Beyond Google Glass: Creating next generation augmented reality
Todd Humphreys and his students at the University of Texas in Austin are tired of waiting for augmented reality that meets the promise of the technology we've been hearing about and seeing in science fiction for years now. So they set out to build it themselves, and presented a very rough prototype for the first time at the South By Southwest Interactive Festival (SXSW) in Austin on Friday.
The prototype isn't nearly as lightweight or wearable as Google's Glass, the augmented reality spectacles that have received lots of buzz lately over their reported consumer release planned for later this year. Instead, it's about the size of a desktop computer with a big, round antenna attached to the top of the box.
But as Humphreys told the SXSW crowd, it's not meant to be an elegantly designed consumer product, but rather a proof of concept for how good augmented reality could (and should) be.
"Google Glass is not the immersive experience I'm looking for," Humphreys told the audience in a conference room of Austin's downtown Omni Hotel.
He then laid out what he calls an augmented reality "Bill of Rights," outlining what he sees as the real promise of AR, which includes a "true three-dimensional immersive experience" with virtual elements that look and behave as if real, creating a "convincing illusion." Further, Humphreys wants to see absolute location registration (geolocation) accurate down to the centimeter, or even millimeter, no matter where a person is on the globe, outdoors and indoors.
Humphreys' final demand for this next generation of augmented reality – it should be available soon; no waiting another decade for the experience Arnold Schwarzenegger and the Terminator series introduced us to in the last century.
The assistant professor of engineering did acknowledge the loftiness of his AR ambitions and said he was willing to make a few concessions, namely that we may have to settle for a handheld AR experience in the short term rather than a totally immersive, wearable one. This is a point where he and his younger students are actually at odds – Humphreys believes it will be difficult for Google Glass and other "somewhat dorky hardware" to become socially acceptable, but his students disagree.
Humphreys also concedes that the suite of sensors in contemporary devices may be inadequate for achieving what he's laid out in his AR Bill of Rights.
There are, of course, some other glaring obstacles to achieving Humphreys' vision for modern augmented reality, including that it would require covering every inch of the earth from multiple angles with something like Google's Street View car or backpack setup. Then there's also the bandwidth demands and latency challenges for a system that would basically require synchronous up and down streaming of 3D video.
Humphreys said that his team's new prototype effectively takes a big step towards working around these issues and delivering on the promise of AR.
The key to this new approach is the fusion of two readily available technologies: carrier-phase differential GPS (CDGPS) and an inertial measurement unit (IMU), like those in smartphones that house a device's accelerometer and gyroscope.
The IMU helps place and track objects in their local environment and, fusing that data with CDGPS (which is accurate down to the centimeter), "scales and anchors [that] reference frame globally," Humpreys explained.
The result is a technique that eliminates the need to map the whole world in advance by essentially mapping just the local area on the fly (with an IMU and camera) and then placing that reference frame into a global context (via CDGPS).
Humphreys' team says that at present IMUs in our phones and tablets are not quite precise enough to get the job done and their antennas don't output the right information to pickup CDGPS. In fact, current antenna limitations present the greatest stumbling blocks in contemporary hardware, which is why the prototype includes such a relatively large and goofy antenna.
When it comes to processor power, Humphreys said that current hardware has sufficient power, but would likely be a major drain on batteries.
"Perhaps this is not the way AR will finally work out," conceded Humphreys. "But we can imagine it tethered to the back of an iPad."
Source: University of Texas