A laser positioning polishing method
By constructing a three-dimensional environment model and using artificial intelligence algorithms to plan the operation path of the robotic polishing equipment, the problem of low efficiency in the path planning of robotic arms in the existing technology has been solved, and a more efficient and safer polishing operation has been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ANHUI YONGCHENG MACHINERY CO LTD
- Filing Date
- 2023-12-29
- Publication Date
- 2026-06-26
AI Technical Summary
Existing technologies in robotic grinding equipment fail to effectively consider path planning for robotic arms in dynamic movement states, resulting in increased computational load and low efficiency, and an inability to effectively avoid obstacles.
By collecting basic data on obstacles and the target grinding range through multi-point laser sensors, a three-dimensional environment model is constructed. Artificial intelligence algorithms are used to reverse plan the joint space trajectory, optimize the grinding path, and generate the optimal operation path in combination with the robotic arm recognition system.
It improves the safety and efficiency of robotic grinding equipment in complex environments, avoids collisions between the robotic arm and obstacles, and extends the service life of the machine.
Smart Images
Figure CN117817445B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of robotic grinding equipment, specifically a laser positioning grinding method. Background Technology
[0002] As robots play an increasingly important role in various manufacturing sectors, robotic polishing equipment is an indispensable piece of automated production equipment in industrial processes. While meeting the needs of modern enterprise production, it also needs to reduce production costs and adapt to complex production environments. Laser positioning polishing technology uses laser sensors to scan the surrounding environment, combined with the robot's recognition system, to plan its operating path and avoid obstacles.
[0003] The prior art (patent application CN109910011A) discloses a multi-sensor-based obstacle avoidance method for robotic arms. The technical solution is as follows: Information about the surrounding environment of the target object is acquired using one or more laser sensors and a vision sensor; it is determined whether there is an obstacle between the robotic arm's end effector and the target object; if an obstacle is present, its position is initially located based on image information; the actual distance information between the robotic arm's end effector and the obstacle is collected using one or more laser sensors; a grid map is constructed based on the preliminary location result and the actual distance information collected by the laser sensors; and path planning is performed using a combination of artificial potential field and rapidly expanding random tree based on the grid map, controlling the robotic arm to move closer to the target object.
[0004] The above scheme improves the accuracy of obstacle analysis and obstacle avoidance of the robotic arm in the working environment. However, the above scheme adopts the fast expanding random tree method when performing path planning, which randomly samples and generates virtual target points in the workspace. At the same time, the target selection rules for trajectory optimization do not take into account the dynamic movement of the robotic arm during operation, and do not take into account the actual situation of the robotic arm, which will lead to an increase in the amount of computation.
[0005] This invention discloses a laser positioning and polishing method to solve the above-mentioned technical problems. Summary of the Invention
[0006] This invention aims to solve at least one of the technical problems existing in the prior art; to this end, this invention proposes a laser positioning grinding method to solve the technical problem that obstacles in the working area affect the safety and efficiency of machine operation during machine operation. This invention solves the above problem by identifying the basic information of the obstacles and using artificial intelligence algorithms to reverse plan the joint space trajectory based on the grinding path of the grinding head.
[0007] To achieve the above objectives, a first aspect of the present invention provides a laser positioning and polishing method, comprising:
[0008] S1: Extract the standard working parameters of the robotic arm, and collect basic obstacle data and target grinding range in the actual working area through multi-point laser sensors; among them, the standard working parameters include the area and volume of the robotic arm grinding head, the distance between each major joint, and the actual working area of the robotic arm; the basic obstacle data includes the position of the obstacle and the obstacle point cloud data;
[0009] S2: Construct a 3D environment model of the actual working environment of the robotic arm based on obstacle basic data; simulate the grinding path based on the target workpiece basic data; and reverse plan the joint space trajectory based on the grinding path and the 3D environment model using artificial intelligence algorithms to obtain several simulated paths; among them, the target workpiece basic data includes the position, dimension and surface area of the target workpiece that need to be ground.
[0010] S3: Evaluate several simulated paths based on work efficiency and energy consumption to obtain corresponding evaluation coefficients; extract the optimal path from several simulated paths according to the evaluation coefficients; control the robotic arm to complete the grinding work according to the optimal path.
[0011] Preferably, the step of collecting basic obstacle data and target grinding range in the actual working area using multi-point laser sensors includes:
[0012] Environmental data of the actual working area is collected by multi-point laser sensors; the multi-point laser sensors include laser sensors at the joints of the robotic arm, laser sensors at the target workpiece, and laser sensors at the four corners of the ground in the actual working area.
[0013] Basic obstacle data and basic target workpiece data are extracted from environmental data, and the target grinding range is determined based on the basic target workpiece data.
[0014] Preferably, the construction of the three-dimensional environment model of the actual working environment of the robotic arm based on obstacle baseline data includes:
[0015] The basic obstacle data is preprocessed to obtain point cloud data of the obstacle from various viewpoints; the preprocessing includes data segmentation, outlier removal and boundary extraction.
[0016] A repulsive potential field is established around the obstacle based on point cloud data from various perspectives; the repulsive potential field, like the obstacle point cloud, is marked as impassable.
