Robot full-coverage path planning method and system based on cattle tilling type movement and application of robot full-coverage path planning method and system based on cattle tilling type movement

Abstract

The invention discloses a robot full-coverage path planning method based on cattle farming movement. The method aims at improving the path coverage efficiency of a robot in a complex environment. According to the method, a two-dimensional plane map of a real environment is obtained and converted into a binary image, and a grid map suitable for the size of a current robot is generated through rasterization processing. And dynamically adjusting a path planning direction according to a main direction and regional characteristics of a map / map region by adopting a motion direction priority judgment method of cattle farming type motion. And when the robot encounters an obstacle and falls into a dead zone, replanning a path by using a backtracking point set construction rule in combination with an A * algorithm, optimizing path selection of the robot for escaping from the dead zone, and ensuring that the robot can cover all reachable areas. The method effectively reduces path overlapping and calculation amount, improves efficiency and coverage rate of path planning, and is especially suitable for environments with complex boundaries and obstacles. The invention further discloses a system and application for implementing the path planning method, and the system and the application have wide application scenes.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot full coverage path planning, and relates to a robot full coverage path planning method, system and application based on ox-plowing motion. Background Art

[0002] Full coverage path planning is a key technology in robot unmanned systems that traverse all reachable areas in the robot's work space. This technology can plan a collision-free path that covers feasible areas in the environment. It has been widely used in agriculture, cleaning, medical care, industrial manufacturing and other fields. Ox-plowing motion is a widely used path planning method in full coverage path planning algorithms. This method is suitable for regular areas such as obstacle-free rectangles, but the environmental maps in real life are usually complex planar areas with irregular boundaries and obstacles inside. Therefore, the unit decomposition method is often used to decompose irregular obstacle maps, and ox-plowing motion is used for the decomposed single area. The traveling salesman problem is used between sub-areas to complete the full coverage path planning of the work area. However, the full coverage path planning method based on unit decomposition and traveling salesman problem solution has the disadvantages of complex implementation, slow solution and large amount of calculation. In addition, the ox-plowing motion has a single priority in the direction of movement during path planning, which aggravates the problems of low path planning efficiency, too many turns and high path repetition rate. These shortcomings lead to the fact that the route obtained by full coverage path planning is not necessarily the best path for the robot. Summary of the invention

[0003] In order to address the deficiencies in the prior art, the purpose of the present invention is to provide a robot full-coverage path planning method, system and application based on ox-plowing motion, so as to solve the technical problems of complex algorithm implementation, slow solution, large amount of calculation, low planning path efficiency, excessive number of turns and high path repetition rate in the existing methods when performing full-coverage path planning in complex planar areas.

[0004] The robot full coverage path planning method based on ox-plowing motion of the present invention comprises the following steps:

[0005] Step 1: Obtain a two-dimensional plane map of the real environment and convert it into a binary image, wherein the parts with pixel values ​​of 0 and 255 in the binary image correspond to the barrier-free and barrier-affected areas in the real environment, respectively;

[0006] Step 2: Based on the binary image map, use the grid method to preprocess the map to obtain a grid map with a grid size equal to the current robot size, using black and white to represent the obstacle grid and free grid respectively;

[0007] Step 3: Take any free grid on the boundary of the grid map as the robot's initial position P0(y0,x0), initialize the global coverage model M and the regional coverage model N to record the coverage path based on the entire grid map and the coverage path based on the sub-region planning of the grid map respectively;

[0008] Step 4: Calculate the main direction of the map based on the Sobel operator or calculate the main direction of the map area by combining the map information and the backtracking point;

[0009] Step 5: Determine the priority of the movement direction of the ox-ploughing movement in the current environment area using the movement direction priority determination method of the ox-ploughing movement based on the main direction of the map or the main direction of the map area;

[0010] Step 6: The robot performs ox-plowing motion, selects adjacent uncovered grids based on the priority of the motion direction of the ox-plowing motion, adds the grids to the regional coverage model N, and updates the robot position;

[0011] Step 7: When the adjacent grids in the southeast, northwest, and northeast directions of the robot's current position are not uncovered, the robot falls into a dead zone, and the global coverage model M is updated based on the regional coverage model N to determine the current free grid coverage; when the robot does not fall into a dead zone, continue to repeat step 6;

[0012] Step 8: Output the global coverage model M and obtain the robot's full coverage path.

[0013] Furthermore, in the step 2, the map of the binary image is processed into a grid map to distinguish between areas containing obstacles and areas without obstacles and to simplify the map information, and the grid size is the current robot size. When an obstacle of the original binary image invades the divided grid, the grid area is recorded as an obstacle area, and the obstacle area and the traversable area are represented by black and white, respectively, and are represented as obstacle grids and free grids in the grid map, respectively. The coordinate origin of the grid map is located in the upper left corner of the two-dimensional plane map, the downward direction is the y-axis, the right direction is the x-axis, and the coordinate (j, i) represents the jth row and ith column grid. In the grid map, black (pixel value is 0) and white (pixel value is 255) are used to represent the obstacle area and the traversable area in the map, and the corresponding values ​​are C(j, i) = 0 and C(j, i) = 255, respectively.

[0014] Furthermore, in step 4, the calculation of the main direction of the map and the main direction of the map area is divided into two cases. When the main direction is calculated for the first time, the Sobel operator is used to calculate the main direction of the map, otherwise the map information and the backtracking point are used to calculate the main direction of the map area.

[0015] In the present invention, the backtracking point refers to the key point used to help the robot replan its path when it encounters an impassable or covered area (i.e., a "dead zone") while performing path planning for ox-plowing motion. The backtracking point records the possible starting point for the robot to return or restart after completing a partial area coverage task, ensuring that the robot can find the uncovered area and continue to perform the path coverage task.

[0016] The steps for calculating the main direction of the map area based on map information and backtracking points are:

[0017] Step 4.1: Calculate the coordinates of the four vertices corresponding to the grid map based on the map information;

[0018] Step 4.2: Based on the backtracking point to which the robot currently locates, find other backtracking points in the horizontal and vertical directions.

