From Mechanical Instability to Virtual Precision: Digital Twin Validation for Next-Generation MEMS-Based Eye-Tracking...

Area of Science:

• Optics and Photonics
• Microelectromechanical Systems (MEMS)
• Biomedical Engineering

Background:

• MEMS-based eye trackers are vital for medical diagnostics and human-computer interfaces.
• Physical prototype development faces limitations due to mechanical instability and recalibration needs.
• Accelerating the design-test-refine cycle is critical for advancing eye-tracking technology.

Purpose of the Study:

• To develop and validate a high-fidelity digital twin (DT) for MEMS eye trackers.
• To accelerate the design-test-refine cycle for MEMS optical systems.
• To demonstrate the DT's utility in developing advanced pupil detection algorithms.

Main Methods:

• Comparative analysis of a physical MEMS scanning system and its digital twin using a USAF 1951 test target.
• Evaluation under static and dynamic conditions to assess accuracy and identify discrepancies.
• Utilizing the validated DT to develop and test a novel pupil detection algorithm.

Main Results:

• The digital twin accurately replicated the physical system's behavior with <30 µm geometric discrepancy.
• Tracking dynamics introduced a 1 µm error in feature shift, which the DT successfully modeled.
• The DT eliminated mechanical vibration artifacts, allowing for precise parameter analysis.
• A pupil detection algorithm developed using the DT achieved 1.80 arc minutes accuracy, outperforming a commercial system.

Conclusions:

• A validated methodology for employing digital twins in rapid prototyping of optical systems has been established.
• Digital twins offer a powerful tool for overcoming physical limitations in MEMS eye tracker development.
• This approach accelerates the development of critical healthcare technologies, including advanced eye-tracking systems.