Automatic cleaning method and system based on plasma and trajectory control

Abstract

This invention provides an automated cleaning method and system based on plasma and trajectory control, which relates to the field of control technology. The method involves acquiring a three-dimensional workpiece model, extracting depth information, and dividing the workpiece surface into planar, convex, and concave regions based on principal curvature values. Differentiated trajectory density parameters are set for each region, along with trajectory transition zones. The robot's posture angle is calculated using normal vectors, ultimately generating trajectory data containing posture information and plasma cleaning process parameters. This invention achieves precise cleaning of complex surfaces, improving cleaning efficiency and quality.

Description

Technical Field

[0001] The present invention relates to the field of control technology, and in particular to an automatic cleaning method and system based on plasma and trajectory control. Background Art

[0002] With the advancement of industrial automation, robot-assisted cleaning technology is gradually replacing traditional manual cleaning methods, improving cleaning efficiency and quality consistency. Plasma cleaning, as an environmentally friendly and efficient cleaning method, is widely used in aerospace, precision manufacturing, and other fields, effectively removing contaminants from workpiece surfaces without damaging the substrate.

[0003] Traditional robotic cleaning systems typically employ simple trajectory planning methods, such as equidistant parallel scanning or preset trajectory templates. However, when working with complex workpiece geometries, these methods struggle to adapt to varying surface curvatures, resulting in uneven cleaning results. Furthermore, most existing systems fail to intelligently partition workpiece surfaces based on their geometric characteristics. Instead, they employ uniform cleaning parameters, failing to differentiate specific areas based on their characteristics, thus impacting cleaning effectiveness. Summary of the Invention

[0004] The embodiments of the present invention provide an automatic cleaning method and system based on plasma and trajectory control, which can solve the problems in the prior art.

[0005] A first aspect of an embodiment of the present invention provides an automatic cleaning method based on plasma and trajectory control, comprising:

[0006] Acquire a three-dimensional digital model of the workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data;

[0007] Determining a principal curvature value based on gradient information corresponding to the depth information, taking a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and dividing the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold;

[0008] Based on the regional division results of the workpiece surface to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the robot's initial cleaning trajectory, where the trajectory density parameter of the planar area is set as a baseline value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas;

[0009] Based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated. The plasma cleaning process parameters corresponding to each trajectory point are set according to the material properties of the workpiece to be cleaned.

[0010] Acquiring a three-dimensional digital model of a workpiece to be cleaned, meshing the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extracting depth information of the surface of the workpiece to be cleaned based on the mesh data includes:

[0011] Calculating the local curvature value of each point on the surface of the three-dimensional digital model, constructing a grid size function based on the local curvature value, wherein the grid size function adopts an exponential decay form, the grid size decreases as the curvature value increases, and the grid division is denser at the location of the curvature mutation; adaptively meshing the workpiece to be cleaned based on the grid size function to obtain grid division data of the surface of the workpiece to be cleaned;

[0012] Establishing a local coordinate system for the surface of the workpiece to be cleaned, projecting the vertex coordinates in the meshing data to the local coordinate system, and calculating the depth value of each vertex;

[0013] Constructing a depth feature map of the surface of the workpiece to be cleaned based on the depth value of each vertex, calculating the gradient distribution of the depth feature map, and determining a depth mutation area according to the gradient distribution;

[0014] The principal curvature information of the surface of the workpiece to be cleaned is extracted, and a shape index is calculated based on the principal curvature information. The shape index is clustered according to the numerical value to obtain shape clustering areas, and a transition zone is established between the shape clustering areas, wherein the width of the transition zone is proportional to the difference in shape indices of adjacent areas.

[0015] Determining a principal curvature value based on gradient information corresponding to the depth information, taking a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and dividing the surface of the workpiece to be cleaned into a plane area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold value includes:

[0016] Calculating an average curvature and a Gaussian curvature of the surface of the workpiece to be cleaned based on the first-order gradient and the second-order gradient corresponding to the depth information, and determining a principal curvature value based on the average curvature and the Gaussian curvature;

[0017] Continuously collecting curvature data of the surface of the workpiece to be cleaned within a preset sampling period, establishing a time-series curvature data set, and using the difference between the curvature data of the time-series curvature data set and the main curvature value as the curvature segmentation degree;

[0018] The surface of the workpiece to be cleaned is divided into regions according to the curvature segmentation degree: when the curvature segmentation degree is less than a first preset threshold, the region where the current curvature data is located is regarded as a plane region, and the current curvature value is updated by using a neighborhood average value based on a distance weight; when the curvature segmentation degree is between the first preset threshold and a second preset threshold, the region where the current curvature data is located is regarded as a convex region, and the current curvature value is updated by using a weighted average value of exponentially decaying time series data; when the curvature segmentation degree is greater than the second preset threshold, the region where the current curvature data is located is regarded as a concave region, and the current curvature value is maintained;

[0019] Adjacent curvature gradients between adjacent regions are calculated, width distribution of transition regions is determined according to the adjacent curvature gradients, and a smooth transition boundary is constructed in the transition region using a B-spline curve.

[0020] According to the area division result of the surface of the workpiece to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the initial cleaning trajectory of the robot, wherein the trajectory density parameter of the flat area is set as the base value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value.

[0021] Collect curvature data of each area of the surface of the workpiece to be cleaned, calculate the average curvature value of all sampling points in each area, calculate the curvature change between adjacent sampling points, and calculate the dispersion degree of the average curvature value to obtain the regional curvature dispersion value;

[0022] Determine a reference track density value of the planar area according to a preset minimum track spacing, and perform a weighted combination of the reference track density value with the average curvature value and the curvature variation, wherein the weight of the average curvature value is proportional to the shape complexity value, and the weight of the curvature variation is proportional to the curvature dispersion value;

[0023] Calculate the trajectory density parameters of each area: for the convex area, divide the maximum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the convex trajectory density parameter; for the concave area, divide the minimum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the concave trajectory density parameter;

[0024] Generate the robot's initial clear trajectory according to the trajectory density parameters of each area.

[0025] Based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated, including:

[0026] Based on the discrete point position coordinate sequence of the initial cleaning trajectory of the robot, calculating the position difference vectors between adjacent trajectory points in the position coordinate sequence, and normalizing the position difference vectors to obtain a trajectory tangent vector sequence;

[0027] Extract the corresponding surface normal vector at each trajectory point position, use the surface normal vector as the Z-axis direction vector of the robot tool coordinate system, and use the trajectory tangent vector sequence at the corresponding position as the Y-axis direction vector of the tool coordinate system; obtain the X-axis direction vector of the tool coordinate system by cross-product operation of the Z-axis direction vector and the Y-axis direction vector, construct a posture rotation matrix based on the X-axis direction vector, the Y-axis direction vector and the Z-axis direction vector, and decompose the posture rotation matrix to obtain the Euler angle of the tool coordinate system relative to the base coordinate system;

[0028] Substitute the position coordinate sequence and posture rotation matrix sequence into the robot kinematic equation, solve the joint angle solution set at each trajectory point, and calculate the joint motion amount and joint median value deviation of each group of solutions in the joint angle solution set.

