Modeling and predicting the beam properties from grating structures using deep neural network

• المؤلفون الرئيسيون: Lim, Yu Dian, Zhao, Peng, Guidoni, Luca, Likforman, Jean-Pierre, Tan, Chuan Seng
• مؤلفون آخرون: School of Electrical and Electronic Engineering
• التنسيق: Conference or Workshop Item
• اللغة: English
• منشور في: 2024
• الموضوعات: Engineering; Particle beam optics; Trapped ions
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/175559
• المؤسسة: Nanyang Technological University

In this study, the optical properties of light beam coupled out from grating structures are simulated, modeled, and predicted. Gratings with various radius of curvature (R) are designed, where optical properties of light coupled out from the grating are modeled with deep neural network (DNN). After omitting the underfitted and overfitted model, the ideally-fitted model accurately predicts the beam waist of light from R = 30 μm with average percentage error (APE) of 7.3%. At the same time, the ideally-fitted model exhibits training/validation loss APE of 2.2%, indicating that the model is well-fitted.