Ultra-Fast Core Mode Prediction in Plasmonic Crystal Fiber Sensor: A Machine Learning Approach

This study introduces a novel approach to improving Photonic Crystal Fiber (PCF) Surface Plasmon Resonance (SPR) sensors by incorporating machine learning (ML) models for predicting core mode index. Traditionally, identifying core phenomena in PCF SPR sensors required manual visual inspection by experts, a process that is both time-consuming and prone to errors. To address this, a dual-core PCF model is developed and used Finite Element Method (FEM) simulations to create extensive training data for ML models. The proposed PCF model shows exceptional sensitivity to wavelength changes, with a recorded sensitivity of $\mathbf{6 0 0 0} \mathrm{nm} /$ RIU, making it highly effective for accurate mode prediction. The study evaluated nine ML models, including XGBoost, Random Forest, and Decision Tree, achieving accuracies of 0.9742,0.9645, and 0.9667, respectively. By integrating ML models, the core mode identification process is significantly faster and more efficient, offering major benefits for PCF SPR research. This work not only advances PCF sensor technology but also enhances efficiency and accuracy in various SPR-based applications.