Earlier this year, Google selected the University of Minnesota’s
as its primary academic partner for Project Tango
, a high-profile indoor mapping initiative that has been compared to Google Maps. The selection came with a $1.35 million grant and a directive to explore—and expand on—the possibilities of a prototype smartphone capable of creating 3D maps of indoor spaces. Google’s only other academic partner on the project, Washington, D.C.’s George Washington University, has a much smaller role.
According to Google, the current prototype device is a “5 inch Android phone containing highly customized hardware and software designed to track the full 3D motion of the device as you hold it while simultaneously creating a map of the environment.”
The phone can take up to 250,000 spatial measurements per second to create an intricate map of its surroundings. While this technology isn’t yet available as an app on regular smartphones, part of MARS Lab’s charge is to create apps and APIs—mobile development platforms—that enable the app to be scaled down and included with non-specialized devices. Within a few years, some form of the technology will be available for download like any other Android app. The U of M lab will have played a central role in making that possible.
A major challenge will involve surmounting the technology’s requirement for two independent cameras. It’s unclear whether future versions will be able to work with a single smartphone camera, or whether devices that use it will need to have at least two vision sensors. A strict non-disclosure agreement, breach of which could jeopardize the lab’s funding, prevents MARS Lab director Stergios I. Roumeliotis from getting into such specifics about Project Tango.
A video released last month
by the MARS Lab team shows the prototype’s capabilities. Although the current version produces a somewhat slow, abstract representation of its surroundings, future iterations will create near-lifelike interior maps. Google and MARS envision three broad areas in which 3D mapping can play a role: virtual/augmented reality video games, internal navigation in unfamiliar buildings (rendering directions in malls and corporate edifices all but obsolete), and navigation aids for the visually impaired. But innovation probably won’t stop there: In a recent interview, Roumeliotis argued that “the list of potential future applications is endless.”
In addition to Roumeliotis, two MARS Lab alums who have since taken positions with Google—Joel Hesch and Esha Nerurkar—are leading the development charge. The building blocks for the project were actually laid about a decade ago, when the MARS Lab team helped create the internal navigation system, known as VINS, for NASA’s Mars landers. A loss of NASA funding for the project proved to be a blessing in disguise, as Roumeliotis’s team found that the system worked just as well for earthbound mapping and navigation.