A double model compensation control method for a six-joint robot

Abstract

This invention discloses a dual-model compensation control method for a six-joint robot, comprising the following steps: Based on the robot's kinematic theory model, a low-precision robot kinematic fitting model and a high-precision robot kinematic actual model are established through a parameter identification process; a specified motion pose is input into the low-precision robot kinematic fitting model, and the rotation angles of each joint axis are solved using inverse motion; the obtained rotation angles of each joint axis are input into the high-precision robot kinematic actual model, and the actual pose corresponding to the specified motion position is solved using forward motion; after obtaining the deviation between the actual pose and the specified motion pose, the specified motion pose is compensated to obtain a corrected target pose; the corrected target pose is input into the low-precision robot kinematic fitting model, and the rotation angles of each joint axis after compensation are solved using inverse motion. This invention can quickly complete control point interpolation calculations within a stable period, laying the foundation for robots to achieve refined operations.

Description

Technical Field

[0001] This invention relates to industrial robots, and in particular to a dual-model compensation control method for a six-joint robot. Background Technology

[0002] Industrial robots are multi-degree-of-freedom, multi-functional, and reprogrammable general-purpose machines that play a vital role in many fields such as electronics, mechanics, and automation. Industrial robots offer advantages such as high precision, good stability, and large load capacity, improving production efficiency and quality while reducing the labor intensity of workers. The development of industrial robots plays a crucial role in improving labor productivity in manufacturing, reducing labor intensity and enterprise production costs, enhancing the international competitiveness and quality of products, improving working conditions and the labor environment, reducing environmental pollution, and achieving energy conservation and emission reduction.

[0003] Industrial robots suffer from poor absolute positioning accuracy due to unavoidable factors such as manufacturing geometric parameter errors during processing and assembly, the flexibility of links and joints, and backlash in reducers. Among these, geometric parameter errors are the primary source of end-effector errors. Calibration and corresponding compensation algorithms are effective ways to improve the absolute positioning accuracy of industrial robots. To enable industrial robots to perform more precise and detailed operations, and to allow offline programming simulation programs to be directly applied in the field, while ensuring consistency between the robot's actual geometric kinematic model and the model in the simulation environment, it is necessary to design a kinematic method to improve the robot's absolute positioning accuracy and compensate for related errors.

[0004] Chinese patent publication number CN109746915A discloses a kinematic method for improving the absolute positioning accuracy of industrial robots. This method primarily involves creating a theoretical model and using an identification method to obtain the robot's DH geometric parameter errors. Then, it solves the inverse kinematics problem according to the theoretical model, substitutes the DH model with geometric errors into the forward kinematics calculation, obtains the Cartesian space pose error, and then uses the Jacobian matrix to inversely solve for and compensate for joint variable deviations. This is achieved through iterative correction of the robot's joint variables by accepting acceptable robot end-effector pose data errors. This patent mainly utilizes the Jacobian matrix to inversely solve for position errors to obtain joint errors and iteratively compensates for them, avoiding the difficulty of finding analytical solutions for DH models with geometric errors, thus effectively improving the absolute positioning accuracy of industrial robots. However, the iterative process of joint error compensation has unpredictable convergence times, which negatively impacts high-frequency point interpolation processes.

[0005] Chinese patent publication number CN114161425A discloses an error compensation method for industrial robots. This method primarily involves solving the inverse kinematics of a theoretical model, substituting it into the actual DH model for forward kinematics calculations to obtain the first compensation position. The robot's motion space is then divided into a grid, and the grid containing the first compensation position is determined. Based on the errors of each vertex in this grid, spatial interpolation is used to obtain the error at the first compensation position within that grid. Finally, combining the first compensation position with the error at that point yields the second compensation position. This second compensation position is then used for error compensation, effectively improving the absolute positioning accuracy of the industrial robot. However, the robot must pre-measure and record the vertex position errors of the spatial grid, placing high demands on the robot parameter calibration process. Summary of the Invention

[0006] Purpose of the invention: The purpose of this invention is to provide a dual-model compensation control method for a six-joint robot, thereby solving the problems of high parameter measurement requirements and uncontrollable calculation iteration in existing error compensation methods.

