Method and system for constructing robot workspace visualization based on digital twin

Abstract

This invention relates to the field of digital twin technology applications, and more particularly to a method and system for constructing a robot workspace visualization based on digital twins. The proposed method for constructing a robot workspace visualization based on digital twins involves simulation verification using a digital twin model and the construction of an initial point cloud workspace using a certain amount of data to reduce computational resource consumption. The method reduces errors and point cloud inhomogeneities caused by random sampling through point cloud spatial contour extraction, precise topology of the point cloud workspace, and boundary point cloud correction. Furthermore, the accuracy of point cloud visualization is improved through rolling triangular mesh generation and multiple passing smoothing modeling. This invention has significant versatility and applicability.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology applications, and in particular to a method and system for visually constructing a robot workspace based on digital twins. Background Technology

[0002] With the rise of intelligent manufacturing, the manufacturing industry is constantly transforming and upgrading towards intelligence, digitalization, and automation. Robots are widely used, and the personalized and diversified nature of upstream product demand leads to frequent production line changes and changes in downstream processes. Changes in the robot's working environment increase potential risks to production, necessitating ensuring the robot's accessibility to its work location. Related technologies utilize robot workspace visualization software to monitor robot operations in real time.

[0003] However, there are the following problems with building a robot visualization workspace using existing technologies:

[0004] 1. Lack of near-physical simulation operation testing and real-time performance; the platform mainly uses MATLAB software to construct the visualization space, but it cannot make corresponding changes to adapt to the production environment, and therefore cannot adjust the actual operation of the robot.

[0005] Second, the robot's visual workspace is three-dimensional, resulting in a large amount of computational data, slow convergence, and high data noise. The workspace formed by a single algorithm, such as the Monte Carlo algorithm, has large errors, high complexity, and high cost.

[0006] Third, the conversion from point cloud to mesh is complex, has low accuracy, and is not smooth enough.

[0007] Therefore, it is necessary to provide a new method for visually constructing robot workspaces to solve the above-mentioned technical problems. Summary of the Invention

[0008] The present invention aims to solve the technical problems of large computational load, low accuracy and poor real-time performance of robot visualization workspaces constructed by existing technologies.

[0009] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for visually constructing a robot workspace based on digital twins, comprising the following steps:

[0010] S1. Construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range.

[0011] S2. Construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is the set of points in the space that can be reached by the motion description reference point of the robot end effector.

[0012] S3. Generate initial point cloud workspace data according to the total kinematic transformation matrix of the robot end effector and a preset point cloud generation method.

[0013] S4. Extract the point cloud spatial contour from the initial point cloud workspace;

[0014] S5. Generate precise point cloud workspace data based on the point cloud spatial contour;

[0015] S6. Use a preset correction method to perform boundary point cloud correction on the precise point cloud workspace data to obtain corrected point cloud workspace data.

[0016] S7. Perform surface reconstruction based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot.

[0017] Furthermore, the total kinematic transformation matrix of the robot's end effector is defined as follows: 0 T m It satisfies the following relation (1):

[0018]

[0019] Where m represents the number of robot axes, i-1 T i This represents the homogeneous transformation matrix that transforms the robot between the i-th axis and the (i-1)-th axis.

[0020] Furthermore, the preset point cloud generation method is the Monte Carlo method, and step S3 includes the following sub-steps:

[0021] S31. Set the preset number of iterations and start the iteration;

[0022] S32. A random joint angle is obtained by randomly selecting and adding values, and the random joint angle is assigned to the joint angle. The random joint angle is defined as θ, the random value is rand, and the range of the joint angle is [θmin, θmax]. Then the random joint angle θ satisfies the following relationship (2):

[0023] θ=θ min +(θ max -θ min )*rand (2);

[0024] S33. Obtain the homogeneous transformation matrix corresponding to the joint rotation angle, and substitute the homogeneous transformation matrix into the total kinematic transformation matrix of the robot end effector to obtain the corresponding end effector pose data;

[0025] S34. Update the current iteration number and determine whether the preset iteration number has been reached. If not, return to step S32; if yes, output all the end pose data as the initial point cloud workspace data.

