Industrial robot digital twin teaching method and system

Abstract

The application provides an industrial robot digital twin teaching method and system, and the industrial robot digital twin teaching method comprises the following steps: S1: proportionally manufacturing an end drag teaching tool according to the shape and size of an end effector; S2: dragging the end drag teaching tool and acquiring the pose information of the end drag teaching tool in real time; S3: constructing a digital twin teaching model, wherein the digital twin teaching model comprises a robot body digital twin model and an end effector digital twin model; S4: taking the pose information as a driving function, driving the end effector digital twin model and dragging the robot body digital twin model to move, so as to acquire joint data in the robot body digital twin model in real time; and S5: obtaining robot inverse kinematics solution according to the joint data. According to the scheme, the teaching work of the industrial robot can be completed by independently dragging the end drag teaching tool, and the traditional drag teaching problem of the industrial robot is overcome.

Description

Technical Field

[0001] This invention relates to the field of industrial robot teaching technology, and in particular to a digital twin teaching method and system for industrial robots. Background Technology

[0002] Currently, industrial robot teaching methods are mainly divided into teach pendant teaching, offline teaching, and drag-and-drop teaching. Teach pendant teaching requires the collection of a large amount of point information, making the teaching task arduous, resulting in low production efficiency. Furthermore, it places high demands on operators using teach pendant control technology, hindering robot adoption. Offline teaching requires operators to be familiar with offline software programming, leading to high operational complexity. Dragging-and-drop teaching offers significant advantages in terms of operation and safety, avoiding the various shortcomings of traditional teaching methods, and is a key technology with promising applications in robot research. However, research on drag-and-drop teaching technology still faces challenges such as high cost, poor scalability, complex models, and difficulties in market application and promotion.

[0003] Therefore, in order to reduce the cost of drag-and-teach robots, current drag-and-teach methods are mainly aimed at lightweight robots, and the application scope of drag-and-teach for actual industrial robots or redundant robots is very limited. Summary of the Invention

[0004] Based on the problems existing in the prior art, the present invention aims to solve the technical problem that the existing drag teaching method is mainly for light robots, while drag teaching of heavy industrial robots or redundant robots is difficult to achieve.

[0005] This invention provides a digital twin teaching method for industrial robots, used for trajectory planning of industrial robots, comprising the following steps:

[0006] S1: Prototype the end effector drag teach pendant according to the external dimensions of the end effector;

[0007] S2: Drag the end effector along the working path of the end effector and use a binocular vision positioning method to obtain the pose information of the end effector during the dragging process in real time;

[0008] S3: Construct a digital twin teaching model, which includes a robot body digital twin model and an end effector digital twin model constructed based on the robot body and the end effector respectively, wherein the end effector digital twin model is integrated into the end of the robot body digital twin model;

[0009] S4: Using the pose information as a driving function, the digital twin model of the end effector reproduces the working path and drags the digital twin model of the robot body to move, and acquires the joint data in the digital twin model of the robot body in real time;

[0010] S5: Process the joint data to obtain the inverse kinematics solution of the robot body, and use the inverse kinematics solution to perform offline programming of the robot body, thus completing the teaching work of the robot body.

[0011] According to an embodiment of the present invention, step S2, obtaining the pose information of the end effector during the dragging process, includes:

[0012] S21: Establish a target surface coordinate system X on the end-effector teaching pendant. M Y M Z M O M Establish a reference coordinate system X at the zero position of the end-drag teach pendant. F Y F Z F O F Establish the workpiece coordinate system X at the workpiece machining position. G Y G Z G O G ;

[0013] S22: Construct the workpiece coordinate system X G Y G Z G O G The base coordinate system X of the robot body W Y W Z W O W The transformation matrix T1, and the reference coordinate system X F Y F Z F O F With respect to the workpiece coordinate system X G Y G Z G O G The transformation matrix T2;

[0014] S23: When dragging the end effector to move, obtain the target surface coordinate system X of the end effector. M Y M Z M O M Relative to the reference coordinate system X F Y F Z F O F The first pose information P i The first pose information P i Includes translation distance and rotation angle information;

[0015] S24: Calculate the first pose information P based on the transformation matrix T2.i In the workpiece coordinate system X G Y G Z G O G The second pose information P in i Then, the second pose information P is calculated based on the transformation matrix T1. i In the base coordinate system X W Y W Z W O W The actual pose information P in i ".