[0019] Step 4.3: Determine the main direction of the map area based on the number of backtracking points and coordinate information.

[0020] When the number of backtracking points is 1, 2 or 3, step 4.3 specifically includes the following steps:

[0021] Step 4.3.1: Based on the vertex coordinates obtained in step 4.1 and the backtracking point coordinate information obtained in step 4.2, construct the four vertex coordinates corresponding to the next uncovered area;

[0022] Step 4.3.2: Based on the size of the rectangle, calculate whether the main direction of the corresponding area is horizontal or vertical.

[0023] The relationship between the size of the rectangle and the main direction of the region is: the direction where the long side of the rectangle is located is the main direction of the corresponding region.

[0024] Furthermore, the step 4.3.1 is calculated using the following method:

[0025] Based on the properties of rectangles, we know that obtaining the coordinates of the three vertices of a rectangle can determine the size of the entire rectangle. Figure 4 The coordinates of the vertices are A(0,0), B(0,m), C(n,0) and D(n,m). The total number of backtracking points V in step 4.2 ranges from 1 to 3, which are discussed in the following three cases:

[0026] (1) When V=3, the coordinates of the four vertices corresponding to the area enclosed by the three backtracking points can be confirmed. The direction corresponding to the long side of the rectangular area enclosed by the three backtracking points is the main direction of the corresponding area.

[0027] (2) When V = 2, the number and coordinates of another backtracking point in the horizontal or vertical direction found based on the robot's initial position P0 (y0, x0) are P1 (y1, x1), and it is determined whether the following conditions are met:

[0028] (a) y1 = y0;

[0029] (b) x1 = x0;

[0030] (c) C(y0-1,x0+1)=255;

[0031] (d) C(y0+1,x0+1)=255;

[0032] (e) C(y0-1,x0-1)=255;

[0033] When conditions (a) and (d) are met, the coordinates of the third vertex are (n, x0); when conditions (a) and (c) are met, the coordinates of the third vertex are (0, x0); when conditions (b) and (e) are met, the coordinates of the third vertex are (y0, 0); when conditions (b) and (c) are met, the coordinates of the third vertex are (y0, m). Based on the two backtracking points and their coordinates obtained in step 4.2 and the coordinates of the third backtracking point obtained after conditional judgment, the rectangular area determined by the above three backtracking points can be obtained, and the direction corresponding to the long side of the rectangle of the rectangular area is the main direction of the corresponding area.

[0034] (3) When V = 1, the uncovered area where the robot's current position P0 is located is a single straight line path. Steps 4.3.1 and 4.3.2 are skipped, and it is determined whether the condition C(y0,x0-1)+C(y0,x0+1)=255 is satisfied. If so, the main direction of this area is horizontal, otherwise it is vertical.

[0035] Furthermore, in step 5, the robot is currently at the initial position P0 (y0, x0), and the main direction calculated in step 4 is assumed to be S, and it is determined whether the following conditions are satisfied:

[0036] (a) S>45 and S<135 or S>225 and S<315;

[0037] (b) C(y0,x0-1)=255 and C(y0,x0+1)=0;

[0038] (c) C(y0,x0+1)=255 and C(y0,x0-1)=0;

[0039] (d) C(y0-1,x0)=255 and C(y0+1,x0)=0;

[0040] (e) C(y0+1,x0)=255 and C(y0-1,x0)=0;

[0041] When the main direction S meets condition (a), there are two cases. When condition (b) or condition (c) is met, the robot starts from the initial position and uses the ox-plowing motion to determine the robot's moving direction in the order of "north-south-west-east" or "north-south-east-west". Otherwise, when the value of S is not in condition (a), but condition (d) or condition (e) is met, the robot uses the order of "east-west-north-south" or "east-west-south-north" to determine the robot's moving direction and perform ox-plowing motion.

[0042] Specifically, when (a) S>45 and S<135 or S>225 and S<315 are satisfied, and (b) C(y0,x0-1)=255 and C(y0,x0+1)=0, the robot starts the ox-plowing motion from the initial position and adopts the priority order of "north-south-west-east" to determine the robot's moving direction;

[0043] When (a) S>45 and S<135 or S>225 and S<315, and (c) C(y0,x0+1)=255 and C(y0,x0-1)=0, the robot starts the ox-plowing motion from the initial position and uses the priority order of "north-south-east-west" to determine the robot's moving direction;

[0044] When (a) S>45 and S<135 or S>225 and S<315 are not satisfied, but (d) C(y0-1,x0)=255 and C(y0+1,x0)=0 is satisfied, the robot starts the ox-plowing motion from the initial position and uses the priority order of "east-west-north-south" to determine the robot's moving direction;

[0045] When (a) S>45 and S<135 or S>225 and S<315 is not satisfied, but (e) C(y0+1,x0)=255 and C(y0-1,x0)=0 is satisfied, the robot starts ox-plowing motion from the initial position and uses the priority order of "east-west-south-north" to determine the robot's moving direction.

[0046] Further, in step 7, the dead zone refers to that the adjacent grids in the four directions of southeast, northwest, and northeast of the current position of the robot are not uncovered grids; and / or,

[0047] If the current free grid coverage is completed, output the global coverage model to obtain the robot's full coverage path; or,

[0048] If the current free grid coverage is not completed, the backtracking point set is calculated and updated based on the backtracking point set construction rule, the backtracking point closest to the robot position is selected in the backtracking point set, and the area coverage model N is reset; and / or,

[0049] Based on the dead zone position, use the A* algorithm to plan a path so that the robot reaches the selected backtracking point and update this path to the global coverage model M. Use the robot position at this time as the new initial position, calculate the main direction of the map area based on step 4, and proceed to subsequent steps.

