A Privacy-preserving Scheme for Passive Monitoring of People’s Flows through WiFi Beacons

Abstract

The proliferation of IoT-based services for smart cities, and especially those related to mobility, are ever becoming more relevant and gaining attention from a number of stake-holders. In our work, we tackle the problem of characterizing people movements in a urban environment by using WiFi sensors connected to the cellular network. In particular, we leverage WiFi probe requests transmitted by people’s smartphones and a machine learning approach to detect people’s flows, while preserving users’ privacy. We validate our approach through a proof-of-concept testbed deployed in the proximity of our campus area. We consider two types of devices, namely, commercial, off-the-shelf WiFi scanners and ad-hoc designed scanners implemented with Raspberry PIs. They provide different levels of visibility of the captured traffic, preserving in different ways the privacy of the people’s movements. In our current work, we investigate the different trade-offs between mobility tracking accuracy and the level of provided people’s privacy.