Scholarlink

Abstract <p>This study designs a refractive index (RI) sensor based on Fano resonance (FR) using plasmonic nanostructure comprising a metal-insulator-metal (MIM) waveguide coupled with square and ring resonators. Using the finite-difference time-domain (FDTD) method, the sensor is shown to realize high sensitivity (1300 nm/RIU) and a high figure of merit, confirming its suitability for RI sensing applications. Beyond the sensor design itself, the main contribution of this work is the development of a machine learning framework, based on an adaptive neuro-fuzzy inference system (ANFIS), for accurately and efficiently predicting the FR wavelength from geometric parameters. Compared with some generally used regression methods through 5-fold cross-validation, the proposed ANFIS model consistently yields lower prediction error and greater stability, indicating its ability to capture the nonlinear relationship between design parameters and sensor response. This machine learning-based approach reduces computational cost by approximately 98% relative to conventional FDTD simulations, which can facilitate the design and optimization of plasmonic sensors.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Neuro-fuzzy modeling and prediction of a high-sensitivity Fano resonance plasmonic sensor

  • 期刊论文

  • 发表日期：

    2026年7月30日

查看全文
  • Mohammad Reza Rakhshani,
  • Fatemeh Kazemi

摘要

Abstract

This study designs a refractive index (RI) sensor based on Fano resonance (FR) using plasmonic nanostructure comprising a metal-insulator-metal (MIM) waveguide coupled with square and ring resonators. Using the finite-difference time-domain (FDTD) method, the sensor is shown to realize high sensitivity (1300 nm/RIU) and a high figure of merit, confirming its suitability for RI sensing applications. Beyond the sensor design itself, the main contribution of this work is the development of a machine learning framework, based on an adaptive neuro-fuzzy inference system (ANFIS), for accurately and efficiently predicting the FR wavelength from geometric parameters. Compared with some generally used regression methods through 5-fold cross-validation, the proposed ANFIS model consistently yields lower prediction error and greater stability, indicating its ability to capture the nonlinear relationship between design parameters and sensor response. This machine learning-based approach reduces computational cost by approximately 98% relative to conventional FDTD simulations, which can facilitate the design and optimization of plasmonic sensors.

Graphical abstract