[0029] The plasma cleaning process parameters corresponding to each trajectory point are set according to the material characteristics of the workpiece to be cleaned, including:

[0030] The material properties of the workpiece to be cleaned include local hardness value and local roughness value;

[0031] Calculating the plasma cleaning power based on the local hardness value and the local roughness value: multiplying the local hardness value by a first weighting coefficient to obtain a first power component, multiplying the local roughness value by a second weighting coefficient to obtain a second power component, and adding the first power component and the second power component to obtain a plasma cleaning power value corresponding to the trajectory point;

[0032] Obtaining a local curvature value at each trajectory point, taking a weighted sum of the local curvature value and the local hardness value as a process parameter modulation coefficient, and multiplying the process parameter modulation coefficient by a preset reference gas flow rate to obtain an actual gas flow rate value corresponding to the trajectory point;

[0033] The angle between the Z-axis direction of the tool coordinate system and the normal vector of the surface of the workpiece to be cleaned is calculated, and the actual working distance of the plasma jet is determined based on the angle. The actual working distance, plasma cleaning power value and actual gas flow value are converted into ion concentration values corresponding to each trajectory point.

[0034] The method further comprises:

[0035] When the difference in plasma cleaning power between adjacent trajectory points is greater than a preset power threshold, a transition point is inserted between the adjacent trajectory points; when the difference in actual gas flow between adjacent trajectory points is greater than a preset flow threshold, linear interpolation is used for transition;

[0036] The plasma cleaning power value, actual gas flow value and ion concentration value of each trajectory point are combined into a process parameter group, and a mapping relationship between the process parameter group and the trajectory point position is established.

[0037] A second aspect of an embodiment of the present invention provides an automatic cleaning system based on plasma and trajectory control, comprising:

[0038] The first unit is configured to obtain a three-dimensional digital model of a workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of a surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data;

[0039] A second unit is configured to determine a principal curvature value based on gradient information corresponding to the depth information, use a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and divide the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold;

[0040] The third unit is configured to set differentiated trajectory density parameters for different types of areas based on the area division result of the workpiece surface to be cleaned, and generate an initial cleaning trajectory for the robot, wherein the trajectory density parameter of the planar area is set as a reference value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; and at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas;

[0041] The fourth unit is used to calculate the posture angle of the robot at each trajectory point based on the initial cleaning trajectory of the robot and the normal vector of each area, generate robot motion trajectory data containing posture information, and set the plasma cleaning process parameters corresponding to each trajectory point according to the material properties of the workpiece to be cleaned.

[0042] According to a third aspect of an embodiment of the present invention, an electronic device is provided, including:

[0043] processor;

[0044] a memory for storing processor-executable instructions;

[0045] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0046] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described above is implemented.

[0047] The beneficial effects of this application are as follows:

[0048] By analyzing the curvature and dividing the workpiece surface into zones, this invention implements differentiated cleaning strategies tailored to different surface features, effectively improving cleaning efficiency and quality. Especially for complex, uneven surfaces, the dynamic adjustment of the track density parameter ensures uniform and thorough cleaning.

[0049] The present invention adopts a track transition zone design to solve the problems of track breakage and discontinuous cleaning between adjacent areas, making the cleaning process smoother and avoiding the generation of cleaning dead corners and missed areas. At the same time, it reduces the number of acceleration and deceleration times during robot movement and extends the service life of the equipment.

[0050] The present invention customizes the plasma cleaning process parameters in combination with the material characteristics of the workpiece, thereby achieving precise control of the cleaning process, ensuring the cleaning effect, and avoiding damage to the workpiece surface caused by excessive cleaning, thereby improving product quality and production efficiency, and reducing production costs and energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Schematic diagram of the process of an automatic cleaning method based on plasma and trajectory control according to an embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of the process of dividing the workpiece surface area based on the curvature segmentation degree of the present invention. DETAILED DESCRIPTION

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0054] The following specific embodiments are used to describe the technical solution of the present invention in detail. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0055] Figure 1 FIG. 1 is a flow chart of an automatic cleaning method based on plasma and trajectory control according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0056] Acquire a three-dimensional digital model of the workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data;

[0057] Determining a principal curvature value based on gradient information corresponding to the depth information, taking a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and dividing the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold;

[0058] Based on the regional division results of the workpiece surface to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the robot's initial cleaning trajectory, where the trajectory density parameter of the planar area is set as a baseline value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas;

[0059] Based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated. The plasma cleaning process parameters corresponding to each trajectory point are set according to the material properties of the workpiece to be cleaned.

[0060] In an optional embodiment, obtaining a three-dimensional digital model of a workpiece to be cleaned, meshing the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extracting depth information of the surface of the workpiece to be cleaned based on the mesh data includes:

[0061] Calculating the local curvature value of each point on the surface of the three-dimensional digital model, constructing a grid size function based on the local curvature value, wherein the grid size function adopts an exponential decay form, the grid size decreases as the curvature value increases, and the grid division is denser at the location of the curvature mutation; adaptively meshing the workpiece to be cleaned based on the grid size function to obtain grid division data of the surface of the workpiece to be cleaned;

[0062] Establishing a local coordinate system for the surface of the workpiece to be cleaned, projecting the vertex coordinates in the meshing data to the local coordinate system, and calculating the depth value of each vertex;

[0063] Constructing a depth feature map of the surface of the workpiece to be cleaned based on the depth value of each vertex, calculating the gradient distribution of the depth feature map, and determining a depth mutation area according to the gradient distribution;

[0064] The principal curvature information of the surface of the workpiece to be cleaned is extracted, and a shape index is calculated based on the principal curvature information. The shape index is clustered according to the numerical value to obtain shape clustering areas, and a transition zone is established between the shape clustering areas, wherein the width of the transition zone is proportional to the difference in shape indices of adjacent areas.

[0065] The present invention provides a method for obtaining a three-dimensional digital model of a workpiece to be cleaned and extracting surface depth information. The method first obtains a three-dimensional digital model of the workpiece to be cleaned, for example by scanning the workpiece with a 3D scanner to obtain point cloud data, or by modeling the workpiece using CAD software. After obtaining the three-dimensional digital model, the model is meshed and surface depth information is extracted.

[0066] When meshing a 3D digital model surface, the local curvature value of each point on the model surface is first calculated. Specifically, for each point on the 3D model surface, the mean curvature and Gaussian curvature at that point are calculated by fitting a local surface around that point. For example, for point P on the surface, a set of points within a 3mm radius around it is selected, and a quadratic surface is fitted to these points to determine the local curvature value at point P.

[0067] Based on the calculated curvature values, a grid size function is constructed. This function uses an exponential decay formula, reducing the grid size as the curvature value increases. Specifically, for a point with a curvature value of K, the corresponding grid size S is calculated as S = Smin + (Smax - Smin) × exp(-αK), where Smin is the minimum grid size (e.g., 0.5mm), Smax is the maximum grid size (e.g., 5mm), and α is the adjustment coefficient (e.g., 10). At locations where curvature changes suddenly, such as edges and corners, the grid is refined by reducing the grid size in that area. For example, if the curvature change rate between two adjacent points exceeds 30%, the grid size in that area is reduced to 1 / 3 of its original size.

