Design and Analysis of 2D Photonic Biosensor with ML for Respiratory Virus Detection

BIOSENSOR WITH ML FOR RESPIRATORY VIRUS DETECTION

Abstract

In this work, the photonic biosensor is designed and envisioned with a Machine learning technique for detecting five viruses commonly affecting the respiratory system. The sensor is based on two-dimensional hexagonal categories in a photonic crystal defect engineering environment. Using the Finite Difference Time Domain (FDTD) method and Plane Wave Expansion (PWE) technique the two structural model designs are illustrated and defined according to the wavelength shift observed throughout the detection process. The analytes can be injected into the pore and get bounded well to attain optimization. The novelty of the sensor is determined by comparing the outcome in terms of sensitivity, quality factor and accuracy with the existing work. The sensitivity and quality factor is 584nm/RIU and 9734 respectively. The Machine Learning algorithm called naïve Bayes classifier is used to calculate accuracy. This algorithmic technique defines distinguishing normal and virus cells that rely on extracted parameters from the sensor design. The high-quality sensor structure combined with the classifier provides a classification of types of virus detection and is compared with the existing effort. The responses are noted with high accuracy. In the end, the graphical user interface has been wrapped over for the better readability of the results attained.