Sensitivity-informed multi-objective adaptive weighting method for automatic alignment of freeform USTP systems.

Area of Science:

• Optical Engineering
• Metrology
• Image Science

Background:

• Six-degree-of-freedom (6DoF) pose errors in freeform mirrors of catadioptric ultra-short-throw projection (USTP) systems create coupled effects on geometric distortion and Modulation Transfer Function (MTF).
• This coupling complicates system alignment by shifting optimization objectives across different alignment stages.

Purpose of the Study:

• To develop a physics-guided optimization framework for the efficient and stable automatic alignment of freeform USTP systems.
• To model and leverage the relationship between pose errors and image quality degradation for improved alignment strategies.

Main Methods:

• Performed full-parameter 6DoF scanning simulations to determine the response laws of geometric distortion and MTF to pose perturbations.
• Developed a physics-guided adaptive weight allocation and stage-wise optimization method incorporating identified optical laws.
• Implemented a physics-guided optimization framework for automatic alignment.

Main Results:

• Identified stage-dependent shifts in dominant responses of geometric distortion and MTF to pose errors.
• Established stable mapping relationships between distortion modes and specific degrees of freedom.
• Achieved stable convergence from coarse to fine alignment, with final mean MTF exceeding 70% in 10 experiments.
• Reduced alignment time to approximately 10 minutes per run with significantly decreased result dispersion.

Conclusions:

• The proposed physics-guided method enables efficient and stable automatic alignment of freeform USTP systems.
• This work provides a valuable reference for modeling pose error coupling and designing alignment strategies in related optical systems.