Black-Box Optimization Method in a Motorized Optomechatronic System for Light Detection and Alignment

This project developed a mechanical system to control laser light for fiber coupling, integrating optomechanics, electronics, and algorithmic optimization for low-light alignment. Using the Gaussian beam profile as a guide, the system systematically adjusts mirrors and lenses to improve coupling efficiency. A probabilistic algorithm predicts which adjustments are most promising, balancing exploration and exploitation to reduce measurements. Using this approach, transmitted power increased from 30nW to 47.3μW during a coarse search, then to 0.48mW with Bayesian optimization, achieving 75% coupling efficiency. This demonstrates an automated method for aligning weakly linked light sources to maximize transmission.