Semantic Trajectories Based Social Relationships Discovery Using WiFi Monitors

Abstract

Smart phones are configured to automatically send WiFi probe message transmissions (latter called WiFi probes) to surrounding environments to search for available networks. Prior studies have provided evidence that it is possible to uncover social relationships of mobile users by studying time and location information contained in these WiFi probes. However, their approaches miss information about transfer patterns between different locations. In this paper, we argue that places mobile users have been to should not be considered in isolation. We propose that semantic trajectory should be used to modeling mobile users and semantic trajectory patterns can well characterize users’ transfer patterns between different locations. Then, we propose a novel semantic trajectory similarity measurement to estimate similarity among mobile users. We deploy WiFi detectors in a university to collect WiFi probes, through which we collect around 20G byte data containing hundreds of millions of records. Through experimental evaluation, we demonstrate that the proposed semantic trajectory similarity measurement is effective. What is more, we experimentally show that the trajectory similarity measurement can be used to exploit underlying social networks exist in the university.

This study is supported by the Fundamental Research Funds for the Central Universities (2014ZD03-1).