Optimal calibration method of cylindrical grinding tool coordinate system based on gradient descent method

Abstract

The application provides a cylindrical grinding tool coordinate system optimization calibration method based on a gradient descent method, and relates to the technical field of robot grinding, which comprises the following steps: firstly, obtaining the position information of a stylus tip in a robot base coordinate system in space; then, driving the grinding tool to contact the stylus tip at different poses by using the robot and recording the joint angle of the robot at this time; obtaining the initial tool matrix, the position information of the stylus tip in the robot base coordinate system and the joint angle by using the SolidWorks software, and obtaining the position information of the stylus tip in the tool coordinate system; calculating the predicted polishing wheel radius; combining the real polishing wheel radius to calculate the root mean square error; and using the gradient descent method with the value of the root mean square error as a loss function to update the tool matrix. After the tool calibration method is used, the root mean square error is reduced from 22.3546 mm to 0.0694 mm compared with that before calibration, and the method is simple in steps, convenient to operate and does not need to rely on additional expensive equipment.

Description

Technical Field

[0001] This invention relates to the field of robotic grinding technology, and in particular to an optimal calibration method for the coordinate system of cylindrical grinding tools based on the gradient descent method. Background Technology

[0002] With the rapid development of science and technology, robots are being used more and more widely in various fields. Robots can undertake many complex and tedious tasks, thereby greatly improving production efficiency and precision. In modern industrial production, robots typically need to be equipped with various tools on their end flanges, such as grippers, welding guns, grinders, and spraying equipment, enabling them to perform a variety of practical operations such as handling, assembly, welding, grinding, and spraying.

[0003] To ensure that robots can perform various tasks accurately and efficiently, the position and orientation of the tools mounted on the robot's end effector flange must be accurately determined before actual operation. This step, known as tool calibration, is a crucial aspect of robot applications. The accuracy of tool calibration directly affects the robot's operational precision and work efficiency.

[0004] With the continuous advancement of technology and the rapid development of various fields, the requirements for tool positioning accuracy are increasing. In high-precision industries such as aerospace, automobile manufacturing, and electronic equipment production, even minute tool deviations can lead to a decline in product quality or even defective products. For example, in the aerospace field, the processing requirements for aircraft parts are extremely stringent. If the calibration of welding or grinding tools is inaccurate, it may lead to dimensional deviations in components, affecting the overall performance and safety of the aircraft. In the automobile manufacturing process, the precision of body welding and painting directly affects the quality and appearance of the vehicle; any minute error can affect the assembly accuracy of parts and the performance of the entire vehicle.

[0005] Currently, various tools are widely used in different scenarios, but existing tool calibration technologies have some obvious shortcomings. For example, some calibration methods are cumbersome and require a lot of manpower and time, which greatly affects work efficiency. Moreover, some traditional calibration technologies are lacking in accuracy and cannot meet the increasingly higher requirements for tool performance precision. Summary of the Invention

[0006] To address the technical problems of low efficiency and complex process in the calibration of industrial robot tools in existing technologies, this invention provides an optimal calibration method for the coordinate system of cylindrical grinding tools based on the gradient descent method.

[0007] The technical solutions provided by the embodiments of the present invention are as follows: First aspect: The coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent provided in this invention includes: S1. Place the stylus at a fixed position on the worktable; S2. Obtain the coordinates of the stylus tip in the robot's base coordinate system; S3. Install the grinding tool, control the robot to drive the grinding tool to contact the needle tip in different positions, and record the joint angle of the robot in different positions; S4. Using SolidWorks software, obtain the initial tool matrix, the joint angles, and the coordinates of the needle tip in the robot base coordinate system to calculate the position information of the needle tip in the tool coordinate system. S5. Calculate the predicted grinding wheel radius using the position information of the needle tip in the tool coordinate system under different postures, and use the predicted grinding wheel radius to calculate the error index at this time by substituting it into the root mean square error formula. S6. Calculate the root mean square error between the predicted grinding wheel radius and the actual grinding wheel radius. The root mean square error result is used as the evaluation standard. If it is less than the threshold, the calibration ends. Otherwise, the gradient descent method is used to update the tool matrix and continue with the S4 operation.