[0016] According to an embodiment of the present invention, when the digital twin model of the end effector and the digital twin model of the robot body are fused in step S3, the target surface coordinate system X... M Y M Z M O M The tool coordinate system X of the end effector of the robot body T Y T Z T O T coincide.

[0017] According to an embodiment of the present invention, when constructing the digital twin teaching model in step S3, the workpiece coordinate system X is used as the basis. G Y G Z G O G and the base coordinate system X W Y W Z W O W The relationship between the robot body and the end effector is established to create a digital twin model.

[0018] According to an embodiment of the present invention, in step S23, the target surface coordinate system X of the end effector is obtained. M Y M Z M O M Relative to the reference coordinate system X F Y F Z F O F The first pose information P i At that time, the means used are visual positioning systems or inertial navigation systems.

[0019] According to an embodiment of the present invention, in step S3, the digital twin teaching model further includes an obstacle digital twin model, which is constructed proportionally or in an expanded manner based on the obstacles in the working areas of the robot body and the end effector.

[0020] According to an embodiment of the present invention, in step S4, when the end effector digital twin model is driven to reproduce the working path and drag the robot body digital twin model to move, the robot body digital twin model needs to avoid the obstacle digital twin model.

[0021] This invention also provides an industrial robot digital twin teaching system for implementing the aforementioned industrial robot digital twin teaching method, comprising:

[0022] A binocular vision positioning module is used to acquire the pose information of the end effector in real time during the dragging process based on a binocular vision positioning method when dragging the end effector along the working path of the end effector; the end effector is manufactured proportionally according to the external dimensions of the end effector.

[0023] The model building module is used to build a digital twin model of the robot body and a digital twin model of the end effector based on the robot body and the end effector, respectively, wherein the digital twin model of the end effector is integrated into the end of the digital twin model of the robot body;

[0024] The path reproduction module can use the pose information obtained by the binocular vision positioning module as a driving function to enable the end effector digital twin model to reproduce the working path and drag the robot body digital twin model to move.

[0025] The data acquisition module is used to acquire joint data of the robot body digital twin model in real time when the end effector digital twin model reproduces the working path and drags the robot body digital twin model to move.

[0026] The programming processing module is used to process the acquired joint data to obtain the inverse kinematics solution of the robot body, and to perform offline programming of the robot body using the inverse kinematics solution.

[0027] According to one embodiment of the present invention, the binocular vision positioning module includes a first camera and a second camera. The first camera is movably disposed on the robot body, and the second camera is fixedly disposed at the end of the robot body and is positioned toward the working direction of the end effector.

[0028] According to one embodiment of the present invention, the first camera is used to acquire multiple images at different positions, and combined with a three-dimensional reconstruction algorithm to construct an obstacle model within the working range of the industrial robot, the obstacle model being used for obstacle avoidance planning of the industrial robot.

[0029] According to one embodiment of the present invention, when the industrial robot is working, the second camera is used to acquire motion trajectory images of the end effector in real time, compare the acquired motion trajectory images with a preset motion trajectory, calculate the trajectory deviation, and adjust the motion trajectory of the end effector according to the trajectory deviation.

[0030] The beneficial effects of this invention are:

[0031] This invention provides a digital twin teaching method and system for industrial robots. After dragging the end effector and acquiring its pose information in real time, the acquired pose information is used as a driving function to drive the digital twin model of the end effector to reproduce the working path. Simultaneously, the robot body digital twin model is dragged to move, thereby acquiring the joint data of the robot body digital twin model. This overcomes the difficulty of drag teaching in traditional industrial robots or redundant robots and improves teaching efficiency. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments or prior art, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of the structure of an industrial robot in the existing technology;

[0034] Figure 2 This is a schematic diagram of a digital twin teaching system for industrial robots provided by the present invention.

[0035] Figure 3 This is a schematic diagram of the end-drive teaching pendant in an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of the structure of the fixing frame in an embodiment of the present invention;

[0037] Figure 5 This is a schematic diagram of the end-effector teaching pendant placed on the fixed frame in an embodiment of the present invention;

[0038] Figure 6 This is a flowchart illustrating a digital twin teaching method for industrial robots provided by the present invention;

[0039] Figure 7 This is a schematic diagram of the working state of writing with a brush in an embodiment of the present invention;

[0040] Figure 8This is an angle data diagram of the six joints of the robot body obtained in an embodiment of the present invention. Detailed Implementation

[0041] The following descriptions of the embodiments are made with reference to the accompanying illustrations to illustrate specific embodiments in which the invention can be implemented.

