A Markov Model for Improving the Performance of COVID-19 Contact Tracing App

Abstract

Nowadays, the world is betting on mobile phone applications to stop the spread of deadly infectious diseases, including the COVID-19 pandemic and its new variants. Since the beginning of the COVID-19 outbreak, a group of countries has launched contact-tracing apps to stem the spread of COVID-19. The app helps health authorities to track the movements of people diagnosed with the virus, which gives a chance to isolate them rather than isolate the whole population. When two users are near each other, their phones exchange tokens via a Bluetooth connection, recording that they’ve had a close contact. However, when two or more phones send their tokens simultaneously to look for another phone, collisions may occur. Therefore, the user may not get the warning notifications even if he was near someone diagnosed with COVID-19. To overcome this problem, we propose a mechanism to improve the Bluetooth network performance. The goal of this improvement is to make the contact tracing apps more efficient. A Markov chain model is then constructed to evaluate the system performance. Numerical results demonstrate that our mechanism can significantly improve Bluetooth performance and improve the contact tracing app’s performance.