A method for generating trajectory of spraying operation on special-shaped surfaces

Abstract

A method for generating a trajectory for spraying special-shaped planes that can be used for automated spraying of shoe sole glue. By obtaining the three-dimensional point cloud of the working plane and extracting the highest point on the contour of the positioning plane, the height change of the plane contour is obtained. The optimized trajectory that adapts to the undulations of the edges of different planes and matches the robot's terminal posture is further generated in combination with the robot's terminal posture. This method can generate adaptive working trajectories for different special-shaped planes, and match them with the current terminal posture of the robot to obtain a trajectory that is continuous in posture.

Description

Technical Field

[0001] The present invention relates to the field of robot control, and in particular to a method for generating adaptive operation trajectories for different working planes in applications such as robot sole spraying and glue coating. Background Art

[0002] Automation technology is being applied to more and more processing and manufacturing work. For example, shoe manufacturers may produce hundreds or thousands of shoe styles on their production lines at the same time, which requires the automated production lines that use robotic technology to have a certain degree of flexibility. Among them, the process of spraying glue on the soles is an important step in the production process. However, due to the different shapes of the soles, it is impossible to complete the spraying with a set of ready-made operation trajectories. Summary of the invention

[0003] The present invention provides a method for generating an operation trajectory that can be used in automated sole gluing and other similar special-shaped plane processing operations, which can achieve self-adaptation to the edge fluctuations of different shoe shapes and generate a more optimal trajectory.

[0004] The method for generating a trajectory for spraying a special-shaped surface provided by the present disclosure comprises the following steps:

[0005] Obtain the three-dimensional point cloud of the working plane, and construct the three-dimensional point cloud coordinate system of the working plane based on the principal component analysis method;

[0006] Perform sliding window scanning along both sides of the working plane in the first main direction of the three-dimensional point cloud coordinate system, extract the highest point in the third main direction in each sliding window, and obtain an initial trajectory;

[0007] Performing sliding window mean filtering on the initial trajectory;

[0008] Based on the current terminal posture of the robot, the posture of each 3D trajectory point in the trajectory obtained after sliding window mean filtering is constructed to obtain a trajectory containing 6 degrees of freedom posture, which is the required trajectory.

[0009] Preferably, the method for obtaining the initial trajectory comprises the following steps:

[0010] Obtain the maximum and minimum values ​​of the three-dimensional point cloud in the three main directions, and subtract the two values ​​to obtain the length, width and height l of the working plane. c , w c ,h c ;

[0011] Assume that the number of sampling points of a working trajectory on the working plane is N, and the sizes of the sampling frame in the three main directions of the three-dimensional point cloud coordinate system are 2l c / N、w c / 2,h c / 2, the sampling frame starts from the second main direction side, from the negative direction extreme point of the first main direction, with 2l c / N is the moving interval, sampling is performed towards the extreme point in the positive direction of the first main direction, and the sampling method is to extract the extreme point in the positive direction of the third main direction from the point cloud in the sampling frame, and obtain P n (n∈[0,N / 2));

[0012] Then, on the negative side of the second main direction, sample from the positive extreme point of the first main direction to the negative extreme point to obtain another point set P n (n∈[N / 2,N));

[0013] P n (n∈[0,N)) is the initial trajectory obtained by sampling.

[0014] Preferably, the method of performing sliding window mean filtering on the initial trajectory comprises the following steps:

[0015] Let the sliding window width be N f , for each point Pn in the initial trajectory, the point set in the sliding window is Find the point set {P n，i The mean of} is P′ n , that is, for P n The points obtained after filtering.

[0016] Preferably, based on the current end position of the robot, constructing the posture of each 3D trajectory point in the trajectory obtained after sliding window mean filtering specifically includes the following steps:

[0017] Suppose the trajectory point P′ obtained after filtering n (n∈[0,N)),

[0018] Let P′ n Pointing to P′ n+1 The unit vector is V n，x ;

[0019] Assume the current robot end posture is {V t，x , V t，y , V t，z}, take its z-direction component V t，z As P′ n The reference vector V of the point n，z′ ;

[0020] Take V n，y =V n，z′ ×V n，x , V n，z =V n，x ×V n，y , that is, construct P′ nPoint posture {V n，x , V n，y , V n，z}.