[0042] This invention provides a digital twin teaching system for industrial robots, used for trajectory planning of industrial robots. The industrial robot 1000 includes a robot body 1001 and an end effector 1002 disposed at the end of the robot body 1001; its structure is as follows: Figure 1 As shown, the system includes a binocular vision positioning module 100, a path reconstruction module 200, a data acquisition module 300, and a programming processing module 400. The binocular vision positioning module 100 is used to acquire, in real time, the pose information of the end effector 2000 during the dragging process based on a binocular vision positioning method when the end effector 2000 is dragged along the working path of the end effector 1002. The structure of the end effector 2000 is as follows... Figure 3 As shown, it is manufactured to scale according to the external dimensions of the end effector 1002, and the surface of the end effector teaching pendant 2000 is provided with an easily identifiable target surface, specifically, the target surface is located on the top surface of the end effector teaching pendant 2000; the path reproduction module 200 can use the pose information acquired by the binocular vision positioning module 100 as a driving function to drive the end effector digital twin model to reproduce the working path and drag the robot body digital twin model to move, and the end effector digital twin model is integrated into the end of the robot body digital twin model; the data acquisition module 300 is used to acquire the joint data of the robot body digital twin model in real time when the end effector digital twin model reproduces the working path and drags the robot body digital twin model to move; the programming processing module 400 is used to process the acquired joint data to obtain the robot inverse kinematics solution, and use the inverse kinematics solution to perform offline programming of the robot body, that is, to complete the teaching work of the industrial robot 1000.

[0043] Specifically, the binocular vision positioning module 100 includes a first camera 101 and a second camera 102. The first camera 101 is movably mounted on the robot body, for example, by means of a linear guide or a rotary guide. The second camera 102 is fixedly mounted at the end of the robot body and is positioned toward the working direction of the end effector 1002.

[0044] More specifically, the robot body 1001 adopts a six-axis serial structure, with the joints sequentially being the first joint (shoulder joint), the second joint (upper arm joint), the third joint (forearm joint), the fourth joint (wrist joint 1), the fifth joint (wrist joint 2), and the sixth joint (wrist joint 3). The rotating guide rail of the first camera 101 is mounted on the housing of the third joint, and the rotation line of the rotating guide rail coincides with the axis of the third joint. The first camera 101 is then mounted on the rotating guide rail. In this embodiment, the axis of the first camera 101 forms a 45° angle with the horizontal plane. The second camera 102 is mounted on the connecting flange between the end effector 1002 and the robot body 1001, and the axis of the second camera 102 is parallel to the axis of the sixth joint. The lens orientation is consistent with the working direction of the end effector 1002, ensuring that the second camera 102 can clearly capture the movement trajectory of the end effector 1002 and the target object within the working area.

[0045] When the industrial robot performs a reset operation, the positions and orientations of the first camera 101 and the second camera 102 reach a preset initial state, at which point they constitute a binocular vision positioning module. By calibrating the first camera 101 and the second camera 102, their internal parameters and relative external parameters are obtained.

[0046] The calibration process is as follows: After the industrial robot is reset, a calibration board with known three-dimensional coordinates is placed in the workspace of the industrial robot. Images of the calibration board are captured by the first camera 101 and the second camera 102 respectively. Using a binocular vision calibration method, the intrinsic parameters (including focal length, principal point coordinates, distortion parameters, etc.) of the first camera 101 and the second camera 102, as well as the relative extrinsic parameters (including translation vector and rotation matrix) between them, are calculated to complete the calibration.

[0047] The external parameters are combined using the rotation matrix R and the translation vector t to represent the points (X, Y, X) in the world coordinate system. w ,Y w Z w Convert ) to a point in the camera coordinate system (X) c ,Y c Z c ):

[0048]

[0049] When it is necessary to locate a target object in space, the first camera 101 and the second camera 102 simultaneously acquire images of the target object. Through image preprocessing, feature extraction, and matching, the corresponding points of the target object in the images of the two cameras are obtained. Based on the disparity formula of binocular vision, the three-dimensional coordinates (X, Y, Z) of the target object are calculated, realizing spatial positioning measurement.

