KnockAround: Location Based Service via Social Knowledge

With the proliferation of location awareness in smartphones, location-based services (LBSs), such as finding nearby Sushi restaurants, have become popular. In traditional LBS applications, the user´s location is sent from her smartphone to an LBS server over the Internet and the LBS server then serves the information from its database. We observe that people who share space-time context may have overlapping needs for location-based information. Based on this observation, in this paper, we propose a pure P2P-based, pull-type LBS application for smartphones, which exploits the location information already uncovered by the surrounding people (including strangers) during their day-to-day visits, to serve the need for location information of a user. Our application is free from a single point of failure. The database of location information in our system is distributed over people´s smartphones and the users themselves update the database dynamically without the need of centralized maintenance. Also our application can serve location information that is volatile and highly idiosyncratic to a particular locality, and thus a central location server is unlikely to contain it. We present the detailed architecture, algorithms, implementation, and survey-based evaluation of our LBS application along with a discussion on the potential caveats and their possible solutions in using our system. We also propose an incentive mechanism to fight against free-riders and evaluate the mechanism using real-life data and simulation.