[0007] Technical solution: The present invention provides a dual-model compensation control method for a six-joint robot, comprising the following steps:

[0008] S1: Based on the robot kinematics theoretical model Kp, a low-precision robot kinematics fitting model Kp1 and a high-precision robot kinematics actual model Kp2 are established through the parameter identification process.

[0009] S2: Input the specified motion pose X into the low-precision robot kinematics fitting model Kp1, and use the inverse motion to solve for the rotation angle Q of each joint axis;

[0010] S3: Input the rotation angle Q of each joint axis into the high-precision robot kinematics model Kp2, and use the forward motion to solve for the actual pose Xr corresponding to the specified motion position;

[0011] S4: After obtaining the deviation Xe between the actual pose Xr and the specified motion pose X, compensate for the specified motion pose to obtain the corrected target pose Xt.

[0012] S5: Use the corrected target pose Xt as input to the low-precision robot kinematic fitting model Kp1, and use inverse motion to solve for the rotation angle Qt of each joint axis after compensation.

[0013] Step S1 specifically involves:

[0014] S11: Establish the robot's kinematic theoretical model Kp using the Standard-DH mode, which includes the following parameters: Z-axis rotation angle θ1~θ6, Z-axis translation distance D1~D6, X-axis translation distance A1~A6, and X-axis rotation angle α1~α6.

[0015] S12: The parameter identification method is used to identify and fit the nine link lengths of A1, A2, A3, A4, A5, D2, D3, D4, and D5, the five link angle parameters of α1, α2, α3, α4, and α5, and the four origin angles of θ2, θ3, θ4, and θ5, to obtain the high-precision robot kinematics model Kp2.

[0016] S13: First, based on the identification results of the Kp2 model, reset and compensate the four robot origin angles θ2, θ3, θ4, and θ5; then, based on the theoretical model Kp, use the parameter identification method to identify and fit the four link lengths A1, A2, A3, and D4.

[0017] Step S2 specifically involves:

[0018] S21: The robot theoretical model Kp corrects the four link lengths of A1, A2, A3, and D4, and the resulting Kp1 model has axes 4, 5, and 6 intersecting at one point, satisfying the Pieper criterion that the serial manipulator has an analytical solution for inverse kinematics.

[0019] S22: The inverse kinematics analytical expression Q=F(X) is established using the low-precision robot kinematics fitting model Kp1;

[0020] S23: Substitute the specified motion pose X into the expression obtained in step S22 to obtain a closed solution;

[0021] S24: When there is more than one valid solution obtained in step S23, the one with the smallest angle difference should be selected as the solution for this time, referring to the robot's current position or the previous solution result.

[0022] The specific steps of step S3 are as follows: Based on the high-precision robot kinematics model Kp2, create the forward kinematics calculation equation Xr=G(Q), substitute the joint angles Q, and calculate the actual pose Xr.

[0023] Step S4 specifically involves:

[0024] S41: The deviation calculation method is Xe=X-Xr. The Euler angle rotation sequence used in the attitude difference calculation is: fixed-axis XYZ rotation.

[0025] S42: The method for calculating the corrected target pose is Xt=Xe+X, which compensates for the deviation between the specified motion pose X and the actual pose Xr, thus obtaining the corrected target pose Xt.

[0026] Step S5 specifically involves:

[0027] S51: Substitute the specified corrected target pose Xt into the expression Qt=F(Xt) obtained in step S22 to obtain a closed solution;

[0028] S52: When there is more than one valid solution obtained in step S51, the one with the smallest angle difference should be selected as the solution for this time, referring to the robot's current position or the previous solution result.

[0029] A computer storage medium storing a computer program that, when executed by a processor, implements the aforementioned dual-model compensation control method for a six-joint robot.

[0030] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described dual-model compensation control method for a six-joint robot.