[0017] The repulsive potential field is transmitted to BIM software for model building, resulting in a three-dimensional environment model.
[0018] Preferably, establishing a repulsive potential field around the obstacle based on point cloud data from various viewpoints includes:
[0019] Read the point cloud data of obstacles within the actual working area and extract the three-dimensional coordinate points of the point cloud data on the obstacle surface;
[0020] Three-dimensional coordinate points within the restricted area are marked as repulsive force coordinate points, and the region formed by the repulsive force coordinate points is marked as the repulsive potential field; the restricted area is obtained by simulating the operation of a robotic arm.
[0021] Preferably, the restricted area is obtained by simulating the operation of a robotic arm, including:
[0022] The joints of the robotic arm are labeled i from near to far; where i = 1, 2, ..., n; and the calculation method from near to far is based on the distance of the joint along the robotic arm to the ground.
[0023] Construct a simulated sphere of level n with the maximum swing distance of the nth joint as the radius, and mark the non-overlapping areas of the simulated spheres of level n and level n-1 as the nth level repulsion area;
[0024] A three-dimensional sphere is constructed with the coordinates of the point cloud on the surface of the obstacle in the area to be repelled as the center and the corresponding joint radius as the radius. The area within the three-dimensional sphere is marked as the restricted area.
[0025] Preferably, the grinding path is simulated based on the basic data of the target workpiece, including:
[0026] The target workpiece is scanned according to the target grinding area to determine whether the target workpiece meets the grinding conditions. If yes, the grinding path is simulated; if no, an alarm is issued to remind the staff to correct the position of the target workpiece. Among them, meeting the grinding conditions includes that the target workpiece must be located within the grinding area and the grinding surface of the target workpiece must be directly facing the grinding head.
[0027] The basic data of the target workpiece is refined to obtain an accurate point cloud of the area to be polished. A polishing surface model corresponding to the target workpiece is then established based on the point cloud of the area to be polished. The refined processing includes point cloud filtering and point cloud registration.
[0028] Determine whether the target workpiece is two-dimensional based on its basic data;
[0029] Yes, then establish a coordinate system Z based on the two-dimensional plane. x,y The grinding area is divided into m coordinate points, and feature points are selected as the grinding start points. The grinding path is simulated based on the pre-selection mode. The feature points include the point closest to the grinding head within the grinding range of the target workpiece, the center point of the target workpiece, and the four corner points of the target workpiece. The pre-selection mode is obtained through computer simulation.
[0030] If not, first obtain the normals and curvatures of all points in the 3D point cloud data, sort the points by curvature from smallest to largest, select the point with the smallest curvature as the grinding starting point, calculate the grinding path using the section method starting from the grinding starting point, and filter the grinding path according to preset rules; among them, the point with the smallest curvature is in a relatively flat area, and starting grinding from a relatively flat area can reduce the amount of calculation when selecting the path.
[0031] Preferably, the step of establishing a coordinate system Z based on a two-dimensional plane... x,y The polishing area is divided into m coordinate points, including:
[0032] Extract the radius of the grinding head and label it as r;
[0033] Establish a coordinate system Z with the four corner points of any polishing area as the center and r as one unit. x,y ;
[0034] Mark the polished area with the corresponding coordinate points according to the horizontal and vertical coordinates.
[0035] Preferably, the step of simulating the polishing path based on the pre-selected mode includes:
[0036] Select feature points sequentially as the starting point for polishing;
[0037] Determine whether there are grinding points adjacent to the grinding starting point; if yes, grind clockwise or counterclockwise around the grinding starting point; otherwise, simulate grinding according to the coordinate point decreasing method.
[0038] Mark the simulated polishing path as the pre-selected path.
[0039] Preferably, the grinding process involving clockwise or counterclockwise rotation around the grinding starting point includes:
[0040] Identify and mark the coordinates of points adjacent to the starting point of the polishing process;
[0041] Determine the position of the grinding start point relative to the center point of the target workpiece to select the second grinding point; if the grinding start point coincides with the center point of the target workpiece, then select the nearest point in any direction (up, down, left, or right) as the second grinding point; if the grinding start point does not coincide with the center point of the target workpiece, then determine the position of the grinding start point relative to the center point of the target workpiece to select the second grinding point.
[0042] With the starting point of the polishing as the center, and r as the step size, rotate clockwise or counterclockwise along the direction of the second polishing point, keeping close to the center. Here, keeping close to the center means shielding the polished points. With the point being polished as the center, and r as the radius, scan and mark the surrounding points, and select the point closest to the starting point of the polishing line as the next polishing point.
[0043] Preferably, the step of determining the position of the grinding starting point relative to the center of the target workpiece and selecting a second grinding point includes:
[0044] Obtain the coordinates of the grinding starting point and the center of the target workpiece;
[0045] Determine the coordinates of the grinding starting point and the center of the target workpiece;
[0046] Obtain the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece.