[0068] The adaptive meshing process begins with a coarse initial mesh and refines it based on the mesh size function. For example, the mesh size for the entire workpiece surface is initially set to 4 mm. The curvature value of each mesh cell center is then iteratively calculated, and the mesh size function is used to determine whether the cell needs further refinement. After multiple iterations (e.g., five), the mesh data for the workpiece surface is obtained, including vertex coordinates and connectivity relationships.

[0069] To extract depth information from the workpiece surface, a local coordinate system must be established for the workpiece surface. For each region on the workpiece surface, the center point of the region is selected as the origin, the surface normal at that point is used as the positive z-axis direction, and the two principal tangent vectors are used as the x- and y-axis directions. In practical applications, the workpiece surface can be divided into several regions, for example, each with an area of approximately 25 square centimeters, and a local coordinate system is established for each region.

[0070] Specifically, for each vertex in the mesh, the local region to which it belongs is first determined. Then, its coordinates are transformed from the global coordinate system to the corresponding local coordinate system. The z coordinate of the transformed vertex is the depth value of that point. For example, for a vertex with global coordinates of (10mm, 15mm, 5mm), if the origin of its local coordinate system is (8mm, 12mm, 3mm) and the coordinate axis rotation matrix is R, then the coordinates of the vertex in the local coordinate system are R × (2mm, 3mm, 2mm)^T, and its depth value is the transformed z coordinate.

[0071] A depth feature map is a two-dimensional image whose grayscale values correspond to the depth values of the vertices. To construct the depth feature map, the xy plane of the local area is divided into a grid (e.g., a 100×100 grid), with each grid cell corresponding to a pixel in the depth feature map. For each grid cell, the average depth value of all vertices falling within the cell is calculated as the grayscale value of the corresponding pixel. If a grid cell does not contain a vertex, the depth value of that cell is interpolated from the surrounding cells with values.

[0072] Calculate the gradient distribution of the constructed depth feature map. The finite difference method is used to calculate the gradient of the depth feature map. For each pixel (i, j) in the depth feature map, calculate its x-direction gradient Gx(i, j) = D(i+1, j) - D(i-1, j) and y-direction gradient Gy(i, j) = D(i, j+1) - D(i, j-1), where D(i, j) represents the depth value at pixel (i, j). The gradient magnitude G(i, j) = sqrt(Gx(i, j)^2 + Gy(i, j)^2) reflects the severity of the depth change. A gradient threshold is set (e.g., if the depth change rate exceeds 20%), and regions with a gradient magnitude greater than the threshold are marked as depth mutation regions.

[0073] Extract the principal curvatures of the workpiece surface, including calculating the maximum and minimum principal curvatures at each point on the surface. Based on this principal curvature information, the shape index (SI) is calculated. The shape index is defined as SI = (2 / π) × arctan((κ1+κ2) / (κ1-κ2)), where κ1 and κ2 are the maximum and minimum principal curvatures, respectively (κ1≥κ2). The shape index ranges from [-1 to 1], with different values corresponding to different surface shape types.

[0074] The calculated shape index is clustered by numerical value. For example, the shape index can be divided into five intervals: [-1, -0.6), [-0.6, -0.2), [-0.2, 0.2), [0.2, 0.6), and [0.6, 1], corresponding to concave, groove, saddle, convex, and convex regions, respectively. Each point on the workpiece surface is assigned to a corresponding region based on its shape index, thereby obtaining shape cluster regions.

[0075] A transition zone is established between shape clusters, with its width proportional to the difference in shape index between adjacent regions. For example, if the shape index center values of two adjacent regions are 0.8 and 0.4, respectively, and the difference is 0.4, the transition zone width is set to 8mm; if the difference is 0.2, the transition zone width is set to 4mm. Points within the transition zone linearly blend the characteristics of the two regions in proportion to their distance from the boundary, ensuring continuity across the entire surface.

[0076] In an optional embodiment, a principal curvature value is determined based on the gradient information corresponding to the depth information, a difference between the principal curvature value and the current curvature data is used as a curvature segmentation degree, and the surface of the workpiece to be cleaned is divided into a plane area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold value, including:

[0077] Calculating an average curvature and a Gaussian curvature of the surface of the workpiece to be cleaned based on the first-order gradient and the second-order gradient corresponding to the depth information, and determining a principal curvature value based on the average curvature and the Gaussian curvature;

[0078] Continuously collecting curvature data of the surface of the workpiece to be cleaned within a preset sampling period, establishing a time-series curvature data set, and using the difference between the curvature data of the time-series curvature data set and the main curvature value as the curvature segmentation degree;

[0079] The surface of the workpiece to be cleaned is divided into regions according to the curvature segmentation degree: when the curvature segmentation degree is less than a first preset threshold, the region where the current curvature data is located is regarded as a plane region, and the current curvature value is updated by using a neighborhood average value based on a distance weight; when the curvature segmentation degree is between the first preset threshold and a second preset threshold, the region where the current curvature data is located is regarded as a convex region, and the current curvature value is updated by using a weighted average value of exponentially decaying time series data; when the curvature segmentation degree is greater than the second preset threshold, the region where the current curvature data is located is regarded as a concave region, and the current curvature value is maintained;

[0080] Adjacent curvature gradients between adjacent regions are calculated, width distribution of transition regions is determined according to the adjacent curvature gradients, and a smooth transition boundary is constructed in the transition region using a B-spline curve.

[0081] In a method for demarcating the surface area of a workpiece to be cleaned, depth information of the workpiece surface is obtained. This depth information can be acquired using equipment such as a depth camera or lidar. Based on this depth information, corresponding gradient information is calculated, including first-order and second-order gradients. The first-order gradient reflects the rate of change of the workpiece surface in each direction, while the second-order gradient indicates the degree of change in the rate of change.

[0082] Figure 2 This is a flow chart illustrating the present invention's process for workpiece surface area segmentation based on curvature segmentation. For each pixel in the depth image, the horizontal and vertical depth changes are calculated to obtain the first-order gradient. The rate of change of the first-order gradient is then calculated to obtain the second-order gradient. For example, for three consecutive points with depths of 10mm, 12mm, and 15mm, the first-order gradient from the first point to the second is 2mm, and the first-order gradient from the second point to the third is 3mm. Therefore, the second-order gradient at the middle point is 1mm.

[0083] The calculated first-order and second-order gradients are used to further calculate the average curvature and Gaussian curvature of the surface of the workpiece being cleaned. The average curvature is the arithmetic mean of the principal curvatures, and the Gaussian curvature is the product of the principal curvatures. For example, if the first-order gradients of an area are 0.05 and 0.03 in the x and y directions, respectively, and the second-order gradients are 0.02, 0.01, and 0.005 in the xx, yy, and xy directions, respectively, the average curvature of the area can be calculated to be 0.015, and the Gaussian curvature can be calculated to be 0.0002.

[0084] The principal curvature is an important metric that describes the degree of curvature of a surface at a specific point. The two principal curvatures can be inversely solved from the mean curvature and Gaussian curvature. Continuing with the above example, when the mean curvature is 0.015 and the Gaussian curvature is 0.0002, the two principal curvatures can be calculated to be 0.02 and 0.01, respectively.