[0026] Furthermore, the preset point cloud generation method is a simulation-driven debugging method, and step S3 includes the following sub-steps:

[0027] S31. Set the preset number of iterations and start the iteration;

[0028] S32. The joint angle is assigned a value by a fixed step size. The joint angle is defined as θ, the step size is step, and the range of the joint angle is [θmin, θmax]. Then the joint angle θ satisfies the following relationship (3):

[0029] θ=(θ max -θ min ) / step (3);

[0030] S33. Obtain the joint rotation angle to drive the physical robot and obtain the corresponding end-effector pose data;

[0031] S34. Update the current iteration number and determine whether the preset iteration number has been reached. If not, return to step S32; if yes, output all the end pose data as the initial point cloud workspace data.

[0032] Furthermore, step S4 includes the following sub-steps:

[0033] S41. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and the points in the initial point cloud workspace data that intersect with the latitude and longitude lines are marked to obtain the first scanning result.

[0034] S42. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from left to right to obtain multiple planes perpendicular to the X-axis in the spatial coordinate system. Point cloud scanning is performed in each layer of latitude and longitude lines to obtain and mark the points in the initial point cloud workspace data that intersect with the latitude and longitude lines to obtain the second scanning result.

[0035] S43. Integrate and deduplicate the first scan result and the second scan result to obtain the point cloud spatial contour.

[0036] Furthermore, step S5 includes the following sub-steps:

[0037] S51. Divide the point cloud spatial contour into multiple point cloud regions of equal volume using a three-dimensional rasterization method.

[0038] S52. Fill each point cloud region with a sphere;

[0039] S53. Traverse the sphere in each point cloud region and determine whether the center position of the current sphere has the corresponding point cloud spatial contour. If not, calculate the analytical joint angle of the center position of the sphere using a preset robot inverse kinematics method. If the analytical joint angle is within the range of the joint angle interval, add the pose data corresponding to the analytical joint angle as new point cloud data into the initial point cloud workspace.

[0040] S54. After completing the sphere traversal of all the point cloud regions, output the current initial point cloud workspace as the precise initial point cloud workspace.

[0041] Furthermore, the preset robot inverse kinematics method is at least one of the algebraic method, geometric method, and numerical method.

[0042] Furthermore, the preset correction method is latitude and longitude scanning, and step S6 specifically includes:

[0043] Based on the latitude and longitude scanning method, the precise point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and points in the precise point cloud workspace data that do not intersect with the latitude and longitude lines are deleted.

[0044] Furthermore, step S7 includes the following sub-steps:

[0045] S71. Randomly select a point cloud data from the precise point cloud workspace data as the starting point pi by a random selection method;

[0046] S72. Using the starting point pi as the center of the sphere, search other point cloud data within a fixed radius a to obtain a set of points in the spherical neighborhood.

[0047] S73. Connect the starting point pi to any point qi in the spherical domain point set, and generate two circumcircles with radius a that pass through both points pi and qi. If there are no other point cloud data in the precise point cloud workspace data in either circumcircle, then connect points pi and qi to generate model edges.

[0048] S74. Select point qi+1 from the set of points in the spherical domain, return to step S73, and continue until all points in the set of points in the spherical domain are selected and step S73 is executed.

[0049] S75. Select the next starting point pi+1, return to step S72, and continue until all points in the precise point cloud workspace data have been selected and steps S72-S74 are executed.

[0050] S76. Output all the model edges as the visualization data model.

[0051] Secondly, the present invention also provides a robot workspace visualization construction system based on digital twins, comprising:

[0052] The digital twin modeling module is used to construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range.

[0053] The kinematic modeling module is used to construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is a set of points in the space that can be reached by the motion description reference point of the robot end effector.

[0054] The point cloud generation module is used to generate initial point cloud workspace data according to the total kinematic transformation matrix of the robot end effector and a preset point cloud generation method.

[0055] The point cloud extraction module is used to extract the point cloud spatial contour from the initial point cloud workspace;

[0056] A point cloud precision module is used to generate precise point cloud workspace data based on the point cloud spatial contour.

[0057] The point cloud correction module is used to perform boundary point cloud correction on the precise point cloud workspace data using a preset correction method to obtain corrected point cloud workspace data.

[0058] The visualization module is used to reconstruct the surface based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot.

[0059] The beneficial effects achieved by this invention lie in proposing a method for constructing a robot workspace visualization based on digital twins. This method is based on simulation verification using a digital twin model and constructs an initial point cloud workspace using a certain amount of data, thereby reducing the consumption of computing resources. The process of extracting the spatial contour of the point cloud, accurately topologically supplementing the point cloud workspace, and correcting the boundary point cloud reduces errors and the unevenness of the point cloud caused by random sampling. Furthermore, the accuracy of point cloud visualization is improved by generating triangular meshes through rolling and smoothing modeling through multiple passing passes. The method of this invention has great versatility and applicability. Attached Figure Description

[0060] Figure 1 This is a flowchart of the steps in the method for visually constructing a robot workspace based on digital twins provided in this embodiment of the invention.