[0047] When both the horizontal and vertical coordinates of the grinding starting point are less than the center coordinates of the target workpiece, the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the horizontal coordinates of the two points is less than the absolute value of the difference between the vertical coordinates, the nearest point to the right is selected as the second grinding point. If the absolute value of the difference between the horizontal coordinates of the two points is greater than the absolute value of the difference between the vertical coordinates, the nearest point to the top is selected as the second grinding point.
[0048] When both the horizontal and vertical coordinates of the grinding starting point are greater than the center coordinates of the target workpiece, the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the horizontal coordinates of the two points is less than the difference between the absolute values of the vertical coordinates, the nearest point to the left is selected as the second grinding point. If the absolute value of the difference between the horizontal coordinates of the two points is greater than the absolute value of the difference between the vertical coordinates, the nearest point to the bottom is selected as the second grinding point.
[0049] When the x-coordinate of the grinding starting point is less than or equal to, and the y-coordinate is greater than or equal to the center coordinate of the target workpiece, the absolute value of the difference between the x-coordinate and y-coordinate of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the x-coordinates of the two points is less than the absolute value of the difference between the y-coordinates, the nearest point to the right is selected as the second grinding point. If the absolute value of the difference between the x-coordinates of the two points is greater than the absolute value of the difference between the y-coordinates, the nearest point to the bottom is selected as the second grinding point.
[0050] When the x-coordinate of the grinding starting point is greater than or equal to the x-coordinate and the y-coordinate is less than or equal to the center coordinate of the target workpiece, the absolute value of the difference between the x-coordinate and y-coordinate of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the x-coordinates of the two points is less than the absolute value of the difference between the y-coordinates, the nearest point to the left is selected as the second grinding point. If the absolute value of the x-coordinates of the two points is greater than the absolute value of the y-coordinates, the nearest point to the top is selected as the second grinding point.
[0051] Preferably, the simulated polishing according to the coordinate point decreasing method includes:
[0052] When it is determined that there is a void defect adjacent to the starting point of the polishing process, the position of that point in the polishing area is obtained, and the point to be polished in the polishing area is identified.
[0053] Starting from the feature point, grinding proceeds in the forward direction along the x-axis. When grinding reaches the edge area, the grinding head moves to the next level and grinds in the opposite direction of the previous level. Here, the next level refers to shifting the vertical coordinate down by one unit.
[0054] When the y-axis exceeds the polishing area, a set of unpolished points is compiled, and the maximum point on the y-axis is selected as a subset. Based on this subset, the minimum point on the x-axis is selected as the second starting point for skip-point polishing. Skip-point polishing means that only unpolished points are polished, while already polished points are not polished.
[0055] Preferably, the step of filtering the polishing path according to preset rules includes:
[0056] The grinding path is optimized based on the interpolation method with equal curvature;
[0057] The optimized grinding paths are filtered by setting filtering conditions, and the filtered grinding paths are sorted from shortest to longest to obtain a path sorting table. The filtering conditions include that the grinding path must cover the grinding area and the grinding head idle path must not exceed one-tenth of the total grinding path.
[0058] Preferably, the step of reverse-planning the joint space trajectory based on the polishing path and the three-dimensional environment model using artificial intelligence algorithms includes:
[0059] Extract the polishing paths sequentially from the path sorting table;
[0060] A kinematic model of the robotic arm is established based on the number of links in the robotic arm; the kinematic model includes a forward kinematic model and an inverse kinematic model.
[0061] The joint fitting trajectory is obtained by inverse calculation based on the grinding path and Jacobi transpose method, and the joint fitting trajectory is solved.
[0062] By combining initial data with an artificial intelligence model, set rule data is obtained, and several simulated trajectories are generated based on the set rule data. The initial data includes the starting position, starting velocity, starting acceleration, ending position, ending velocity, and ending acceleration of each joint; the set rule data includes the shortest movement time, minimum spatial distance of movement, and maximum joint torque.
[0063] Preferably, the evaluation of several simulated paths based on work efficiency and energy consumption to obtain corresponding evaluation coefficients includes:
[0064] Obtain the total energy consumption during the polishing process and label it as NH;
[0065] Obtain the time required for the polishing process and mark it as T;
[0066] The evaluation coefficient px is obtained using the formula PX = α × HN × T; where α is a proportionality coefficient, obtained by fitting the planned joint space trajectory, and is a real number greater than 0. The priority of the evaluation coefficient px is as follows: the smaller the px value, the higher the priority of the corresponding simulation path.
[0067] Preferably, α is a scaling factor, obtained by fitting the planned joint space trajectory, including:
[0068] In this simulated joint trajectory, obtain the number of times each joint offset angle exceeds the limit offset angle, and record the total number as A; where the limit offset angle is 90% of the maximum offset angle of the joint; where the number of times each joint offset angle exceeds the limit offset angle is 0, the total number is recorded as 1.
[0069] In this simulated joint trajectory, the number of times each joint offset angle is lower than 20% of the maximum offset angle of the joint is obtained, and the total number is recorded as B; if the number of times each joint offset angle is lower than 20% of the maximum offset angle of the joint is 0, the total number is recorded as 1.
[0070] Through formula Obtain the proportionality coefficient α.
