Organic indoor location : infrastructure and applications

Abstract

We describe OIL, a system that uses the existing wireless infrastructure of a building to enable a mobile device to discover its indoor location. One of the main goals behind OIL is to enable non-expert users to contribute the data that is required for localization. Toward this we have developed (1) a server-client architecture for aggregating and distributing data; (2) a caching scheme that enables client devices to estimate indoor location; (3) a simple user interface for contributing data; and (4) a way to indicate how uncertain localization estimates are. We evaluate our system with a nine-day, nineteen-person user study that took place on campus as well as a deployment of the system at an off-campus long-term specialized care facility. We also describe how to use a person's indoor location trace (i.e. the rooms they had visited and the times of each visit) to build a content-based recommendation system for academic seminars. Such a system would learn about a user's preferences implicitly, placing no burden on the user. We evaluate a prototype recommendation system based on data gathered from a user study in which participants ranked seminars.