Augmented reality: a long way off?

Professor Steve Feiner teaches a class on AR each spring at Columbia University, New York. Every year, at the beginning of the course - a course which has the words “augmented reality” in the title, he asks his students how many of them have actually heard of the phrase before. In 2008, there were maybe two or three hands in the air out of a total of 15. Last spring, the number went up to almost half - proof, Pocket-lint suggests, that AR’s moment has finally arrived. Not so, says the augmented reality guru down the line from his desk in New York City.

“People are getting excited about handheld AR but we were excited years ago,” says the man who practically invented the discipline along with the likes of Blair MacIntyre, his colleague on the KARMA project in 1991 - one of the first solid examples of augmented reality in action.

“If you asked people on the street who own a smartphone about Layar and Wikitude, they’d look at you like you were crazy,” he giggles down the phone as we half wonder whether that’s exactly what Feiner’s experiment of the morning will be. He softens.

“Well, it’s not the watershed moment for AR right now but it could be a watershed moment. Have you ever heard of VisiCalc?”

VisiCalc, as it turns out, was the original “killer app”. It was a spreadsheet application for the computers, the one which made people finally realise that these machines were worth owning.

“VisiCalc came out in the 70s,” he continues, “but it still wasn’t until well into the 90s when everyone had a computer in their own home.”

It’s not that Feiner has no faith in augmented reality and its development but, like many of the top experts in the field that we’ve spoken to in AR Week, he does harbour a certain level of distaste for the wave it around and look at the dots version of apps currently available.

“The model of smartphone AR is not good enough at the moment,” he explains. “You have to switch it on, go to an app, wait for the GPS to kick in, hold it up at arms length, look stupid, get tired and, when it finally kicks in, the camera’s field of view is not right anyway. Those adverts you see for Wiktude and Layar with what’s on the screen perfectly lining up like a frame around the real world just don’t tell the story.”



What he’s referring to is that the camera of a smartphone is just that. It’s a camera, for taking pictures, not for providing a 50mm type, 1:1 scale relation with the world. If they were built that way, it’d be too hard to fit the things in that you needed. One of the side effects of this in the use of AR on the hoof, is that it brings a certain degree of distortion to your surroundings when viewed through your phone, something which adds yet another problem to the large pile of issues around getting the real and the virtual to line up on demand and in real time. Just in case we weren’t depressed enough, Feiner chucks on a few more.

“The GPS system isn’t good enough either. You can get better equipment with backpacks that surveyors use. It brings the accuracy down to a few cm rather 5m or so and that software is freely available. It’s just that it costs in excess of $10,000.”

Not something we'll be finding in the next Samsung Bada then.

“We could also use better gyros, compasses and accelerometers too but the likes of Qualcomm and Nokia have labs working on and targeting AR all over the world and they’ve realised this. So, what they’re actually doing is making their chip technology more tailored for these features to better support what’s already there.”

Thankfully, Feiner’s outlook is much cheerier than his reservations might suggest. Within the iron clad practicality of this scientific researcher beats the heart of a man with a dream.

“I honesty believe that at some point in the future we’re going to have AR eyewear that’s sufficiently light weight, comfortable, visually appealing, high quality enough and at the right price that people will want to wear while walking around. It has to be socially acceptable and desirable.”

While it may sound unlikely to those with 20-20 vision or people who’ve chosen to wear contact lenses instead of spectacle frames to correct their myopia or hyperopia, as Feiner points out, it has become the norm to wear little bits of plastic in our ears - for some they’re even a badge of status. However, he does appreciate that there’s a difference.

“We need to be very careful with eyewear, though. It’s a much higher bar to pass when we’re dealing with someone's face. It’s the first place people look. We need to get around the whole Borg 'we will assimilate you’ look. We need to have good industrial design to make the glasses appealing and comfortable enough, and still be able to see into people’s eyes, but industrial designers are very good and I’m confident that they’ll find a way of doing it.”



As Feiner points out, there are a certain number of prerequisites for such AR glasses. They need to wirelessly connect to the user’s mobile or at least have tiny CPUs and GPUs of their own and not “instant on” but “always on”, but with an easy way of turning them off when you want a break.

"The interesting problem is the software."

As the professor points out, it’s all very well having the kit and the infrastructure in place but making it usable, understandable, helpful and non-intrusive is a whole other task. Aside the ongoing work with the US Marines for the ARMAR (Augmented Reality for Maintenance and Repair) project, software design and development is the primary focus of the work of Feiner and his students at Columbia when it comes to AR.

“It has to be a low level interface,” he stresses. “We don’t want people to get run over while totally immersed in the sky or the trees or something else.”

The key is that while we’re normally undertaking a number of jobs at any one time, there’s always a prime task to which we’re attending and the research is about how to narrow what’s out there down to just the information that we require for what’s in hand. One technique is simply to filter out everything else that’s going on but another is to keep most of the input there, but only highlighting that which is germane to the moment.

Rather like how some calendars work, the UI can focus on the primary task and make the information displayed for that purpose much more obvious than anything else. The secondary, tertiary and next level data can still be there but slightly smaller or not highlighted in the same way, and the AR below that can either still be present or not at all.

“It’s what we’re referring to as a fish-eye view and proximity is one way to do it. The idea is that things physically closer to the user are the ones that are more important so the information for these objects are the ones the system would prioritise.

“We can even work with haptic, visual and sonic cues as well to bring more of the user’s senses into play, but what we really need for it all to work is a system that shows us the world by sensing our responses to it, as well as an element of us telling it what we want it to do. That’s the heart of what it needs before we can have good, usable AR.”



If Feiner and the department at Columbia didn’t already have their work cut our for them, there’s also plenty of issues if you want to start rendering the augmentations on these glasses in 3D, as will doubtless become important when augmentations have to be objects rather than just text. So given that it’s going to be tricky enough even to get good AR spectacles going, how about the nirvana of the augmented reality contact lenses?

“Contact lenses? Yep, it will happen. It’s very delicate and there’s lots of problems including radiation to your eyeball and how people are going to be able to see a sharp image so close to their retina rather than a wash of colour. It’s hard but it will be doable.

“But my question to you is this - why have it washing around on the surface of your eye when you can have it implanted inside your head? Sure there are social and ethical issues but these things will change with each generation as it becomes more acceptable.

“And then if you can have it implanted in your head as an adult, then why not have it done at birth? And if that can happen, then why not into our genes? And at that point, we would have changed the human species altogether but beyond that, I’ve no idea where we’ll go.”

For more information on what Qualcomm is doing with Augmented Reality please click: http://www.qualcomm.co.uk/products/augmented-reality

And for more on AR Week head over to our AR Week homepage on Pocket-lint



>