[0050] In step 7, to determine whether the robot is trapped in the dead zone, assume that the robot position at this time is C(y,x), and any adjacent grid in the four directions of the robot is C(j,i). At this time, the following judgment is made on C(j,i):

[0051] (a) C(j,i) = 0;

[0052] (b) C(j,i)∈M;

[0053] When the robot's four adjacent grids in the southeast, northwest, and northeast directions C(y,x+1), C(y+1,x), C(y,x-1), and C(y-1,x) all meet condition (a) or (b), it means that the robot's four adjacent grids are all obstacle areas or covered areas, and the robot is now trapped in a dead zone.

[0054] Furthermore, in step seven,

[0055] Step 7.1: After the robot falls into the dead zone, use the backtracking point set to record the optional destinations for the robot to escape from the dead zone, and calculate the new backtracking points generated after covering the current area based on the backtracking point set construction rules.

[0056] Step 7.2: Update the backtracking point set: add the new backtracking point to the backtracking point set, and remove the backtracking points in the backtracking point set that do not conform to the backtracking point set construction rules.

[0057] Step 7.3: Traverse the backtracking point set and select the nearest backtracking point as the starting point of the next ox-plowing motion area based on the current robot position.

[0058] Step 7.4: Reset the area coverage model N.

[0059] Furthermore, the steps for constructing the backtracking point set rule in step 7.1 are as follows:

[0060] (a) Grid C has 8 adjacent grids denoted by C n (c) = {c1, c2, c3, c4, c5, c6, c7, c8}, where the walkable grid is denoted by c i The value is 255, otherwise it is 0. For any two grids c i ∈C n and c j ∈C n ,(i,j=1,2,…,8), establish the mathematical formula:

[0061]

[0062] k(c i ,c j ) is used to represent the traffic relationship between any two adjacent grids of grid C;

[0063] (b) Based on the four movement direction priorities of the ox-plowing movement in step 5, 16 points of different situations are constructed as backtracking points. Based on these 16 situations, the sum function is constructed as follows:

[0064] k(c i ,c j )=k(c1,c8)+k(c5,c6)+k(c1,c2)+k(c5,c4)+k(c7,c6)+k(c7,c8)+k(c5,c7)+k(c5,c3 )+k(c1,c7)+k(c1,c3)+k(c3,c5)+k(c7,c5)+k(c3,c1)+k(c7,c1)+k(c3,c4)+k(c3,c2)

[0065] When the sum function k(c i ,c j ) satisfies k(c i ,c j )≥1, grid C meets the condition of the backtracking point and is added to the backtracking point set.

[0066] Furthermore, in step 7.3, the distance measurement of the nearest backtracking point selected based on the current robot position uses the Euclidean distance, which has the advantages of small calculation amount, simple formula and easy understanding. The calculation formula is:

[0067]

[0068] The current position of the robot is P0(y0,x0), and the coordinates of a backtracking point are C0(y i ,x i ), d(P0,C0) is the distance between the two points.

[0069] Furthermore, the A* algorithm is an improved heuristic A-star algorithm, which is specifically:

[0070] The A* algorithm uses the current position as a benchmark, uses the cost function to score the surrounding points, and selects the point with the minimum score as the robot's next arrival point. Its cost function finds the optimal path by combining the actual cost (the path cost from the starting point to the current node) and the heuristic estimate (the estimated cost from the current node to the target node). The cost function is defined as:

[0071] f(n)=g(n)+h(n)

[0072] Where n is the current node under investigation, f(n) is the total cost, g(n) and h(n) are the actual path cost from the starting point to the current node and the heuristic estimated cost from the current node to the target node, respectively. The heuristic estimated cost is calculated using Euclidean distance.

[0073] The present invention also provides a robot full coverage path planning system, the path planning system comprises: a map acquisition and processing module, a rasterization preprocessing module, a path planning module, a backtracking point management module, a path optimization module, and a control and output module;

[0074] The map acquisition and processing module acquires a two-dimensional plane map of the real environment and converts it into a binary image;

[0075] The grid preprocessing module divides the map information into multiple sub-areas based on the grid map; the module can adjust the grid size according to the size of the robot and preprocess the irregular obstacle map to provide an optimized map structure for subsequent path planning;

[0076] The path planning module performs path planning using a movement direction priority determination method of ox-plowing motion according to the main direction of the map and the main direction of the area; the module determines the movement direction priority of the robot in the current environment area and controls the robot to perform ox-plowing motion, thereby planning a coverage path;

[0077] The backtracking point management module constructs and updates the backtracking point set through the backtracking point set construction rules; the module records the next uncovered area position when the robot escapes the dead zone in a complex map environment, and uses the Euclidean distance to select the nearest backtracking point as the new starting point to ensure that the robot can continue to execute the full coverage path planning;

[0078] The path optimization module plans the path through the improved heuristic A* algorithm when the robot falls into the dead zone; the module calculates the optimal path through the cost function so that the robot can efficiently reach the selected backtracking point, and updates the area coverage model to ensure that the robot continues to perform ox-plowing movement to achieve coverage of all free areas;

[0079] The control and output module controls and outputs the robot's full coverage path; this module generates the final full coverage path based on the regional coverage model, outputs the robot's movement trajectory, and can also work in conjunction with other modules to achieve effective control and scheduling of each step in the path planning process.

[0080] The present invention also discloses the application of the above-mentioned path planning method or the above-mentioned path planning system in scenarios such as farmland sowing, household cleaning, and industrial parts transportation.

[0081] The beneficial effects of the present invention include: the robot full coverage path planning method and system based on ox-plowing motion of the present invention can complete the planning of a highly efficient full coverage path in a plane area with complex regional boundaries and obstacles. Among them, the proposed method for determining the priority of the movement direction of ox-plowing motion takes the map or the main direction of the area as the primary movement direction of path planning, which solves the problem that the ox-plowing motion has a single priority of the movement direction during path planning, which leads to low efficiency of path planning, too many turns, and high path repetition rate. In addition, the backtracking point set construction rule proposed by the present invention can obtain the starting points of other uncovered areas and be applied to the calculation of the starting points of uncovered areas when dynamic full coverage path planning is performed in an unknown environment. In addition, based on the backtracking point set, the Euclidean distance is used to select the starting point of the next uncovered area and the A* algorithm is used to escape the dead zone and reach the selected area to continue the ox-plowing motion. This method ensures that the robot reaches 100% coverage in practical applications and greatly reduces the proportion of calculation and repeated paths. In addition, the above key technologies have the advantages of simple algorithm implementation, fast solution, and small calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0083] Figure 1 It is an overall flow chart of the path planning method of the present invention.

