Kinematical error analysis and autonomous calibration of a 5PUS-RPUR parallel robot

Area of Science:

• Robotics
• Mechanical Engineering
• Control Systems

Background:

• Parallel robots require precise kinematic calibration for absolute accuracy.
• Conventional methods struggle with complex error parameter coupling, leading to suboptimal calibration.
• Existing techniques may converge to local optima, limiting calibration effectiveness.

Purpose of the Study:

• To propose a novel self-calibration methodology for parallel robots using a global optimization strategy.
• To enhance the absolute accuracy and performance of parallel robots.
• To address the limitations of conventional kinematic calibration methods.

Main Methods:

• Established inverse kinematics using screw theory for a 5PUS-RPUR parallel robot.
• Performed sensitivity analysis via finite difference to identify critical error sources.
• Utilized farthest point sampling for uniform measurement point selection.
• Constructed a genetic algorithm (GA) objective function integrating actuator and platform pose errors.
• Applied a penalty function approach for non-linear constraints.

Main Results:

• The proposed global optimization method significantly improved robot positional accuracy across the workspace.
• Sensitivity analysis effectively screened and eliminated negligible error sources.
• Farthest point sampling ensured comprehensive workspace coverage for measurements.
• The GA successfully integrated multiple error sources for robust calibration.

Conclusions:

• The novel self-calibration methodology offers superior performance compared to existing state-of-the-art approaches.
• The global optimization strategy effectively overcomes local optima issues in parallel robot calibration.
• This method provides a robust and accurate solution for enhancing parallel robot kinematic accuracy.