[0071] Compared with the prior art, the beneficial effects of the present invention are:
[0072] 1. In complex operating environments, this invention can actively identify the working environment within the working area of the robotic arm, generate an identification map using the robotic arm identification system, and plan the operation path to avoid operational accidents caused by the robotic arm accidentally touching surrounding objects.
[0073] 2. Existing robotic arms operate according to fixed programs. This invention uses lasers to identify objects and plans operating schemes adapted to the surrounding environment to avoid obstacles, thereby improving the efficiency and accuracy of the robot during polishing.
[0074] 3. This invention can generate the most suitable operating scheme for the environment, avoid useless operation of the machine, and increase the service life of the machine. Attached Figure Description
[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or 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 creative effort.
[0076] Figure 1 This is a schematic diagram of the method steps of the present invention;
[0077] Figure 2 This is a schematic diagram illustrating the establishment of a three-dimensional environment model for this invention;
[0078] Figure 3 This is a schematic diagram of obtaining the polishing path according to the present invention. Detailed Implementation
[0079] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0080] Please see Figures 1-3 The first aspect of the present invention provides a laser positioning and polishing method, comprising:
[0081] S1: Extract the standard working parameters of the robotic arm, and collect basic obstacle data and target grinding range in the actual working area through laser sensors.
[0082] The standard operating parameters in this application include the area and volume of the robotic arm's grinding head, the distance between each major joint, and the actual working range of the robotic arm; the basic obstacle data includes the location of the obstacles and obstacle point cloud data.
[0083] This application directly measures the standard operating parameters of the robotic arm using an image dimension measuring instrument (I-DIM) and stores them locally. When needed, the system directly calls upon these parameters. Specific steps include:
[0084] The maximum swing diameter of the robotic arm is obtained through actual measurement, and a square working area is established with the diameter as the side.
[0085] Position the robotic arm at the center of the actual work area;
[0086] The robotic arm and the actual working area environment are captured by multi-directional high-definition cameras; the multi-directional high-definition cameras include those located three meters above the center of the actual working area and at the four corners of the ground in the actual working area.
[0087] The acquired images are imported into the computer and processed; image processing includes noise removal, contrast enhancement, and determination of object boundaries.
[0088] Once the object's boundaries are determined, I-DIM performs geometric calculations to determine the object's actual size and stores the resulting standard operating parameters of the robotic arm locally.
[0089] The multi-point laser sensor distribution described in this application includes:
[0090] Pre-set laser sensor locations were selected at the robotic arm joints, the target workpiece, and the four corners of the actual work area. The laser sensor at the robotic arm joints was placed with its face facing the swing direction. The laser sensor at the target workpiece was placed one meter directly above the target workpiece area with its face facing down. The laser sensors at the four corners of the actual work area were placed at a 45° angle upwards.
[0091] S2: Construct a 3D environment model of the actual working environment of the robotic arm based on obstacle data; simulate the grinding path based on the target workpiece data; and reverse plan the joint space trajectory based on the grinding path and the 3D environment model using artificial intelligence algorithms to obtain several simulated paths.
[0092] The basic data of the target workpiece in this application includes the location, dimensions, and surface area of the target workpiece that need to be polished; wherein, the dimension of the target workpiece that needs to be polished is two-dimensional or three-dimensional, and this method is only discussed in terms of these two dimensions.
[0093] This application constructs a three-dimensional environment model of the actual working environment of the robotic arm based on obstacle baseline data, including:
[0094] The basic obstacle data is preprocessed to obtain point cloud data of the obstacle from various viewpoints; the preprocessing includes data segmentation, outlier removal and boundary extraction.
[0095] A repulsive potential field is established around the obstacle based on point cloud data from various perspectives; the repulsive potential field, like the obstacle point cloud, is marked as impassable.
[0096] The repulsive potential field is transmitted to BIM software for model building, resulting in a three-dimensional environment model.
[0097] It should be noted that the repulsive potential field, like the obstacle point cloud, is marked as impassable. This is so that the software can recognize the repulsive potential field as an obstacle during modeling, thus reducing the probability of errors during modeling.
[0098] It is worth noting that the establishment of a repulsive potential field around the obstacle based on point cloud data from various viewpoints is to incorporate the volume of the robotic arm joints as an influencing factor into the obstacle point cloud data during modeling, allowing for joint space trajectory planning of the robotic arm joints using points as the basis. This reduces the robotic arm modeling steps, improving modeling efficiency. Furthermore, joint space trajectory planning using points eliminates the need to consider volume factors, reducing the computational load of the simulation. The specific steps for establishing a repulsive potential field around the obstacle based on point cloud data from various viewpoints include:
[0099] Read the point cloud data of obstacles within the actual working area and extract the three-dimensional coordinate points of the point cloud data on the obstacle surface;
[0100] Three-dimensional coordinate points within the restricted area are marked as repulsive coordinate points, and the region formed by the repulsive coordinate points is marked as a repulsive potential field.