[0031] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0032] 1. Based on the characteristic that the spatial pose error of the low-precision model Kp1 and the high-precision model Kp2 is continuously and uniformly distributed, this invention uses a one-step iterative compensation method to effectively correct the actual end position of the robot. When the robot corrects Kp1 and Kp2 in real time through the built-in algorithm, the inverse kinematics of the Kp1 model still has an analytical solution, while the Kp2 model always records the true state of the robot model. This process can quickly complete the interpolation calculation of control points within a stable period, laying the foundation for the robot to achieve refined operations.

[0033] 2. During robot operation, only the parameters of the Kp2 model are compensated and corrected in real time. Since the Kp2 model only participates in forward kinematics calculation, the compensation process does not need to consider whether the inverse kinematics of the Kp2 model has an analytical solution. The parameters of the Kp2 model can be adjusted arbitrarily, increasing the applicability of the real-time compensation and correction process. Attached Figure Description

[0034] Figure 1 This is a flowchart of the steps of the method described in this invention;

[0035] Figure 2 This is a structural diagram of a six-joint robot;

[0036] Figure 3 This is a comparison chart of the robot's end-effector position error before and after kinematic compensation. Detailed Implementation

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, a dual-model compensation control method for a six-joint robot includes the following steps:

[0039] S1: Based on the robot kinematics theoretical model Kp, a low-precision robot kinematics fitting model Kp1 and a high-precision robot kinematics actual model Kp2 are established through the parameter identification process.

[0040] S2: Input the specified motion pose X into the low-precision robot kinematics fitting model Kp1, and use the inverse motion to solve for the rotation angle Q of each joint axis;

[0041] S3: Input the rotation angle Q of each joint axis into the high-precision robot kinematics model Kp2, and use the forward motion to solve for the actual pose Xr corresponding to the specified motion position;

[0042] S4: After obtaining the deviation Xe between the actual pose Xr and the specified motion pose X, compensate for the specified motion pose to obtain the corrected target pose Xt.

[0043] S5: Use the corrected target pose Xt as input to the low-precision robot kinematic fitting model Kp1, and use inverse motion to solve for the rotation angle Qt of each joint axis after compensation.

[0044] This example demonstrates an end-effector position compensation method for industrial robots. The experimental system used a six-axis industrial robot, employed a laser tracker to identify geometric error parameters, and validated the method using a robot controller. The robot used in the experiment is shown below. Figure 2 As shown, the standard DH parameters Kp and the low-precision model Kp1 and high-precision model Kp2 used are shown in Tables 1, 2 and 3.

[0045] Table 1 Standard DH parameters Kp of the six-axis industrial robot used in the experiment

[0046] Linkage number Theta: Z-axis rotation angle D: Z-axis translation distance A: Translation distance along the X-axis Alpha: X-axis rotation angle 1 0 445 195 90 2 90 0 560 0 3 0 0 130 90 4 0 600 0 -90 5 -90 0 0 90 6 0 160.500000 0 0

[0047] Table 2 Low-precision model Kp1 of the six-axis industrial robot used in the experiment

[0048] Linkage number Theta: Z-axis rotation angle D: Z-axis translation distance A: Translation distance along the X-axis Alpha: X-axis rotation angle 1 0 445 195.591014 90 2 90 0 560.530835 0 3 0 0 130.279716 90 4 0 600.342724 0 -90 5 -90 0 0 90 6 0 160.500000 0 0

[0049] Table 3. High-precision model Kp2 of the six-axis industrial robot used in the experiment.

[0050] Linkage number Theta: Z-axis rotation angle D: Z-axis translation distance A: Translation distance along the X-axis Alpha: X-axis rotation angle 1 0 445 195.671015 90.023161 2 90 0 560.199309 -0.007673 3 0 -0.847293 129.786853 89.970650 4 0 600.237519 0.339429 -89.581658 5 -90 0.043541 -0.339365 89.967962 6 0 160.500000 0.391965 -0.001542

[0051] The kinematic steps for improving the absolute positioning accuracy of the six-axis industrial robot in the experiment are as described above. The actual effects of using the dual-model compensation control method for the six-joint robot are shown in Table 4. Figure 3 As shown.

