Abstract

Coronavirus disease 2019 (COVID-19) virus is an infectious disease which has spread globally since 2019, resulting in an ongoing pandemic. Since it is a new virus, it takes some time to develop a vaccine against it. Until then, the best way to prevent the fast spread of the virus is to enable the proper social distancing and isolation or containment to identify potential patients. Since the virus has up to 14 days of the incubation period, it is important to identify all the social interactions during this period and enforce social isolation for such potential patients. However, proper social interaction tracking methods and patient prediction methods based on such data are missing for the moment. This paper focuses on tracking the social interaction of users and predict the infection possibility based on social interactions. We first developed a BLE (Bluetooth Low Energy) and GPS based social interaction tracking system. Then, we developed an algorithm to predict the possibility of being infected with COVID-19 based on the collected data. Finally, a prototype of the system is implemented with a mobile app and a web monitoring tool. In addition, we performed a simulation of the system with a graph-based model to analyze the behaviour of the proposed algorithm and it verifies that self-isolation is important in slowing down the disease progression.