[0050] Utilizing the parallax principle of binocular vision, target objects in space can be located and measured to obtain their three-dimensional coordinate information. When the initial state of the binocular vision positioning module 100 cannot meet the detection range, a new detection range can be achieved through the movement of the robot body 1001. At this time, the external parameters of the binocular vision module at the new camera position, as well as the motion relationship between the first camera 101 and the second camera 102, can be determined by the joint displacement θi and the motion of the camera on the guide rail. Then, the point (X) in the world coordinate system is... w ,Y w Z w Convert ) to a point in the camera coordinate system (X) c ,Y c Z c ):

[0051]

[0052] Where T=f(θ1,θ2,...,θ n θi represents the pose matrix of the camera coordinate system; θi is the angle (applicable to the rotational joints of the robot body) or displacement (applicable to the translational joints of the robot body); α represents the rotation angle or displacement distance of the first camera 101. The essential matrix, fundamental matrix, and baseline length of the new position can be calculated based on the positional relationship between the first camera 101 and the second camera 102 and the robot body.

[0053] Please continue to refer to Figure 3 Preferably, in order to facilitate dragging the end-effector teach pendant 2000, the end-effector teach pendant 2000 is made of a lightweight material, such as foam, plastic, or wood.

[0054] Please refer to Figure 4 In some embodiments, to facilitate the placement of the end effector 2000, the industrial robot digital twin teaching system further includes a mounting bracket 500 for placing the end effector 2000. The mounting bracket 500 has a placement slot or hole 501 whose shape matches the shape of the end effector. Figure 5 The diagram shows the end-effector teaching pendant 2000 placed on the mounting bracket 500.

[0055] Please continue to refer to Figures 3-5 In some embodiments, to facilitate determining the initial position of the end effector 2000, the industrial robot digital twin teaching system further includes a positioning block 600 and a positioning sensor 700. The positioning block 600 is fixedly disposed on the surface of the end effector 2000 and is fixedly disposed in the placement slot or placement hole 501. The positioning sensor 700 is used to identify the positioning block 600, thereby identifying the end effector 2000.

[0056] It is understood that when the end effector 2000 is placed on the mounting bracket 500 and the positioning sensor 700 detects the positioning block 600, it indicates that the end effector 2000 is in the initial position. Through the recognition function of the positioning sensor 700, the end effector 2000 can be quickly adjusted to the initial position, that is, the zeroing operation of the end effector 2000 is completed.

[0057] This invention also provides a digital twin teaching method for industrial robots, implemented based on the aforementioned digital twin teaching system for industrial robots, the process of which is as follows: Figure 6 As shown, it includes the following steps:

[0058] S1: Make an end effector teaching pendant proportionally to the external dimensions of the end effector of the robot body;

[0059] S2: Drag the end effector teaching pendant 1 along the working path of the end effector, and use binocular vision positioning to obtain the pose information of the end effector teaching pendant in real time during the dragging process;

[0060] S3: Construct a digital twin teaching model, which includes a robot body digital twin model and an end effector digital twin model constructed based on the robot body and the end effector, respectively. The end effector digital twin model is fused to the end of the robot body digital twin model, i.e., the position of the end effector. The fusion model of the robot body digital twin model and the end effector digital twin model is as follows: Figure 5 As shown.

[0061] S4: Using the pose information as a driving function, the digital twin model of the end effector reproduces the working path and drags the digital twin model of the robot body to move, and acquires the joint data in the digital twin model of the robot body in real time;

[0062] S5: Process the joint data to obtain the inverse kinematics solution of the robot body, and use the inverse kinematics solution to perform offline programming of the robot body, thus completing the teaching work of the industrial robot.

[0063] Compared to existing technologies, the present invention provides a digital twin teaching method for industrial robots. After dragging the end effector teach pendant 2000 and acquiring its pose information in real time, the acquired pose information of the end effector teach pendant 2000 is used as a driving function to drive the end effector digital twin model to reproduce the working path in a virtual environment. At the same time, the robot body digital twin model is dragged to move, thereby acquiring the joint data of the robot body digital twin model. This overcomes the difficulty of dragging and teaching the traditional robot body and improves teaching efficiency.

[0064] The following combination Figures 6-8 The above process will be explained in detail.

[0065] S1: The end effector is manufactured in proportion to the external dimensions of the end effector of the robot body; the end effector is designed and manufactured in a simplified 1:1 scale according to the external features of the end effector of the robot body.

