Audio Based Location Fingerprinting for Mobile Devices

Abstract

The main goal of the project is to enhance the context-awareness of cellular handsets by exploiting commonly available technologies/capabilities present on the devices. For instance, most cellular handsets today come with a variety of sensors and capabilities including location sensors, accelerometers, light sensors, cameras and high-fidelity speakers and microphones. By fusing information gathered from multiple sensors, cellular handsets can tailor their services based on the user's context. Location is a major component of user's context. In open outdoor environments, Global Positioning System (GPS) can provide the needed location information to infer user's context. However, location information is not easy to obtain when the user is either indoors or in obstruction-laden outdoor terrain. Our goal in this project is to exploit other resources in cellular handsets to improve the location context of users in environments where GPS is not readily available. GPS is currently the most precise estimation of location. Wi-Fi and Cell-Tower based location services are also currently available. They help determine the approximate location to the accuracy of approximately 100 meters. This is not very helpful when if we need to determine the user context within a building likes office or homes. The need to precisely determine location in an indoor environment by exploring readily available information from the surrounding is the main idea of the project. There can be several uses of trying to estimate location indoors. The mobile devices can switch user modes when the user moves across a building from office suite to a conference suite, without the intervention of the user. Also other features like automatically connecting to wireless networks or syncing up emails to the local server in the building are some of the compelling advantages.