[0101] It is worth noting that the restricted area was obtained by simulating the operation of a robotic arm, and includes:
[0102] Label the joints of the robotic arm from proximal to distal as i; where i = 1, 2, ..., n;
[0103] Construct a simulated sphere of level n with the maximum swing distance of the nth joint as the radius, and mark the non-overlapping areas of the simulated spheres of level n and level n-1 as the nth level repulsion area;
[0104] A three-dimensional sphere is constructed with the coordinates of the point cloud on the surface of the obstacle in the area to be repelled as the center and the corresponding joint radius as the radius. The area within the three-dimensional sphere is marked as the restricted area.
[0105] It should be noted that the joints of the robotic arm are labeled i from near to far, and the distance between near and far is calculated as the distance from the joint along the robotic arm to the ground. Specifically, when constructing the first-level simulated sphere with the maximum swing distance of the first joint as the radius, the first-level repulsion area is the range encompassed by the first-level simulated sphere.
[0106] It is worth noting that the purpose of establishing the exclusion region in this application is to construct different 3D models for different joints, thereby reducing the amount of simulation computation. Specifically, by adding the volume of the corresponding joint to the obstacle point cloud data in different exclusion regions for joint modeling, the maximum joint volume is avoided in the fixed modeling, invalid modeling is reduced, and the selectivity during joint path simulation is increased.
[0107] This application simulates the grinding path based on the basic data of the target workpiece, including:
[0108] The target workpiece is scanned according to the target grinding range to determine whether it meets the grinding conditions. If yes, the grinding path is simulated; otherwise, an alarm is issued to remind the staff to correct the position of the target workpiece.
[0109] The grinding conditions in this application include that the target workpiece must be located within the grinding area and that the grinding surface of the target workpiece must be directly facing the grinding head. It should be noted that this application uses a laser scanner above the workpiece to scan the target workpiece to determine whether the target workpiece is located within the grinding area and whether the grinding surface is directly facing the grinding head.
[0110] After determining that the grinding conditions are met, the basic data of the target workpiece is refined to obtain an accurate point cloud of the area to be ground. Based on the point cloud of the area to be ground, a grinding surface model corresponding to the target workpiece is established. The refined processing in this application includes point cloud filtering and point cloud registration processing.
[0111] Determine whether the target workpiece is two-dimensional based on its basic data;
[0112] Yes, then establish a coordinate system Z based on the two-dimensional plane. x,y The grinding area is divided into m coordinate points, and feature points are selected as the grinding start points. The grinding path is obtained by simulation based on the pre-selected mode. In this application, the feature points include the point closest to the grinding head within the grinding range of the target workpiece, the center point of the target workpiece, and the four corner points of the target workpiece. In this application, the pre-selected mode is obtained by computer simulation.
[0113] It is worth noting that, in the selection of feature points, the four corner points of the irregular grinding surface target workpiece are the points with the maximum x-coordinate, the minimum x-coordinate, the maximum y-coordinate, and the minimum y-coordinate; if several maximum points exist simultaneously, the farthest point closest to the grinding head is selected.
[0114] For example, when selecting the point with the maximum horizontal coordinate, if two points exist simultaneously at the position of maximum horizontal coordinate, compare the distances between the two points and the nearest point of the grinding head, and select the point with the larger distance as the target point. The center point of the target workpiece is determined by connecting the four corner points pairwise, and the point with the smallest absolute value of the difference between the horizontal and vertical coordinates among the intersection points is selected as the center point of the target workpiece.
[0115] No, first obtain the normals and curvatures of all points in the 3D point cloud data, sort the points by curvature from smallest to largest, select the point with the smallest curvature as the grinding starting point, calculate the grinding path using the section method starting from the grinding starting point, and filter the grinding path using preset rules.
[0116] In this application, the point with the smallest curvature is located in a relatively flat area. Starting the polishing process from a relatively flat area can reduce the amount of computation required for path selection.
[0117] The coordinate system Z established based on a two-dimensional plane as described in this application x,y The polishing area is divided into m coordinate points, including:
[0118] Extract the radius of the grinding head and label it as r;
[0119] Establish a coordinate system Z with the four corner points of any polishing area as the center and r as one unit. x,y ;
[0120] Mark the polished area with the corresponding coordinate points according to the horizontal and vertical coordinates.
[0121] The grinding path obtained by simulating a pre-selected mode as described in this application includes:
[0122] Select feature points sequentially as the starting point for polishing;
[0123] Determine whether there are grinding points adjacent to the grinding starting point: if yes, grind clockwise or counterclockwise around the grinding starting point; if no, simulate grinding by decreasing the coordinate points step by step.
[0124] Mark the simulated polishing path as the pre-selected path.
[0125] It is worth noting that determining whether there are grinding points adjacent to the grinding starting point is to determine whether the point is inside the target workpiece or at its edge. When the grinding point is inside the target workpiece, clockwise or counterclockwise rotation grinding can prioritize grinding the central area, reducing the swing amplitude of the robotic arm; when the grinding point is at the edge of the target workpiece, the coordinate point decreasing stepwise method can grind from top to bottom, reducing useless grinding paths.
