Fiber Optic Sensor-based VOCs Prediction using Machine Learning Techniques

The detection of volatile organic compounds (VOCs) using fiber optic sensors has become an active area of research in recent years due to its potential for real-time monitoring in a variety of applications, including environmental monitoring, industrial process control, and healthcare. The application of machine learning (ML) techniques for the prediction of VOCs will intelligently automate the process. In the proposed approach, the use of a subset of ML i.e., a deep learning (DL) technique namely, one-dimensional Convolutional Neural Network (1D CNN) for the classification of VOCs namely, ethanol, isopropyl alcohol (IPA) and methanol at two different regions of wavelengths of spectral data have been explored. The spectroscopic data represent the spectra of the different concentrations of VOCs obtained through the real-time experiment. The model is subjected to the real-time spectroscopic data of VOCs generated with the help of the developed clad modified fiber optic sensor (CMFOS). The performance of the model in classifying various VOCs is validated using performance metrics and confusion matrices.