[0084] Figure 2 This is a schematic diagram of map preprocessing. Figure 2 (a) is a binary plane map. Figure 2 (b) is a grid map.

[0085] Figure 3 This is a flowchart for calculating the main direction of the map and the main direction of the map area.

[0086] Figure 4 Schematic diagram of the main direction of the region calculated based on the backtracking points.

[0087] Figure 5 A flowchart of ox-plowing movement adapted to the map environment.

[0088] Figure 6 Select the conditional diagram for the backtracking point.

[0089] Figure 7 This is a schematic diagram of the full coverage path planning steps of the present invention. Figure 7(a) indicates the first ox-plowing movement based on the main direction of the map. Figure 7 (b) represents the calculated backtracking point, Figure 7 (c) indicates using the A* algorithm to reach the nearest backtracking point. Figure 7 (d) represents the ox-plowing motion based on the main direction of the region, Figure 7 (e) shows the full coverage moving path calculated by the present invention.

[0090] Figure 8 A schematic diagram of the fully covered moving path using traditional ox-plowing locomotion. DETAILED DESCRIPTION

[0091] The present invention is further described in detail with reference to the following specific examples and drawings. The process, conditions, experimental methods, etc. for implementing the present invention, except for the contents specifically mentioned below, are all common knowledge and common common sense in the art and are not particularly limited by the present invention.

[0092] The present invention provides a robot full coverage path planning method based on ox-plowing motion, and the method aims to improve the path coverage efficiency of the robot in a complex environment. The method first obtains and converts a two-dimensional plane map of the real environment into a binary image, and then generates a grid map suitable for the current robot size through rasterization processing. In the path planning process, the motion direction priority determination method of ox-plowing motion is adopted, and the path planning direction is dynamically adjusted according to the main direction of the map / map area and the regional characteristics. When the robot encounters an obstacle and falls into a dead zone, the path is replanned by combining the backtracking point set construction rule with the A* algorithm to optimize the path selection of the robot to escape the dead zone, and ensure that the robot can cover all accessible areas. The method effectively reduces path overlap and calculation amount, improves the efficiency and coverage of path planning, and is particularly suitable for environments with complex boundaries and obstacles. The present invention also provides a system and application for implementing the above-mentioned path planning method, which has a wide range of application scenarios.

[0093] In a specific embodiment, the size of the current robot used is 23*23, the size of the two-dimensional plane map based on the real environment is 1611*1482, and the number of grids in the corresponding grid map is 70*64, that is, the y-axis direction of the grid map is composed of 64 grids, and the x-axis direction has 70 grids. Taking this as an example, the present invention performs full coverage path planning for binary images based on the real environment, and outputs the planned path coordinate values ​​and the visual map coverage.

[0094] like Figure 1 As shown, the robot full coverage path planning method based on ox-plowing motion proposed in the present invention includes the following steps:

[0095] Step 1: Obtain a two-dimensional plane map of the real environment and convert it into a binary image, wherein the parts of the binary image with pixel values ​​of 0 and 255 respectively represent the barrier-free and barrier-affected areas in the real environment corresponding to the pixel.

[0096] Step 2: Based on the binary image map, use the grid method to preprocess the map to obtain a grid map with a grid size equal to the current robot size, using black and white to represent the obstacle grid and the free grid respectively.

[0097] Specifically, the binary image map corresponding to the two-dimensional plane map of the real environment is Figure 2 As shown in (a), it is processed into a grid map to distinguish between areas with obstacles and areas without obstacles and simplify the map information. The grid size is the current robot size. The obstacle area and the traversable area are represented by black and white respectively, and are represented as obstacle grids and free grids in the grid map. Figure 2 As shown in (b), the coordinate origin of the grid map is located in the upper left corner of the two-dimensional plane map, the downward direction is the y-axis, and the right direction is the x-axis. The coordinate (j, i) represents the j-th row and i-th column grid. In the grid map, black (pixel value is 0) and white (pixel value is 255) are used to represent the obstacle area and traversable area in the map, and the corresponding values ​​are C(j, i) = 0 and C(j, i) = 255, respectively.

[0098] Step 3: Take any free grid on the boundary of the grid map as the robot's initial position P0(y0,x0), initialize the global coverage model M and the regional coverage model N to record the full coverage path based on the entire grid map and the coverage path based on the sub-region planning of the grid map, respectively.

[0099] Specifically, Figure 7 The first bold red path point shown in (a) is the grid corresponding to the robot's initial position P0 (y0, x0) in this example.

[0100] Step 4: Calculate the main direction of the map based on the Sobel operator or calculate the main direction of the map area by combining map information and backtracking points.

[0101] Specifically, the main direction calculation is divided into two cases, such as Figure 3 As shown, the Sobel operator is used to calculate the main direction of the map when the main direction is calculated for the first time, otherwise the map information and the backtracking points are used to calculate the main direction of the map area. The main direction of the map area is calculated based on the map information and the backtracking points. Based on the cases where the number of backtracking points is 1, 2, and 3, the main direction of the map area is calculated using a single backtracking point, a double backtracking point, and three backtracking points, respectively. The corresponding steps are:

[0102] Step 4.1: Calculate the coordinates of the four vertices corresponding to the grid map based on the map information;

[0103] Step 4.2: Based on the backtracking point to which the current initial position of the robot P0 (y0, x0) belongs, find other backtracking points in the horizontal and vertical directions of the backtracking point.

[0104] Step 4.3: Determine the main direction of the map area based on the number of backtracking points and coordinate information.

