Privacy-Preserving BLE Scanning for Population Estimation to Mitigate the Spread of COVID-19

Abstract

Real-time monitoring of population density at specific locations while ensuring anonymity could contribute to slowing the spread of COVID-19 through reducing densely areas. In this paper, we design and deploy sensors and base stations at specific locations to monitor the communication from nearly devices installed COCOA App and count the number of devices to estimate the population density. Our sensors can also measure population with distances via signal strengths and the estimation accuracy is increasing as the increase of COCOA App users. Note that we count the number of devices only, while neither concerning the communication content nor collecting personal information. Our system has been widely accepted and deployed at more than 200 places both on campus and off campus. Finally, we propose a machine learning based population prediction method with high population prediction accuracy through expanding supervising dataset via Newton interpolation.