A tcp calibration method of industrial robot based on uniform attitude distribution

Abstract

The invention discloses a TCP calibration method of an industrial robot based on the uniform distribution of postures, which belongs to the technical field of calibration of end tools of industrial robots. The invention uses the "six-point method" of the industrial robot as the preliminary calibration basis to create the initial measurement point configuration; uses the mechanical repulsion iterative method to construct virtual points evenly distributed on the spherical surface with the fixed reference point as the center of the sphere, and adjusts the virtual points group by group. Robot, so that the center of the sixth axis end of the virtual robot is at each virtual point; finally adjust the actual robot to the configuration of each measurement point after screening, and apply the least square spherical fitting method to solve the final TCP calibration result. This method solves the randomness and non-uniformity of the robot attitude selection in the calibration of the current fixed reference point method, makes the robot attitude evenly distributed around the fixed reference point at each measurement point, and maximizes the robot attitude between the measurement points. The degree of difference can effectively improve the calibration accuracy.

Description

technical field [0001] The invention belongs to the technical field of industrial robot terminal tool calibration, and in particular relates to a TCP calibration method for industrial robots based on uniform posture distribution. Background technique [0002] With the rapid development of my country's automated manufacturing industry, labor costs continue to rise, and industrial robots are more and more widely used in the field of industrial automated manufacturing. In practical applications, the robot completes various tasks through tools installed on the end flange, and the offset of the Tool Central Point (TCP) relative to the end position is mostly unknown or inaccurate Yes, robot tool TCP calibration is the process of calculating the position identification of the tool end relative to the robot end coordinate system. Usually, the offset is accurately calculated by a certain method without the use of external measuring tools. The accuracy of tool calibration directly af...

AI Technical Summary

Problems solved by technology

During the implementation of the fixed reference point method, the key factor affecting the calibration accuracy is the difference between the robot attitude when multiple TCP points coincide with the fixed reference point. However, the current "four-point method" or "six-point method" is practical During operation, the selection distribution of the robot posture by the staff is random, especially when the four points are distributed on the side of the reference point, the concentration of the distribution positions limits the range of the robot’s posture, resulting in a reduction in the distance between the measurement points. The degree of difference between attitudes, thus leading to an increase in the calibration error

[0006] In order to solve the calibration error problem of the "four-point method", the Chinese invention patent application "a TCP calibration method for industrial robots" (application publication number CN106502208A) discloses a multi-point TCP calibration method. Determine the equations, and then use the least squares method to achieve tool calibration. The algorithm is simple to implement, and the calibration error is reduced to a certain extent. However, this method still does not solve the problem that the difference between the attitudes cannot be optimized due to the randomness of the robot attitude selection distribution. , since the coefficient matrix of the overdetermined equation system is composed of the difference between the attitude matrix of the robot end at different measurement points, when the attitude distribution of the robot is uneven, the condition number of the matrix will be increased to form an ill condition, which will cause errors in the calibration results

Images