[0066] S2: Drag the end effector teaching pendant 2000 along the working path of the end effector, and use binocular vision positioning to obtain the pose information of the end effector teaching pendant in real time during the dragging process, including:

[0067] First, establish the target surface coordinate system X on the end-effector teaching pendant. M Y M Z M O M Establish a reference coordinate system X at the initial position after the teach pendant is dragged back to zero at the end. F Y F Z F O F Establish the workpiece coordinate system X at the machining position of workpiece 200. G Y G Z G O G Specifically, in this embodiment, the target surface coordinate system X M Y M Z M O M The X coordinate system of the end effector tool relative to the robot body T Y T Z T O T Coincidence; the initial position of the end effector is the zero point position detected by the end effector in the fixture 500, calibrating the target surface coordinate system X. M Y M Z M O MRelative to the reference coordinate system X F Y F Z F O F The pose information consists of spatial geometric position and attitude information, eliminating the initial error after the system starts.

[0068] Secondly, construct the workpiece coordinate system X. G Y G Z G O G The robot's own base coordinate system X W Y W Z W O W The transformation matrix T1, and the reference coordinate system X F Y F Z F O F With respect to the workpiece coordinate system X G Y G Z G O G The transformation matrix T2; specifically, in this embodiment, let the origin of the reference coordinate system O be... F Relative to the origin O of the workpiece coordinate system G If the distances the object moves in the X, Y, and Z directions are a, b, and c, respectively, then:

[0069]

[0070] Then, when the end-capture teach pendant is moved, the reference coordinate system X is used. F Y F Z F O F Using the static coordinate system, with the target surface coordinate system X... M Y M Z M O M Using a moving coordinate system, obtain the target surface coordinate system X of the end-effector dragging teach pendant. M Y M Z M O M Relative to the reference coordinate system X F Y F Z F O F The first pose information P in i The first pose information P i This includes translation distance and rotation angle information. Specifically, in this embodiment, the target surface coordinate system X... M Y M Z M O M Origin MFor reference, translation distance information can be obtained using the target surface coordinate system X. M Y M Z M O M The rotation angle information can be obtained by using the coordinate axes as a reference;

[0071] Next, the first pose information P is calculated based on the transformation matrix T2. i In the workpiece coordinate system X G Y G Z G O G The second pose information P in i Then, the second pose information P is calculated based on the transformation matrix T1. i In the base coordinate system X W Y W Z W O W The actual pose information P in i ".

[0072] S3: Construct a digital twin teaching model, which includes a robot body digital twin model and an end effector digital twin model constructed based on the robot body and the end effector, respectively. The end effector digital twin model is fused to the end of the robot body digital twin model, i.e., the position of the end effector. The fusion model of the robot body digital twin model and the end effector digital twin model is as follows: Figure 5 As shown.

[0073] After obtaining the actual pose information P i Subsequently, in the dynamic analysis software, digital twin models of the robot body, workpiece, obstacles, and end effector are established based on the actual relationship between the robot body and the workpiece under actual working conditions, thus constructing the digital twin teaching model. The obstacle digital twin model is constructed proportionally or at an expanded scale based on the obstacles within the working areas of the robot body and the end effector. In this digital twin teaching model, the end effector digital twin model is integrated into the end of the robot body digital twin model, i.e., the origin O of the tool coordinate system of the robot body digital twin model. T The origin O of the target surface coordinate system of the digital twin model of the end effector M coincide.

[0074] During operation, the first camera 101 rotates on the robot body 1001 around the axis of the robot's second joint at a preset rotation speed and angle range, acquiring real-time images of the surrounding environment. Image processing techniques are used to detect the presence of new obstacles in the robot's reachable area. For example, inter-frame difference or optical flow methods are used to compare two consecutive frames to detect moving areas. For detected obstacles, multiple images acquired by the first camera 101 at different positions are combined with 3D reconstruction algorithms, such as stereo vision matching algorithms or structure-from-motion (Sfm) algorithms, along with the camera's intrinsic and extrinsic parameters (the extrinsic parameters of the first camera 101 can be calculated from the joint position information of the industrial robot during rotation), to calculate the 3D coordinates of each point on the obstacle, constructing a dynamic 3D model of the obstacle. This 3D obstacle model can be used for obstacle avoidance planning of the industrial robot, for example, by using collision detection algorithms to determine the relative position of the obstacle and the industrial robot, and planning a safe movement path. The specific process is as follows: The first camera 101 acquires images at position one and position two respectively, and then performs image preprocessing and feature extraction (such as the center points of the two ends of the obstacle). The position of the obstacle is obtained based on the parallax principle. Then, the image processing technology is used to obtain the central axis and boundary line dimensions of the obstacle in the image. The sweep volume is constructed with the central axis as the sweep curve and the boundary dimensions as the appropriately enlarged size as the diameter, thereby constructing a digital twin model of the obstacle.