[0085] To enhance the system's stability and anti-interference capabilities, curvature data of the workpiece surface being cleaned is continuously collected within a preset sampling period to create a time-series curvature dataset. For example, curvature data is collected 10 times every 500 milliseconds, forming a time-series curvature dataset containing 10 data points. This curvature data is compared with the previously calculated principal curvature value, and the difference between them is calculated, which is defined as the curvature segmentation degree.

[0086] Specifically, two preset thresholds are set, the first preset threshold is 0.005, and the second preset threshold is 0.02. When the curvature segmentation degree is less than the first preset threshold of 0.005, the area where the current curvature data is located is divided into a plane area. For example, the curvature segmentation degree of a certain area is 0.003, which is less than 0.005, so the area is divided into a plane area. For the plane area, the current curvature value is updated using the neighborhood average value based on the distance weight, that is, the curvature values of the surrounding points are taken into account, different weights are assigned according to the distance, and the weighted average value is calculated as the updated curvature value. For example, the curvature value of the center point of a plane area is 0.01, and the curvature values of the four surrounding points are 0.012, 0.009, 0.011, and 0.008, respectively, and the distances are 1, 2, 1.5, and 2.5 units, respectively. The updated curvature value can be calculated to be 0.0102.

[0087] When the curvature segmentation degree is between the first preset threshold value of 0.005 and the second preset threshold value of 0.02, the area where the current curvature data is located is divided into a convex area. For example, the curvature segmentation degree of a certain area is 0.01, which is greater than 0.005 but less than 0.02, so the area is divided into a convex area. For convex areas, the current curvature value is updated using the weighted average of the exponentially decaying time series data, that is, considering the historical curvature data, giving more recent data a greater weight and more distant data a smaller weight, and calculating the weighted average as the updated curvature value. For example, the curvature values of the last five measurements of a convex area are 0.025, 0.028, 0.027, 0.026, and 0.029, respectively. Using a decay factor of 0.8, the updated curvature value is 0.0267.

[0088] When the curvature segmentation degree exceeds the second preset threshold of 0.02, the area containing the current curvature data is classified as a concave area. For example, if the curvature segmentation degree of a region is 0.03, which is greater than 0.02, the region is classified as a concave area. For concave areas, the current curvature value remains unchanged. This is because concave areas often have complex geometric shapes, and simple smoothing may result in the loss of important features.

[0089] After the region division is completed, the adjacent curvature gradient between adjacent regions is calculated, that is, the rate of change of the curvature values of two adjacent regions. For example, at the intersection of a flat region and a convex region, the curvature value of the flat region is 0.005, the curvature value of the convex region is 0.015, and the spatial distance between the two regions is 10 units, then the adjacent curvature gradient is 0.001 / unit.

[0090] The width distribution of the transition region is determined based on the calculated adjacent curvature gradient. A larger adjacent curvature gradient indicates a more dramatic curvature change, requiring a narrower transition region. A smaller adjacent curvature gradient indicates a more gradual curvature change, requiring a wider transition region. For example, when the adjacent curvature gradient is 0.001 / unit, the transition region width can be set to 15 units; when the adjacent curvature gradient is 0.005 / unit, the transition region width can be set to 5 units.

[0091] Within a defined transition region, a B-spline curve is used to construct a smooth transition boundary. B-spline curves offer excellent local controllability and high computational efficiency, making them suitable for constructing smooth transition boundaries. By selecting appropriate control points within the transition region, a B-spline curve is generated that smoothly connects adjacent regions. For example, within a transition region with a width of 10 units, five control points are selected to generate a third-order B-spline curve as a smooth transition boundary.

[0092] In an optional embodiment, based on the regional division result of the surface of the workpiece to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the robot's initial cleaning trajectory, wherein the trajectory density parameter of the planar area is set as a baseline value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value, including:

[0093] Collect curvature data of each area of the surface of the workpiece to be cleaned, calculate the average curvature value of all sampling points in each area, calculate the curvature change between adjacent sampling points, and calculate the dispersion degree of the average curvature value to obtain the regional curvature dispersion value;

[0094] Determine a reference track density value of the planar area according to a preset minimum track spacing, and perform a weighted combination of the reference track density value with the average curvature value and the curvature variation, wherein the weight of the average curvature value is proportional to the shape complexity value, and the weight of the curvature variation is proportional to the curvature dispersion value;

[0095] Calculate the trajectory density parameters of each area: for the convex area, divide the maximum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the convex trajectory density parameter; for the concave area, divide the minimum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the concave trajectory density parameter;

[0096] Generate the robot's initial clear trajectory according to the trajectory density parameters of each area.

[0097] In this implementation, the system first performs a three-dimensional scan of the workpiece surface to be cleaned, acquiring point cloud data. Using a point cloud data processing algorithm, the system then divides the workpiece surface into flat, convex, and concave regions. Once the regions are divided, the system collects curvature data for each region, preparing for the subsequent setting of differentiated trajectory density parameters.

[0098] The system uses a high-precision laser scanner to scan the workpiece surface, with a sampling interval of 0.5 mm to ensure sufficiently dense surface information is collected. For each collected point, the system uses the least squares method to fit the local surface of its surrounding neighborhood points and calculate the principal curvature value at that point. The system groups the curvature data for all points by region and calculates the average curvature value for all sampled points within each region. For example, the average curvature value of a convex region is calculated to be 0.025, indicating the degree of surface curvature in that area.

[0099] The system further calculates the curvature change between adjacent sampling points within each region. Based on the adjacent points in the grid, the system calculates the absolute value of the curvature difference between each pair of adjacent points. Then, all curvature differences across the entire region are averaged to obtain the average curvature change for that region. For example, if the average curvature change for a region is 0.003, it indicates that the curvature change in that region is relatively gentle.

[0100] To assess the dispersion of regional curvature, the system calculates the standard deviation of the curvature values at all points within the region, yielding the regional curvature dispersion value. A larger curvature dispersion value indicates a more uneven distribution of curvature within the region and a more complex shape. For example, a complex convex region has a curvature dispersion value of 0.018, while a simple convex region has a curvature dispersion value of only 0.004.

[0101] The system presets a minimum track spacing of 10 mm as the baseline track density value for planar areas. For each area, the system weights the baseline track density value with the area's average curvature value and curvature variation. The weighting process is proportional to the area's shape complexity, and the weighting of the curvature variation is proportional to the curvature dispersion value. The shape complexity value is calculated by combining the complexity of the area boundary and the unevenness of the curvature distribution within the area.

[0102] For example, if the shape complexity value of a region is 0.75 and the curvature dispersion value is 0.012, the weight of the average curvature value can be set to 0.75, and the weight of the curvature variation can be set to 0.6. Through this weighted combination method, the system can more accurately reflect the impact of regional characteristics on trajectory density.

[0103] For convex areas, the system divides the area's maximum curvature by the preset baseline curvature value of 0.01, then raises the value to the second power to calculate the convex track density parameter. For example, if the maximum curvature value of a convex area is 0.035, the track density parameter is calculated as (0.035 / 0.01)² = 12.25, meaning the track density in this area is 12.25 times the baseline density, and the actual track spacing is approximately 0.82 mm. This calculation method ensures that convex areas with greater curvature have higher track density, resulting in more thorough cleaning.