[0008] Preferably, the step S2 of obtaining the coordinates of the stylus tip in the robot's base coordinate system specifically includes: the DH modeling method can obtain the transformation matrix of the end effector coordinate system relative to the base coordinate system. The transformation matrix includes a rotation matrix and a position vector. The position vector represents the position of the origin of the end effector coordinate system relative to the base coordinate system. Then, the stylus tip can be contacted by the center point of the end effector flange. By using the DH modeling method in combination with the robot's joint angle at this time, the position of the stylus tip in the base coordinate system can be obtained.

[0009] Preferably, the step S5 to determine the predicted grinding wheel radius specifically includes: the grinding tool is cylindrical, and after modeling the grinding tool, the radius of the cylindrical grinding tool is calculated using the Y-axis and Z-axis coordinates of the needle tip in the tool coordinate system.

[0010] Preferably, in S6, gradient descent is an iterative method used in optimization and machine learning algorithms, mainly for finding the minimum (or maximum) value of a function. It minimizes (or maximizes) the objective function by gradually adjusting the parameters along the direction of gradient descent.

[0011] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: (1) In this invention, after using the optimal calibration method for the coordinate system of cylindrical grinding tools based on gradient descent proposed in this invention, the root mean square error is reduced from 22.3546 mm to 0.0694 mm, and the attitude error and position error of the grinding tool coordinate system are greatly reduced. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating the coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent, provided in an embodiment of the present invention. Figure 2 A schematic diagram of a tool calibration system for a coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent, provided in an embodiment of the present invention. Figure 3 A schematic diagram of a grinding tool model for a coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent, provided in an embodiment of the present invention. Figure 4 A schematic diagram of a grinding wheel model for a coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent, provided in an embodiment of the present invention. Figure 5 The tool tip point cloud information in the tool coordinate system before and after calibration is provided by the gradient descent method for the optimal calibration method of the cylindrical grinding tool coordinate system in the embodiments of the present invention. Detailed Implementation

[0014] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0015] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0016] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0017] Reference manual attached Figure 1 The diagram shows a flowchart of the coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent provided in an embodiment of the present invention.

[0018] This invention provides a method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent. This method may include the following steps: S1. Place the stylus at a fixed position on the worktable; S2. Obtain the coordinates of the stylus tip in the robot's base coordinate system; S3. Install the grinding tool, control the robot to drive the grinding tool to contact the needle tip in different positions, and record the robot joint angles in different positions; S4. Using SolidWorks software, obtain the initial tool matrix, joint angles, and the coordinates of the needle tip in the robot base coordinate system, and calculate the position information of the needle tip in the tool coordinate system. S5. Calculate the predicted grinding wheel radius using the position information of the needle tip in the tool coordinate system under different postures, and use the predicted grinding wheel radius to calculate the error index at this time by substituting it into the root mean square error formula. S6. Calculate the root mean square error between the predicted grinding wheel radius and the actual grinding wheel radius. The root mean square error result is used as the evaluation standard. If it is less than the threshold, the calibration ends. Otherwise, the gradient descent method is used to update the tool matrix and continue with the S4 operation.

[0019] S2 specifically includes: S21. Without installing grinding tools, control the robot to move until the center point of the end flange contacts the tip of the stylus, and record the joint angle. S22. Using the DH parameter modeling method, substitute the joint angle at this time to obtain the position information of the needle tip in the base coordinate system; S23. Obtain the robot end effector coordinate system from the DH parameter table. Relative to the base coordinate system homogeneous transformation matrix The coordinate system transformation relationship between adjacent links of the robot is as follows: In the above formula, Indicates rotation about the X-axis , Similarly, Indicates translation along the X-axis , Similarly; S24, calculated using the following formula , In the above formula, Can be written as The position of the stylus tip in the base coordinate system is .

[0020] The root mean square error formula in S5 specifically includes: In the above formula, The radius of the grinding wheel is calculated from the tip position information in the tool coordinate system at each robot pose. This indicates the actual radius of the grinding wheel.

