S115012 Kinematic Calibration for Six-Degree-of-freedom Parallel Kinematic Mechanism (11th report) : Kinematic Calibration Based on Position Measurement Error
Takaaki OIWA, Naoya TAZAWA, Junichi Asama
Abstract
Open-access reader
Takaaki OIWA, Naoya TAZAWA, Junichi Asama
Abstract
Open-access reader
This paper examines parameter identification for six-degree-of-freedom (6-DOF) parallel manipulators, based on coordinate error of a reference artifact measured by a touch trigger probe. A redundant passive chain with a displacement sensor, installed between the moving stage and the machine frame, is passively expanded and contracted while an end-effector or a touch trigger probe measures a three-dimensional ball plate. Moreover, balls' position can be calculated from the forward kinematics of a 6-DOF parallel manipulator consisting of six of the seven chains. Consequently, the least-squares method using a Jacobian matrix corrects 40 kinematic parameters so that seven sets of the position errors at each pose are minimized because seven combinations are possible. The above calculations were repeated until convergence in numerical simulation. The simulation showed that the parameter identification using relative coordinate errors reduced the parameter errors by 92 % from 0.826 mm_ to 0.066 mm_ even without using the redundant chain. Moreover, increase of the ball combination improved the identification accuracy.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper examines parameter identification for six-degree-of-freedom (6-DOF) parallel manipulators, based on coordinate error of a reference artifact measured by a touch trigger probe. A redundant passive chain with a displacement sensor, installed between the moving stage and the machine frame, is passively expanded and contracted while an end-effector or a touch trigger probe measures a three-dimensional ball plate. Moreover, balls' position can be calculated from the forward kinematics of a 6-DOF parallel manipulator consisting of six of the seven chains. Consequently, the least-squares method using a Jacobian matrix corrects 40 kinematic parameters so that seven sets of the position errors at each pose are minimized because seven combinations are possible. The above calculations were repeated until convergence in numerical simulation. The simulation showed that the parameter identification using relative coordinate errors reduced the parameter errors by 92 % from 0.826 mm_ to 0.066 mm_ even without using the redundant chain. Moreover, increase of the ball combination improved the identification accuracy.
Key concepts: Kinematics, Jacobian matrix and determinant, Control theory (sociology), Position (finance), Kinematic chain, Ball (mathematics), Coordinate-measuring machine, Calibration