[0061] Figure 2 This is a schematic diagram of the initial point cloud workspace data obtained using the Monte Carlo method according to an embodiment of the present invention;

[0062] Figure 3 This is a schematic diagram of the initial point cloud workspace data obtained using the simulation-driven debugging method provided in an embodiment of the present invention;

[0063] Figure 4 Is with Figure 2 A schematic diagram of the corresponding latitude and longitude line scanning results;

[0064] Figure 5 Is with Figure 3 A schematic diagram of the corresponding latitude and longitude line scanning results;

[0065] Figure 6 This is a schematic diagram of boundary point cloud correction provided in an embodiment of the present invention. Detailed Implementation

[0066] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0067] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of a robot workspace visualization construction method based on digital twins provided in this invention. The robot workspace visualization construction method based on digital twins includes the following steps:

[0068] S1. Construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range.

[0069] Specifically, for different types of robots, the embodiments of the present invention use 3D modeling software to establish corresponding digital twin models of the robots during implementation, and the models should maintain the kinematic relationship corresponding to the physical objects, that is, each joint of the robot needs to be modeled independently.

[0070] S2. Construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is the set of points in the space that can be reached by the motion description reference point of the robot end effector.

[0071] The total kinematic transformation matrix of the robot's end effector is defined as follows: 0 T m It satisfies the following relation (1):

[0072]

[0073] Where m represents the number of robot axes, i-1 T i This represents the homogeneous transformation matrix that transforms the robot between the i-th axis and the (i-1)-th axis.

[0074] S3. Generate initial point cloud workspace data according to the total kinematic transformation matrix of the robot end effector and a preset point cloud generation method.

[0075] The purpose of step S3 is to change the pose of each joint of the robot by changing the joint rotation angle of each joint, and to obtain the pose of the end effector by matrix transformation, thereby obtaining the initial reachable workspace of the robot.

[0076] This invention uses two methods to generate point clouds. The first method is the Monte Carlo method, and correspondingly, step S3 includes the following sub-steps:

[0077] S31. Set the preset number of iterations and start the iteration;

[0078] S32. A random joint angle is obtained by randomly selecting and adding values, and the random joint angle is assigned to the joint angle. The random joint angle is defined as θ, the random value is rand, and the range of the joint angle is [θmin, θmax]. Then the random joint angle θ satisfies the following relationship (2):

[0079] θ=θ min +(θ max -θ min )*rand (2);

[0080] S33. Obtain the homogeneous transformation matrix corresponding to the joint rotation angle, and substitute the homogeneous transformation matrix into the total kinematic transformation matrix of the robot end effector to obtain the corresponding end effector pose data;

[0081] S34. Update the current iteration number and determine whether the preset iteration number has been reached. If not, return to step S32; if yes, output all the end pose data as the initial point cloud workspace data.

[0082] The initial point cloud workspace data obtained using the Monte Carlo method is as follows: Figure 2 As shown.

[0083] The second method is the simulation-driven debugging method. Correspondingly, step S3 includes the following sub-steps:

[0084] S31. Set the preset number of iterations and start the iteration;

[0085] S32. The joint angle is assigned a value by a fixed step size. The joint angle is defined as θ, the step size is step, and the range of the joint angle is [θmin, θmax]. Then the joint angle θ satisfies the following relationship (3):

[0086] θ=(θ m ax-θ min ) / step (3));

[0087] S33. Obtain the joint rotation angle to drive the physical robot and obtain the corresponding end-effector pose data;

[0088] S34. Update the current iteration number and determine whether the preset iteration number has been reached. If not, return to step S32; if yes, output all the end pose data as the initial point cloud workspace data.

[0089] The initial point cloud workspace data obtained using the simulation-driven debugging method is as follows: Figure 3 As shown.

[0090] S4. Extract the point cloud spatial contour from the initial point cloud workspace.