[0075] S4: Using the pose information as a driving function, the digital twin model of the end effector reproduces the working path and drags the digital twin model of the robot body to move, and acquires the joint data in the digital twin model of the robot body in real time;

[0076] Using the base coordinate system of the robot's main digital twin model as the static coordinate system, and the origin O of the tool coordinate system of the robot's main digital twin model... T As a moving point, apply pose information P to it. i "As a driving function, when the digital twin model of the end effector reproduces the trajectory of the end effector dragging the teach pendant 2000 in the digital twin teach model, the joint data of each joint of the robot body digital twin model can be obtained."

[0077] It is understandable that since a digital twin model of the obstacle is constructed in step S3, the digital twin model of the obstacle should not be included in the working path when step S4 is executed, so as to realize obstacle avoidance planning.

[0078] S5: Process the joint data to obtain the inverse kinematics solution of the robot body, and use the inverse kinematics solution to perform offline programming of the robot body, thus completing the teaching work of the industrial robot.

[0079] By using the data of each joint of the robot body digital twin model as a sample for digital twin data processing, the inverse kinematics solution of the robot body is obtained. This solution is then used as the joint input of the real robot body. Under actual working conditions, the robot is programmed according to the workpiece coordinate system, which enables the reproduction of the processing motion trajectory of the robot body, thereby realizing the digital twin teaching of the industrial robot.

[0080] During actual operation of the industrial robot, the second camera 102 can be used for trajectory monitoring or trajectory guidance planning. During trajectory monitoring, the second camera 102 acquires real-time images of the motion trajectory of the end effector 1002. Through image recognition and processing technology, feature points (e.g., marker points installed on the end effector 1002) are extracted. The real-time positions of the feature points are compared with their theoretical positions on the preset trajectory to calculate the trajectory deviation. When the deviation exceeds a preset threshold, an adjustment signal is sent to the industrial robot's control system. The control system adjusts the movement of each joint according to the deviation value, causing the end effector to return to the preset trajectory.

[0081] Before the industrial robot starts working, the second camera 102 acquires images of the target position in the working environment. The two-dimensional coordinates of the target position are determined by image processing and target recognition technology. Combined with the kinematic model of the industrial robot and the spatial coordinate information obtained by the binocular vision positioning module 100, the motion trajectory of the end effector 1002 from the current position to the target position is planned. The planned trajectory is sent to the control system of the industrial robot to guide the industrial robot to perform the work task.

[0082] In this embodiment of the invention, a flexible binocular vision positioning module 100 is constructed by setting a movable first camera 101 on the robot body 1001 and a second camera 102 at the end of the robot body 1001. When reset, the internal and external parameters of the binocular vision positioning module 100 can be acquired, and binocular vision positioning modules with different detection areas can be constructed by moving the first camera 101 to different specific positions, thereby realizing high-precision spatial positioning measurement, and realizing the acquisition of digital twin teaching data of industrial robots and their digital twin teaching.

[0083] On the one hand, the movable first camera 101 enables comprehensive monitoring of the industrial robot's surrounding environment and the reconstruction of 3D models of dynamic obstacles, providing favorable support for obstacle avoidance and path planning in complex environments, thus improving the industrial robot's environmental adaptability and safety. On the other hand, the second camera 102 can perform trajectory monitoring or trajectory guidance planning during operation, ensuring that the industrial robot operates accurately along a predetermined trajectory, thereby improving the reliability and stability of the operation.

[0084] Verification process

[0085] To verify an industrial robot digital twin teaching method provided by an embodiment of the present invention, in this embodiment, an example of a robot body driving a writing brush to write is used for illustration. The corresponding working state schematic diagram is as shown in Figure 7 Figure [missing number]. It is not difficult to understand that the writing brush is equivalent to the end effector in the present invention, and the paper is equivalent to the workpiece to be processed. The specific teaching process is as follows:

[0086] First step, make a writing brush teaching device according to the external dimensions of the writing brush in equal proportion.

