All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
The technology industry’s hyperactive search for the next big thing has a new object of affection. Forget voice assistants; at least for today, augmented reality is the smartphone’s heir apparent. “It's definitely having a moment,” says Brian Mullins ruefully. “It's probably having two different moments at the same time, because the truth is, AR is here, right now, today. People use AR at work. We ship to almost 200 customers.”
Getting augmented reality to customers became a great deal easier last year, after Apple and Android launched kits to help developers make AR apps on the world’s dominant mobile operating systems. But DAQRI, the AR startup which Mullins co-founded in 2010, goes far beyond the phone-in-front-of-the-face Pokemon Go variety of the technology. Its $15,000 (£11,000) AR helmets can be worn for hours, entirely hands-free – a must for the soldiers, sailors and industrial workers who are its primary users.
Now the LA-based startup is moving into sales mode – a shift signalled by a change in leadership. In October, 2017, Mullins stepped down as CEO, to be replaced by chief product officer and former Qualcomm executive Roy Ashok. (Mullins remains on the board.) “We have been R&D and somewhat visionary,” Ashok says. “We're moving now into a more mass-market, execution-focused company.” In other words: AR better be here, right now, because DAQRI is betting its future on it.
In June 2017, the US Space Naval Warfare Lab released a video from on board the USS Bunker Hill. In it, the cruiser’s machine gunners fired live rounds into the sea, kicking up bursts of spray. Such training exercises are hardly unusual – but this time, something was different. The command to “fire” wasn’t shouted across the deck, but transmitted visually, via “GunnAR”, a custom version of DAQRI’s hard hat-style helmet.
The “Smart Helmet”, as it is known, was DAQRI’s first product. It was designed for industrial and military uses: “a foreman at a heavy manufacturing site,” was the imagined ideal user. Clients included German internet of things giant Siemens, which handed the device to teams of gas turbine inspectors. “You can see flow rates and temperatures on top of the real turbine,” Mullins explains. “In industrial environments things tend to get hot before they break, so you can use it both as a safety sensor and to detect problems.”
Mullins, a burly 42-year-old with the precise diction and erect posture of a US Merchant Marine Academy graduate, is in a cluttered research lab on a business park just outside Milton Keynes, showing off DAQRI’s second, and latest, product. Launched in August 2017, the black-and-white Smart Glasses are a smaller, lighter version of the Smart Helmet – AR for the boss, rather than the worker. This is where DAQRI thinks it can make AR mass-market. But even with the battery and processing units squeezed into a beltpack, the new device is a long way from comfortable.
The problem is the one that always undermines AR: the narrow rectangular field of view, that cuts off images when you move, as if you’re peering at the world from inside a post box. “It’s twice the field of view of a Hololens,” Mullins says defensively, referring to Microsoft’s rival headset. But he knows that with this limitation – the result of painful but necessary compromises between size, speed and quality of display – AR is a long way from replacing smartphones. “That's not a thing. That's not going to happen. It will eventually, but not in the next three to five years. That's just too insurmountable an adoption rate.”
Mullins first saw something that looked like AR in 1997, during his service obligation at the Department of Transportation. He worked on simulations for training pilots. “There was literally a model of a ship inside a big building. If you looked through the windows you'd see a giant simulation projected in the distance, but also a model of the boat.” He saw similar concepts over the next few years, working as a robotics consultant. But it wasn’t until the iPhone 4 came out in 2010 that Mullins realised devices were now small and powerful enough to make AR. He started at exactly the point as developers for ARKit and ARCore today: with see-through displays using the camera on the phone.
If AR is in its youthful stages now, back then it was positively infantile. Mullins initial pitch was rejected by 123 investors. “It was a little hard to convince everyone,” he recalls, mildly. Even when the firm had successes – animated magazine covers; partnerships Sony and Crayola – it felt gimmicky rather than game-changing. Then, in early 2014, just as DAQRI was raising a $15 million Series A round, Google released Glass.
Years after they were last seen in public, the memory of glassholes lingers over AR like a bad smell. Lately that odour has started to fade: in July 2017, Alphabet, Google’s parent company, even released a Glass Enterprise Edition (competition for DAQRI in years to come). But even when disgust at Glass was at its most vituperative, Mullins had reason to thank it. Google’s failure convinced him to stop waiting for the big companies and build the hardware for AR himself. With $200 million in funding, around 250 employees and six facilities worldwide, the bet is just about paying off, despite workforce cuts of 25 per cent in February 2017. But the question remains: when will AR escape the post box? This is what DAQRI is working on in Milton Keynes.
In one corner of the laboratory, DAQRI’s British CTO, Jamieson Christmas, stands next to small transparent screen showing a road as it appears to a driver. As the car advances, directions pop up, apparently on top of junctions, as if they are appearing in the road itself. The device making this happen is the laser holographic unit Christmas invented at his University of Cambridge spin-off Two Trees Photonics, which DAQRI acquired in March 2016 for an undisclosed sum. (“It was substantive!” says Mullins. “But the best investment we've made.”) Two Trees – named for Christmas and his co-founder Peter Woodland – sold the first version of the device to Jaguar Land Rover, who used it to project dashboard information onto the windscreens of their cars. But that was still a flat display. It was only when the holographics were combined with DAQRI’s computer vision technology, which mapped the world in real time, that they truly deserved the title of AR.
Christmas, a slim, softly spoken 43-year-old, developed the holographic unit for a PhD he finished while working full time as chief UK engineer of Japanese electronics manufacturer Alps: “only OK if you don't enjoy sleeping,” he laughs. Rather than traditional LED backlit LCD displays, which use pixels to push out light, it reconstructs images by steering existing light energy into the right shape, a technique known as phase only holography. The basis of the technology is a light-manipulating device called a spatial light modulator, but the real advance is the complex mathematical algorithms which dictate how the light is shaped. In this way, problems such as the field of view are pushed from hardware towards software – a realm where, even if the calculations are extremely hard, they have the advantage of not needing to fight against fundamental limits of physics.
Other AR startups are looking for different answers to the same question, most notably secretive Google-backed Magic Leap, whose long-awaited headset is rumoured to project images directly onto retina. But Mullins is confident in Christmas’s solution. “It solves fundamental optical challenges,” he says. “You don't need to resort to things like retinal projection that add a computational burden.”
Microsoft appears to agree: in July 2017, it revealed a pair of prototype AR glasses that employed a very similar technique. “Near-eye holographic displays… offer high resolution, full colour, good image quality, variable focus, and vision correction,” Microsoft researchers wrote in an accompanying paper, adding that it “enabled a wide field of view in a very compact device.” It didn’t take AR out of the post box. (“There's still a lot of work to be done,” stressed the researchers.) But, at long last, it showed a way it could happen.
DAQRI, however, is thinking beyond headsets. Two Trees’ holographic unit has completed the gruelling process of qualifying for use in cars, giving the firm a route into the gigantic automotive market. “It's very clear that some time in the next three or four years automotive will be the largest customers,” says Mullins. “One car sells millions and that gets very interesting very quickly.” The possibilities abound: safety warnings; animated directions; personalised billboards in driverless vehicles, an idea which could be replicated in buildings across the world. Will DAQRI be there to see it? Only time – and sales of Smart Glasses – will tell.
This article was originally published by WIRED UK
