Prediction of Fiber Bragg Gratings Characteristics from Its Design Parameters Using Deep Learning

Abstract

This research addresses the complexities and inefficiencies encountered in fabricating fiber Bragg gratings (FBGs), which are crucial for applications in optical communications, lasers, and sensors. The core challenge lies in the intricate relationship between fabrication parameters and the FBG's physical properties, making optimization time-consuming. To circumvent these obstacles, the study introduces an artificial intelligence-based approach, utilizing a neural network to predict FBG physical parameters from transmission spectra, thereby streamlining the fabrication process. The neural network demonstrated exceptional predictive accuracy, significantly reducing the parameter prediction time from days to seconds. This advancement offers a promising avenue for enhancing the efficiency and precision of FBG sensor design and fabrication. The research not only showcases the potential of artificial intelligence in revolutionizing FBG production but also contributes to the broader field of optical technology by facilitating more rapid and informed design decisions, ultimately paving the way for developing more sophisticated and sensitive FBG-based applications.