[0104] For concave areas, the system takes the absolute value of the area's minimum curvature (negative value), divides it by the preset baseline curvature value of 0.01, and then raises the value to a power of two to obtain the concave track density parameter. For example, if the minimum curvature value of a concave area is -0.042, the track density parameter is calculated as (0.042 / 0.01)² = 17.64, meaning the track density in this area is 17.64 times the baseline density, resulting in an actual track spacing of approximately 0.57 mm. Because concave areas are prone to dirt accumulation, increasing the track density exponentially ensures more thorough cleaning.

[0105] The system generates the robot's initial cleaning trajectory based on the calculated trajectory density parameters for each zone. This trajectory generation process utilizes an improved ZigZag algorithm, first determining the primary trajectory direction and then generating parallel trajectory lines based on the calculated trajectory spacing. To ensure trajectory continuity, the system applies trajectory smoothing at zone boundaries, fitting the trajectory with a cubic spline curve to eliminate sudden changes in the trajectory.

[0106] In actual application, the system divides an industrial part with various surface morphologies into five flat areas, three convex areas, and two concave areas. Calculated track spacing in the flat areas is maintained at 10 mm, while track spacings in the convex areas are 1.25 mm, 0.82 mm, and 1.67 mm, respectively, and in the concave areas are 0.57 mm and 0.63 mm, respectively. The total length of the initial cleaning track generated by the system is 4.73 meters, achieving a cleaning coverage rate of 99.8%.

[0107] Through the above-mentioned differentiated trajectory density parameter setting method, the system can adaptively adjust the cleaning trajectory density according to the geometric characteristics of the workpiece surface, which not only ensures the cleaning quality but also improves the cleaning efficiency. It is particularly suitable for precision cleaning scenarios of industrial parts with complex shapes.

[0108] In an optional embodiment, based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated, including:

[0109] Based on the discrete point position coordinate sequence of the initial cleaning trajectory of the robot, calculating the position difference vectors between adjacent trajectory points in the position coordinate sequence, and normalizing the position difference vectors to obtain a trajectory tangent vector sequence;

[0110] Extract the corresponding surface normal vector at each trajectory point position, use the surface normal vector as the Z-axis direction vector of the robot tool coordinate system, and use the trajectory tangent vector sequence at the corresponding position as the Y-axis direction vector of the tool coordinate system; obtain the X-axis direction vector of the tool coordinate system by cross-product operation of the Z-axis direction vector and the Y-axis direction vector, construct a posture rotation matrix based on the X-axis direction vector, the Y-axis direction vector and the Z-axis direction vector, and decompose the posture rotation matrix to obtain the Euler angle of the tool coordinate system relative to the base coordinate system;

[0111] Substitute the position coordinate sequence and posture rotation matrix sequence into the robot kinematic equation, solve the joint angle solution set at each trajectory point, and calculate the joint motion amount and joint median value deviation of each group of solutions in the joint angle solution set.

[0112] This embodiment provides a method for calculating the robot posture angle based on the robot's initial cleaning trajectory. This method is suitable for motion planning when the robot performs cleaning operations on complex curved surfaces, ensuring that the robot can complete the cleaning task with an appropriate posture.

[0113] In this embodiment, the robot's initial cleaning trajectory typically consists of a series of discrete position coordinate points, which can be obtained through a pre-planned path planning algorithm or manual teaching. These position coordinate points are arranged in sequence to form the spatial distribution of the cleaning trajectory. To calculate the robot's posture angle during the cleaning process, it is necessary to combine the surface normal vectors of each area.

[0114] The robot's initial cleaning trajectory can be represented as a sequence of discrete points in three-dimensional space. For example, when the robot needs to clean a spherical surface, the cleaning trajectory may contain 100 points. For example, the coordinates of the first point may be (100, 200, 300), the coordinates of the second point may be (105, 202, 302), and so on.

[0115] To calculate the tangent vector sequence for the trajectory, we need to perform a differential calculation on the positions of adjacent trajectory points. Specifically, for the i-th point and the i+1-th point, the position difference vector is calculated as (x_{i+1}-x_i, y_{i+1}-y_i, z_{i+1}-z_i). For example, for the first and second points in the example above, the position difference vector is (5, 2, 2).

[0116] Next, normalize the position difference vector, i.e., normalize its length to 1, to obtain the trajectory tangent vector. Using the example above, the length of the position difference vector is √(5²+2²+2²) = √33. Therefore, the normalized trajectory tangent vector is (5 / √33, 2 / √33, 2 / √33), which is approximately equal to (0.87, 0.35, 0.35). This calculation is applied to every point on the trajectory, ultimately obtaining the complete trajectory tangent vector sequence.

[0117] After determining the trajectory tangent vector, the next step is to extract the corresponding surface normal at each trajectory point. Surface normals are typically perpendicular to the surface being cleaned and can be calculated from the surface's geometric model or point cloud data. For example, for a sphere, the normal at any point points toward the center of the sphere. Assume the surface normal extracted at the first trajectory point is (0.2, 0.3, 0.93), which is already a unit vector.

[0118] When constructing the robot's tool coordinate system, the surface normal vector is used as the Z-axis direction vector of the tool coordinate system, and the trajectory tangent vector is used as the Y-axis direction vector of the tool coordinate system. Taking the first point in the above example as an example, the Z-axis direction vector is (0.2, 0.3, 0.93) and the Y-axis direction vector is (0.87, 0.35, 0.35).

[0119] To ensure the tool coordinate system is orthogonal, a cross product is required to calculate the X-axis direction vector. The X-axis direction vector is equal to the cross product of the Z-axis direction vector and the Y-axis direction vector, that is, X = Z × Y. For the above example, the X-axis direction vector is calculated as follows: X = (0.2, 0.3, 0.93) × (0.87, 0.35, 0.35), which is approximately (-0.06, 0.76, -0.65).

[0120] At this point, the Y-axis direction vector needs to be corrected so that it is orthogonal to the X and Z axes. The new Y-axis direction vector is equal to the cross product of the Z and X axes, that is, Y' = Z × X. For the above example, the corrected Y-axis direction vector is (0.2, 0.3, 0.93) × (-0.06, 0.76, -0.65), which is approximately (-0.98, 0.14, 0.11).

[0121] Based on the X-axis and the corrected Y-axis and Z-axis direction vectors, we can construct a posture rotation matrix. This matrix consists of three columns containing the X-axis, Y-axis, and Z-axis direction vectors, respectively. For the example above, the rotation matrix is: [[-0.06, -0.98, 0.2], [0.76, 0.14, 0.3], [-0.65, 0.11, 0.93]].

[0122] By decomposing the pose rotation matrix into Euler angles, we can obtain the pose representation of the tool coordinate system relative to the base coordinate system. For example, for the above rotation matrix, we can decompose the Euler angles (α, β, γ), which represent the rotation angles around the X, Y, and Z axes, respectively. This calculation process is applied to each point on the trajectory, ultimately obtaining a complete robot motion trajectory consisting of position coordinates and pose angles. For example, the complete information for the first trajectory point might be: position (100, 200, 300) and pose angles (30°, 45°, 60°).