[0105] When the number of backtracking points is 1, 2 or 3, step 4.3 specifically includes the following steps:

[0106] Step 4.3.1: Based on the vertex coordinates obtained in step 4.1 and the backtracking point coordinate information obtained in step 4.2, the four vertex coordinates corresponding to the next uncovered area can be constructed.

[0107] Step 4.3.2: Based on the size of the rectangle, calculate whether the main direction of the corresponding area is horizontal or vertical.

[0108] The main direction of the region is consistent with the direction of the long side of the corresponding rectangle.

[0109] Furthermore, the step 4.3.1 is calculated using the following method:

[0110] Assume that the grid in step 4.1 Figure 4 The coordinates of the vertices are A(0,0), B(0,m), C(n,0) and D(n,m). The total number of backtracking points V in step 4.2 ranges from 1 to 3, and is discussed in the following three cases:

[0111] (1) When V=3, three vertices can determine a rectangle. Based on the three backtracking points, it can be confirmed that the area enclosed by the backtracking points corresponds to the coordinates of the four vertices.

[0112] (2) When V = 2, the number and coordinates of another backtracking point in the horizontal or vertical direction found based on the robot's initial position P0 (y0, x0) are P1 (y1, x1), and it is determined whether the following conditions are met:

[0113] (a) y1 = y0;

[0114] (b) x1 = x0;

[0115] (c) C(y0-1,x0+1)=255;

[0116] (d) C(y0+1,x0+1)=255;

[0117] (e) C(y0-1,x0-1)=255;

[0118] When conditions (a) and (d) are met, the coordinates of the third vertex are (n, x0); when conditions (a) and (c) are met, the coordinates of the third vertex are (0, x0); when conditions (b) and (e) are met, the coordinates of the third vertex are (y0, 0); when conditions (b) and (c) are met, the coordinates of the third vertex are (y0, m).

[0119] (3) When V = 1, the uncovered area where the robot's current position P0 is located is a single straight line path. Steps 4.3.1 and 4.3.2 are skipped, and it is determined whether the condition C(y0,x0-1)+C(y0,x0+1)=255 is satisfied. If so, the main direction of this area is horizontal, otherwise it is vertical.

[0120] Based on the properties of rectangles, we know that obtaining the coordinates of the three vertices of a rectangle can determine the size of the entire rectangle, and the main direction of the region can be calculated based on the size of the rectangle. The schematic diagram of the main direction of the region calculated based on single backtracking point, double backtracking point and three backtracking points is as follows: Figure 4 As shown, the main directions obtained by the calculation results are horizontal direction, vertical direction and vertical direction respectively; Figure 4 In the figure, the red points represent the backtracking points, the blue frame represents the rectangular area, and the green points are the ox-ploughing path grids.

[0121] Step 5: Determine the priority of the movement direction for ox-plowing movement in the current environment area using the movement direction priority determination method for ox-plowing movement based on the main direction of the map or the main direction of the map area.

[0122] Specifically, the robot is currently at the initial position P0 (y0, x0). Let the main direction calculated in step 4 be S. Determine whether it satisfies:

[0123] (a) S>45 and S<135 or S>225 and S<315;

[0124] (b) C(y0,x0-1)=255 and C(y0,x0+1)=0;

[0125] (c) C(y0,x0+1)=255 and C(y0,x0-1)=0;

[0126] (d) C(y0-1,x0)=255 and C(y0+1,x0)=0;

[0127] (e) C(y0+1,x0)=255 and C(y0-1,x0)=0;

[0128] When the main direction S meets condition (a), there are two cases. When condition (b) or condition (c) is met, the robot starts from the initial position and uses the ox-plowing motion to determine the robot's moving direction in the order of "north-south-west-east" or "north-south-east-west". Otherwise, when the value of S is not in condition (a), but condition (d) or condition (e) is met, the robot uses the order of "east-west-north-south" or "east-west-south-north" to determine the robot's moving direction and perform ox-plowing motion.

[0129] The flowchart of the ox-plowing movement adapted to the map environment is as follows Figure 5 Its input is: the main direction of the map, the robot configuration, the initial area coverage model N of the workspace, and the output is: the updated area coverage model N. The steps of ox-plowing motion are:

[0130] (a) According to the map environment information, select the priority order of "North-South-West-East" or "North-South-East-West" or "East-West-North-South" or "East-West-South-North" to find the first direction for the robot to move. If the dead zone is reached, jump out of the current loop;

[0131] (b) Move one unit step according to the currently selected direction;

[0132] (c) Create a coverage area based on the robot size (robot configuration size);

[0133] (d) adding the coverage area in step (c) to the regional coverage model N and continuing with step (a);

[0134] Step 6: The robot performs ox-plowing motion, selects adjacent uncovered grids based on the priority of the motion direction of the ox-plowing motion, and adds the grids to the regional coverage model N. Update the robot position.

[0135] Specifically, to improve the efficiency of path planning, the ox-ploughing movement is performed based on the priority of the movement direction calculated in step 5, which is divided into the ox-ploughing movement based on the main direction of the map and the main direction of the map area in step 4, and the coverage path planned based on the grid map sub-area is recorded using the regional coverage model N. Among them, the schematic diagram of the first ox-ploughing movement based on the main direction of the map is as follows Figure 7 As shown in (a), at this time, the main direction of the map calculated based on the Sobel operator is horizontal, and the path formed by the green path points is the planned path for the first path planning. The schematic diagram of the ox-plowing movement based on the main direction of the map area is as follows Figure 7 As shown in (d), at this time, based on the backtracking points and map information, it is calculated that the main direction of the blue box area is vertical, and the path planning result is the path composed of green path points in the blue box area.

[0136] Step 7: When the adjacent grids in the four directions of the robot's current position are not uncovered, the robot falls into a dead zone, and the global coverage model M is updated based on the regional coverage model N, and the process jumps to step 8. Otherwise, repeat step 6.

