SafeMobility: An IoT- based System for safer mobility using machine learning in the age of COVID-19

ABSTRACT

In the face of the COVID-19 pandemic and the absence of a vaccine or an effective treatment against the virus, the available studies show that today, the most effective measure for prevention continues to be social distancing. In this sense, in this article, we focus implementing an IoT-based System for safer mobility in the age of COVID-19 using machine learning called SafeMobility. This system has been designed to monitor in real-time the social distancing between people and control the capacity in common interior spaces via a multilayer architecture that integrates IoT, fog, and cloud solutions. To control the capacity safely, we have detected the location of people using machine learning models. We have trained and evaluated these models from a data set containing the RSSI signals of the different surrounding WiFi networks obtained via a portable IoT device. Besides, this portable device integrated with a high precision laser sensor has also been used to detect the distance between people, thus avoiding potential infections. Also, we have exploited the advantages of fog computing to perform data processing and analysis in a fog node using the machine learning model that presented the highest accuracy in the evaluation. In case of non-compliance with the allowed social distance or the established peak capacity, alert messages are sent via a lightweight and optimal protocol in using IoT applications. A web application hosted on a cloud server receives the information from the fog node in real-time and dynamically displays the congestion sites in the environment. Our experiments demonstrate the effectiveness of the system to determine the position of the people with an accuracy of 91%. (C) 2021 The Authors. Published by Elsevier B.V.