Towards Exploiting Wi-Fi Signals From Low Density Infrastructure for Crowd Estimation

The ubiquity of wireless devices such as smartphones, tablets and laptops, has enabled sensing large crowds. This was made possible with numerous methods available that mostly listen to Bluetooth or Wi-Fi channels to observe traffic diversity, sources, and destinations. On one hand, it is clearly useful to create crowd awareness, for example to estimate the number of people and assess people flows inside buildings or in areas, with applications in disaster management, network evaluation, and human mobility modeling, as well as for individual mobile devices to assess their context. At the same time, most of these network activity monitoring methods risk compromising the privacy of the individuals being counted and possibly-deliberately or inadvertently-tracked. That is, they may leak private information about people's individual mobility patterns without their consent or even awareness. In this paper, we take a first stab at addressing the problem of privacy-preserving crowd (density) estimation by utilizing the received signal strength (RSS) of Wi-Fi signals from stationary beacons. We use management frames as an approximation of ground truth to validate our observations. We evaluate this method in a real world measurement, observing very strong correlations between the presence of over 35,000 mobile devices in a large building and Wi-Fi RSS values from stationary devices.