[0123] After calculating the pose information, the robot's joint angles need to be determined. Substituting the position coordinate sequence and the pose rotation matrix sequence into the robot's kinematic equations yields a solution set for the joint angles at each trajectory point. For a six-degree-of-freedom robot, multiple sets of solutions are often available, requiring further selection of the optimal solution.

[0124] To evaluate the quality of each solution, the joint motion and joint median deviation are calculated. The joint motion represents the sum of the changes in joint angles between two adjacent trajectory points, while the joint median deviation represents the sum of the deviations between the joint angles and the joint median. For example, if there are two solutions for the first trajectory point: [10°, 20°, 30°, 40°, 50°, 60°] and [15°, 25°, 35°, 45°, 55°, 65°], their joint motions and joint median deviations are calculated, and the solution with the best overall score is selected as the final solution.

[0125] Through the above method, a robot motion trajectory containing posture information can be generated, so that the robot can perform cleaning tasks with an appropriate posture along the predetermined trajectory, thereby improving cleaning efficiency and quality.

[0126] In an optional embodiment, setting the plasma cleaning process parameters corresponding to each trajectory point according to the material properties of the workpiece to be cleaned includes:

[0127] The material properties of the workpiece to be cleaned include local hardness value and local roughness value;

[0128] Calculating the plasma cleaning power based on the local hardness value and the local roughness value: multiplying the local hardness value by a first weighting coefficient to obtain a first power component, multiplying the local roughness value by a second weighting coefficient to obtain a second power component, and adding the first power component and the second power component to obtain a plasma cleaning power value corresponding to the trajectory point;

[0129] Obtaining a local curvature value at each trajectory point, taking a weighted sum of the local curvature value and the local hardness value as a process parameter modulation coefficient, and multiplying the process parameter modulation coefficient by a preset reference gas flow rate to obtain an actual gas flow rate value corresponding to the trajectory point;

[0130] The angle between the Z-axis direction of the tool coordinate system and the normal vector of the surface of the workpiece to be cleaned is calculated, and the actual working distance of the plasma jet is determined based on the angle. The actual working distance, plasma cleaning power value and actual gas flow value are converted into ion concentration values corresponding to each trajectory point.

[0131] The present invention relates to a method for setting plasma cleaning process parameters corresponding to trajectory points based on the material properties of the workpiece being cleaned. This method takes into account factors such as the local hardness, roughness, and curvature of the workpiece being cleaned, as well as the angle between the tool coordinate system and the workpiece surface normal vector, to accurately set the plasma cleaning power, actual gas flow rate, and ion concentration corresponding to each trajectory point.

[0132] In practical applications, the material property data of the workpiece to be cleaned is first acquired. This data includes the local hardness and roughness values at each track point on the workpiece surface. For example, for a metal part, a hardness tester can be used to measure the Vickers hardness value at each track point. For example, assume that the local hardness value of track point A is 250 HV. A surface roughness tester can be used to measure the roughness value at that point, assuming it is 3.5 μm. This data will serve as the basis for the subsequent calculation of plasma cleaning process parameters.

[0133] Based on the acquired local hardness value and local roughness value, the plasma cleaning power value corresponding to each trajectory point is calculated. During the calculation process, the local hardness value is multiplied by the preset first weighting coefficient to obtain the first power component, and the local roughness value is multiplied by the preset second weighting coefficient to obtain the second power component, and then the two power components are added to obtain the final plasma cleaning power value. Specifically, assuming that the first weighting coefficient is 0.5 and the second weighting coefficient is 20, for trajectory point A, the first power component is 250HV×0.5=125W, the second power component is 3.5μm×20=70W, and the final plasma cleaning power value is 125W+70W=195W. This calculation method can adaptively adjust the cleaning power according to the hardness and surface roughness characteristics of the material. Areas with higher hardness or greater roughness will obtain higher cleaning power to ensure the cleaning effect.

[0134] Next, obtain the local curvature value at each trajectory point. The local curvature value can be calculated from the three-dimensional model data of the workpiece, or directly measured by a curvature measuring instrument. Assume that the local curvature value of trajectory point A is 0.05mm^-1. The local curvature value and the local hardness value are weighted to obtain the process parameter modulation coefficient. Specifically, the local curvature value can be multiplied by a weight of 0.3, and the local hardness value can be multiplied by a weight of 0.002, and then added to obtain the process parameter modulation coefficient. For trajectory point A, the process parameter modulation coefficient is 0.05mm^-1×0.3+250HV×0.002=0.515. Multiply the modulation coefficient by the preset reference gas flow rate to obtain the actual gas flow rate value. Assuming that the reference gas flow rate is 10L / min, the actual gas flow rate value of trajectory point A is 10L / min×0.515=5.15L / min. In this way, by considering the combined effects of local curvature and hardness, the gas flow rate during the plasma cleaning process can be controlled more accurately to adapt to changes in the workpiece surface shape and material properties.

[0135] Calculating the angle between the Z-axis of the tool coordinate system and the surface normal vector of the workpiece to be cleaned is also an important part of setting process parameters. This angle can be calculated through the CAD model of the workpiece and the kinematic analysis of the robot. Assume that at trajectory point A, the angle between the Z-axis of the tool coordinate system and the surface normal vector is 15 degrees. Based on this angle value, the actual working distance of the plasma jet is determined. For example, when the angle is 0 degrees, the standard working distance is set to 10mm; as the angle increases, the working distance needs to be appropriately reduced to ensure the cleaning effect. Specifically, the following relationship can be used: actual working distance = standard working distance × (1-angle / 90). For trajectory point A, the actual working distance is 10mm×(1-15 / 90)=8.33mm.

[0136] Finally, the actual working distance, plasma cleaning power, and actual gas flow rate are converted to ion concentration values for each trajectory point. The ion concentration value can be converted using the following relationship: ion concentration = K × power × gas flow rate / (working distance^2), where K is an empirical constant related to the equipment, assuming K = 0.05.

[0137] For trajectory point A, the ion concentration is 0.05 × 195W × 5.15L / min / (8.33mm²) = 6.04. As a direct indicator of cleaning effectiveness, the ion concentration can be used to evaluate and predict cleaning quality and guide further optimization of process parameters.

[0138] This method systematically considers the impact of workpiece material properties, surface shape characteristics, and tool posture on the plasma cleaning process. Precise process parameters, including plasma cleaning power, actual gas flow rate, and actual working distance, are customized for each trajectory point and ultimately converted into ion concentration values. This parameterized cleaning strategy can significantly improve the quality and efficiency of plasma cleaning and is particularly suitable for precision cleaning of workpieces with complex shapes and widely varying material properties.

[0139] In an optional embodiment, the method further includes:

[0140] When the difference in plasma cleaning power between adjacent trajectory points is greater than a preset power threshold, a transition point is inserted between the adjacent trajectory points; when the difference in actual gas flow between adjacent trajectory points is greater than a preset flow threshold, linear interpolation is used for transition;

[0141] The plasma cleaning power value, actual gas flow value and ion concentration value of each trajectory point are combined into a process parameter group, and a mapping relationship between the process parameter group and the trajectory point position is established.

