Greedy-Face-Greedy Routing based Human Tracking in Mobile Social Networks

Abstract Mobile phones are increasingly being equipped with hardware and software services allowing them to determine their locations, however, support for building location-based applications remains a challenging problem. The most widely used localization technology in mobile-phones is GPS, but it rarely works indoors and provides low energy efficiency. Cell-tower based localization is widely available, but can provide very poor accuracy without a fingerprint profile. Wi-Fi localization, provides reasonable accuracy, but is also much less effective in other areas. Constandache et al. proposed an Escort system to assist localizing and tracking others in a public place without requiring either GPS, Wi-Fi, war-driving, maps, or floor plans. However, the Escort system may route one person on a long path even though the person being tracked may be close by. In this paper, we will investigate the problem of better tracking paths in the Escort system. We propose a Greedy-Face-Greedy routing based human tracking algorithm to reduce the length of tracking path for every pair of users using mobile phones in mobile social networks. Through adding one seeker in the Escort, whose main work is to find the better paths between any pair of two intersections by applying Greedy-Face-Greedy routing algorithm, the localization and tracking algorithm in the Escort system is more effective than the original one. Finally, we conduct simulations of our proposed algorithm at the main campus of Temple University with different number of mobile users and duration time. The simulation results show that the human tracking performance has been greatly enhanced.