[0126] The grinding rotation around the grinding starting point as described in this application includes:
[0127] Identify and mark the coordinates of points adjacent to the starting point of the polishing process;
[0128] Determine the position of the grinding start point relative to the center point of the target workpiece to select the second grinding point; if the grinding start point coincides with the center point of the target workpiece, then select the nearest point in any direction (up, down, left, or right) as the second grinding point; if the grinding start point does not coincide with the center point of the target workpiece, then determine the position of the grinding start point relative to the center point of the target workpiece to select the second grinding point.
[0129] With the starting point of the polishing as the center, and r as the step size, rotate clockwise or counterclockwise along the direction of the second polishing point, keeping close to the center. Here, keeping close to the center means shielding the polished points. With the point being polished as the center, and r as the radius, scan and mark the surrounding points, and select the point closest to the starting point of the polishing line as the next polishing point.
[0130] The method described in this application for determining the position of the grinding starting point relative to the center of the target workpiece and selecting a second grinding point includes:
[0131] Obtain the coordinates of the grinding starting point and the center of the target workpiece;
[0132] Determine the coordinates of the grinding starting point and the center of the target workpiece;
[0133] Obtain the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece.
[0134] When both the horizontal and vertical coordinates of the grinding starting point are less than the center coordinates of the target workpiece, the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the horizontal coordinates of the two points is less than the absolute value of the difference between the vertical coordinates, the nearest point to the right is selected as the second grinding point. If the absolute value of the difference between the horizontal coordinates of the two points is greater than the absolute value of the difference between the vertical coordinates, the nearest point to the top is selected as the second grinding point.
[0135] When both the horizontal and vertical coordinates of the grinding starting point are greater than the center coordinates of the target workpiece, the absolute value of the difference between the horizontal and vertical coordinates of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the horizontal coordinates of the two points is less than the absolute value of the difference between the vertical coordinates, the nearest point to the left is selected as the second grinding point. If the absolute value of the difference between the horizontal coordinates of the two points is greater than the absolute value of the difference between the vertical coordinates, the nearest point to the bottom is selected as the second grinding point.
[0136] When the x-coordinate of the grinding starting point is less than or equal to, and the y-coordinate is greater than or equal to the center coordinate of the target workpiece, the absolute value of the difference between the x-coordinate and y-coordinate of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the x-coordinates of the two points is less than the absolute value of the difference between the y-coordinates, the nearest point to the right is selected as the second grinding point. If the absolute value of the difference between the x-coordinates of the two points is greater than the absolute value of the difference between the y-coordinates, the nearest point to the bottom is selected as the second grinding point.
[0137] When the x-coordinate of the grinding starting point is greater than or equal to the x-coordinate and the y-coordinate is less than or equal to the center coordinate of the target workpiece, the absolute value of the difference between the x-coordinate and y-coordinate of the grinding starting point and the center point of the target workpiece is determined. If the absolute value of the difference between the x-coordinates of the two points is less than the absolute value of the difference between the y-coordinates, the nearest point to the left is selected as the second grinding point. If the absolute value of the difference between the x-coordinates of the two points is greater than the absolute value of the difference between the y-coordinates, the nearest point to the top is selected as the second grinding point.
[0138] For example: if the selected grinding starting point coordinates are (60, 103), the system determines that both the horizontal and vertical coordinates of the grinding starting point are less than the center coordinates of the target workpiece (150, 160); then the absolute value of the difference between the horizontal coordinates of the grinding starting point and the center point of the target workpiece is obtained as 90, and the absolute value of the difference between the vertical coordinates is 57. If the absolute value of the difference between the horizontal coordinates is greater than the absolute value of the difference between the vertical coordinates, then the second grinding point is determined to be (60, 104).
[0139] The filtering of grinding paths through preset rules in the three-dimensional grinding surface described in this application includes:
[0140] The grinding path is optimized based on the interpolation method with equal curvature;
[0141] The optimized polishing paths are filtered by setting the filtering conditions, and the filtered polishing paths are sorted from shortest to longest to obtain a path sorting table.
[0142] The selection criteria in this application include that the grinding path must cover the grinding area and the grinding head idle path must not exceed one-tenth of the total grinding path.
[0143] It is worth noting that optimizing the grinding path is to obtain a smoother grinding path, which facilitates subsequent calculations. The specific process of optimization using the equal curvature interpolation method is as follows: For two adjacent path points P(ui) and P(ui+1), Li is a linear interpolation, that is, the line connecting the two adjacent path points, P is the radius of curvature of the corresponding path curve, and δ is the bow height error; the bow height error is judged by a judgment principle to ensure that the bow height error is less than the set value in places where the curvature changes significantly; the judgment principle is: the shorter the interpolation step size, the smaller the bow height error; the longer the interpolation step size, the larger the bow height error.
[0144] The method described in this application for reverse planning of joint space trajectories based on polishing paths and 3D environment models using artificial intelligence algorithms includes:
[0145] Extract the polishing paths sequentially from the path sorting table;
[0146] A kinematic model of the robotic arm is established based on the number of links in the robotic arm; the establishment of the kinematic model is based on existing technology (invention patent application with publication number CN103235513A);
[0147] The joint fitting trajectory is obtained by inverse calculation based on the grinding path and Jacobi transpose method, and the joint fitting trajectory is solved.