[0021] Preferably, the spatial pose of each point of the final trajectory is expressed as

[0022]

[0023] T′ n (n∈[0,N)) is the desired trajectory.

[0024] The method for generating a spraying trajectory for special-shaped surfaces such as shoe soles provided by the present invention obtains the three-dimensional point cloud of the working plane and extracts and locates the highest point on the contour of the working trajectory to obtain the height change of the contour, and further combines it with the terminal posture of the robot to obtain an optimized trajectory that adapts to the edge fluctuations of different special-shaped planes and matches the terminal posture of the robot.

[0025] Compared with the prior art, the beneficial effects of the present invention are: ① generating an adaptive trajectory according to the ups and downs of different working planes; ② matching the current terminal posture of the robot, thereby obtaining a trajectory that is continuous in posture. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present disclosure will become more apparent through a more detailed description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present disclosure.

[0027] Figure 1 Shows an exemplary robot sole gluing work scene and related coordinate system;

[0028] Figure 2 Displaying a three-dimensional point cloud data set of a shoe sole according to an exemplary embodiment;

[0029] Figure 3 Displaying three orthogonal main directions of a three-dimensional point cloud of a shoe sole according to an exemplary embodiment;

[0030] Figure 4 Display the process of sliding window scanning along both sides of the sole and the initial trajectory point set obtained;

[0031] Figure 5 Schematic diagram representing the unit vectors between adjacent trajectory points;

[0032] Figure 6 Indicates the introduction of the robot end posture at each trajectory point;

[0033] Figure 7 Represents the relationship between the attitude components at the trajectory points. DETAILED DESCRIPTION

[0034] The preferred embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.

[0035] Figure 1 An exemplary operation scenario of a robot applying glue to shoe soles and the related coordinate system are shown.

[0036] According to the method for generating a trajectory for spraying a special-shaped plane described in the present disclosure, the trajectory generation for gluing a shoe sole includes the following steps:

[0037] 1. Use a depth camera to shoot the sole and obtain a 3D point cloud data set of the sole {P c}, as attached Figure 2 As shown, other feasible means can also be used to obtain the three-dimensional point cloud of the sole.

[0038] 2. Apply principal component analysis to the sole point cloud to obtain three orthogonal principal directions {V1, V2, V3}, as shown in the attached figure. Figure 3 As shown, all point clouds are averaged to obtain the center point of the point cloud (x c ,y c , z c ), the orthogonal basis formed by the main direction and the point cloud center are combined to form the pose expression of the sole point cloud

[0039]

[0040] 3. Obtain the maximum and minimum values ​​of the sole point cloud in the three main directions, and subtract the two values ​​to obtain the length, width and height l of the sole c , w c ,h c ;

[0041] Assuming that the number of sampling points of the entire sole trajectory is N, we can get the number of half-edge points to be N / 2, and let the size of the sampling frame in the three main directions be 2l respectively. c / N、w c / 2,h c / 2, start from the extreme point along the negative direction of the first main direction on the second main direction side, with 2l c / N is the moving interval, sampling towards the extreme point in the positive direction, as shown in the attached Figure 4 As shown, the sampling method is to extract the extreme point of the point cloud in the sampling frame in the positive direction of the third main direction, and obtain P n (n∈[0,N / 2));

[0042] Then, on the negative side of the second main direction, sample from the positive extreme point of the first main direction to the negative extreme point to obtain another point set P n (n∈[N / 2,N));

[0043] Then P n (n∈[0,N)) is the initial trajectory sampled.

[0044] 4. n}The trajectory points are filtered by sliding window mean, and the sliding window width is N f , for each point P n Take the point set in the sliding window as Find the point set {P n，i The mean of} is P′ n , which is P n Point after filtering.

[0045] 5. For trajectory point P′ n (n∈[0,N)), let P′ n Pointing to P′ n+1 The unit vector is V n，x , as attached Figure 5 shown.

