ProTrack: Detecting Proximity and Trajectory from Passive Wireless Traces of Mobile Devices

In this paper, we propose the ProTrack system to detect proximity levels and extract movement patterns to infer social relationships, which are key factors in many urban computing applications. In ProTrack, the system passively collects wireless broadcasting packets from the WiFi or Bluetooth-enabled mobile devices by multiple wireless scanners deployed in the region of interest. The received signal strength extracted from these packets is used to infer the devices' spatial and movement features. The proposed ProTrack mobility framework investigates the trajectory level, namely companion, follow, independent, and the proximity level, namely immediate, near, far for a pair of users. Based on the output results, the system can generate a graph-based proximity map and trajectory map to infer social distance and social relationship without depending on complicated infrastructure and localization.