[0091] After completing step S3, since the obtained workspace is an initial workspace and a collection of point clouds, it cannot accurately represent the robot's workspace. In order to accurately topologically complete the point cloud workspace, contour extraction needs to be performed on the initial workspace to provide a basic boundary for the topology. Contour extraction of point cloud space involves three-dimensional space, and it is necessary to scan the x, y, and z directions respectively, compare the coordinate data of the point cloud, and obtain the boundary contour. At the same time, due to the special nature of the robot's workspace, it may have an inner contour. Therefore, it is necessary to extract both the inner and outer contours simultaneously during extraction.

[0092] Step S4 includes the following sub-steps:

[0093] S41. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and the points in the initial point cloud workspace data that intersect with the latitude and longitude lines are marked to obtain the first scanning result.

[0094] S42. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from left to right to obtain multiple planes perpendicular to the X-axis in the spatial coordinate system. Point cloud scanning is performed in each layer of latitude and longitude lines to obtain and mark the points in the initial point cloud workspace data that intersect with the latitude and longitude lines to obtain the second scanning result.

[0095] S43. Integrate and deduplicate the first scan result and the second scan result to obtain the point cloud spatial contour.

[0096] Specifically, during the scanning process, the minimum bounding box of the point set is first established in the two-dimensional plane, and the maximum and minimum values ​​of the point cloud data on the X, Y, and Z axes are calculated respectively. Let the meridians start from the minimum value of X, take a step size of dx, calculate all the points where each meridian intersects within the range of X values, and put their index values ​​into a newly created linked list, and scan the entire X interval. Based on the same method, the parallels are scanned.

[0097] The latitude and longitude scan results obtained by scanning the initial point cloud workspace data obtained using the Monte Carlo method are as follows: Figure 4 As shown, the latitude and longitude line scanning results obtained by scanning the initial point cloud workspace data obtained using the simulation-driven debugging method are as follows: Figure 5 As shown.

[0098] S5. Generate precise point cloud workspace data based on the point cloud spatial contour.

[0099] Step S5 includes the following sub-steps:

[0100] S51. Divide the point cloud spatial contour into multiple point cloud regions of equal volume using a three-dimensional rasterization method.

[0101] S52. Fill each point cloud region with a sphere;

[0102] S53. Traverse the spheres in each point cloud region and determine whether the center position of the current sphere has a corresponding point cloud spatial contour. If not, calculate the analytical joint angle of the center position using a preset robot inverse kinematics method. If the analytical joint angle is within the range of the joint angle interval, add the pose data corresponding to the analytical joint angle as new point cloud data to the initial point cloud workspace. The preset robot inverse kinematics method is at least one of algebraic method, geometric method, and numerical method.

[0103] S54. After completing the sphere traversal of all the point cloud regions, output the current initial point cloud workspace as the precise initial point cloud workspace.

[0104] Preferably, in step S53, analytical methods such as algebraic methods and geometric methods are preferred for calculation and solution to obtain a faster solution speed.

[0105] S6. Apply a preset correction method to the precise point cloud workspace data to perform boundary point cloud correction, thereby obtaining corrected point cloud workspace data.

[0106] The preset correction method is latitude and longitude line scanning, and step S6 specifically involves:

[0107] Based on the latitude and longitude scanning method, the precise point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and points in the precise point cloud workspace data that do not intersect with the latitude and longitude lines are deleted.

[0108] A schematic diagram of boundary point cloud correction for the precise point cloud workspace data is shown below. Figure 6 As shown, the points in the lower left part are the points that need to be deleted.

[0109] S7. Perform surface reconstruction based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot.

[0110] This invention uses the rolling ball method to reconstruct the surface of a visual data model. Step S7 includes the following sub-steps:

[0111] S71. Randomly select a point cloud data from the precise point cloud workspace data as the starting point pi by a random selection method;

[0112] S72. Using the starting point pi as the center of the sphere, search other point cloud data within a fixed radius a to obtain a set of points in the spherical neighborhood.

[0113] S73. Connect the starting point pi to any point qi in the spherical domain point set, and generate two circumcircles with radius a that pass through both points pi and qi. If there are no other point cloud data in the precise point cloud workspace data in either circumcircle, then connect points pi and qi to generate model edges.

[0114] S74. Select point qi+1 from the set of points in the spherical domain, return to step S73, and continue until all points in the set of points in the spherical domain are selected and step S73 is executed.

[0115] S75. Select the next starting point pi+1, return to step S72, and continue until all points in the precise point cloud workspace data have been selected and steps S72-S74 are executed.

[0116] S76. Output all the model edges as the visualization data model.

