Optimization of integrated optic micro ring resonator gas sensor using machine learning technique

Abstract

Integrated Optic Micro ring Resonator-based sensors are suitable for lab-on-chip applications due to their smaller footprint. Optical sensors are sensitive to detecting small changes in external parameters. Simultaneous detection of multiple gases present in the atmosphere is crucial for several civilian and military applications. Integrated optic micro ring resonators are promising sensing devices. In this paper, machine learning techniques are used in the classification and detection of gases for a sensor of a Micro Ring Resonator (MRR) array . In this paper machine learning techniques are used to reduce the data to be used for the analysis and improve accuracy of the sensor. Three target gases in the proposed model are Ammonia, Methane and Carbon on each ring simulated in this work. The features and influences on wavelength, transmittance, concentration of gases, and ring radius have also been analyzed. Principal Component Analysis (PCA) and K-Clustering algorithm has been used for the classification and detection of different gases. The Davies Boulden Index is calculated as 0.57 which shows the distance between the clusters. The sensor has a sensitivity of 0.35 nm/ppm.