[0137] Specifically, the step seven includes the following specific steps:

[0138] Assume that the robot position at this time is C(y,x), and any adjacent grid in the four directions of the robot is C(j,i). At this time, the following judgment is made on C(j,i):

[0139] (a) C(j,i) = 0;

[0140] (b) C(j,i)∈M;

[0141] When the robot's four adjacent grids in the southeast, northwest, and northeast directions C(y,x+1), C(y+1,x), C(y,x-1), and C(y-1,x) all meet condition (a) or (b), it means that the robot's four adjacent grids are all obstacle areas or covered areas, and the robot is now trapped in a dead zone.

[0142] Step 8: Determine whether the current free grid has been covered. If so, jump to step 11, otherwise jump to step 9.

[0143] Step 9: Calculate and update the backtracking point set based on the backtracking point set construction rules, select the backtracking point closest to the robot position in the backtracking point set, and reset the area coverage model N.

[0144] Specifically, the backtracking point set in step nine is composed of multiple backtracking points, wherein the backtracking points have the following two characteristics:

[0145] (a) Its global coverage model M;

[0146] (b) It is the starting point of the next uncovered area;

[0147] The step nine comprises the following specific steps:

[0148] Step 9.1: After the robot falls into the dead zone, use the backtracking point set to record the optional destinations for the robot to escape from the dead zone, and calculate the new backtracking points generated after covering the current area based on the backtracking point set construction rules.

[0149] Step 9.2: Update the backtracking point set: add the new backtracking point to the backtracking point set, and remove the backtracking points in the backtracking point set that do not conform to the backtracking point set construction rules.

[0150] Step 9.3: Traverse the backtracking point set and select the nearest backtracking point as the starting point of the next ox-plowing motion area based on the current robot position.

[0151] Step 9.4: Reset the area coverage model N.

[0152] Furthermore, the steps for constructing the backtracking point set rule in step 9.1 are as follows:

[0153] (a) Grid C has 8 adjacent grids denoted by C n (c) = {c1, c2, c3, c4, c5, c6, c7, c8}, where the walkable grid is denoted by c i The value is 255, otherwise it is 0. For any two grids c i ∈C n and c j ∈C n ,(i,j=1,2,…,8), establish the mathematical formula:

[0154]

[0155] (b) Based on the four movement direction priorities of the ox-plowing movement in step 5, 16 points of different situations are constructed as backtracking points. Figure 6 As shown. For grid C, the white adjacent areas in the figure are walkable areas, the blue areas are unimportant areas that are not considered, and the black adjacent areas are areas that cannot be walked (there are obstacles or grids that have been traversed). Based on these 16 cases, the sum function is constructed as:

[0156] k(c i ,c j )=k(c1,c8)+k(c5,c6)+k(c1,c2)+k(c5,c4)+k(c7,c6)+k(c7,c8)+k(c5,c7)+k(c5,c3 )+k(c1,c7)+k(c1,c3)+k(c3,c5)+k(c7,c5)+k(c3,c1)+k(c7,c1)+k(c3,c4)+k(c3,c2)

[0157] When the sum function k(c i ,c j ) satisfies k(c i ,c j )≥1, then grid C meets the condition of the backtracking point and is added to the backtracking point set. After the first ox-plowing movement based on the main direction of the map, the backtracking point set is calculated and updated based on the backtracking point set construction rule. The calculated backtracking points are as follows: Figure 7 As shown in (b), except for the first red bold path point which is the initial position of the robot, the remaining red path points are all backtracking points that meet the backtracking point selection conditions. At this time, the backtracking point set contains a total of 21 backtracking points.

[0158] Furthermore, in order to reduce the amount of calculation and improve efficiency, the distance measurement of the nearest backtracking point selected based on the current robot position in step 9.3 uses the Euclidean distance, and the calculation formula is:

[0159]

[0160] The current position of the robot is P0(y0,x0), and the coordinates of a backtracking point are C0(y i ,x i ), d(P0,C0) is the distance between the two points. Traverse the backtracking point set and use the Euclidean distance to select the backtracking point closest to the dead zone where the robot is located as the starting point of the next ox-plowing motion area.

[0161] Step 10: Based on the dead zone position, use the A* algorithm to plan a path so that the robot reaches the selected backtracking point and update this path to the global coverage model M. Use the robot position at this time as the new initial position P0 (y0, x0), and jump to step 4.

[0162] Specifically, the A* algorithm is an improved heuristic A-star algorithm, which is specifically:

[0163] The A* algorithm uses the current position as a benchmark, uses the cost function to score the surrounding points, and selects the point with the minimum score as the robot's next arrival point. Its cost function finds the optimal path by combining the actual cost (the path cost from the starting point to the current node) and the heuristic estimate (the estimated cost from the current node to the target node). The cost function is defined as:

[0164] f(n)=g(n)+h(n)

[0165] Where n is the current node under investigation, f(n) is the total cost, g(n) and h(n) are the actual path cost from the starting point to the current node and the heuristic estimated cost from the current node to the target node, respectively. The heuristic estimated cost is calculated using the Euclidean distance in step nine.

[0166] The A* algorithm is used to plan the path to the nearest backtracking point and update this path to the global coverage model M. The nearest backtracking point at this time is the initial position P0 (y0, x0) of the robot for the next area of ​​ox-plowing motion. When the robot first falls into the dead zone, the path planned by the A* algorithm is as follows Figure 7 As shown in (c), the path in the blue box area is the path planned by the A* algorithm. The two red path points from bottom to top are the path point where the robot falls into the dead zone and the nearest backtracking point. The middle path point is the path planned using the A* algorithm to escape from the dead zone.

[0167] Step 11: Output the global coverage model M and obtain the robot's full coverage path.

[0168] Specifically, the global coverage model M is mainly composed of the regional coverage model N that is updated by ox-plowing motion and the path planned by the A* algorithm to escape the dead zone and reach the starting point of the next uncovered area. After completing the traversal of all non-obstruction grids, the full coverage movement path calculated by the present invention is obtained based on the real environment binary map, such as Figure 7 As shown in (e), the path formed by the green path points is the planned full coverage path.

