DUKE UNIVERSITY, NORTH CAROLINA, USA and EGYPT-JAPAN UNIVERSITY OF SCIENCE AND TECHNOLOGY, ALEXANDRIA, EGYPT. UnLoc (Unsupervised indoor Localization) application that can be accessed through smart phones. This application maximizes the use of invisible landmarks in indoor environments that smart phones can sense through their built-in sensors. The use of GPS technology is very advantageous in outdoor environments. Therefore, it lacks in providing GPS-like accuracy and consistency in indoor environments.

The telecommunications technology has been constantly searching for a solution in providing a GPS-like accuracy in indoor environments. Achieving an accurate indoor localization has been a major challenge by the telecommunication industry. This led to a vigilant campaign in finding a solution for this technological need that will benefit a lot of people.

Romit Roy Choudhury and his research team at Duke’s Pratt School of Engineering developed an application that could help parents locate their children in malls or for shoppers to easily locate stores in malls. This application can also help people in navigating hospitals or any complex indoor environments.

The application uses the same concept on how ordinary people utilize landmarks in outdoor environments when giving directions. The smart phones’ built-in sensors can easily take advantage of invisible landmarks in indoor environments. These landmarks are possibly distinct motion signatures made by stairwells or elevators. Since the smart phones can detect dead spots or motion where 3G or WiFi are missing.

When an invisible landmark is sensed, the smart phone can now infer its location. Using motion sensors through compasses, accelerometers and gyroscopes the path is being tracked. There will be occasions that the tracking is inaccurate but as the smart phones hit new landmarks, they automatically correct their current location.

Unlike GPS, UnLoc is less battery-hungry and allows users to continuously track locations throughout the day. It does not require wardriving where every location has to be visited and calibrated in order to create a database of per-location fingerprints – a process that is costly since periodic updates are required. Most importantly, the application is recursive. Meaning, it starts with zero knowledge but as it learns over time it becomes more and more accurate in a given indoor environment.