[0021] In S6, gradient descent is an iterative algorithm used to optimize and find the extrema of a function. The main idea of ​​gradient descent is to iteratively update the parameters, causing the value of the objective function to gradually decrease until it converges to a local minimum or a global minimum. Specifically, it includes: S61. Randomly initialize parameter vector The initial values ​​of the parameters are randomly selected or initialized through other methods; S62. Calculate the loss function For parameters gradient , The gradient represents the rate of change of the loss function at the current parameter value, and points in the direction in which the loss function increases; S63. Based on gradient and learning rate Update the parameters; the learning rate is a pre-set small positive number used to control the step size of each update. In the above formula, Indicates the learning rate. Indicates the current parameter value gradient at; S64. If the change in the loss function is small enough or the maximum number of iterations is reached, stop iterating; otherwise, return to S62. The convergence criterion is that the change in the loss function value is less than a certain threshold, or the norm of the gradient is less than a certain threshold.

[0022] The predicted radius of the grinding wheel specifically includes: The grinding tool is cylindrical. After modeling the grinding tool, the radius of the cylindrical grinding tool is calculated using the Y-axis and Z-axis coordinates of the needle tip in the tool coordinate system.

[0023] Reference manual attached Figure 2 The diagram shows a tool calibration system based on the gradient descent method for optimizing the coordinate system of a cylindrical grinding tool, as provided in an embodiment of the present invention.

[0024] The ABB robot used in the experiment was model IRB6700, and the cylindrical grinding tool was mounted on the robot's flange.

[0025] First, without installing the grinding tool, control the robot's movement until the center point of the end flange contacts the stylus tip, and record the joint angle at this point. Using the DH parameter modeling method and substituting the robot's joint angle at this time, the position information of the stylus tip in the base coordinate system can be obtained. The DH parameter table of the IRB6700 robot is shown below: Experiments were conducted using 90 sets of robot path information. Whenever the grinding tool touched the needle tip, the joint angle at that moment was saved.

[0026] according to Figure 2 It can be seen that the stylus tip is in the tool coordinate system With robot base coordinate system The location information conversion can be explained by the following formula:

[0027] Therefore, the position information of the needle tip in the tool coordinate system is as follows:

[0028] To solve for the radius of the grinding wheel using the position information of the needle tip in the tool coordinate system, it is necessary to model the grinding tool. The grinding tools of common tool-type grinding robots are usually cylindrical. Based on this feature, this invention designs a grinding tool consisting of three cylindrical wheels and an abrasive belt for research.

[0029] Reference manual attached Figure 3 The diagram shows a grinding tool model based on the gradient descent method for optimizing the coordinate system calibration of a cylindrical grinding tool, as provided in an embodiment of the present invention.

[0030] The drive wheel in the diagram is the main power source of the grinding tool, responsible for transmitting power to the abrasive belt and moving it along the set path; the tension wheel maintains a certain tension during the grinding process to ensure that the abrasive belt does not slip off during high-speed movement; the grinding wheel is the working part of the grinding tool, and the abrasive belt forms a working surface on the contact wheel.

[0031] Reference manual attached Figure 4 The diagram shows a grinding wheel model of the coordinate system optimization calibration method for cylindrical grinding tools based on gradient descent provided in an embodiment of the present invention.

[0032] The coordinate system of the grinding tool can be The model is built on the grinding wheel. The midpoint of the grinding wheel's generatrix is ​​taken as the origin of the grinding tool's coordinate system. The direction of the cylinder's axial vector is taken as the X-axis, and its Z-axis is orthogonal to the cylinder's central axis. The Y-axis can be obtained using the right-hand rule. After establishing the grinding tool model, [the following steps are taken]. Figure 4It can be seen that the radius of the grinding wheel is determined by the Y-axis and Z-axis coordinates of the point, which can be established by the following formula: in, Let y be the radius of the grinding wheel, and z be the position information of the needle tip relative to the coordinate system of the grinding tool under different robot poses.

[0033] Using the calculated grinding wheel radius as the measured value, the root mean square error (RMSE) is calculated in conjunction with the actual grinding wheel radius. The RMS error is used as a criterion: if the value is less than a set threshold, the experiment ends; otherwise, the gradient descent method is used to update the tool matrix and continue the experiment. After the experiment, the calibrated tool matrix is ​​obtained. The Euler angles, position vectors, and RMS errors before and after calibration are shown in the table below: As shown in the table above, the root mean square error value before and after calibration decreased from 22.3546 mm to 0.0694 mm, which greatly improved the positioning accuracy of the grinding tool.