[0169] When only ox-plowing motion is used for full coverage path planning, its ox-plowing motion uses a single motion direction priority of "east-west-south-north" for path planning. When the robot falls into a dead zone, it uses the traditional backtracking method to reach the starting point of the next uncovered area. When the robot falls into a dead zone, the traditional backtracking method will return to the last explorable bifurcation point along the original path and try other paths again. This method can complete the full coverage path planning task of a given map, but it has the disadvantage of being less efficient. The final path planning obtained by the traditional backtracking method is shown in the figure below. Figure 8 As shown, the path formed by the green path points is the planned full coverage path. Compared with the robot full coverage path planning method and system based on ox-plowing motion proposed in the present invention, although this method can also achieve 100% coverage, when the algorithm is evaluated using path planning length, calculation time, and number of turns, the situation is shown in Table 1.

[0170] Path planning length Calculation time Path repetition rate The algorithm of the present invention plans the path 3164 1.02 seconds 6.6％ Path planning for ox-plowing motion 5342 4.93 seconds 180％

[0171] The total number of grids in the grid map calculated based on the real environment map is 4480, of which the grids occupied by the obstacle area are 1513, and the reachable grids, i.e., free grids, are 2967. Based on the path planning length values ​​in Table 1, it can be calculated that the path repetition rates of the path planned by the algorithm of the present invention and the path planned by the ox-ploughing motion are 6.6% and 180%, respectively. In addition, the efficiency of the path planned by the algorithm of the present invention is much better than that of the path planned by the ox-ploughing motion in terms of calculation time and number of turns.

[0172] The robot full-coverage path planning method and system based on ox-plowing motion in the present invention can be used in scenarios such as farmland sowing, home cleaning, and industrial parts transportation.

[0173] In the farmland seeding scenario, the robot needs to ensure that every piece of farmland is evenly covered to complete the seeding task. The method first obtains a two-dimensional map of the farmland, converts it into a binary image, and distinguishes walkable areas from obstacle areas (such as trees and irrigation ditches). Then, the system rasterizes the map, and the grid size matches the size of the robot. The robot uses ox-plowing movement to cover the farmland in the priority order of "north-south-west-east" or "east-west-north-south" or "north-south-east-west" or "east-west-south-north". When the robot encounters an obstacle (such as a field obstacle), the system will dynamically adjust the path planning, use backtracking points and the A* algorithm to find a new path to ensure that no area is missed during the seeding work. During the seeding process, the robot can effectively avoid overlapping seeding, reduce seeding time and improve efficiency through such path planning.

[0174] In home cleaning, the robot first generates a binary image through the home map, automatically identifies obstacles such as furniture and walls, and generates a grid map based on the indoor layout. The ox-plowing movement helps the robot plan its cleaning path, giving priority to covering the ground in a straight line, and adjusting the direction of movement according to the furniture in the room (such as sofas, tables and chairs, etc.). When the robot encounters an area that cannot be covered or falls into a "dead zone" (for example, when it is surrounded by furniture), the system will plan a new path by calculating the backtracking points and the A* algorithm to ensure that every corner of the room is effectively cleaned. The entire process minimizes path overlap, ensuring that the robot completes the cleaning task efficiently and can intelligently avoid obstacles.

[0175] In the industrial parts delivery scenario, robots often need to shuttle through complex industrial workshops or warehouses to perform item transportation tasks. The system first obtains a two-dimensional plane map of the factory or warehouse, converts it into a binary image, and marks obstacles in the transportation route (such as large equipment, shelves). When the robot is transporting parts, ox-plowing motion planning will guide it along the optimal route to ensure that it covers all designated delivery areas. When encountering obstacles or falling into a "dead zone" due to complex environmental layout, the system will enable the backtracking point mechanism, re-plan the shortest path through the A* algorithm, and guide the robot to the next unfinished delivery area. This path planning method not only ensures that the robot can transport parts efficiently, but also reduces stagnation caused by obstacles and complex environments.

[0176] Through the full-coverage path planning method of ox-plowing motion, the robots in these scenarios can achieve efficient and complete tasks without omissions in different environments, significantly improving productivity and operational efficiency.

[0177] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0178] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0179] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0180] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0181] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0182] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0183] The protection content of the present invention is not limited to the above embodiments. Without departing from the spirit and scope of the present invention, changes and advantages that can be thought of by those skilled in the art are included in the present invention and are protected by the attached claims.

Claims

1. A robot full coverage path planning method based on ox-plowing motion, characterized in that: The path planning method comprises the following steps: Step 1: Obtain a two-dimensional plane map of the real environment and perform binary conversion according to the presence or absence of obstacle areas to obtain a binary image; Step 2: Divide the binary image obtained in step 1 into a grid map with a grid size equal to the size of the robot; in the grid map, black represents obstacle grids and white represents free grids; Step 3: Use any free grid on the grid map boundary as the robot's initial position to initialize the global coverage model and the regional coverage model; Step 4: Calculate the main direction of the map based on the Sobel operator or calculate the main direction of the map area by combining the map information and the backtracking point; Step 5: Determine the priority of the movement direction for ox-plowing movement in the current environment area; Step 6: Select adjacent uncovered grids based on the priority of the movement direction, include them in the area coverage model, and update the robot position; Step 7: When the robot falls into a dead zone, the global coverage model is updated based on the regional coverage model to determine the current free grid coverage completion status; when the robot does not fall into a dead zone, repeat step 6; Step 8: Output the global coverage model to obtain the robot's full coverage path.

2. The path planning method according to claim 1, characterized in that: In step 1, in the binary image, the portion with a pixel value of 0 corresponds to an obstacle-free area in the real environment, and the portion with a pixel value of 255 corresponds to an obstacle-affected area in the real environment; and / or, In step 2, the coordinate origin of the grid map is located at the upper left corner of the two-dimensional plane map, the downward direction is the y-axis, the right direction is the x-axis, and the coordinate (j, i) represents the j-th row and i-th column grid; C(j, i) = 0 represents the obstacle area in the map, and C(j, i) = 255 represents the barrier-free area in the map; and / or, In step 3, the global coverage model is used to record the full coverage path based on the entire grid map, and the regional coverage model is used to record the coverage path based on the sub-region planning of the grid map; and / or, Define the robot's initial position P0(y0,x0).