[0148] By combining initial data with an artificial intelligence model, set rule data is obtained, and several simulated trajectories are generated based on the set rule data.
[0149] It should be noted that the kinematic model established in this application is based on the invention patent application CN103235513A published on August 7, 2013, which discloses a trajectory planning and optimization method for a mobile robotic arm based on a genetic algorithm. The technical solution is as follows: First, establish the forward and inverse kinematic models of the multi-degree-of-freedom mobile robotic arm; then, use a combination of fourth-order and fifth-order polynomial mathematical models to fit the joint trajectory, and obtain the solution of the corresponding mathematical model based on its linear constraint equations; then, select the trajectory optimization target based on the principles of minimizing the movement time, minimizing the spatial distance of movement, and not exceeding the maximum set joint torque; finally, use a genetic algorithm to perform global optimization of the optimization target to obtain the optimal trajectory curve of the robotic arm's end effector.
[0150] In this application, the initial data includes the starting position, starting velocity, starting acceleration, ending position, ending velocity, and ending acceleration of each joint; the set rule data includes the shortest movement time, minimum spatial distance of movement, and maximum joint torque.
[0151] S3: Evaluate several simulated paths based on work efficiency and energy consumption to obtain corresponding evaluation coefficients; extract the optimal path from several simulated paths according to the evaluation coefficients; control the robotic arm to complete the grinding work according to the optimal path.
[0152] The evaluation of several simulated paths based on work efficiency and energy consumption described in this application, to obtain corresponding evaluation coefficients, includes:
[0153] Obtain the total energy consumption during the polishing process and mark it as 0.53; where the unit of total energy consumption HN is kilowatts;
[0154] Obtain the time required for the polishing process and mark it as 0.31; where time T is in hours;
[0155] The evaluation coefficient px is obtained using the formula PX=α×HN×T=0.61×0.53×0.31=0.1;
[0156] The scaling factor α in this application is obtained by fitting the planned joint space trajectory. Different joint space trajectories correspond to different scaling factors. The steps for obtaining the scaling factor α include:
[0157] In this simulated joint trajectory, obtain the number of times each joint offset angle exceeds the limit offset angle, and record the total number as 6; where the limit offset angle is 90% of the maximum offset angle of the joint; where the number of times each joint offset angle exceeds the limit offset angle is 0, the total number is recorded as 1.
[0158] In this simulated joint trajectory, the number of times each joint offset angle is lower than 20% of the maximum offset angle of the joint is obtained, and the total number is recorded as 16; if the number of times each joint offset angle is lower than 20% of the maximum offset angle of the joint is 0, the total number is recorded as 1.
[0159] Through formula Obtain the proportionality coefficient α.
[0160] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.
[0161] Working principle of the invention:
[0162] Extract the standard working parameters of the robotic arm, and collect basic obstacle data and target grinding range in the actual working area through multi-point laser sensors.
[0163] A three-dimensional environment model of the actual working environment of the robotic arm is constructed based on the obstacle data; the grinding path is simulated based on the target workpiece data.
[0164] Based on the polishing path and 3D environment model, the joint space trajectory is reverse-planned using artificial intelligence algorithms to obtain several simulated paths.
[0165] Several simulated paths are evaluated based on work efficiency and energy consumption to obtain corresponding evaluation coefficients; the optimal path is extracted from several simulated paths based on the evaluation coefficients; and the robotic arm is controlled to complete the grinding work based on the optimal path.
[0166] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A laser positioning and polishing method, characterized in that: Extract the standard working parameters of the robotic arm, and collect basic obstacle data and target grinding range in the actual working area through multi-point laser sensors; A three-dimensional environment model of the actual working environment of the robotic arm is constructed based on obstacle data; a grinding path is simulated based on the target workpiece data; and several simulated paths are obtained by reverse planning the joint space trajectory using artificial intelligence algorithms based on the grinding path and the three-dimensional environment model. Several simulated paths are evaluated based on work efficiency and energy consumption to obtain corresponding evaluation coefficients; the optimal path is then extracted from these simulated paths based on the evaluation coefficients. The robotic arm is controlled according to the optimal path to complete the grinding work; The construction of a three-dimensional environment model of the robotic arm's actual working environment based on obstacle data includes: The basic obstacle data is preprocessed to obtain point cloud data of the obstacle from various viewpoints; the preprocessing includes data segmentation, outlier removal and boundary extraction. A repulsive potential field is established around the obstacle based on point cloud data from various perspectives; the repulsive potential field, like the obstacle point cloud, is marked as impassable. The repulsive potential field is transmitted to BIM software for model building, resulting in a three-dimensional environment model. The establishment of a repulsive potential field around the obstacle based on point cloud data from various viewpoints includes: Read the point cloud data of obstacles within the actual working area and extract the three-dimensional coordinate points of the point cloud data on the obstacle surface; Three-dimensional coordinate points within the restricted area are marked as repulsive force coordinate points, and the region formed by the repulsive force coordinate points is marked as a repulsive potential field; the restricted area is obtained by simulating the operation of a robotic arm. The restricted area is obtained through simulated robotic arm operation and includes: The joints of the robotic arm are labeled i from near to far, where i = 1, 2, ..., n; the calculation method from near to far is based on the distance of the joint along the robotic arm to the ground; Construct a simulated sphere of level n with the maximum swing distance of the nth joint as the radius, and mark the non-overlapping areas of the simulated spheres of level n and level n-1 as the nth level repulsion area; A three-dimensional sphere is constructed with the coordinates of the point cloud on the surface of the obstacle in the area to be repelled as the center and the corresponding joint radius as the radius. The area within the three-dimensional sphere is marked as the restricted area.