[0046] 6. In order to make the gluing trajectory match the current end position of the robot as much as possible, so as to obtain a trajectory that is continuous in posture, the end position of the robot needs to be referred to during the trajectory generation process. This posture is recorded as {V t，x , V t，y , V t，z}. Take the z-component V of the robot's end posture t，z As P′ n The reference vector V of the point n，z′ , as attached Figure 6 shown.

[0047] 7. Let V n，y =V n，z′ ×V n，x , V n，z =V n，x ×V n，y , thus constructing P′ based on the reference vector n Point posture {V n，x , V n，y , V n，z}, as attached Figure 7 shown.

[0048] In summary, we get P′ n The corresponding spatial pose expression T′ n :

[0049]

[0050] Then T′ n (n∈[0,N)) is the obtained trajectory result.

[0051] It can be seen that according to the method for generating a glue coating trajectory of the exemplary embodiment, by obtaining the three-dimensional point cloud of the sole and extracting and locating the highest point on the sole contour, the height change of the sole contour is obtained, and further combined with the robot terminal posture, an optimized trajectory is generated that adapts to the edge fluctuations of different shoe types and matches the robot terminal posture. Compared with the prior art, the beneficial effects of the present disclosure are: ① generating adaptive trajectories for different shoe types; ② matching the current terminal posture of the robot, thereby obtaining a trajectory that is continuous in posture.

[0052] The trajectory generation method disclosed in the present invention can also be applied to robot processing operations similar to spraying glue on shoe soles.

[0053] The above technical scheme is only an exemplary embodiment of the present invention. For those skilled in the art, it is easy to make various types of improvements or modifications based on the application methods and principles disclosed in the present invention, and it is not limited to the method described in the above specific embodiment of the present invention. Therefore, the method described above is only preferred and does not have a restrictive meaning.

Claims

1. A method for generating a trajectory of a special-shaped plane spraying operation, comprising the following steps: S1, obtain the three-dimensional point cloud of the working plane, and construct the three-dimensional point cloud coordinate system based on the principal component analysis method; S2, performing sliding window scanning along both sides of the working plane in the first main direction of the three-dimensional point cloud coordinate system, extracting the highest point in the third main direction in each sliding window, and obtaining an initial trajectory; S3, performing sliding window mean filtering on the initial trajectory; S4, based on the current terminal posture of the robot, construct the posture of each 3D trajectory point in the trajectory obtained after sliding window mean filtering, and obtain a trajectory containing 6 degrees of freedom posture, which is the desired trajectory; in: In step S2, the method for obtaining the initial trajectory includes the following steps: Obtain the maximum and minimum values ​​of the three-dimensional point cloud in the three main directions, and subtract the two values ​​to obtain the length, width and height of the working plane. , , ; Assume that the number of sampling points of a working trajectory on the working plane is N, and the sizes of the sampling frame in the three main directions of the three-dimensional point cloud coordinate system are , , , the sampling frame starts from the second main direction side, from the negative direction extreme point of the first main direction, and The moving interval is sampled towards the extreme point in the positive direction of the first main direction. The sampling method is to extract the extreme point in the positive direction of the third main direction from the point cloud in the sampling frame, and obtain ; Then, on the negative side of the second main direction, sample from the positive extreme point of the first main direction to the negative extreme point to obtain another point set ; That is, the initial trajectory obtained by sampling; In step S4, based on the current end position of the robot, the posture at each 3D trajectory point in the trajectory obtained after sliding window mean filtering is constructed to obtain a trajectory containing a 6-DOF posture, specifically including: Assume that the trajectory point obtained after filtering is , make point to The unit vector is ; Assume the current robot end posture is { }, take its z-direction component As The reference vector of the point ; Pick , , that is, construct Point gesture ; The spatial pose of each point of the final trajectory is expressed as This is the desired trajectory.

2. According to the trajectory generation method of claim 1, the method of performing sliding window mean filtering on the initial trajectory comprises the following steps: Let the sliding window width be , for each point in the initial trajectory Take the point set in the sliding window as , find the point set The mean value of , that is, The points obtained after filtering.

Images