[0052] Table 4 Comparison of absolute positioning accuracy of the six-axis robot used in the experiment before and after improvement

[0053] Position error RMS mean Maximum value Single-model Kp1 control 1.194 1.158 1.957 Dual-model Kp1 and Kp2 compensation control 0.248 0.226 0.596

Claims

1. A dual-model compensation control method for a six-joint robot, characterized in that, Includes the following steps: S1: Based on the robot kinematics theoretical model Kp, a low-precision robot kinematics fitting model Kp1 and a high-precision robot kinematics actual model Kp2 are established through the parameter identification process; S11: Establish the robot's kinematic theoretical model Kp using the Standard-DH mode, which includes the following parameters: Z-axis rotation angle θ1~θ6, Z-axis translation distance D1~D6, X-axis translation distance A1~A6, and X-axis rotation angle α1~α6. S12: The parameter identification method is used to identify and fit the nine link lengths of A1, A2, A3, A4, A5, D2, D3, D4, and D5, the five link angle parameters of α1, α2, α3, α4, and α5, and the four origin angles of θ2, θ3, θ4, and θ5, to obtain the high-precision robot kinematics model Kp2. S13: First, based on the identification results of the Kp2 model, reset and compensate the four robot origin angles θ2, θ3, θ4, and θ5; then, based on the theoretical model Kp, use the parameter identification method to identify and fit the four link lengths A1, A2, A3, and D4. S2: Input the specified motion pose X into the low-precision robot kinematics fitting model Kp1, and use the inverse motion to solve for the rotation angle Q of each joint axis; S21: The robot theoretical model Kp corrects the four link lengths of A1, A2, A3, and D4, and the resulting Kp1 model has axes 4, 5, and 6 intersecting at one point, satisfying the Pieper criterion that the serial manipulator has an analytical solution for inverse kinematics. S22: The inverse kinematics analytical expression Q=F(X) is established using the low-precision robot kinematics fitting model Kp1; S23: Substitute the specified motion pose X into the expression obtained in step S22 to obtain a closed solution; S24: When there is more than one valid solution obtained in step S23, the robot's current position or the previous solution result should be referenced, and the one with the smallest angle difference should be selected as the solution result for this time. S3: Input the rotation angle Q of each joint axis into the high-precision robot kinematics model Kp2, and use the forward motion to solve for the actual pose Xr corresponding to the specified motion position; S4: After obtaining the deviation Xe between the actual pose Xr and the specified motion pose X, compensate for the specified motion pose to obtain the corrected target pose Xt. S5: Use the corrected target pose Xt as input to the low-precision robot kinematic fitting model Kp1, and use inverse motion to solve for the compensated rotation angles Qt of each joint axis.

2. The dual-model compensation control method for a six-joint robot according to claim 1, characterized in that, The specific steps of step S3 are as follows: Based on the high-precision robot kinematics model Kp2, create the forward kinematics calculation equation Xr=G(Q), substitute the joint angles Q, and calculate the actual pose Xr.

3. The dual-model compensation control method for a six-joint robot according to claim 1, characterized in that, Step S4 specifically involves: S41: The deviation is calculated as Xe=X-Xr, and the Euler angle rotation order used in the attitude difference calculation is: fixed-axis XYZ rotation; S42: The method for calculating the corrected target pose is Xt=Xe+X, which compensates for the deviation between the specified motion pose X and the actual pose Xr, thus obtaining the corrected target pose Xt.

4. The dual-model compensation control method for a six-joint robot according to claim 1, characterized in that, Step S5 specifically involves: S51: Substitute the specified corrected target pose Xt into the expression Qt=F(Xt) obtained in step S22 to obtain a closed solution; S52: When there is more than one valid solution obtained in step S51, the one with the smallest angle difference should be selected as the solution for this time, referring to the robot's current position or the previous solution result.

5. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a dual-model compensation control method for a six-joint robot as described in any one of claims 1-4.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a dual-model compensation control method for a six-joint robot as described in any one of claims 1-4.

Images