2. The laser positioning and polishing method according to claim 1, characterized in that, The process of collecting basic obstacle data and target grinding range in the actual working area using multi-point laser sensors includes: Environmental data of the actual working area is collected by multi-point laser sensors; the multi-point laser sensors include laser sensors at the joints of the robotic arm, laser sensors at the target workpiece, and laser sensors at the four corners of the ground in the actual working area. Basic obstacle data and basic target workpiece data are extracted from environmental data, and the target grinding range is determined based on the basic target workpiece data.
3. The laser positioning and polishing method according to claim 1, characterized in that, The grinding path is simulated based on the target workpiece's basic data, including: The target workpiece is scanned according to the target grinding area to determine whether the target workpiece meets the grinding conditions. If yes, the grinding path is simulated; if no, an alarm is issued to remind the staff to correct the position of the target workpiece. Among them, meeting the grinding conditions includes that the target workpiece must be located within the grinding area and the grinding surface of the target workpiece must be directly facing the grinding head. The basic data of the target workpiece is refined to obtain an accurate point cloud of the area to be polished. A polishing surface model corresponding to the target workpiece is then established based on the point cloud of the area to be polished. The refined processing includes point cloud filtering and point cloud registration. Determine whether the target workpiece is two-dimensional based on its basic data; Yes, then establish a coordinate system based on the two-dimensional plane. The grinding area is divided into m coordinate points, and feature points are selected as the grinding start points. The grinding path is simulated based on the pre-selected mode. The feature points include the point closest to the grinding head within the grinding range of the target workpiece, the center point of the target workpiece, and the four corner points of the target workpiece. If not, first obtain the normals and curvatures of all points in the 3D point cloud data, sort the points by curvature from smallest to largest, select the point with the smallest curvature as the grinding starting point, calculate the grinding path using the section method starting from the grinding starting point, and filter the grinding path using preset rules; the preset rules are to optimize the grinding path based on the interpolation method of equal curvature; and filter the optimized grinding path using the set filtering conditions.
4. The laser positioning and polishing method according to claim 3, characterized in that, The grinding path obtained by simulation based on the pre-selected mode includes: Select feature points sequentially as the starting point for polishing; Determine whether there are grinding points adjacent to the grinding starting point; if yes, grind clockwise or counterclockwise around the grinding starting point; otherwise, simulate grinding according to the coordinate point decreasing method. Mark the simulated polishing path as the pre-selected path.
5. The laser positioning and polishing method according to claim 4, characterized in that, The simulated polishing process, which uses a step-by-step decreasing coordinate point method, includes: When it is determined that there is a void defect adjacent to the starting point of the polishing process, the position of that point in the polishing area is obtained, and the point to be polished in the polishing area is identified. Starting from the feature point, grinding proceeds in the forward direction along the x-axis. When grinding reaches the edge area, the grinding head moves to the next level and grinds in the opposite direction of the previous level. Here, the next level refers to shifting the vertical coordinate down by one unit. When the ordinate exceeds the grinding area, the set of ungrinded points is counted, the maximum point on the y-axis is selected as the subset, and the minimum point on the x-axis is selected as the second starting point for skip grinding based on the subset; skip grinding means that only the ungrinded points are ground, and the ground points are not ground.
6. The laser positioning and polishing method according to claim 1, characterized in that, The method of reverse-planning joint space trajectory based on polishing path and 3D environment model using artificial intelligence algorithm includes: Extract the polishing paths sequentially from the path sorting table; A kinematic model of the robotic arm is established based on the number of links in the robotic arm; The joint fitting trajectory is obtained by inverse calculation based on the grinding path and Jacobi transpose method, and the joint fitting trajectory is solved. By combining initial data with an artificial intelligence model, set rule data is obtained, and several simulated trajectories are generated based on the set rule data. The initial data includes the starting position, starting velocity, starting acceleration, ending position, ending velocity, and ending acceleration of each joint. The set rule data includes the shortest movement time, minimum spatial distance of movement, and maximum joint torque.
7. The laser positioning and polishing method according to claim 1, characterized in that, The evaluation of several simulated paths based on work efficiency and energy consumption yields corresponding evaluation coefficients, including: Obtain the total energy consumption during the polishing process and label it as follows: ; Obtain the time required for the polishing process and mark it as... ; Through formula Obtain the evaluation coefficient px; where, The scaling factor is obtained by fitting the planned joint space trajectory.