The commercialization of GPS technology has been a boon for those navigating unfamiliar city streets, highways and byways, but head inside out of sight of the GPS satellite signals and the limitations of the technology can quickly become evident. Other efforts to solve the problem involve the use of accelerometers, sometimes combined with magnetic field sensors, but a new system developed at Duke University promises to provide precise indoor localization using a different approach – detecting “invisible” landmarks.

Similar to the way in which humans provide directions using easily recognizable landmarks – turn left at the McDonalds and keep going until you get to the fountain, for example – the new system called UnLoc (short for unsupervised indoor localization) relies on “invisible” landmarks in indoor environments that can be detected by a smartphone’s sensors.

"Example landmarks could be distinct motion signatures created by elevators or stairwells, because the phone can detect motion, or certain dead spots where WiFi or 3G signals are absent,” says Roy Choudhury, associate professor of computer engineering at Duke’s Pratt School of Engineering and principal investigator for the Duke research team.

Like other indoor localization efforts, UnLoc would also make use of the smartphone’s accelerometers, compasses and gyroscopes to track the user’s path from a detected landmark. While this tracking will usually become inaccurate over time, the system will recalibrate to correct its location whenever it detects another invisible landmark.

The developers say the system doesn’t require any mapping of locations before being deployed to create a database of fingerprints – a process that would be time consuming, expensive, and need to be done periodically.

“The best part of the application is that it is recursive, which means that it starts with zero knowledge but ‘learns’ over time,” said He Wang, the lead Ph.D. student on the project. “Therefore, it becomes more and more accurate the more it is used in a given building.”

The team says the system’s reliance on energy-efficient inertial sensors also makes it less battery-hungry than GPS, allowing it to be used to track a user’s location continuously throughout the day.

After testing UnLoc in the Northgate shopping mall in Durham, North Carolina, and in the computer science and engineering buildings on Duke’s campus, the system was found to achieve an accuracy of 1.6 meters (5.2 ft) on average.

Choudhury says precise indoor localization capabilities could open up a host of mobile apps, such as shoppers getting more information about products in front of them or parents being able to locate their children in sprawling shopping malls. Locating the car in underground car parks is another potential use that comes to mind.

In addition to researchers from Duke, UnLoc’s development team included members from the Egypt-Japan University of Science and Technology and was supported by the National Science Foundation and Google.