[0142] In the plasma cleaning process, to ensure the stability and consistency of cleaning quality, it is necessary to properly transition the parameters between trajectory points. This embodiment provides an optimized trajectory point parameter transition method to ensure smooth parameter changes during the plasma cleaning process and improve the cleaning effect.

[0143] When the system detects that the difference in plasma cleaning power between adjacent track points is greater than a preset power threshold, the system inserts a transition point between the two track points. For example, when the preset power threshold is set to 50 watts, if the plasma cleaning power of track point A is 300 watts and the plasma cleaning power of the adjacent track point B is 400 watts, the difference is 100 watts, exceeding the preset threshold of 50 watts. The system then inserts a transition point C between A and B. The power of transition point C can be set to 350 watts, and its position can be set midway between A and B. By inserting transition point C, the original power surge from A to B is broken down into two smaller surges, A to C and C to B, making the power change smoother and avoiding uneven cleaning or substrate damage caused by power surges.

[0144] For gas flow transition processing, when the difference in actual gas flow between adjacent trajectory points exceeds the preset flow threshold, the system uses linear interpolation to transition. Taking argon as an example, assuming the preset flow threshold is 10 standard cubic centimeters per minute (sccm), the argon flow at trajectory point D is 50 sccm, and the argon flow at the adjacent trajectory point E is 70 sccm, the difference is 20 sccm, which exceeds the preset threshold. At this time, the system will linearly distribute the gas flow along the path between D and E in proportion to the distance between the two points. If the path from D to E is divided into five equal segments, the corresponding flow values will be 54 sccm, 58 sccm, 62 sccm, 66 sccm, and finally reach 70 sccm at point E. This linear interpolation method ensures that the gas flow changes smoothly during trajectory execution, avoiding plasma instability caused by sudden changes in flow.

[0145] In order to achieve accurate correspondence between parameters and trajectory positions, the system combines the plasma cleaning power value, actual gas flow value, and ion concentration value of each trajectory point into a process parameter group, and establishes a mapping relationship between the process parameter group and the trajectory point position. This mapping relationship is stored in the system control unit in the form of a data table. For example, for a cleaning path containing 10 trajectory points, the system will create a data table with 10 rows and multiple columns, with each row corresponding to a trajectory point, and the columns including the coordinate information (X, Y, Z coordinate values) of the trajectory point and the corresponding plasma cleaning power value, gas flow value, and ion concentration value.

[0146] In a specific example, the coordinates of trajectory point 1 are (10mm, 20mm, 0mm), and the corresponding process parameter set might be (Power: 300W, Ar flow: 50sccm, Oxygen flow: 10sccm, Ion concentration: 5×10^10 cm^-3). The coordinates of trajectory point 2 are (15mm, 20mm, 0mm), and the corresponding process parameter set might be (Power: 350W, Ar flow: 55sccm, Oxygen flow: 12sccm, Ion concentration: 5.5×10^10 cm^-3). When executing a cleaning task, the system uses this mapping relationship to query or interpolate the appropriate process parameters based on the actual position of the cleaning head, and adjusts the plasma generator output power and gas control valve opening in real time.

[0147] In practice, the power and flow threshold settings are closely related to the workpiece material, contaminant type, and cleaning requirements. For cleaning precision electronic components that are sensitive to parameter changes, the power threshold can be set lower, such as 30 watts, and the flow threshold can be set to 5 sccm to ensure extremely gradual parameter changes. For cleaning more tolerant metal workpieces, the thresholds can be relaxed appropriately, such as 80 watts for power and 15 sccm for flow, to improve cleaning efficiency.

[0148] Mapping relationships can be established through offline programming or online learning. With offline programming, technicians pre-set parameters for each trajectory point based on experience and test data. With online learning, the system dynamically adjusts parameter mapping relationships using sensor data that monitors cleaning performance in real time. For example, the system could be equipped with a spectrum analyzer to monitor plasma luminescence intensity. If it detects poor cleaning performance in a particular area, the system would automatically increase the power or gas flow rate for that trajectory point and update the mapping data table.

[0149] By optimizing parameter transitions and establishing mapping relationships, the system can achieve smooth and precise control of process parameters during the cleaning process of complex workpieces, significantly improving the uniformity and reliability of plasma cleaning. Practice has proven that this approach improves workpiece surface cleaning uniformity by 25% and cleaning efficiency by 30%, while significantly reducing the problem of over- or under-cleaning of parts caused by sudden parameter changes.

[0150] The automatic cleaning system based on plasma and trajectory control according to the embodiment of the present invention includes:

[0151] The first unit is configured to obtain a three-dimensional digital model of a workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of a surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data;

[0152] A second unit is configured to determine a principal curvature value based on gradient information corresponding to the depth information, use a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and divide the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold;

[0153] The third unit is configured to set differentiated trajectory density parameters for different types of areas based on the area division result of the workpiece surface to be cleaned, and generate an initial cleaning trajectory for the robot, wherein the trajectory density parameter of the planar area is set as a reference value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; and at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas;

[0154] The fourth unit is used to calculate the posture angle of the robot at each trajectory point based on the initial cleaning trajectory of the robot and the normal vector of each area, generate robot motion trajectory data containing posture information, and set the plasma cleaning process parameters corresponding to each trajectory point according to the material properties of the workpiece to be cleaned.

[0155] According to a third aspect of an embodiment of the present invention, an electronic device is provided, including:

[0156] processor;

[0157] a memory for storing processor-executable instructions;

[0158] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0159] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described above is implemented.

[0160] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic cleaning method based on plasma and trajectory control, characterized in that: include: Acquire a three-dimensional digital model of the workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data; Determining a principal curvature value based on gradient information corresponding to the depth information, taking a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and dividing the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold; Based on the regional division results of the workpiece surface to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the robot's initial cleaning trajectory, where the trajectory density parameter of the planar area is set as a baseline value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas; Based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated. The plasma cleaning process parameters corresponding to each trajectory point are set according to the material properties of the workpiece to be cleaned.

2. The method according to claim 1, characterized in that Acquiring a three-dimensional digital model of a workpiece to be cleaned, meshing the three-dimensional digital model to obtain mesh data of the surface of the workpiece to be cleaned, and extracting depth information of the surface of the workpiece to be cleaned based on the mesh data includes: Calculating the local curvature value of each point on the surface of the three-dimensional digital model, constructing a grid size function based on the local curvature value, wherein the grid size function adopts an exponential decay form, the grid size decreases as the curvature value increases, and the grid division is denser at the location of the curvature mutation; adaptively meshing the workpiece to be cleaned based on the grid size function to obtain grid division data of the surface of the workpiece to be cleaned; Establishing a local coordinate system for the surface of the workpiece to be cleaned, projecting the vertex coordinates in the meshing data to the local coordinate system, and calculating the depth value of each vertex; Constructing a depth feature map of the surface of the workpiece to be cleaned based on the depth value of each vertex, calculating the gradient distribution of the depth feature map, and determining a depth mutation area according to the gradient distribution; The principal curvature information of the surface of the workpiece to be cleaned is extracted, and a shape index is calculated based on the principal curvature information. The shape index is clustered according to the numerical value to obtain shape clustering areas, and a transition zone is established between the shape clustering areas, wherein the width of the transition zone is proportional to the difference in shape indices of adjacent areas.

