Rapid calibration of optical phased array chips via an improved stochastic parallel gradient descent algorithm

Le résumé fourni par la source

Optical phased arrays (OPAs) have emerged as promising beam steering technologies for advanced spatial perception and free-space optical (FSO) communications. However, inherent fabrication errors inevitably introduce stochastic initial phase variations, which must be precisely calibrated to enable high-quality beam steering. In this study, we propose an improved stochastic parallel gradient descent (SPGD) algorithm, integrated with Nesterov-accelerated adaptive moment estimation (Nadam) and termed NASPGD, to achieve rapid and robust initial phase calibration. The proposed NASPGD algorithm was experimentally validated on a 32-element OPA chip. Comparative analysis against conventional SPGD, Adam-SPGD, and other modified SPGD variants demonstrates the superior performance of our approach. Notably, NASPGD effectively suppresses noise interference and escapes local optima, achieving a higher side-lobe suppression ratio with a significantly reduced number of iterations.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.