3. The path planning method according to claim 1, characterized in that: In step 4, the Sobel operator is used for calculation only when the main direction of the map is calculated for the first time; and / or, In subsequent calculations, the main direction of the map area is calculated by combining the map information and the backtracking points, including the following steps: Step 4.1, calculate the coordinates of four vertices corresponding to the grid map based on the map information; Step 4.2, based on the backtracking point to which the current initial position of the robot belongs, find other backtracking points in the horizontal and vertical directions of the backtracking point; Step 4.3: Determine the main direction of the map area based on the number of backtracking points and coordinate information.

4. The path planning method according to claim 3, characterized in that: The total number of backtracking points obtained in step 4.2 is 1, 2, or 3; and / or, When the total number of backtracking points obtained in step 4.2 is 1, 2 or 3, construct the four vertex coordinates corresponding to the next uncovered area according to the vertex coordinates obtained in step 4.1 and the backtracking point coordinates obtained in step 4.2; and / or, Based on the size of the rectangle, the main direction of the corresponding area is calculated as horizontal or vertical.

5. The path planning method according to claim 1, characterized in that: In step 5, the priority of the movement direction of the ox-ploughing movement in the current environment area is determined based on the main direction of the map or the main direction of the map area using the movement direction priority determination method of the ox-ploughing movement; the determination method of the movement direction priority is as follows: The robot is currently at the initial position P0 (y0, x0). Let the main direction calculated in step 4 be S. Determine whether it satisfies: (a) S>45 and S<135 or S>225 and S<315; (b) C(y0,x0-1)=255 and C(y0,x0+1)=0; (c) C(y0,x0+1)=255 and C(y0,x0-1)=0; (d) C(y0-1,x0)=255 and C(y0+1,x0)=0; (e) C(y0+1,x0)=255 and C(y0-1,x0)=0; When the main direction S satisfies condition (a) and condition (b) or condition (c) respectively, the robot starts from the initial position and uses the ox-plowing motion to determine the robot's moving direction in the priority order of "north-south-west-east" or "north-south-east-west"; when the main direction S value does not meet condition (a) but meets condition (d) or condition (e), the robot uses the priority order of "east-west-north-south" or "east-west-south-north" to determine the robot's moving direction and performs ox-plowing motion.

6. The path planning method according to claim 1, characterized in that: In step 7, the dead zone refers to the adjacent grids in the four directions of the southeast, northwest and northeast of the current position of the robot are not uncovered grids; and / or, If the current free grid coverage is completed, output the global coverage model to obtain the robot's full coverage path; or, If the current free grid coverage is not completed, the backtracking point set is calculated and updated based on the backtracking point set construction rule, the backtracking point closest to the robot position is selected in the backtracking point set, and the area coverage model is reset; and / or, Based on the dead zone position, use the A* algorithm to plan a path so that the robot reaches the selected backtracking point and update this path to the global coverage model. Use the robot position at this time as the new initial position, calculate the main direction of the map area based on step 4, and proceed to subsequent steps.

7. The path planning method according to claim 6, characterized in that: When the robot falls into a dead zone, the following steps are included: Step 7.1, use the backtracking point set to record the optional destinations of the robot when escaping the dead zone, and build rules based on the backtracking point set to calculate the new backtracking points generated after covering the current area; Step 7.2, adding the new backtracking point to the backtracking point set, and removing the backtracking points in the backtracking point set that do not conform to the backtracking point set construction rules; Step 7.3, traverse the backtracking point set, and select the nearest backtracking point as the starting point of the next ox-plowing motion area based on the current robot position; Step 7.4: Reset the area coverage model.

8. The path planning method according to claim 7, characterized in that: In step 7.1, the backtracking point set construction includes the following steps: Step 7.1.1: Let the eight adjacent grids of grid C be C n (c) = {c1, c2, c3, c4, c5, c6, c7, c8}, where the walkable grid is denoted by c i The value is 255, otherwise it is 0. For any two grids c i ∈C n and c j ∈C n , establish the mathematical formula: Step 7.1.2, based on the four movement direction priorities of the ox-plowing movement, 16 points of each case are constructed as backtracking points. Based on these 16 cases, the sum function is constructed as follows: k(c i ,c j )=k(c1,c8)+k(c5,c6)+k(c1,c2)+k(c5,c4)+k(c7,c6)+k(c7,c8)+k(c5,c7)+k(c5,c3)+k(c1,c7)+k(c1,c3)+k(c3,c5)+k(c7,c5)+k(c3,c1)+k(c7,c1)+k(c3,c4)+k(c3,c2), When the sum function k(c i ,c j ) satisfies k(c i ,c j )≥1, then grid C meets the backtracking point condition and is added to the backtracking point set; and / or, In step 7.3, the distance metric for selecting the nearest backtracking point based on the current robot position uses the Euclidean distance, as shown in the following formula: Among them, the current position of the robot is P0(y0,x0), and the coordinates of a backtracking point are C0(y i ,x i ), d(P0,C0) is the distance between the two points.

9. A robot full coverage path planning system, characterized in that: The path planning system includes: a map acquisition and processing module, a rasterization preprocessing module, a path planning module, a backtracking point management module, a path optimization module, and a control and output module; The map acquisition and processing module acquires a two-dimensional plane map of the real environment and converts it into a binary image; The rasterization preprocessing module divides the map information into a plurality of sub-areas based on the raster map; The path planning module performs path planning according to the main direction of the map and the main direction of the area using a movement direction priority determination method of ox-plowing movement; The backtracking point management module constructs and updates the backtracking point set according to the backtracking point set construction rules; The path optimization module plans the path through the improved heuristic A* algorithm when the robot falls into the dead zone; The control and output module controls and outputs the robot's full coverage path.

10. Application of the path planning method according to any one of claims 1 to 8, or the path planning system according to claim 9 in scenarios such as farmland sowing, household cleaning, and industrial parts transportation.

Images