Increasing optical pose estimation accuracy via freeform design and its application to hand-eye calibration

Area of Science:

• Optics and Photonics
• Robotics and Automation
• Manufacturing Engineering

Background:

• Accurate component pose estimation is crucial for robot-assisted assembly of complex optical systems.
• Conventional optical components present ill-conditioned wavefront-based pose estimation due to their geometry.
• Existing methods struggle with the precision required for advanced optical manufacturing.

Purpose of the Study:

• To enhance wavefront-based pose estimation accuracy for optical components using freeform optics design.
• To develop a method for parameterizing component surfaces with a freeform surface model for optimization.
• To demonstrate the practical feasibility and improved accuracy of the proposed approach in robotic applications.

Main Methods:

• Derivation of an optimization problem to design freeform optical surfaces.
• Parameterization of component surfaces using a predetermined freeform surface model.
• Simulation of various optical designs to compare pose estimation accuracy.
• Application to hand-eye calibration of a wavefront sensor in robotics.

Main Results:

• Demonstrated significant improvement in wavefront-based pose estimation accuracy through freeform optics.
• Simulation results show superior performance compared to conventional optical designs.
• Successful hand-eye calibration of a wavefront sensor using the enhanced pose estimation.
• Validation of practical feasibility with additively manufactured freeform lenses and industrial robots.

Conclusions:

• Freeform optics design is an effective strategy to overcome limitations in wavefront-based pose estimation.
• The proposed method offers a pathway to higher precision in robot-assisted optical assembly.
• The approach is validated through simulations and practical implementation, showing its potential for industrial applications.