Hierarchical Temporal Mobility Analysis with Semantic Labeling

By recording observed WiFi signals over time, it is possible for the mobile phone to deduce salient locations and mobility patterns of the user. To complicate matters, fluctuations of WiFi hotspots and unreliability of mobile phone antenna create false or missed readings. We propose a family of efficient and robust algorithms to perform the tasks of location identification, mobility inference, and localization. Our algorithms are able to handle real-life noisy readings, and construct a hierarchy of salient locations providing a multiresolution model of the environment.