Stereo vision-based Kinematic calibration method for the Stewart platforms.

Area of Science:

• Robotics and Control Systems
• Mechanical Engineering
• Metrology

Background:

• Stewart platforms are crucial for industrial applications requiring high precision.
• Kinematic calibration is essential for ensuring the accuracy and performance of Stewart platforms.
• Existing methods for identifying geometric parameter errors have limitations in effectiveness and accuracy.

Purpose of the Study:

• To improve the effectiveness of the least squares algorithm for identifying geometric parameter errors in Stewart platforms.
• To develop a novel dimensionless error model tailored to the structural characteristics of Stewart platforms.
• To propose and validate a stereo vision-based measurement method for determining the 6-degree-of-freedom (DOF) pose of the moving platform.

Main Methods:

• Investigation of an applicable dimensionless error model based on Stewart platform structural characteristics.
• Proposal of a novel stereo vision-based measurement technique for 6-DOF pose acquisition.
• Simulation and experimental validation of the dimensionless error model and calibration method on a prototype.

Main Results:

• The dimensionless error model demonstrated efficiency in identification simulations.
• Kinematic calibration experiments on a prototype yielded significant error reduction.
• Position error decreased to 0.261 mm (89.720% improvement), and orientation error decreased to 0.051° (90.351% improvement).

Conclusions:

• The proposed dimensionless error model and stereo vision method effectively improve the kinematic calibration accuracy of Stewart platforms.
• The enhanced accuracy significantly boosts the motion orbit performance for industrial applications.
• This research provides a robust solution for achieving high-precision industrial robotic systems.