[0034] Reference manual attached Figure 5 This paper illustrates the tip point cloud information in the tool coordinate system before and after calibration using the gradient descent-based coordinate system optimization calibration method for cylindrical grinding tools provided in this embodiment of the invention.

[0035] Before calibration, the needle tip cloud was scattered and showed obvious layering. After calibration, the needle tip cloud was compact and had a smooth transition, resembling a cylindrical surface, which proves the effectiveness and accuracy of the method in the embodiments of the present invention.

[0036] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: (1) In this invention, after using the optimal calibration method for the coordinate system of cylindrical grinding tools based on gradient descent proposed in this invention, the root mean square error is reduced from 22.3546 mm to 0.0694 mm, and the attitude error and position error of the grinding tool coordinate system are greatly reduced.

[0037] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0038] The following points need to be explained: (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0039] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0040] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0041] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent, characterized in that, include: S1. Place the stylus at a fixed position on the worktable; S2. Obtain the coordinates of the stylus tip in the robot's base coordinate system; S3. Install the grinding tool, control the robot to drive the grinding tool to contact the needle tip in different positions, and record the joint angle of the robot in different positions; S4. Using SolidWorks software, obtain the initial tool matrix, the joint angles, and the coordinates of the needle tip in the robot base coordinate system to calculate the position information of the needle tip in the tool coordinate system. S5. Calculate the predicted grinding wheel radius using the position information of the needle tip in the tool coordinate system under different postures, and use the predicted grinding wheel radius to calculate the error index at this time by substituting it into the root mean square error formula. S6. Calculate the root mean square error between the predicted grinding wheel radius and the actual grinding wheel radius. The root mean square error result is used as the evaluation standard. If it is less than the threshold, the calibration ends. Otherwise, the gradient descent method is used to update the tool matrix and continue with the S4 operation.

2. The method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent as described in claim 1, characterized in that, S2 specifically includes: S21. Without installing grinding tools, control the robot to move until the center point of the end flange contacts the tip of the stylus, and record the joint angle. S22. Using the DH parameter modeling method, substitute the joint angle at this time to obtain the position information of the needle tip in the base coordinate system; S23. Obtain the robot end-effector coordinate system from the DH parameter table. Relative to the base coordinate system homogeneous transformation matrix The coordinate system transformation relationship between adjacent links of the robot is as follows: In the above formula, Indicates rotation about the X-axis , Similarly, Indicates translation along the X-axis , Similarly; S24, calculated using the following formula , In the above formula, Can be written as The position of the stylus tip in the base coordinate system is: .

3. The method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent as described in claim 1, characterized in that, The root mean square error formula in S5 specifically includes: In the above formula, The radius of the grinding wheel is calculated from the tip position information in the tool coordinate system at each robot pose. This indicates the actual radius of the grinding wheel.

4. The method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent as described in claim 1, characterized in that, The gradient descent method in S6 specifically includes: S61. Randomly initialize parameter vector The initial values ​​of the parameters are randomly selected or initialized through other methods; S62. Calculate the loss function For parameters gradient , The gradient represents the rate of change of the loss function at the current parameter value, and points in the direction in which the loss function increases; S63. Based on gradient and learning rate Update the parameters, where the learning rate is a pre-set small positive number. Used to control the step size of each update. In the above formula, Indicates the learning rate. Indicates the current parameter value gradient at; S64. If the change in the loss function is small enough or the maximum number of iterations is reached, stop iterating; otherwise, return to S62. The convergence criterion is that the change in the loss function value is less than a certain threshold, or the norm of the gradient is less than a certain threshold.

5. The method for optimizing and calibrating the coordinate system of a cylindrical grinding tool based on gradient descent according to claim 3, characterized in that, The predicted grinding wheel radius specifically includes: The grinding tool is cylindrical. After modeling the grinding tool, the radius of the cylindrical grinding tool is calculated using the Y-axis and Z-axis coordinates of the needle tip in the tool coordinate system.

Images