Real-Time Learning and Monitoring System in Fighting Against Sars-Cov in a Private Indoor Environment

Abstract

The SARS-CoV-2 virus has posed formidable challenges that must be tackled through scientific and technological investigations on each environmental scale. This research aims to learn and report about the current state of user activities, in real-time, in a specially designed private indoor environment with sensors in infection transmission control of SARS-CoV-2. Thus, a real-time learning system that evolves and updates with each incoming piece of data from the environment is developed to predict user activities categorized for remote monitoring. Accordingly, various experiments are conducted in the private indoor space. Multiple sensors, with their inputs, are analyzed through the experiments. The experiment environment, installed with microgrids and Internet of Things (IoT) devices, has provided correlating data of various sensors from that special care context during the pandemic. The data is applied to classify user activities and develop a real-time learning and monitoring system to predict the IoT data. The microgrids were operated with the real-time learning system developed by comprehensive experiments on classification learning, regression learning, Error-Correcting Output Codes (ECOC), and deep learning models. With the help of machine learning experiments, data optimization, and the multilayered-tandem organization of the developed neural networks, the efficiency of this real-time monitoring system increases in learning the activity of users and predicting their actions, which are reported as feedback on the monitoring interfaces. The developed learning system predicts the real-time IoT data, accurately, in less than 5 milliseconds and generates big data that can be deployed for different usages in larger-scale facilities, networks, and e-health services.