Remote real-time wastewater surveillance by the optical sensor supported by machine learning

Abstract

In this study, a remote monitoring sewage system based on optical method is presented for the first time. The built-in wastewater surveillance system can perform autonomous monitoring with no requirement of sample collection and its transport to the laboratory. The measurement results can be obtained in real time, the work operation allows continuous or on-demand mode. The study includes design, development and application of a biofunctionalized fiber-optic sensor, as well as engaging machine learning algorithms for measured signals classification. The validated SARS-CoV-2 antibodies sensor has a measurement range of 10⁻¹² mg/mL to 10⁻¹ mg/mL. Known concentrations of Immunoglobulin G (IgG) from 10⁻⁶ mg/mL to 10⁻¹ mg/mL were added to the tested wastewater samples and then detected by the prepared optical probe. The data obtained were then processed and classified by traditional method and selected machine learning algorithms; the results obtained for the KNeighbors algorithm are Balanced Accuracy of 92.97 % and F1-score of 94.19 %. This study is of significance to establish a system that can effectively monitor the outbreak of potential infections in society.