[0087] Second step, first, establish a base coordinate system X W Y W Z W O W at the position of the industrial robot, establish a reference coordinate system X F Y F Z F O F at the initial position of the writing brush teaching device, and establish a workpiece coordinate system X G Y G Z G O G ;

[0088] Secondly, construct the transformation matrix T1 between the workpiece coordinate system X G Y G Z G O G and the base coordinate system X W Y W Z W O W , and the transformation matrix T2 between the reference coordinate system X F Y F Z F O F and the workpiece coordinate system X G Y G Z G O G ;

[0089] Then, drag the writing brush teaching device along the working path of the writing brush. For example, when the actual work of the writing brush is to write the character "中" on the paper, the working path of the writing brush is the moving trajectory of the writing brush when writing the character "中"; and while dragging the writing brush teaching device, real-time obtain the pose information of the writing brush teaching device. The partial first pose information of the obtained writing brush teaching device in the reference coordinate system X F Y F Z F O F is shown in Table 1:

[0090] Table 1. Partial pose information of the calligraphy pendant

[0091]

[0092] Then, based on transformation matrices T1 and T2, the brush pendant in the base coordinate system X is calculated. W Y W Z W O W The actual pose information obtained is shown in Table 2:

[0093] Table 2. Partial actual posture information of the calligraphy pendant.

[0094]

[0095]

[0096] The third step involves constructing a digital twin model of the robot body and a digital twin model of the brush pendant in the dynamic analysis software based on the actual positional relationship between the robot body and the brush under actual working conditions, and then integrating the digital twin model of the brush pendant into the end of the robot body.

[0097] The fourth step involves using the actual pose information of the calligraphy pendant as a driving function to drive the digital twin model of the calligraphy pendant to reproduce the working path of the calligraphy brush, and dragging the movement of the robot body's digital twin model. This allows the acquisition of each joint data of the robot body's digital twin model through the industrial robot digital twin teaching system. Specifically, in this embodiment, the robot body is a six-axis robot, and the obtained joint data is as follows: Figure 8 As shown.

[0098] In summary, the industrial robot digital twin teaching method and system provided by the embodiments of the present invention independently drags the end effector and records its pose information, and drives the digital twin model of the end effector to drag the robot body at the same time with the recorded pose information, thereby obtaining the angle data of each joint of the robot body, overcoming the drag teaching problem of large or heavy robot bodies or redundant robots, and improving teaching efficiency.

[0099] It should be noted that although the present invention has been disclosed above with specific embodiments, the above embodiments are not intended to limit the present invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the scope defined in the claims.

Claims

1. A digital twin teaching method for an industrial robot, used for trajectory planning of the industrial robot, the industrial robot comprising a robot body and an end effector disposed at the end of the robot body; characterized in that, Includes the following steps: S1: An end effector teaching pendant is manufactured proportionally to the external dimensions of the end effector, and the surface of the end effector teaching pendant is provided with a target surface for easy identification; the industrial robot digital twin teaching system also includes a mounting frame for placing the end effector teaching pendant, and the mounting frame has a placement slot or placement hole with a shape matching the shape of the end effector teaching pendant; the industrial robot digital twin teaching system also includes a positioning block and a positioning sensor, the positioning block is fixedly disposed on the surface of the end effector teaching pendant, the positioning block is fixedly disposed in the placement slot or placement hole, and the positioning sensor is used to identify the positioning block, thereby identifying the end effector teaching pendant; when the end effector teaching pendant is placed on the mounting frame and the positioning sensor identifies the positioning block, it indicates that the end effector teaching pendant is in the initial position; S2: Drag the end effector along the working path of the end effector and use a binocular vision positioning method to obtain the pose information of the end effector during the dragging process in real time; S3: Construct a digital twin teaching model, which includes a robot body digital twin model and an end effector digital twin model constructed based on the robot body and the end effector respectively, wherein the end effector digital twin model is integrated into the end of the robot body digital twin model; S4: Using the pose information as a driving function, the digital twin model of the end effector reproduces the working path and drags the digital twin model of the robot body to move, and acquires the joint data in the digital twin model of the robot body in real time; S5: Process the joint data to obtain the inverse kinematics solution of the robot body, and use the inverse kinematics solution to perform offline programming of the robot body, thus completing the teaching work of the industrial robot. Step S2 involves obtaining the pose information of the end effector during the dragging process, including: S21: Establish a target surface coordinate system X on the end-effector teaching pendant. M Y M Z M O M Establish a reference coordinate system X at the zero position of the end-drag teach pendant. F Y F Z F O F Establish the workpiece coordinate system X at the workpiece machining position. G Y G Z G O G ; S22: Construct the workpiece coordinate system X G Y G Z G O G With respect to the robot's main body base coordinate system X W Y W Z W O W The transformation matrix T1, and the reference coordinate system X F Y F Z F O F With respect to the workpiece coordinate system X G Y G Z G O G The transformation matrix T2; S23: When dragging the end effector to move, obtain the target surface coordinate system X of the end effector. M Y M Z M O M Relative to the reference coordinate system X F Y F Z F O F The first pose information P i The first pose information P i This includes the translation distance and rotation angle information for the i-th sampling period; S24: The first pose information P is calculated based on the transformation matrix T2. i In the workpiece coordinate system X G Y G Z G O G The second pose information P in i Then, the second pose information P is calculated based on the transformation matrix T1. i In the base coordinate system X W Y W Z W O W The actual pose information P in i ''.

