth of the seed path and its accompanying paths is carried out, specifically as follows: A neighborhood is defined centered on the grid cell containing the starting and ending points of the seed path that stopped growing in the previous cycle. Each neighborhood is a square region comprising multiple grid cells, and the side length of each neighborhood is l. se =2d pl ; Calculate the neighborhood of the starting point respectively and the endpoint neighborhood The middle part belongs to the area yet to be explored. The grid area will be the starting neighborhood. and the endpoint neighborhood The middle part belongs to the area yet to be explored. The neighborhood with the larger grid area is used as the starting neighborhood of the seed path for this loop, and the endpoints within the starting neighborhood are denoted as Q. n ; The initial neighborhood belongs to the area to be explored. The grid cell is selected as the candidate grid cell containing the starting point of the seed path for this loop, with the candidate grid cell as the center and a side length of l. ss Determine candidate square regions and classify them as areas to be explored. The region with the largest grid area is taken as the starting region, and the center point Q of the starting region is set as the region to which the starting region belongs. ns As the starting point of the seed path for this iteration, d pl <l ss <2d pl ; The initial growth direction of the seed path in this cycle is direction.

7. The coverage path planning method based on 3D terrain geometry features according to claim 1, characterized in that, In step S4, obtaining all path endpoints and generating the optimal connection of the path endpoints based on the minimum travel cost includes: The two endpoints of a single path are defined as follows: Constructing the travel cost equation: F=min(d Q [1,p],d Q [2,p]) Where, d Q [1, p] are the endpoints of the connecting path P. The travel cost of path P-1, d Q [2,p] are the endpoints of the connecting path P. The travel cost of path P-1 Let p be the distance between the two endpoints of path P, p be the index of path P, p ≥ 2, and F be the fitness. Solving the travel cost equation yields the endpoint connection sequence with the minimum travel cost, i.e., the optimal connection of the path endpoints.

Images