[0117] The beneficial effects achieved by this invention lie in proposing a method for constructing a robot workspace visualization based on digital twins. This method is based on simulation verification using a digital twin model and constructs an initial point cloud workspace using a certain amount of data, thereby reducing the consumption of computing resources. The process of extracting the spatial contour of the point cloud, accurately topologically supplementing the point cloud workspace, and correcting the boundary point cloud reduces errors and the unevenness of the point cloud caused by random sampling. Furthermore, the accuracy of point cloud visualization is improved by generating triangular meshes through rolling and smoothing modeling through multiple passing passes. The method of this invention has great versatility and applicability.

[0118] This invention also provides a robot workspace visualization and construction system based on digital twins, comprising:

[0119] The digital twin modeling module is used to construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range.

[0120] The kinematic modeling module is used to construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is a set of points in the space that can be reached by the motion description reference point of the robot end effector.

[0121] The point cloud generation module is used to generate initial point cloud workspace data according to the total kinematic transformation matrix of the robot end effector and a preset point cloud generation method.

[0122] The point cloud extraction module is used to extract the point cloud spatial contour from the initial point cloud workspace;

[0123] A point cloud precision module is used to generate precise point cloud workspace data based on the point cloud spatial contour.

[0124] The point cloud correction module is used to perform boundary point cloud correction on the precise point cloud workspace data using a preset correction method to obtain corrected point cloud workspace data.

[0125] The visualization module is used to reconstruct the surface based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot.

[0126] The robot workspace visualization construction system based on digital twins can implement the steps in the robot workspace visualization construction method based on digital twins in the above embodiments, and can achieve the same technical effect. Refer to the description in the above embodiments, which will not be repeated here.

[0127] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0128] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0130] The embodiments of the present invention have been described above with reference to the accompanying drawings. The disclosed embodiments are merely preferred embodiments of the present invention. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many equivalent changes in form without departing from the spirit and scope of the claims of the present invention, and all such changes are within the protection scope of the present invention.

Claims

1. A method for visually constructing a robot workspace based on digital twins, characterized in that, Includes the following steps: S1. Construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range. S2. Construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is the set of points in the space that can be reached by the motion description reference point of the robot end effector. S3. Based on the total kinematic transformation matrix of the robot end effector, generate initial point cloud workspace data according to a preset point cloud generation method, wherein the preset point cloud generation method is a Monte Carlo method or a simulation-driven debugging method. S4. Extract the point cloud spatial contour from the initial point cloud workspace; S5. Generate precise point cloud workspace data based on the point cloud spatial contour; S6. Use a preset correction method to perform boundary point cloud correction on the precise point cloud workspace data to obtain corrected point cloud workspace data. S7. Perform surface reconstruction based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot. Step S5 includes the following sub-steps: S51. Divide the point cloud spatial contour into multiple point cloud regions of equal volume using a three-dimensional rasterization method. S52. Fill each point cloud region with a sphere; S53. Traverse the sphere in each point cloud region and determine whether the center position of the current sphere has the corresponding point cloud spatial contour. If not, calculate the analytical joint angle of the center position of the sphere using a preset robot inverse kinematics method. If the analytical joint angle is within the range of the joint angle interval, add the pose data corresponding to the analytical joint angle as new point cloud data into the initial point cloud workspace. S54. After completing the sphere traversal of all the point cloud regions, output the current initial point cloud workspace as the precise point cloud workspace data.

2. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, The total kinematic transformation matrix of the robot's end effector is defined as follows: It satisfies the following relation (1): （1）； in, Indicates the number of robot axes. The robot's first Axis and the first Homogeneous transformation matrix for transformations between axes.

3. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, The preset point cloud generation method is the Monte Carlo method, and step S3 includes the following sub-steps: S31. Set the preset number of iterations and start the iteration; S32. Obtain a random joint angle by randomly selecting and adding values, and assign the random joint angle to the joint angle, defining the random joint angle as... The random value is The range of the joint rotation angle is [ ， ], then the random joint angle The following relation (2) must be satisfied: （2）； S33. Obtain the homogeneous transformation matrix corresponding to the joint rotation angle, and substitute the homogeneous transformation matrix into the total kinematic transformation matrix of the robot end effector to obtain the corresponding end effector pose data; S34. Update the current iteration count and determine whether the preset iteration count has been reached. If not, return to step S32. If so, output all the end pose data as the initial point cloud workspace data.

4. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, The preset point cloud generation method is a simulation-driven debugging method, and step S3 includes the following sub-steps: S31. Set the preset number of iterations and start the iteration; S32. Assign a value to the joint angle using a fixed step size, and define the joint angle as... The step size is The range of the joint rotation angle is [ ， The joint angle is then... The following relation (3) must be satisfied: （3）； S33. Obtain the joint rotation angle to drive the physical robot and obtain the corresponding end-effector pose data; S34. Update the current iteration count and determine whether the preset iteration count has been reached. If not, return to step S32. If so, output all the end pose data as the initial point cloud workspace data.

5. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, Step S4 includes the following sub-steps: S41. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and the points in the initial point cloud workspace data that intersect with the latitude and longitude lines are marked to obtain the first scanning result. S42. Based on the latitude and longitude line scanning method, the initial point cloud workspace data is layered from left to right to obtain multiple planes perpendicular to the X-axis in the spatial coordinate system. Point cloud scanning is performed in each layer of latitude and longitude lines to obtain and mark the points in the initial point cloud workspace data that intersect with the latitude and longitude lines to obtain the second scanning result. S43. Integrate and deduplicate the first scan result and the second scan result to obtain the point cloud spatial contour.

6. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, The preset robot inverse kinematics method is at least one of the algebraic method, geometric method, and numerical method.

7. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, The preset correction method is latitude and longitude line scanning, and step S6 specifically involves: Based on the latitude and longitude scanning method, the precise point cloud workspace data is layered from bottom to top to obtain multiple planes perpendicular to the Z-axis in the spatial coordinate system. Point cloud scanning is performed in each plane, and points in the precise point cloud workspace data that do not intersect with the latitude and longitude lines are deleted.

8. The method for constructing a robot workspace visualization based on digital twins according to claim 1, characterized in that, Step S7 includes the following sub-steps: S71. Randomly select a point cloud data from the precise point cloud workspace data as the starting point pi by a random selection method; S72. Using the starting point pi as the center of the sphere, search other point cloud data within a fixed radius a to obtain a set of points in the spherical neighborhood. S73. Connect the starting point pi to any point qi in the spherical domain point set, and generate two circumcircles with radius a that pass through both points pi and qi. If there are no other point cloud data in the precise point cloud workspace data in either circumcircle, then connect points pi and qi to generate model edges. S74. Select point qi+1 from the set of points in the spherical domain, return to step S73, and continue until all points in the set of points in the spherical domain are selected and step S73 is executed. S75. Select the next starting point pi+1, return to step S72, and continue until all points in the precise point cloud workspace data have been selected and steps S72-S74 are executed. S76. Output all the model edges as the visualization data model.

9. A robot workspace visualization construction system based on digital twins, characterized in that, include: The digital twin modeling module is used to construct a corresponding digital twin model of the robot based on the motion joint structure of the physical robot. The motion joint structure includes joint rotation angle, link offset, link length, link torsion angle and joint rotation angle range. The kinematic modeling module is used to construct the total kinematic transformation matrix of the robot end effector based on the workspace of the robot digital twin model. The workspace of the robot digital twin model is a set of points in the space that can be reached by the motion description reference point of the robot end effector. The point cloud generation module is used to generate initial point cloud workspace data according to the total kinematic transformation matrix of the robot end effector and a preset point cloud generation method, wherein the preset point cloud generation method is the Monte Carlo method or the simulation-driven debugging method. The point cloud extraction module is used to extract the point cloud spatial contour from the initial point cloud workspace; A point cloud precision module is used to generate precise point cloud workspace data based on the point cloud spatial contour. The point cloud correction module is used to perform boundary point cloud correction on the precise point cloud workspace data using a preset correction method to obtain corrected point cloud workspace data. The visualization module is used to perform surface reconstruction based on the corrected point cloud workspace data to obtain a visualized data model of the workspace of the physical robot. Specifically, the point cloud precision module is used to perform the following steps: S51. Divide the point cloud spatial contour into multiple point cloud regions of equal volume using a three-dimensional rasterization method. S52. Fill each point cloud region with a sphere; S53. Traverse the sphere in each point cloud region and determine whether the center position of the current sphere has the corresponding point cloud spatial contour. If not, calculate the analytical joint angle of the center position of the sphere using a preset robot inverse kinematics method. If the analytical joint angle is within the range of the joint angle interval, add the pose data corresponding to the analytical joint angle as new point cloud data into the initial point cloud workspace. S54. After completing the sphere traversal of all the point cloud regions, output the current initial point cloud workspace as the precise point cloud workspace data.

Images