2. The industrial robot digital twin teaching method according to claim 1, characterized in that, When the digital twin model of the end effector and the digital twin model of the robot body are fused in step S3, the target surface coordinate system X M Y M Z M O M The tool coordinate system X of the end effector of the robot body T Y T Z T O T coincide.

3. The industrial robot digital twin teaching method according to claim 1, characterized in that, In step S23, the target surface coordinate system X of the end effector is obtained. M Y M Z M O M Relative to the reference coordinate system X F Y F Z F O F The first pose information P i At that time, the means used are visual positioning systems or inertial navigation systems.

4. The industrial robot digital twin teaching method according to claim 1, characterized in that, In step S3, the digital twin teaching model further includes an obstacle digital twin model, which is constructed proportionally or in an expanded manner based on the obstacles in the working areas of the robot body and the end effector.

5. The industrial robot digital twin teaching method according to claim 4, characterized in that, In step S4, when driving the end effector digital twin model to reproduce the working path and drag the robot body digital twin model to move, the robot body digital twin model must avoid the obstacle digital twin model.

6. An industrial robot digital twin teaching system for implementing the industrial robot digital twin teaching method according to any one of claims 1 to 5, characterized in that, include: A binocular vision positioning module is used to acquire the pose information of the end effector in real time during the dragging process based on a binocular vision positioning method when dragging the end effector along the working path of the end effector; the end effector is manufactured proportionally according to the external dimensions of the end effector. The model building module is used to build a digital twin model of the robot body and a digital twin model of the end effector based on the robot body and the end effector, respectively, wherein the digital twin model of the end effector is integrated into the end of the digital twin model of the robot body; The path reproduction module can use the pose information obtained by the binocular vision positioning module as a driving function to enable the end effector digital twin model to reproduce the working path and drag the robot body digital twin model to move. The data acquisition module is used to acquire joint data of the robot body digital twin model in real time when the end effector digital twin model reproduces the working path and drags the robot body digital twin model to move. The programming processing module is used to process the acquired joint data to obtain the inverse kinematics solution of the robot body, and to perform offline programming of the robot body using the inverse kinematics solution.

7. The industrial robot digital twin teaching system according to claim 6, characterized in that, The binocular vision positioning module includes a first camera and a second camera. The first camera is movably mounted on the robot body, and the second camera is fixedly mounted at the end of the robot body and is positioned toward the working direction of the end effector.

8. The industrial robot digital twin teaching system according to claim 7, characterized in that, The first camera is used to acquire multiple images at different positions and, combined with a 3D reconstruction algorithm, to construct an obstacle model within the working range of the industrial robot. The obstacle model is used for obstacle avoidance planning of the industrial robot.

9. The industrial robot digital twin teaching system according to claim 7, characterized in that, When the industrial robot is working, the second camera is used to acquire motion trajectory images of the end effector in real time, compare the acquired motion trajectory images with the preset motion trajectory, calculate the trajectory deviation, and adjust the motion trajectory of the end effector according to the trajectory deviation.

Images