3. The method according to claim 1, characterized in that Determining a principal curvature value based on gradient information corresponding to the depth information, taking a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and dividing the surface of the workpiece to be cleaned into a plane area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold value includes: Calculating an average curvature and a Gaussian curvature of the surface of the workpiece to be cleaned based on the first-order gradient and the second-order gradient corresponding to the depth information, and determining a principal curvature value based on the average curvature and the Gaussian curvature; Continuously collecting curvature data of the surface of the workpiece to be cleaned within a preset sampling period, establishing a time-series curvature data set, and using the difference between the curvature data of the time-series curvature data set and the main curvature value as the curvature segmentation degree; The surface of the workpiece to be cleaned is divided into regions according to the curvature segmentation degree: when the curvature segmentation degree is less than a first preset threshold, the region where the current curvature data is located is regarded as a plane region, and the current curvature value is updated by using a neighborhood average value based on a distance weight; when the curvature segmentation degree is between the first preset threshold and a second preset threshold, the region where the current curvature data is located is regarded as a convex region, and the current curvature value is updated by using a weighted average value of exponentially decaying time series data; when the curvature segmentation degree is greater than the second preset threshold, the region where the current curvature data is located is regarded as a concave region, and the current curvature value is maintained; Adjacent curvature gradients between adjacent regions are calculated, width distribution of transition regions is determined according to the adjacent curvature gradients, and a smooth transition boundary is constructed in the transition region using a B-spline curve.

4. The method according to claim 1, wherein According to the area division result of the surface of the workpiece to be cleaned, differentiated trajectory density parameters are set for different types of areas to generate the initial cleaning trajectory of the robot, wherein the trajectory density parameter of the flat area is set as the base value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value. Collect curvature data of each area of the surface of the workpiece to be cleaned, calculate the average curvature value of all sampling points in each area, calculate the curvature change between adjacent sampling points, and calculate the dispersion degree of the average curvature value to obtain the regional curvature dispersion value; Determine a reference track density value of the planar area according to a preset minimum track spacing, and perform a weighted combination of the reference track density value with the average curvature value and the curvature variation, wherein the weight of the average curvature value is proportional to the shape complexity value, and the weight of the curvature variation is proportional to the curvature dispersion value; Calculate the trajectory density parameters of each area: for the convex area, divide the maximum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the convex trajectory density parameter; for the concave area, divide the minimum curvature value of the area by the preset reference curvature value and perform a quadratic operation to obtain the concave trajectory density parameter; Generate the robot's initial clear trajectory according to the trajectory density parameters of each area.

5. The method according to claim 1, wherein Based on the initial cleaning trajectory of the robot, the posture angle of the robot at each trajectory point is calculated in combination with the normal vector of each area, and the robot motion trajectory data containing posture information is generated, including: Based on the discrete point position coordinate sequence of the initial cleaning trajectory of the robot, calculating the position difference vectors between adjacent trajectory points in the position coordinate sequence, and normalizing the position difference vectors to obtain a trajectory tangent vector sequence; Extract the corresponding surface normal vector at each trajectory point position, use the surface normal vector as the Z-axis direction vector of the robot tool coordinate system, and use the trajectory tangent vector sequence at the corresponding position as the Y-axis direction vector of the tool coordinate system; obtain the X-axis direction vector of the tool coordinate system by cross-product operation of the Z-axis direction vector and the Y-axis direction vector, construct a posture rotation matrix based on the X-axis direction vector, the Y-axis direction vector and the Z-axis direction vector, and decompose the posture rotation matrix to obtain the Euler angle of the tool coordinate system relative to the base coordinate system; Substitute the position coordinate sequence and posture rotation matrix sequence into the robot kinematic equation, solve the joint angle solution set at each trajectory point, and calculate the joint motion amount and joint median value deviation of each group of solutions in the joint angle solution set.

6. The method according to claim 1, characterized in that The plasma cleaning process parameters corresponding to each trajectory point are set according to the material characteristics of the workpiece to be cleaned, including: The material properties of the workpiece to be cleaned include local hardness value and local roughness value; Calculating the plasma cleaning power based on the local hardness value and the local roughness value: multiplying the local hardness value by a first weighting coefficient to obtain a first power component, multiplying the local roughness value by a second weighting coefficient to obtain a second power component, and adding the first power component and the second power component to obtain a plasma cleaning power value corresponding to the trajectory point; Obtaining a local curvature value at each trajectory point, taking a weighted sum of the local curvature value and the local hardness value as a process parameter modulation coefficient, and multiplying the process parameter modulation coefficient by a preset reference gas flow rate to obtain an actual gas flow rate value corresponding to the trajectory point; The angle between the Z-axis direction of the tool coordinate system and the normal vector of the surface of the workpiece to be cleaned is calculated, and the actual working distance of the plasma jet is determined based on the angle. The actual working distance, plasma cleaning power value and actual gas flow value are converted into ion concentration values corresponding to each trajectory point.

7. The method according to claim 6, characterized in that The method further comprises: When the difference in plasma cleaning power between adjacent trajectory points is greater than a preset power threshold, a transition point is inserted between the adjacent trajectory points; when the difference in actual gas flow between adjacent trajectory points is greater than a preset flow threshold, linear interpolation is used for transition; The plasma cleaning power value, actual gas flow value and ion concentration value of each trajectory point are combined into a process parameter group, and a mapping relationship between the process parameter group and the trajectory point position is established.

8. An automatic cleaning system based on plasma and trajectory control, for implementing the method according to any one of claims 1 to 7, characterized in that: include: The first unit is configured to obtain a three-dimensional digital model of a workpiece to be cleaned, mesh the three-dimensional digital model to obtain mesh data of a surface of the workpiece to be cleaned, and extract depth information of the surface of the workpiece to be cleaned based on the mesh data; A second unit is configured to determine a principal curvature value based on gradient information corresponding to the depth information, use a difference between the principal curvature value and current curvature data as a curvature segmentation degree, and divide the surface of the workpiece to be cleaned into a flat area, a convex area, and a concave area according to a comparison result of the curvature segmentation degree with a preset threshold; The third unit is configured to set differentiated trajectory density parameters for different types of areas based on the area division result of the workpiece surface to be cleaned, and generate an initial cleaning trajectory for the robot, wherein the trajectory density parameter of the planar area is set as a reference value, the trajectory density parameter of the convex area increases linearly according to the curvature value, and the trajectory density parameter of the concave area increases exponentially according to the curvature value; and at the same time, a trajectory transition zone is set between adjacent areas, and the density parameter of the trajectory transition zone is determined by the weighted average of the density parameters of the adjacent areas; The fourth unit is used to calculate the posture angle of the robot at each trajectory point based on the initial cleaning trajectory of the robot and the normal vector of each area, generate robot motion trajectory data containing posture information, and set the plasma cleaning process parameters corresponding to each trajectory point according to the material properties of the workpiece to be cleaned.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images